I think the posted title need to be improved and updated to include function or waveform generator instead of just More.
Perhaps because of that most of the comments here just focusing in the oscilloscope part.
As any who has dabbled with electronics know your typical workbench is normally consist of oscilloscope, logic analyzer, waveform generator, digital multi-meter and power supply normally 12V DC.
It seems to me this little $18 hybrid signals scope/generator can do almost all of these functions, and since they are integrated on the same device you don't need to synchronize them which is a big plus.
For old school HP/Agilent/Keysight electronics workbench discrete solutions can easily cost tens of thousands dollars, no kidding.
The modern version of this is the Digilent Discovery Pro 3000/4000/5000 series with price tags of several thousands dollars. They are described as the All-In-One High-Speed Mixed Signal Oscilloscope, Function Generator, Power Supply, and DMM. The cheaper version of this Digilent Discovery 1/2/3 series that cost around $400. But the latter is quite bandwidth and resources limited thus more comparable with the Flea-Scope [1].
[1] Digilent Analog Discovery 3 vs. Digilent Pro ADP2230 [video]:
One post below might be a bit misleading -- it's just 18 Msps, not 100 -- so think of it as good for signals up to 1 MHz or so with decent fidelity... And it uses a PIC32, not STM32. It is 12 bit ADC, but the noise floor makes it more like 10 or 11. A cool thing is it also does digital capture and waveform generation at the same time, but clearly it's all pretty simple...
My goal was to be able to get these in a classroom at every seat, where there's already a computer or tablet or chromebook, and have it "just work"... The thing I like the best by far, though, is javascript access -- you only need a webpage for the GUI -- no software install or app store or anything!
To go deeper for the classroom experiments, you can use a different "deep dive" webpage with a command-line UI and log in interactively and take control of all the pins programmatically in BASIC, and even configure the board to autorun a BASIC program on power-up... You can make a "simon" game with 4 switches, 4 LEDs, a buzzer, and about 100 lines of code... Or with a thermocouple and an op-amp and a solid-state relay, you can make a cool toaster oven temperature controller for reflow soldering -- which is how I built all the prototypes... The nice thing is the javascript access on the host comes for free still, and so BASIC can talk directly to the host computer (e.g., running python or powershell)! Examples are here: https://rtestardi.github.io/pages/
I have also had fun with automotive applications for Flea-Scope -- it is so easy to just bring out a phone or tablet and measure a crankshaft sensor -- and without it you're just guessing and replacing components from some diagnostics flowchart...
As for licensing, it is all open source, and anyone can rebuild it (and even sell it). There is a patent on the internal BASIC, but there is also a perpetual license to the software builds I have made and tested and released -- conceivably you could build the scope without BASIC at all -- it is unneeded except when you use the board in "deep dive" mode for BASIC programming.
If there is interest, I have toyed with the idea of making a Mosquito-Scope, with 2 channels and selectable input amplification (for sub-millivolt signals) based on the still-brand-new dsPIC33A, which is actually cheaper and faster than the PIC32MK.
I know you are looking at this for education but I've actually ordered two to toss in my travel bag for when I visit factories to handle manufacturing bringup.
I think the low cost of the flea-scope justifies sticking to just one analog input per device, even with a successor, provided that the interface gets a little bit of an upgrade to be able to handle multiple scopes at once.
This would retain the small form factor and would keep the design free from compromises.
Wow, I love this idea -- I bet I could make an alternate UI display two analog waveforms from two devices instead of analog and digital from one... I don't think it would even require threading complexity, since the second device will only trigger after the first already has!
My goal was to be able to get these in a classroom at every seat
This is really amazing and you are to be commended. And you've clearly thought hard about what can make that successful. There's just no replacement for every student having their own setup to experiment with. My first exposure to o-scopes, the whole class was basically sharing one giant Tek 7k series scope, one of the best of the time (with commiserate cost), and while I did learn the basics, it wasn't till I had my own much, much less feature filled scope did I learn it well enough to be useful.
N.B.: Since you're new here, I'll just warn you that it's SoP around here that any time anyone posts an "I built this", the comments will be filled with nitpicks about trivial issues presented as massive errors, people telling you how much better they would have done it if they ever got around to doing it and how they got something cheaper on alibaba so why bother. And, of course, you should have written it in Rust. Just ignore that stuff.
For about the same price ($13), I'm happy with this logic analyzer I bought through amazon[1] a month ago. I spent more on probes[2] ($24) than I did the analyzer. ;-)
To power it, I'm using SigRock/Pulseview[3], which sees it as a compatible clone. It's wild to be able to see signals a few tens of nanoseconds wide so cheaply.
I have this exact one, and I don't reccomend using it with the Saleae Logic software, it constantly drops out and has issues enumerating. Haven't tried sigrock yet.
> I don't reccomend using it with the Saleae Logic software
The Saleae software is designed for use only with Saleae products. Anything else is a clone.
I wouldn’t be surprised if they intentionally made the software frustrating and buggy to use with clones. The software (which is great) is funded by hardware purchases.
Yeah, those are awesome. I found that a great use was to profile microcontroller code. Setting various GPIO pins high / low around critical sections of code. Then you can see the results in SigRock.
Mandatory caveat: beware of ground loops when using a USB oscilloscope! Most budget USB scopes share ground with your PC via the USB port. If you're probing high-voltage circuits (e.g., mains power, switching supplies), this can create a dangerous potential difference, risking damage to your computer or worse. Some scopes offer isolated inputs, but if yours doesn’t, use differential probes or an isolation transformer. Otherwise, you might end up debugging your laptop’s fried USB controller instead of your circuit.
On a related note, I once fried the built in audio hardware of my laptop. I even saw smoke coming out one of the vents. The rest of the computer survived.
I think projects like that would have been a godsend 2-3 decades ago, when even a basic oscilloscope costed as much as a used car.
Nowadays, very good oscilloscopes with 200 MHz bandwidth, good user interfaces, and responsive displays are selling for $300 - I'm talking Siglent, Rigol, UNI-T. So the merits of DIYing something much worse just aren't quite there. It's that one piece of equipment you use to troubleshoot all your other designs, so you want it to be dependable, easy to use, and accurate.
This is not to say it's not a fun, geeky project to work on and publish... but you know, only once you have a real oscilloscope. If you're just setting up, do yourself a favor and spend a bit more money on this. The remaining equipment is not nearly as critical.
> very good oscilloscopes with 200 MHz bandwidth, good user interfaces, and responsive displays are selling for $300
This is a very strong first world POV. Those U$D 300 rank pretty high across world's minimum salaries [1], not to mention that prices can very likely double in countries like my own.
Moreover, a large % of people that can really get a very good use of this (Kids in high school) usually have a lot less money at their disposal, and we are not getting into collaterals like building your own and learning how it works.
Want to keep a distractable 14 year old busy for an hour? Give him or her some wire, a cheap meter, a battery, some switches and LEDs. Want to possibly change that kid’s life? Give him the above, plus a scope, and hook it up to a music player playing his favorite song.
That's a very fair point, but it's also fair to discuss things with a 'first world' point of view. The same way it can be humourous or eye-opening to point out something's a 'first world problem', but if that's where you live then that is a problem, that is something you're looking to address.
We're all looking to do the best we can from different baselines.
Note: the intention of this comment is just to be informative.
I did a quick search in my country's Mercado Libre (the local Amazon, we could say) and the cheapest available one, Fnirsi-138 Pro [1] goes for a price of aprox. 50 USD [2]
what makes the Flea-Scope so particularly interesting to me is that you get I think 9 channels worth of digital inputs all at once. That is hard to beat for the price.
Regarding raw specs, it seems to be much better in most regards compared to the mentioned Fnirsi-138 Pro.
So starting with nothing at all, what is the total cost of ownership for the flea -cope? My guess, its over $300, (if $300 is your annual salary, your existing phone will probably not cut it). And for comparison, standalone toy oscilloscopes (<10Mhz bandwidth) go for $30 and under (some $11), with screen, on AliExpress. And a "decent" one, like OWON, is around $150.
The problem is, with a toy scope, you aren't really gonna know if what you're measuring is real. This might be useful as a kit to build to learn how to program microcontrollers, or measure audio signals (the $11 one, or the mic on your phone can do this), but a bad scope will generally cause more problems for the hobbyist. When you get near the limit of a scope's bandwidth, the signals get really messy and full of artifacts. Arduinos run at 8-16Mhz, so you're gonna hit a wall really quickly and once you can't rely on the output, the investment will be lost.
Where's that guess coming from? The scope is $18, the optional probe from elecrow is $5 extra. You most definitely have either an android phone or a laptop/PC with a browser from the last ten years already. As such the cost is pretty minimal.
Say you bought a set of these kind of test probes with hooks at the front. A set would maybe cost you $10. Now you can use all your 9 channels.
I don't understand your price calculations.
Every scope out there has an upper limit in terms of speed. If you have no use for it, don't buy it. This oscilloscope is first and foremost a learning resource.
"This is a very strong first world POV. Those U$D 300 rank pretty high across world's minimum salaries "
My point, since I was replying to a poster who said the above, was if you are making $300 a year, you don't have a laptop. So, if you read my assumption, that you don't have a laptop, android phone or laptop, and you are making < $1000 USD a year, the Flea-Scope will cost more than a decent one with a built in display and much more than a ~2mhz toy intro oscilloscope ($11 on Aliexpress) [1]
I could also very well reply with just the quote above.
I think there's a difference in options at $200/$20/$2 (see elsewhere in this thread) just in lowering the bar to entry to the point of triviality. I'm not spending $300 without knowing exactly what I want it for, $20 is easy, $2 is an impulse buy.
Fair enough, but on the flip side - you either want to get into electronics or not. If not, there's no real point in spending the money, even if it's $2. If you do, then you probably don't yet have an understanding of what features you need, and the cheapest option will actually hold you back.
Some specific issues: first, the number of inputs. A lot of circuit debugging is about "let's see how signal A looks like when signal B happens" (B might be a bus clock or something like that). So, a lot of the time, you need two inputs, not one.
Second issue: even hobby MCUs generally run faster than 4 MHz, so you might need more bandwidth to monitor I/O, even for old-school Arduino stuff - let alone RP2040.
Third issue: for anything analog, from audio equipment to household appliances, the 0-6 V input range just doesn't cut it.
I'm not trying to dunk on this project: I think it's about as good as you can do for the price, and it's clearly a passion thing for the author. But if you can afford it, and if you want to learn electronics, a "real" oscilloscope is almost certainly a better deal.
> you either want to get into electronics or not. If not, there's no real point in spending the money, even if it's $2
It's fuzzier than that; there are hobbies that look fun but I can't decide if I want to fully invest into them, but having a cheap, but not-great equipment is a good stepping stone. Besides, beginners won't have the skill to fully use fully-featured pro-level-gear. A cheap soldering iron is great for those looking to get their feet wet; sure the lack of precise temperature control can be a hinderance, but investing in an expensive soldering station is wasteful if one decides that it's not really for them.
Ironically, beginners would benefit the most from a good soldering station, proper liquid flux and leaded solder. Seasoned electronics guys can probably solder anything with a cigarette lighter and a scrap piece of metal, but using proper tools makes a huge difference when you're just starting and might mean not dropping the hobby altogether because of a lousy first experience..
> Seasoned electronics guys can probably solder anything with a cigarette lighter and a scrap piece of metal
This made me laugh. In college in the late 80's I repaired a roommate's not-quite vintage C64 fastloader cartridge with a bad wire bodge using a lighter and the tine of a dining hall fork...
> having a cheap, but not-great equipment is a good stepping stone.
I don't agree for something like this. The problem is that beginners often don't understand how to deal with the limitations of equipment--all they understand is that something is wonky and they don't know how to fix it.
In addition, bad equipment can make your initial impressions so painful that they prevent you from going further. Oscilloscopes, cheap soldering tools, etc. all have this effect. It's not unique to electrical engineering though--super cheap guitars are painful to play while something just $150 more is dramatically easier to deal with.
Where cheap equipment shines is when a beginner has been at it long enough and becomes "dangerous". They know just enough to do things but not enough to avoid things that will destroy their equipment. That's when the $20 "whatever" is a blessing because you'll get annoyed but you won't be distraught if you blow it up. Even as someone experienced, I semi-regularly destroy a $10 logic analyzer because the circuit I was analyzing wasn't doing what I thought it was (of course, this is mostly laziness--since the analyzer is merely $10 I don't take the time to double check things that I would if I were to put my Agilent analyzer on it)
> Besides, beginners won't have the skill to fully use fully-featured pro-level-gear.
Quite often, pro gear is better for beginners. For a beginner, an "Auto Scale" button on their scope that does the right thing is a godsend. Even I use it all the time. Sure, I can dial the signals in quickly, but "Auto Scale" does the right thing 99% of the time with a single button press.
Even with soldering, a Metcal system maintains the temperature exactly where you need it. You don't have to worry about having too much/not enough heat on the component you are working with. You don't have to worry about the calibration on your soldering station. A Metcal will pump a lot more energy into a big hunk of transformer and a lot less energy into a QFP integrated circuit. Sure, someone experienced can adjust for this on a Hakko, but a beginner is having a tough enough time getting solder down and the soldering iron in the right place without knocking the IC all over the place and doing it all through a microscope.
> investing in an expensive soldering station is wasteful if one decides that it's not really for them.
Used Metcal soldering systems are "expensive" ($250-$500), but they also resell for almost the exact same amount you bought them. Cheap soldering stations won't resell for anything. So, all the money you spend on a cheap station is lost if you decide not to go further while the expensive station cost all gets recovered.
Regarding soldering, I think a Pinecil [1] is a suitable soldering iron for beginners. It's affordable and since the heating element is in the tip it will make soldering much easier compared to older irons or soldering stations.
Could someone recommend one starting out (something beginner could use and has enough, good functions for continued use into later advanced projects without running into limitations for 'most' projects).
Would be helpful to understand the what limitations might be encountered such at the frequency.
Buy a Pinecile[1] from Pine64. It's currently discounted to $25.99 (excluding shipping), and it's both cheap and has a great feature list that makes soldering a breeze.
Rich designed it with trigger IO pins to be able to easily chain them. While the UI doesn't support managing multiple Flea-Scopes at once, you can just open the UI multiple times, which does seem like a reasonable option to me.
As as others have pointed out before, there's just no alternatives in this form factor. The scope is tiny, and for what it is it packs quite a punch. 3.3V and 5V are standard logic levels which will suffice for quite a lot of basic probing. Being made for students, it does likely everything that would be required of it in an educational environment. How well this pans out in real-life we'll see soon.
Granted that with enough money and space at hand there are much better options out there.
I got one recently, and it's been a pleasure to use. My only regret is not getting the 4 channel version, though I've only needed 2 channels so far. But at my workplace, I've found that an extra channel is the thing that helps me solve a problem, more frequently than higher performance does.
I'm an old timer so I'm familiar with the traditional scope front panel, but have come to like the touch screen for changing settings.
Perhaps my only gripe is how long it takes to boot.
I get around the long boot time by never turning mine off lol
I used to thing touch screens on scopes were a gimmick. Then I used one to name a signal quickly and easily for a trace capture. After that I was hooked.
Also - be sure to upgrade the firmware if you haven’t yet. They fixed that annoying bug where it doesn’t remember your probe settings between power cycles!
Edit: Updated the firmware. Works. First time I noticed the problem, it was at my desk, and I assumed it was because the scope was plugged into a switched power outlet, and I was habitually turning off the lights when I left the room. Only after a couple of tries did I discover that it was an actual bug.
HN: Come for the arcane programming topics that are over my head, stay for fixing my scope!
I started out buying little cheapy toy scopes when I started tinkering with electronics and I wish I’d just waited and saved for a proper one.
Way more useable, especially having proper probes.
I get it - I’m in the first world, unlimited money and privilege, yada yada - but if you think you realistically can save $400 USD for a scope like this, you should.
In his presentation video [0] Rich explicitly states that his goal is to make electronics approachable for the younger generation. A low barrier of entry in almost all directions is a key component in that in my opinion.
It’s a matter of perspective, though. For a white collar professional in the US? Sure. For most of the world’s population, the distinction between $300 and $18 is extremely consequential.
Right. $18 is 6% of $300. For that to be a difference of little consequence, $300 must be an expenditure of little consequence. In the other direction $300 is 6% of $5000. There are people— a number of whom likely frequent this forum— for whom a $5000 expenditure isn’t particularly consequential, and getting an OK thing for $300 doesn’t make sense when you can get a pretty good thing for $5000. That certainly doesn’t mean someone being able to offer a useful version of a $5000 thing for $300 has no merit.
I see this sort of thing in amateur radio all the time. “Why on earth would anyone just starting out consider getting a Baofeng radio for $20 when you can get a better quality name brand equivalent for $200?” Well because $180 is a whole lot of money for a whole lot of people.
I think there's something to be said for the form factor here though: I can see myself owning one of these just to take as part of a kit if I expect to need to do some troubleshooting away from my typical workspace.
I have a cheap handheld oscilloscope (ZOYI ZT703S, ~$80) that I find very convenient to have around. It's the size of a multimeter, runs on a battery and works well enough that I don't feel too limited by it.
Nowadays I have a fairly decent Rigol scope, which suffices for my modest and infrequent needs. But back when I was a teenager I had very little money but plenty of time. A cheap DIY option would have made a lot of difference to me. I hope project this makes a difference to people on a budget, of whatever age.
edit: parent was heavily edited, so this comment makes less sense now.
Note that calling this FOSS is completely inaccurate. Some parts are all rights reserved, some are public domain and some are only available for non-commercial use.
Public domain is not open source, and especially not free (as in freedom). Restricting commercial use is also not free. At best, this project is partially source-available.
> Public domain is not open source, and especially not free (as in freedom).
The Free Software Foundation (FSF) and Open Source Initiative (OSI) recognize public domain software as "free software," but not as open source, but as I understand it, it's not that different from MIT licensed software. Someone can take MIT licensed code, make changes, sell the resulting binary as proprietary code and no one else is allowed to take that binary and sell it as theirs. There are some subtle other differences, but MIT licensing doesn't require other people's changes get contributed back out as open, same as public domain.
The OSI does not recognize source code released to the public domain as free [1], because this "will depend where the software was written, where you are located, who the author is and where the people you are sharing the software with are located."
The only thing more frustrating than a rustacaen redefining "safety" to suit their purpose is a Stallman discipline redefining freedom to mean something totally weird. Something in the public domain is not free? Give me a break.
Copyright law is tricky. Under US law, "public domain" means "not copyrighted", the author waives all rights and the law allows them to release their works to the public domain. This you can consider free.
In other jurisdictions, say The Netherlands, the "Auteurswet" does not have a way for you to release your source code into the public domain, making any such statement void. Other countries such as Germany do not allow the author to waive all rights ("Urheberrecht"), etc etc.
Without knowing where the software was written, by whom the software was written and with whom it is shared, you cannot guarantee anything from "public domain" software.
"Memory safety" is a concept that predates rust, but was certainly popularized by it. It refers to statically eliminating a category of bugs - those related to misusing memory - and all sources of undefined behavior.
Rust code without unsafe blocks guarantees memory safety, and when you're writing unsafe blocks you're culturally expected to expose a safe-api that upholds rust's memory safety guarantees for users who don't themselves use unsafe blocks.
Rust programmers (rustaceans, a pun on crustacean because the language mascot is a crustacean) tend to extol the virtues of memory safety [1], and are sometimes lax with using the word "safety" to refer to "memory safety". There are obviously non-memory-safety safety considerations as well (e.g. the other 1/3rds of security vulnerabilities. Or all the non-software uses of the word like making buildings that don't fall down).
Personally I think it's usually pretty clear how the word is being used, and don't find it annoying. But then I'm a rust programmer.
[1] E.g. pointing out that roughly 2/3rds of security vulnerabilities are caused by the lack of it.
> How so? There are no restrictions on use or distribution.
It's not open source because it does not use an open source license (public domain is not a license). It's not free because this depends on your and the author's jurisdiction (if you both live in the US, then yes there are no restrictions on use or distribution)
> I would include neutering any commercial uses, in which case GPL is non-free.
In what way does GPL neuter commercial uses? GPLv3 simply states that you have to make your code available to your users (note: users, not everyone), not that you cannot sell your product.
Jurisdictions always matter, other OS licenses are also subject to local laws and its not a given they will be valid or interpreted as intended. Arguably it's the intention of the creator of the software that matters in how licenses are actually enforced (or not). So in that sense PD is fine. There are also plenty of easy cures for PD, such as making changes and then copyrighting and licensing that derived work as required.
The GPL forces you to license all your relevant software under restrictive terms (ie under the terms of the GPL). How people can declare that as 'free' I'll never understand, it's just ideological nonsense.
The GPL makes it hard or impossible to make money from your software since anyone else can duplicate and sell it. That's obviously neutering commercial use.
Strange patent. At first glance you could do something similar with an intel 8052AH BASIC (a 8051 type microcontroller with integrated basic interpreter) back in the 90ies.
Indeed, I built simple industrial controllers from 8052AH-BASIC chips. And lots of patents are strange. Sometimes the examination process results in a narrowing of the claims to the point where they don't really read on anything.
My advice is to stay away from "FOSS" unless you're GPL, and "open source" unless you're an OSI-certified license. It's not that I don't think those terms apply to anything other than that, but using those terms invites debate about licensing and attempts to control language instead of discussion of your project.
"Source available" is a term that usually no one gets mad at.
The thread title as of this moment (subject to moderator edit) says "source available" which seems reasonable. So I don't think you made a mistake. "With published design" might also work.
I've got my own little GitHub page with projects that I share. They usually involve both hardware and code. I'm also not an engineer, but a scientist and electronics hobbyist, both of which have cultures of open design that predate the software industry. What I do with my projects is attach a MIT license, attribute the sources of any code or ideas that I borrow from elsewhere, and wish you good luck. ;-)
I can recommend this gizmo for people just starting to learn about electronics. I got one in Hackerbox 102[1] and found it to be a good value in terms of bandwidth vs cost.
In talking about it to my engineering friends I got the somewhat expected dismissive "well its just a toy" kinds of comments, and to be clear it has a lot of limitations, but within those limitations it also has a lot of opportunity to teach people about things like analog electronics which are best learned by looking at the signals. And given that so many people starting out don't even know why they might need an oscilloscope, its a pretty good way to introduce the tool without the huge learning curve of features you may not use for the first year of your learning process :-).
And then when it comes time to pick one of the many really excellent choices of bench tools, you won't be completely clueless about what the functions of the oscilloscope do. So for $20 I consider it a good value and have recommended it to people who were just starting to learn about these things.
The flea-scope's hybrid FPGA/MCU architecture for USB streaming is clever - using FPGA pipelining to handle 100MS/s capture paired with an STM32 for protocol translation is sweet cost-wise. BUT, the 8-bit ADC resolution and lack of input protection networks (compared to Rigol's 1MΩ//20pF frontends with overvoltage clamping) make it risky in case of unattenuated signals.
The Python analysis toolkit using NumPy/SciPy for FFTs instead of baked-in DSP shows cool resource partitioning - could see Jupyter soon.
I'm wondering if we looked at the same document... there is no FPGA and it is PIC32MK0512GPK064 instead of STM32. It's also 12 bits at nowhere near 100 Msps, being only 18 Msps.
Did you use the aid from AI to write the comment, or are you referring to another device?
We’ve had cheap, tiny USB oscilloscope and logic analyzers for well over a decade. They’re all over Amazon and Aliexpress.
The cool thing about this one is that it’s open source and well documented.
There are a lot of downsides: The front end has poor input protection and the entire unit is much less resilient, calibrated, and trustworthy than what you’d use in the Navy.
Have you tried using a NanoVNA? Those things are amazing. For << $100, you get a device that can measure SWR, attenuation, impedance, generate Smith charts, from about 10 Khz to 4 Ghz+ to a reasonable extent.
Oh, and RTLsdr dongles are fantastic deals too, when coupled with the GNU Radio project.
The Amazon letter soup of the month is FNIRSI - I am questioning whether they are actually a letter soup brand. They have a bunch of impossibly cheap hand-held tools, including oscilloscopes. I picked up their multi-meter and power supply and they actually work.
I've seen FNIRSI products in videos by Matthias Wandel (woodgears.ca) on his second channel after the company sent some of them for free to test. They compare relatively well, especially for the price but are be no means perfect. I assume they're still a rebadged letter soup product, but they seem to be trying to get some recognition for the name.
Does the software provide a way to export recorded waveforms to allow analysis in other programs? And what is the maximum number of samples it can record?
Recording, exporting and interpreting waveforms doesn't seem to be amongs the objectives of the scope interface.
Over on hackaday [0] Rich explains that you can however easily grab the data directly from the serial port using python and thus record it:
> (07/28/2024) Sorry, there is no protocol analysis -- just digital display. You can invoke a "scope" command by python or something, and get the data yourself in binary -- I have an example of that in StickOS2 repo python/scope.py, but it is very rough. (Analog has additional challenges of having to do calibration, but digital is easy.)
Using an oscilloscope to find a signal, then writing custom DSP code to decode the signal. I'm thinking of stuff like 1 or n-wire binary protocols, NTSC, and such. Or am I supposed to be looking at software-defined radio (SDR) hardware?
Looks like the schematic and PCB layout are in "Diptrace", which isn't a piece of software I'd heard of before. I'd much prefer Kicad for anything claiming to be hacker-friendly or source-available. It appears this Diptrace software does have a pretty much unlimited 30-day trial, which is preferrable to Altium or Orcad, and I appreciate that they actually sell their software with a perpetual license like buying old shrink-wrapped CDs instead of some subscription hell. But Kicad is so compelling these days.
The Diptrace trial does support exporting various Altium/Orcad/Kicad/Eagle/PADS/Protel/Mentor netlists and file formats, but it would almost be easier to re-draw and re-route the design in your PCB software of choice.
I downloaded the trial and opened it up, the schematic is remarkably simple:
The un-named nets and complete lack of concern for crossing other nets is certainly one way to draw a design. On the one hand, it all fits in a single image and is pretty beginner-friendly, on the other hand, a hierarchical schematic with named nets makes it much easier to take in and reason about a single section of the design.
The PCB is simple as well, it's just a 2-layer design with all traces routed at 8 mils and generous vias at 10 mil drill/25 mil pad. Unfortunately, being a compact 2-layer design, it doesn't have good power planes, and worse yet it doesn't even try to have a ground pour at all. Here's the ground/VSS routing:
I know performance isn't the primary goal of this device, but these decisions are probably leaving a lot on the table. (And here's the bottom routing for completeness: https://i.imgur.com/qxwJvpU.png)
Your eyes do not deceive you, the pin headers aren't in a straight line, they wobble up and down with every other pin 9 mils higher or lower than the previous pin. This probably makes assembly a little simpler because the pin headers will comfortably fit in the 0.0354" holes, but the insulators will flex and hold them in place even when the board is inverted to solder the bottom half, but might put some tension on the solder joints if you wanted to stack a daughterboard on it.
The main U5/U6 connectors are spaced out by 0.9" in Y and aligned in X, the -5V power connector is conveniently aligned in Y with U6 and skips two 0.1" pins in X if you wanted to include that in the design (or try to stick the whole thing in a breadboard). The J3/J4 jumpers/headers by the USB connector are on the same 0.1" grid, but inconveniently not aligned with U6 for deep integration in a breadboard or daughterboard and power isolation through a lab power supply. I do appreciate the inclusion of J10 as an extra ground pin J10 off in the corner - you weren't doing anything with that space anyways, and that's sure to come in useful for someone poking at a circuit with a meter or wiring something up!
A notable feature at this price point is the inclusion of a charge pump to generate -5V, I expected a 0-5V range with nominal rail-to-rail opamp rather than positive and negative voltages. I suppose this is persuasive when working with audio signals, which is probably a significant use case for this device. But again, the circuitous (ha) route to the opamp by the noisy VS- output signal and the decoupling and power supply to the charge pump chip U4 are not helping the performance. I doubt it would pass FCC certification, but it doesn't really have to.
All in all, a pretty neat little device at an impressive price point! Don't take my criticisms as complaining, I meant it to be constructive - well done making cool stuff and putting it on the internet!
For the price this small thing is really awesome.
I think the posted title need to be improved and updated to include function or waveform generator instead of just More.
Perhaps because of that most of the comments here just focusing in the oscilloscope part.
As any who has dabbled with electronics know your typical workbench is normally consist of oscilloscope, logic analyzer, waveform generator, digital multi-meter and power supply normally 12V DC.
It seems to me this little $18 hybrid signals scope/generator can do almost all of these functions, and since they are integrated on the same device you don't need to synchronize them which is a big plus.
For old school HP/Agilent/Keysight electronics workbench discrete solutions can easily cost tens of thousands dollars, no kidding.
The modern version of this is the Digilent Discovery Pro 3000/4000/5000 series with price tags of several thousands dollars. They are described as the All-In-One High-Speed Mixed Signal Oscilloscope, Function Generator, Power Supply, and DMM. The cheaper version of this Digilent Discovery 1/2/3 series that cost around $400. But the latter is quite bandwidth and resources limited thus more comparable with the Flea-Scope [1].
[1] Digilent Analog Discovery 3 vs. Digilent Pro ADP2230 [video]:
https://youtu.be/yr9SGxiBAnI
I'm at the character limit, so I'll take suggestions for a new title!
There's a lot of unique, interesting and useful features this offers, all of which are deserving of a spot in the title...
Edit: Seems like I have lost the ability to change the post title. Unfortunate
Perhaps Dang can help on that
How does it compare to the $100 "YouTube-famous" Zoyi ZT-703S?
Thanks for the info never heard it before, it seems to be an improved 2 channel with 50 MHz version of the older ZT702S single channel version [1].
The Flea-Scope is only 4 MHz single channel so this more similar to the older ZT702S 10 MHz version.
[1] EEVblog 1597 – Zoyi ZT-703S $80 2CH 50MHz Oscilloscope/Multimeter Review:
https://www.eevblog.com/2024/02/07/eevblog-1597-zoyi-zt-703s...
Hi all, new here.
One post below might be a bit misleading -- it's just 18 Msps, not 100 -- so think of it as good for signals up to 1 MHz or so with decent fidelity... And it uses a PIC32, not STM32. It is 12 bit ADC, but the noise floor makes it more like 10 or 11. A cool thing is it also does digital capture and waveform generation at the same time, but clearly it's all pretty simple...
My goal was to be able to get these in a classroom at every seat, where there's already a computer or tablet or chromebook, and have it "just work"... The thing I like the best by far, though, is javascript access -- you only need a webpage for the GUI -- no software install or app store or anything!
To go deeper for the classroom experiments, you can use a different "deep dive" webpage with a command-line UI and log in interactively and take control of all the pins programmatically in BASIC, and even configure the board to autorun a BASIC program on power-up... You can make a "simon" game with 4 switches, 4 LEDs, a buzzer, and about 100 lines of code... Or with a thermocouple and an op-amp and a solid-state relay, you can make a cool toaster oven temperature controller for reflow soldering -- which is how I built all the prototypes... The nice thing is the javascript access on the host comes for free still, and so BASIC can talk directly to the host computer (e.g., running python or powershell)! Examples are here: https://rtestardi.github.io/pages/
I have also had fun with automotive applications for Flea-Scope -- it is so easy to just bring out a phone or tablet and measure a crankshaft sensor -- and without it you're just guessing and replacing components from some diagnostics flowchart...
As for licensing, it is all open source, and anyone can rebuild it (and even sell it). There is a patent on the internal BASIC, but there is also a perpetual license to the software builds I have made and tested and released -- conceivably you could build the scope without BASIC at all -- it is unneeded except when you use the board in "deep dive" mode for BASIC programming.
If there is interest, I have toyed with the idea of making a Mosquito-Scope, with 2 channels and selectable input amplification (for sub-millivolt signals) based on the still-brand-new dsPIC33A, which is actually cheaper and faster than the PIC32MK.
-- Rich
I know you are looking at this for education but I've actually ordered two to toss in my travel bag for when I visit factories to handle manufacturing bringup.
Great tool. Thank you.
What you've done is awesome. Ignore the haters and build your mosquito scope.
I think the low cost of the flea-scope justifies sticking to just one analog input per device, even with a successor, provided that the interface gets a little bit of an upgrade to be able to handle multiple scopes at once.
This would retain the small form factor and would keep the design free from compromises.
Wow, I love this idea -- I bet I could make an alternate UI display two analog waveforms from two devices instead of analog and digital from one... I don't think it would even require threading complexity, since the second device will only trigger after the first already has!
My goal was to be able to get these in a classroom at every seat
This is really amazing and you are to be commended. And you've clearly thought hard about what can make that successful. There's just no replacement for every student having their own setup to experiment with. My first exposure to o-scopes, the whole class was basically sharing one giant Tek 7k series scope, one of the best of the time (with commiserate cost), and while I did learn the basics, it wasn't till I had my own much, much less feature filled scope did I learn it well enough to be useful.
N.B.: Since you're new here, I'll just warn you that it's SoP around here that any time anyone posts an "I built this", the comments will be filled with nitpicks about trivial issues presented as massive errors, people telling you how much better they would have done it if they ever got around to doing it and how they got something cheaper on alibaba so why bother. And, of course, you should have written it in Rust. Just ignore that stuff.
For about the same price ($13), I'm happy with this logic analyzer I bought through amazon[1] a month ago. I spent more on probes[2] ($24) than I did the analyzer. ;-)
To power it, I'm using SigRock/Pulseview[3], which sees it as a compatible clone. It's wild to be able to see signals a few tens of nanoseconds wide so cheaply.
[1] https://www.amazon.com/dp/B077LSG5P2
[2] https://www.amazon.com/dp/B0CLB63GL3
[3] https://sigrok.org/wiki/PulseView
I have this exact one, and I don't reccomend using it with the Saleae Logic software, it constantly drops out and has issues enumerating. Haven't tried sigrock yet.
> I don't reccomend using it with the Saleae Logic software
The Saleae software is designed for use only with Saleae products. Anything else is a clone.
I wouldn’t be surprised if they intentionally made the software frustrating and buggy to use with clones. The software (which is great) is funded by hardware purchases.
I too have the same one, and had endless problems till I switched to a shorter USB cable and suddenly everything was fine.
I think the firmware might have some timing issue with the USB bus, and just a few extra nanoseconds of a 2 foot long cable seems to break it.
Thanks, all of the mini usb cables I've tried are fairly long (~1m). I'll try and find a short one and see if it fixes it.
Yeah, those are awesome. I found that a great use was to profile microcontroller code. Setting various GPIO pins high / low around critical sections of code. Then you can see the results in SigRock.
Mandatory caveat: beware of ground loops when using a USB oscilloscope! Most budget USB scopes share ground with your PC via the USB port. If you're probing high-voltage circuits (e.g., mains power, switching supplies), this can create a dangerous potential difference, risking damage to your computer or worse. Some scopes offer isolated inputs, but if yours doesn’t, use differential probes or an isolation transformer. Otherwise, you might end up debugging your laptop’s fried USB controller instead of your circuit.
On a related note, I once fried the built in audio hardware of my laptop. I even saw smoke coming out one of the vents. The rest of the computer survived.
On page 9 of the linked document Rich also addresses this and calls for the use of a full-speed usb isolator
it can also be good... only 1 wire to connect instead of two to see whats happening on a pin.
Seems like magic.
For comparison, this is what $2 USB scope (800ksps, 2 voltage ranges) look like: https://files.catbox.moe/0t0dq3.jpg
For yet another comparison, scoppy is a similar project: https://github.com/fhdm-dev/scoppy
I made something like this in college with a spare ADC, resistor divider, and an 8052.
*sigh*
Link doesn't work.
I'll put it on GitHub tomorrow
https://github.com/dvhx/arduino-blue-pill-scope
I think projects like that would have been a godsend 2-3 decades ago, when even a basic oscilloscope costed as much as a used car.
Nowadays, very good oscilloscopes with 200 MHz bandwidth, good user interfaces, and responsive displays are selling for $300 - I'm talking Siglent, Rigol, UNI-T. So the merits of DIYing something much worse just aren't quite there. It's that one piece of equipment you use to troubleshoot all your other designs, so you want it to be dependable, easy to use, and accurate.
This is not to say it's not a fun, geeky project to work on and publish... but you know, only once you have a real oscilloscope. If you're just setting up, do yourself a favor and spend a bit more money on this. The remaining equipment is not nearly as critical.
> very good oscilloscopes with 200 MHz bandwidth, good user interfaces, and responsive displays are selling for $300
This is a very strong first world POV. Those U$D 300 rank pretty high across world's minimum salaries [1], not to mention that prices can very likely double in countries like my own.
Moreover, a large % of people that can really get a very good use of this (Kids in high school) usually have a lot less money at their disposal, and we are not getting into collaterals like building your own and learning how it works.
[1] https://en.wikipedia.org/wiki/List_of_countries_by_minimum_w...
I totally agree.
Want to keep a distractable 14 year old busy for an hour? Give him or her some wire, a cheap meter, a battery, some switches and LEDs. Want to possibly change that kid’s life? Give him the above, plus a scope, and hook it up to a music player playing his favorite song.
That's a very fair point, but it's also fair to discuss things with a 'first world' point of view. The same way it can be humourous or eye-opening to point out something's a 'first world problem', but if that's where you live then that is a problem, that is something you're looking to address.
We're all looking to do the best we can from different baselines.
I get what you’re saying, but there are also fully functioning oscilloscopes for far less.
For the purposes you are talking about you can get what you need with one of the $20-40 scopes.
Those ones also tend to be portable and more rugged which is what you need if you don’t have a full blown lab to keep a nice oscilloscope in.
Note: the intention of this comment is just to be informative.
I did a quick search in my country's Mercado Libre (the local Amazon, we could say) and the cheapest available one, Fnirsi-138 Pro [1] goes for a price of aprox. 50 USD [2]
[1] https://www.amazon.com/Digital-Oscilloscope-Portable-Samplin...
https://www.mercadolibre.com.ar/osciloscopio-digital-fnirsi-...
Isn’t that an effect of taxes and logistics/shipping to Argentina?
It seems like the price of anything would scale similarly. If you imported this $18 device to At, it would probably double in price.
what makes the Flea-Scope so particularly interesting to me is that you get I think 9 channels worth of digital inputs all at once. That is hard to beat for the price.
Regarding raw specs, it seems to be much better in most regards compared to the mentioned Fnirsi-138 Pro.
The future has been here for a while but remains unevenly distributed.
Probably my favourite Gibson phrase.
So starting with nothing at all, what is the total cost of ownership for the flea -cope? My guess, its over $300, (if $300 is your annual salary, your existing phone will probably not cut it). And for comparison, standalone toy oscilloscopes (<10Mhz bandwidth) go for $30 and under (some $11), with screen, on AliExpress. And a "decent" one, like OWON, is around $150.
The problem is, with a toy scope, you aren't really gonna know if what you're measuring is real. This might be useful as a kit to build to learn how to program microcontrollers, or measure audio signals (the $11 one, or the mic on your phone can do this), but a bad scope will generally cause more problems for the hobbyist. When you get near the limit of a scope's bandwidth, the signals get really messy and full of artifacts. Arduinos run at 8-16Mhz, so you're gonna hit a wall really quickly and once you can't rely on the output, the investment will be lost.
Where's that guess coming from? The scope is $18, the optional probe from elecrow is $5 extra. You most definitely have either an android phone or a laptop/PC with a browser from the last ten years already. As such the cost is pretty minimal. Say you bought a set of these kind of test probes with hooks at the front. A set would maybe cost you $10. Now you can use all your 9 channels. I don't understand your price calculations.
Every scope out there has an upper limit in terms of speed. If you have no use for it, don't buy it. This oscilloscope is first and foremost a learning resource.
"This is a very strong first world POV. Those U$D 300 rank pretty high across world's minimum salaries "
My point, since I was replying to a poster who said the above, was if you are making $300 a year, you don't have a laptop. So, if you read my assumption, that you don't have a laptop, android phone or laptop, and you are making < $1000 USD a year, the Flea-Scope will cost more than a decent one with a built in display and much more than a ~2mhz toy intro oscilloscope ($11 on Aliexpress) [1]
I could also very well reply with just the quote above.
https://www.aliexpress.us/w/wholesale-oscilloscope-.html?spm...
I think there's a difference in options at $200/$20/$2 (see elsewhere in this thread) just in lowering the bar to entry to the point of triviality. I'm not spending $300 without knowing exactly what I want it for, $20 is easy, $2 is an impulse buy.
Fair enough, but on the flip side - you either want to get into electronics or not. If not, there's no real point in spending the money, even if it's $2. If you do, then you probably don't yet have an understanding of what features you need, and the cheapest option will actually hold you back.
Some specific issues: first, the number of inputs. A lot of circuit debugging is about "let's see how signal A looks like when signal B happens" (B might be a bus clock or something like that). So, a lot of the time, you need two inputs, not one.
Second issue: even hobby MCUs generally run faster than 4 MHz, so you might need more bandwidth to monitor I/O, even for old-school Arduino stuff - let alone RP2040.
Third issue: for anything analog, from audio equipment to household appliances, the 0-6 V input range just doesn't cut it.
I'm not trying to dunk on this project: I think it's about as good as you can do for the price, and it's clearly a passion thing for the author. But if you can afford it, and if you want to learn electronics, a "real" oscilloscope is almost certainly a better deal.
> you either want to get into electronics or not. If not, there's no real point in spending the money, even if it's $2
It's fuzzier than that; there are hobbies that look fun but I can't decide if I want to fully invest into them, but having a cheap, but not-great equipment is a good stepping stone. Besides, beginners won't have the skill to fully use fully-featured pro-level-gear. A cheap soldering iron is great for those looking to get their feet wet; sure the lack of precise temperature control can be a hinderance, but investing in an expensive soldering station is wasteful if one decides that it's not really for them.
Ironically, beginners would benefit the most from a good soldering station, proper liquid flux and leaded solder. Seasoned electronics guys can probably solder anything with a cigarette lighter and a scrap piece of metal, but using proper tools makes a huge difference when you're just starting and might mean not dropping the hobby altogether because of a lousy first experience..
> Seasoned electronics guys can probably solder anything with a cigarette lighter and a scrap piece of metal
This made me laugh. In college in the late 80's I repaired a roommate's not-quite vintage C64 fastloader cartridge with a bad wire bodge using a lighter and the tine of a dining hall fork...
> having a cheap, but not-great equipment is a good stepping stone.
I don't agree for something like this. The problem is that beginners often don't understand how to deal with the limitations of equipment--all they understand is that something is wonky and they don't know how to fix it.
In addition, bad equipment can make your initial impressions so painful that they prevent you from going further. Oscilloscopes, cheap soldering tools, etc. all have this effect. It's not unique to electrical engineering though--super cheap guitars are painful to play while something just $150 more is dramatically easier to deal with.
Where cheap equipment shines is when a beginner has been at it long enough and becomes "dangerous". They know just enough to do things but not enough to avoid things that will destroy their equipment. That's when the $20 "whatever" is a blessing because you'll get annoyed but you won't be distraught if you blow it up. Even as someone experienced, I semi-regularly destroy a $10 logic analyzer because the circuit I was analyzing wasn't doing what I thought it was (of course, this is mostly laziness--since the analyzer is merely $10 I don't take the time to double check things that I would if I were to put my Agilent analyzer on it)
> Besides, beginners won't have the skill to fully use fully-featured pro-level-gear.
Quite often, pro gear is better for beginners. For a beginner, an "Auto Scale" button on their scope that does the right thing is a godsend. Even I use it all the time. Sure, I can dial the signals in quickly, but "Auto Scale" does the right thing 99% of the time with a single button press.
Even with soldering, a Metcal system maintains the temperature exactly where you need it. You don't have to worry about having too much/not enough heat on the component you are working with. You don't have to worry about the calibration on your soldering station. A Metcal will pump a lot more energy into a big hunk of transformer and a lot less energy into a QFP integrated circuit. Sure, someone experienced can adjust for this on a Hakko, but a beginner is having a tough enough time getting solder down and the soldering iron in the right place without knocking the IC all over the place and doing it all through a microscope.
> investing in an expensive soldering station is wasteful if one decides that it's not really for them.
Used Metcal soldering systems are "expensive" ($250-$500), but they also resell for almost the exact same amount you bought them. Cheap soldering stations won't resell for anything. So, all the money you spend on a cheap station is lost if you decide not to go further while the expensive station cost all gets recovered.
Regarding soldering, I think a Pinecil [1] is a suitable soldering iron for beginners. It's affordable and since the heating element is in the tip it will make soldering much easier compared to older irons or soldering stations.
[1]: https://pine64.com/product/pinecil-smart-mini-portable-solde...
Could someone recommend one starting out (something beginner could use and has enough, good functions for continued use into later advanced projects without running into limitations for 'most' projects).
Would be helpful to understand the what limitations might be encountered such at the frequency.
Buy a Pinecile[1] from Pine64. It's currently discounted to $25.99 (excluding shipping), and it's both cheap and has a great feature list that makes soldering a breeze.
1. https://pine64.com/product/pinecil-smart-mini-portable-solde...
Rich designed it with trigger IO pins to be able to easily chain them. While the UI doesn't support managing multiple Flea-Scopes at once, you can just open the UI multiple times, which does seem like a reasonable option to me.
As as others have pointed out before, there's just no alternatives in this form factor. The scope is tiny, and for what it is it packs quite a punch. 3.3V and 5V are standard logic levels which will suffice for quite a lot of basic probing. Being made for students, it does likely everything that would be required of it in an educational environment. How well this pans out in real-life we'll see soon.
Granted that with enough money and space at hand there are much better options out there.
The new Rigol DHO800 series are fabulous bargains for the money, and extremely well designed in a lot of ways:
- 12 bit ADC frontends
- Actually usable touchscreen inputs
- Super simple to drive remotely over a network connection (though getting the IP address with just the touchscreen is a little tricky)
- USB-C power input - you can run it from a battery pack if you like!
- All the essential decoders built in for free (I2C, SPI, UART)
- VESA Mountable - I got a monitor arm to gain back some desk space (awesome feature)
- Very, very easy to hack one of the budget models and upgrade it to a 250MHz model :D
I got one recently, and it's been a pleasure to use. My only regret is not getting the 4 channel version, though I've only needed 2 channels so far. But at my workplace, I've found that an extra channel is the thing that helps me solve a problem, more frequently than higher performance does.
I'm an old timer so I'm familiar with the traditional scope front panel, but have come to like the touch screen for changing settings.
Perhaps my only gripe is how long it takes to boot.
I get around the long boot time by never turning mine off lol
I used to thing touch screens on scopes were a gimmick. Then I used one to name a signal quickly and easily for a trace capture. After that I was hooked.
Also - be sure to upgrade the firmware if you haven’t yet. They fixed that annoying bug where it doesn’t remember your probe settings between power cycles!
Awesome. I've cursed that bug. Thanks!
Edit: Updated the firmware. Works. First time I noticed the problem, it was at my desk, and I assumed it was because the scope was plugged into a switched power outlet, and I was habitually turning off the lights when I left the room. Only after a couple of tries did I discover that it was an actual bug.
HN: Come for the arcane programming topics that are over my head, stay for fixing my scope!
Dude me too. Such a goofy bug oversight. But, fortunately, an easy fix with a firmware upgrade.
That bug was my only complaint with the DHO924 as purchased. I was already extremely satisfied with it before that bugfix.
Hate to say it as a proud American and fanboy of HP’s legacy, but these folks are coming for Keysight on the low end.
I also got one last year. It's a hundred times better than the little ones discussed in this test.
I started out buying little cheapy toy scopes when I started tinkering with electronics and I wish I’d just waited and saved for a proper one.
Way more useable, especially having proper probes.
I get it - I’m in the first world, unlimited money and privilege, yada yada - but if you think you realistically can save $400 USD for a scope like this, you should.
These start from 450 euros.
Well worth the money if you work with electronics with any kind of regularity.
Maybe a bit pricy for the hobbyist.
In his presentation video [0] Rich explicitly states that his goal is to make electronics approachable for the younger generation. A low barrier of entry in almost all directions is a key component in that in my opinion.
[0]: https://www.youtube.com/watch?v=sbRXDkHS_V0
It’s a matter of perspective, though. For a white collar professional in the US? Sure. For most of the world’s population, the distinction between $300 and $18 is extremely consequential.
For the majority of humanity, it's several days' worth of wages vs several weeks or months.
Right. $18 is 6% of $300. For that to be a difference of little consequence, $300 must be an expenditure of little consequence. In the other direction $300 is 6% of $5000. There are people— a number of whom likely frequent this forum— for whom a $5000 expenditure isn’t particularly consequential, and getting an OK thing for $300 doesn’t make sense when you can get a pretty good thing for $5000. That certainly doesn’t mean someone being able to offer a useful version of a $5000 thing for $300 has no merit.
I see this sort of thing in amateur radio all the time. “Why on earth would anyone just starting out consider getting a Baofeng radio for $20 when you can get a better quality name brand equivalent for $200?” Well because $180 is a whole lot of money for a whole lot of people.
I think there's something to be said for the form factor here though: I can see myself owning one of these just to take as part of a kit if I expect to need to do some troubleshooting away from my typical workspace.
I have a cheap handheld oscilloscope (ZOYI ZT703S, ~$80) that I find very convenient to have around. It's the size of a multimeter, runs on a battery and works well enough that I don't feel too limited by it.
There is still a WIDE gap between $18 and $300.
Nowadays I have a fairly decent Rigol scope, which suffices for my modest and infrequent needs. But back when I was a teenager I had very little money but plenty of time. A cheap DIY option would have made a lot of difference to me. I hope project this makes a difference to people on a budget, of whatever age.
You can put, for the same price, almost 20 fleascope in a classroom instead of one profeasional one.
I think it does have its place even today.
Edit: Licensing consists of ARR, Public Domain and MIT.
https://github.com/rtestardi/StickOS2/blob/main/license.txt
https://github.com/rtestardi/usbte/blob/master/LICENSE
edit: parent was heavily edited, so this comment makes less sense now.
Note that calling this FOSS is completely inaccurate. Some parts are all rights reserved, some are public domain and some are only available for non-commercial use.
Public domain is not open source, and especially not free (as in freedom). Restricting commercial use is also not free. At best, this project is partially source-available.
> Public domain is not open source, and especially not free (as in freedom).
The Free Software Foundation (FSF) and Open Source Initiative (OSI) recognize public domain software as "free software," but not as open source, but as I understand it, it's not that different from MIT licensed software. Someone can take MIT licensed code, make changes, sell the resulting binary as proprietary code and no one else is allowed to take that binary and sell it as theirs. There are some subtle other differences, but MIT licensing doesn't require other people's changes get contributed back out as open, same as public domain.
The OSI does not recognize source code released to the public domain as free [1], because this "will depend where the software was written, where you are located, who the author is and where the people you are sharing the software with are located."
[1] https://opensource.org/blog/public-domain-is-not-open-source
[dead]
The only thing more frustrating than a rustacaen redefining "safety" to suit their purpose is a Stallman discipline redefining freedom to mean something totally weird. Something in the public domain is not free? Give me a break.
> Something in the public domain is not free?
Copyright law is tricky. Under US law, "public domain" means "not copyrighted", the author waives all rights and the law allows them to release their works to the public domain. This you can consider free.
In other jurisdictions, say The Netherlands, the "Auteurswet" does not have a way for you to release your source code into the public domain, making any such statement void. Other countries such as Germany do not allow the author to waive all rights ("Urheberrecht"), etc etc.
Without knowing where the software was written, by whom the software was written and with whom it is shared, you cannot guarantee anything from "public domain" software.
Can I get some context for the rustacian safety thing? I'm honestly interested if that's real.
"Memory safety" is a concept that predates rust, but was certainly popularized by it. It refers to statically eliminating a category of bugs - those related to misusing memory - and all sources of undefined behavior.
Rust code without unsafe blocks guarantees memory safety, and when you're writing unsafe blocks you're culturally expected to expose a safe-api that upholds rust's memory safety guarantees for users who don't themselves use unsafe blocks.
Rust programmers (rustaceans, a pun on crustacean because the language mascot is a crustacean) tend to extol the virtues of memory safety [1], and are sometimes lax with using the word "safety" to refer to "memory safety". There are obviously non-memory-safety safety considerations as well (e.g. the other 1/3rds of security vulnerabilities. Or all the non-software uses of the word like making buildings that don't fall down).
Personally I think it's usually pretty clear how the word is being used, and don't find it annoying. But then I'm a rust programmer.
[1] E.g. pointing out that roughly 2/3rds of security vulnerabilities are caused by the lack of it.
> Public domain is not open source, and especially not free (as in freedom)
How so? There are no restrictions on use or distribution.
> Restricting commercial use is also not free.
I agree, and I would include neutering any commercial uses, in which case GPL is non-free.
> How so? There are no restrictions on use or distribution.
It's not open source because it does not use an open source license (public domain is not a license). It's not free because this depends on your and the author's jurisdiction (if you both live in the US, then yes there are no restrictions on use or distribution)
> I would include neutering any commercial uses, in which case GPL is non-free.
In what way does GPL neuter commercial uses? GPLv3 simply states that you have to make your code available to your users (note: users, not everyone), not that you cannot sell your product.
Jurisdictions always matter, other OS licenses are also subject to local laws and its not a given they will be valid or interpreted as intended. Arguably it's the intention of the creator of the software that matters in how licenses are actually enforced (or not). So in that sense PD is fine. There are also plenty of easy cures for PD, such as making changes and then copyrighting and licensing that derived work as required.
The GPL forces you to license all your relevant software under restrictive terms (ie under the terms of the GPL). How people can declare that as 'free' I'll never understand, it's just ideological nonsense.
The GPL makes it hard or impossible to make money from your software since anyone else can duplicate and sell it. That's obviously neutering commercial use.
All Rights Reserved (ARR)
Also, looks like some parts are patented.
https://patents.google.com/patent/US8117587B1
Strange patent. At first glance you could do something similar with an intel 8052AH BASIC (a 8051 type microcontroller with integrated basic interpreter) back in the 90ies.
Indeed, I built simple industrial controllers from 8052AH-BASIC chips. And lots of patents are strange. Sometimes the examination process results in a narrowing of the claims to the point where they don't really read on anything.
Correct. What should I replace "FOSS" with so that the title stays short and concise but correctly reflects the licensing/rights situation?
I'm very new to posting on this platform.
My advice is to stay away from "FOSS" unless you're GPL, and "open source" unless you're an OSI-certified license. It's not that I don't think those terms apply to anything other than that, but using those terms invites debate about licensing and attempts to control language instead of discussion of your project.
"Source available" is a term that usually no one gets mad at.
"partially source available" is maybe a bit long, but something along those lines?
Good question, and I sympathize.
The thread title as of this moment (subject to moderator edit) says "source available" which seems reasonable. So I don't think you made a mistake. "With published design" might also work.
I've got my own little GitHub page with projects that I share. They usually involve both hardware and code. I'm also not an engineer, but a scientist and electronics hobbyist, both of which have cultures of open design that predate the software industry. What I do with my projects is attach a MIT license, attribute the sources of any code or ideas that I borrow from elsewhere, and wish you good luck. ;-)
I can recommend this gizmo for people just starting to learn about electronics. I got one in Hackerbox 102[1] and found it to be a good value in terms of bandwidth vs cost.
In talking about it to my engineering friends I got the somewhat expected dismissive "well its just a toy" kinds of comments, and to be clear it has a lot of limitations, but within those limitations it also has a lot of opportunity to teach people about things like analog electronics which are best learned by looking at the signals. And given that so many people starting out don't even know why they might need an oscilloscope, its a pretty good way to introduce the tool without the huge learning curve of features you may not use for the first year of your learning process :-).
And then when it comes time to pick one of the many really excellent choices of bench tools, you won't be completely clueless about what the functions of the oscilloscope do. So for $20 I consider it a good value and have recommended it to people who were just starting to learn about these things.
[1] https://hackerboxes.com/collections/past-hackerboxes/product... and the experiment guide https://www.instructables.com/HackerBox-0102-Flea-Scope/
Reminds me of the similarly priced EspoTek Labrador
https://espotek.com/labrador/
The flea-scope's hybrid FPGA/MCU architecture for USB streaming is clever - using FPGA pipelining to handle 100MS/s capture paired with an STM32 for protocol translation is sweet cost-wise. BUT, the 8-bit ADC resolution and lack of input protection networks (compared to Rigol's 1MΩ//20pF frontends with overvoltage clamping) make it risky in case of unattenuated signals.
The Python analysis toolkit using NumPy/SciPy for FFTs instead of baked-in DSP shows cool resource partitioning - could see Jupyter soon.
I'm wondering if we looked at the same document... there is no FPGA and it is PIC32MK0512GPK064 instead of STM32. It's also 12 bits at nowhere near 100 Msps, being only 18 Msps.
Did you use the aid from AI to write the comment, or are you referring to another device?
The specs say 12 bit though.
Not as capable as Flea-Scope, but buck50 is another MCU-based oscilloscope and logic analyzer. https://github.com/thanks4opensource/buck50
This is crazy. The ones we used in the Navy were heavy and delicate. Now it fits into the palm of our hand. Just incredible how far things have come.
We’ve had cheap, tiny USB oscilloscope and logic analyzers for well over a decade. They’re all over Amazon and Aliexpress.
The cool thing about this one is that it’s open source and well documented.
There are a lot of downsides: The front end has poor input protection and the entire unit is much less resilient, calibrated, and trustworthy than what you’d use in the Navy.
Have you tried using a NanoVNA? Those things are amazing. For << $100, you get a device that can measure SWR, attenuation, impedance, generate Smith charts, from about 10 Khz to 4 Ghz+ to a reasonable extent.
Oh, and RTLsdr dongles are fantastic deals too, when coupled with the GNU Radio project.
Anything I can buy for $50?
The Amazon letter soup of the month is FNIRSI - I am questioning whether they are actually a letter soup brand. They have a bunch of impossibly cheap hand-held tools, including oscilloscopes. I picked up their multi-meter and power supply and they actually work.
I've seen FNIRSI products in videos by Matthias Wandel (woodgears.ca) on his second channel after the company sent some of them for free to test. They compare relatively well, especially for the price but are be no means perfect. I assume they're still a rebadged letter soup product, but they seem to be trying to get some recognition for the name.
Those look pretty cool! I like how nice the price point is nowadays for 10-20mhz bw scopes.
lol @ highend
"Keysigh Bandwidth 6 GHz 8 channels maximum storage depth 1.6G EXR608A Infiniium EXR series oscilloscope"
137,999.86, I like that 86 cents on the end, maybe the cents are a covert communications channel.
Does the software provide a way to export recorded waveforms to allow analysis in other programs? And what is the maximum number of samples it can record?
Recording, exporting and interpreting waveforms doesn't seem to be amongs the objectives of the scope interface.
Over on hackaday [0] Rich explains that you can however easily grab the data directly from the serial port using python and thus record it:
> (07/28/2024) Sorry, there is no protocol analysis -- just digital display. You can invoke a "scope" command by python or something, and get the data yourself in binary -- I have an example of that in StickOS2 repo python/scope.py, but it is very rough. (Analog has additional challenges of having to do calibration, but digital is easy.)
[0]: https://hackaday.io/project/192598-flea-scope-usb-oscillosco...
What’s the use case you’re trying to solve for here?
Using an oscilloscope to find a signal, then writing custom DSP code to decode the signal. I'm thinking of stuff like 1 or n-wire binary protocols, NTSC, and such. Or am I supposed to be looking at software-defined radio (SDR) hardware?
I've always been fascinated by oscilloscopes, but I don't dabble in electronics. Is there any use of it for normal homeowner?
A Fnrsy three in one oscilloscope is 50 bucks on AliExpress
Looks like the schematic and PCB layout are in "Diptrace", which isn't a piece of software I'd heard of before. I'd much prefer Kicad for anything claiming to be hacker-friendly or source-available. It appears this Diptrace software does have a pretty much unlimited 30-day trial, which is preferrable to Altium or Orcad, and I appreciate that they actually sell their software with a perpetual license like buying old shrink-wrapped CDs instead of some subscription hell. But Kicad is so compelling these days.
The Diptrace trial does support exporting various Altium/Orcad/Kicad/Eagle/PADS/Protel/Mentor netlists and file formats, but it would almost be easier to re-draw and re-route the design in your PCB software of choice.
I downloaded the trial and opened it up, the schematic is remarkably simple:
https://i.imgur.com/8EudXIZ.png
The un-named nets and complete lack of concern for crossing other nets is certainly one way to draw a design. On the one hand, it all fits in a single image and is pretty beginner-friendly, on the other hand, a hierarchical schematic with named nets makes it much easier to take in and reason about a single section of the design.
The PCB is simple as well, it's just a 2-layer design with all traces routed at 8 mils and generous vias at 10 mil drill/25 mil pad. Unfortunately, being a compact 2-layer design, it doesn't have good power planes, and worse yet it doesn't even try to have a ground pour at all. Here's the ground/VSS routing:
https://i.imgur.com/BcKm8vL.png
VDD/power makes a literal loop around much of the waveform output and near the scope input pins:
https://i.imgur.com/E2UK9Ot.png
I know performance isn't the primary goal of this device, but these decisions are probably leaving a lot on the table. (And here's the bottom routing for completeness: https://i.imgur.com/qxwJvpU.png)
Your eyes do not deceive you, the pin headers aren't in a straight line, they wobble up and down with every other pin 9 mils higher or lower than the previous pin. This probably makes assembly a little simpler because the pin headers will comfortably fit in the 0.0354" holes, but the insulators will flex and hold them in place even when the board is inverted to solder the bottom half, but might put some tension on the solder joints if you wanted to stack a daughterboard on it.
The main U5/U6 connectors are spaced out by 0.9" in Y and aligned in X, the -5V power connector is conveniently aligned in Y with U6 and skips two 0.1" pins in X if you wanted to include that in the design (or try to stick the whole thing in a breadboard). The J3/J4 jumpers/headers by the USB connector are on the same 0.1" grid, but inconveniently not aligned with U6 for deep integration in a breadboard or daughterboard and power isolation through a lab power supply. I do appreciate the inclusion of J10 as an extra ground pin J10 off in the corner - you weren't doing anything with that space anyways, and that's sure to come in useful for someone poking at a circuit with a meter or wiring something up!
A notable feature at this price point is the inclusion of a charge pump to generate -5V, I expected a 0-5V range with nominal rail-to-rail opamp rather than positive and negative voltages. I suppose this is persuasive when working with audio signals, which is probably a significant use case for this device. But again, the circuitous (ha) route to the opamp by the noisy VS- output signal and the decoupling and power supply to the charge pump chip U4 are not helping the performance. I doubt it would pass FCC certification, but it doesn't really have to.
All in all, a pretty neat little device at an impressive price point! Don't take my criticisms as complaining, I meant it to be constructive - well done making cool stuff and putting it on the internet!