Should be easy to compare - sand down both the cheapo and a legit one (with a similar datecode) with fine sandpaper, clean with isoprop and shove under an optical microscope - you can then tell how similar/dissimilar the dies are, you don't even have to look for markings or anything super legible, or even sand them to the same metal layer.
Clones tend to be vastly different - different technology node, architecture, die size, etc. - that's because they are generally functional clones, not mask clones.
Err, the title isn't really correct. The genuine part is still only $4. This is a 60 cent knockoff. Not having done design in a while but still having a rough feeling for costs I was a bit confused.
> So what’s going on with the cheap parts? My best guess is that these are either quite good copies, or failed parts that somehow made their way into the hobbyist supply chain.
The vast majority of counterfeit chips I've seen were from ghost shifts but IIRC TI fabs all their analog parts in house, I doubt they're ghost shift parts or failed QC.
The price difference could be explained by LCSC purchasing in bulk directly from TI or similar and selling them at lower margins because their volumes are larger. I've seen "clone" chips sold at LCSC, but they're listed under a different brand (I can't recall one off memory unfortunately).
> These are cheap, relatively fast, and not particularly good. While they sport impressive-sounding 12- and 16-bit readouts, the effective number of bits (ENOB) is usually around 8 or 9.
I don't think that's quite accurate for reasonably modern MCUs. You can typically shake 10+ bits out of them, but you need to take a lot of precautions, such as providing very stable external reference voltage and shutting down unneeded subsystems of the chip.
They're still not as good as standalone ADCs, but they're at a point where you can actually use them for 90% of things that require an ADC.
In cases where you need more bits, there's a lot more that must go into the design, which is what gives me a pause about the article. There's nothing about the PSU the author is using or how he managed the MCU noise and RFI. So I don't know if the findings here are that these are knock-off devices with worse specs, or if his overhead LED lamp is causing a lot of interference.
Ooh this is interesting! I've seen some big price differences between Digikey and LCSC at small volumes; not sure! You can also sometimes buy direct from TI.
Of interest from early in the article: I'm curious how these external ones compare to onboard, e.g. STM32's. Btw, the TI one listed is actually pretty simple to use in comparison. The ST integrated ones have more config and hardware considerations, e.g. complicated calibration procedures, external VREF (mentioned) etc. So, if you do app the config, is the integrated one as good?
The integrated ones usually have nice ways to integrate with timers and other onboard periphs.
Well, I guess a whole-three-genuine-US-dollars is actually pretty expensive for an ADC, and that the person-in-charge-of-your-BOM in one of the countries that can actually still manufacture things can get one for way less than that.
Does it work? Well, does your design power up during factory testing, and then pass whatever things your rig (hope you made a few!) has in mind? Well, then, yes, in fact it does...
I'd like to know more about the world of ADCs. I've used the ADS1115 with success but only at very slow speeds.
On the current project we started with an MCP3208 via SPI. It did the job but only has 8 channels and it's slow (100K samples per sec).
To get something faster we switched to ADS7953. It has 16 channels and runs 10 times faster. It's somewhat more complex to code, and you can only get the highest sample rate if you scan the inputs in a predictable order. But it sure flies.
To me, these chips feel like cars. The ADS7953 is somewhat of a ferrari, whereas the MCP3208 feels like a Toyota, simple to use, unimpressive performance.
I'd love to know the industry background about how these varieties of ADC chips came to be and carved their own space in the world, and how widely they are used (millions? billions?).
Digikey's markup is the issue.. most people in Asia buy from lcsc or agents, that can get parts 2-10 times cheaper.. In this case both parts are genuine just different batch/year/production location probably..
Most companies have region-specific price lists. My Spotify subscription costs about 60% of what the same plan costs in the US (2200 HUF vs. 11 USD). In the electronics industry, everyone has a separate price list for China.
Btw. the western list price is just an indicative at-most number anyway. Even a small-sized project gets discounted prices when you start talking to a sales rep.
Why would it be bad for that price? Highly accurate ADCs are pretty easy to make, provided you don't need a high sample rate, thanks to the magic of signal processing. Delta-sigma ADCs, and ramp ADCs can use a single bit of digital input + some pattern hold circuitry to do incredibly accurate measurements, it's just they take some time.
If you want a flash ADC that can do 16 bit (and can do 16 bit at 100Mhz), however you'll have to probably mortgage your house.
I thought I sucked at debouncing until I tried a rotary encoder that wasn’t like 10 for $4 on aliexpress. Lesson learned. Then I found out they sell a lot of decent related stuff at Micro center.
$3 ADCs are not cheap ADCs. Cheap ADCs are the one embedded inside cheap MCU. These days, TI sell cheap ADC/DAC that is artificially limited to 10/12-bit, but with linearity at more than 16-bit. The cost of producing bad ones versus good ones are not much different. They go through the same semiconductor process. So long as the fundamental design is sound, cheap ADC don’t perform any worse.
If these are genuine TI parts, I highly suspect the issue to be a measurement error by the user. Whether it's $0.60 or $4 does not matter, that's due to the order size as mentioned in the article.
> These are cheap, relatively fast, and not particularly good. While they sport impressive-sounding 12- and 16-bit readouts, the effective number of bits (ENOB) is usually around 8 or 9.
The edge case of the 1-bit conversion scheme used in SACD format is compelling from a few perspectives. The idea is to run the sampling rate in the megahertz region. SACD achieves 120dB of dynamic range with an extended frequency response up to ~100kHz. CD audio only achieves 96dB of range up to 20kHz with its 16-bit PCM scheme. From the analog hardware complexity standpoint, a bitstream converter is much simpler than a multi-bit converter. The 16-bit ADC might be cheaper due to the insane manufacturing volumes.
Trading bit depth for sample rate is a very compelling offer in many cases. The 3d graphics version of this is SSAA where you sample more pixels than your monitor needs in order to resolve higher frequency information.
Weird follow up without apologizing for smearing LCSC with no evidence whatsoever. Imagine someone writing a blog post like
"LLMs produce AI slop on regular basis. This piece of code I found somewhere on the web sure is AI slop, I never saw any code by James Bowman but _I very much think these are made with the_ AI slop generators _or similar_". And the follow up is "Remember James Bowman? _I’m still curious about_ his code. I might look at it in something or other at some undetermined point in the future to see how it performs." Does that sound even remotely ok?
Its not even guilt by association. Its guilt by gut feeling? prejudice?
25 comments
[ 2.0 ms ] story [ 47.4 ms ] threadClones tend to be vastly different - different technology node, architecture, die size, etc. - that's because they are generally functional clones, not mask clones.
(also, as a general shoutout to the low tech sandpaper technique for exploratory work, here's a sanded down RP2350 thrown under a clapped out SEM: https://object.ceph-eu.hswaw.net/q3k-personal/484e7b33dbdbd9... https://object.ceph-eu.hswaw.net/q3k-personal/3290eef9b6b9ad... )
Weirdly honest deal, haha.
The vast majority of counterfeit chips I've seen were from ghost shifts but IIRC TI fabs all their analog parts in house, I doubt they're ghost shift parts or failed QC.
I think its probably a relabeled ADS1015.
Non-linear as hell - and evil side effects once you use the calibration curves.
I don't think that's quite accurate for reasonably modern MCUs. You can typically shake 10+ bits out of them, but you need to take a lot of precautions, such as providing very stable external reference voltage and shutting down unneeded subsystems of the chip.
They're still not as good as standalone ADCs, but they're at a point where you can actually use them for 90% of things that require an ADC.
In cases where you need more bits, there's a lot more that must go into the design, which is what gives me a pause about the article. There's nothing about the PSU the author is using or how he managed the MCU noise and RFI. So I don't know if the findings here are that these are knock-off devices with worse specs, or if his overhead LED lamp is causing a lot of interference.
Of interest from early in the article: I'm curious how these external ones compare to onboard, e.g. STM32's. Btw, the TI one listed is actually pretty simple to use in comparison. The ST integrated ones have more config and hardware considerations, e.g. complicated calibration procedures, external VREF (mentioned) etc. So, if you do app the config, is the integrated one as good?
The integrated ones usually have nice ways to integrate with timers and other onboard periphs.
Does it work? Well, does your design power up during factory testing, and then pass whatever things your rig (hope you made a few!) has in mind? Well, then, yes, in fact it does...
Takes an analog signal from something like light or sound and convert it into a digital signal.
Acronyms introduced in an article should be spelled out at least once please.
On the current project we started with an MCP3208 via SPI. It did the job but only has 8 channels and it's slow (100K samples per sec).
To get something faster we switched to ADS7953. It has 16 channels and runs 10 times faster. It's somewhat more complex to code, and you can only get the highest sample rate if you scan the inputs in a predictable order. But it sure flies.
To me, these chips feel like cars. The ADS7953 is somewhat of a ferrari, whereas the MCP3208 feels like a Toyota, simple to use, unimpressive performance.
I'd love to know the industry background about how these varieties of ADC chips came to be and carved their own space in the world, and how widely they are used (millions? billions?).
Single cycle readings defeat the point of sigma delta ADC setups.
You're taking many high noise samples and averaging them over time to get a better picture of the average voltage.
Btw. the western list price is just an indicative at-most number anyway. Even a small-sized project gets discounted prices when you start talking to a sales rep.
If you want a flash ADC that can do 16 bit (and can do 16 bit at 100Mhz), however you'll have to probably mortgage your house.
The edge case of the 1-bit conversion scheme used in SACD format is compelling from a few perspectives. The idea is to run the sampling rate in the megahertz region. SACD achieves 120dB of dynamic range with an extended frequency response up to ~100kHz. CD audio only achieves 96dB of range up to 20kHz with its 16-bit PCM scheme. From the analog hardware complexity standpoint, a bitstream converter is much simpler than a multi-bit converter. The 16-bit ADC might be cheaper due to the insane manufacturing volumes.
Trading bit depth for sample rate is a very compelling offer in many cases. The 3d graphics version of this is SSAA where you sample more pixels than your monitor needs in order to resolve higher frequency information.
https://en.wikipedia.org/wiki/1-bit_DAC
https://en.wikipedia.org/wiki/Direct_Stream_Digital
There you are. Assumptions. Nothing to do with LCSC.
And here: https://news.ycombinator.com/item?id=45658163
"LLMs produce AI slop on regular basis. This piece of code I found somewhere on the web sure is AI slop, I never saw any code by James Bowman but _I very much think these are made with the_ AI slop generators _or similar_". And the follow up is "Remember James Bowman? _I’m still curious about_ his code. I might look at it in something or other at some undetermined point in the future to see how it performs." Does that sound even remotely ok?
Its not even guilt by association. Its guilt by gut feeling? prejudice?