What are you doing with Padauk chips you bought?

[bentyler]

I know lots of people got excited about low cost chips and how that would somehow democratize embedded development. The effort to reverse engineer the tools was impressive.
However people like me actually want to make stuff with them, and it's a PITA sourcing them. I'm willing to bet most people gave up almost immediately after programming OTP chips got old. The open source tools brought out the best from the community, but there is no compatible ICE, the chips don't have a UART for serial comms, and having a full C compiler is kind of silly for a chip this limited. Let me know if I'm wrong, but I will be shocked if someone out there has used the open source tools, burning through a bunch of chips while debugging, for more than one or two trivial designs. I don't mean to disparage anyone, but I'm just annoyed that people bought up a bunch of chips they'll never use from the distributor that's easiest to deal with.

[westfw]

it's a PITA sourcing them

I don’t think that we anticipated the the difficulties in “sourcing” them.  They seemed to be available in bulk from lcsc, and we didn’t think that that would change.


We should have known better.  It ought to take more than brief availability from a relatively unknown manufacturer to inspire any confidence in a single-source part.  :-(

[SiliconWizard]

I know lots of people got excited about low cost chips and how that would somehow democratize embedded development. (...)

Not commenting on the rest, but this part? Who in their right mind ever thought that they would democratize embedded developement?
The whole Arduino environment did democratize embedded development. The RP2040 is also doing that great. ($1 chip this powerful and easy to use, really!)

But how anyone could ever think that a couple-cents MCU, this crippled, supported by a language that is not even C, and rather poorly documented to the point of requiring reverse engineering, would "democratize" anything is beyond me. Anyway.

[thm_w]

Where is the option "Didn't buy because I knew they'd be a huge pain in the ass to program and debug".

[chickenHeadKnob]

Where is the option "Didn't buy because I knew they'd be a huge pain in the ass to program and debug".

Yes, my feeling on them. For me, a more interesting  question is what is the next lowest cost microcontroller which has reprogram-able flash and a decent tool-chain.

[james_s]

At least pre-covid there were AVR and PIC flash micros for <$1 each. I'm not building anything high volume enough to care about going any cheaper than that. If I'd wanted to build a million of something then maybe those Padauks would make sense. 

[woofy]

They are not a chip I would ever use. Just too limited in every way from technology to support.

[rooppoorali]

I am very sorry. Never heard of these microcontrollers before

[Benta]

I am very sorry. Never heard of these microcontrollers before

Neither have I.

But to the original question: as they supposedly cost 3 cents, dump them in the waste bin.

[hans]

I know lots of people got excited about low cost chips and how that would somehow democratize embedded development. (...)

But how anyone could ever think that a couple-cents MCU, this crippled, supported by a language that is not even C, and rather poorly documented to the point of requiring reverse engineering, would "democratize" anything is beyond me. Anyway.

Well going even as far as people building their own programmers and reverse engineering basic tools, it really throws me back to the PIC/AVR programmers running from a serial or parallel port with sketchy circuits. And even that started to die out 15+ years ago! I'm glad that in my student time, I could get a PICKIT for 30-40$ and get my feet wet with official manufacturer tools. And if you start now, you don't need to spend more than 10$ to get boards that talk USB, WiFi, and run 100+MHz at 1-2$ per chip. It's crazy if you think about it.

I think the enthusiasm was more about the mere existence of a 4ct chip that can be reprogrammed [once], and the potential to replace basic glue logic (like 555s or HC series logic) for very high volume applications.
Perhaps this is why the LCSC supply dried up. It's probably one of the few channels these parts were easily facing the western market, and if some full-time engineer spent some time to put a small amount of logic into a product with multi-10k's of volume per year, then it's very easy to saturate that supply bridge (in terms of depletion).

[martinribelotta]

I have worked with padaux MCUs using them as a true random number generators.

The client needed a source of true random numbers and did not want to pay for the real generators (also, shortage)

I put two padaux chips running with their internal RC oscillator and using each other as entropy source... strangely it worked very well and the tests with the NIST suite gave excellent results...

For 0.06 USD it was a great little project

[PCB.Wiz]

 I noticed  LCSC have these parts in stock

https://lcsc.com/search?q=Nyquest

https://www.nyquest.com.tw/en/product/HomeAppliances/NY8_Series/download

They are in a similar slot, to Padauk parts, 14b opcodes and 8 level stack,  and OTP with HV Vpp in SO8 and SO14 and SO16  - from Taiwan, so the English support looks good.
They also look to have a "C compiler " and some MTP ones are coming.

Some have 12b ADC and can support crystals - which is rare for bottom-feeder MCUs, so they seem well suited to simple peripheral slaves and timer tasks, around a larger FLASH MCU.

They are similar price to the venerable HC4060, and cheaper than more niche parts like HC5555 or HC6323  and cheaper than Oscillator/dividers like Nisshinbo NJU63xx series

[Doctorandus_P]

I never understood the craze about the Padauk controllers.

Why would it even matter if a uC is 3ct or 30ct, for production runs of less than 10.000 or so.
Both prices are negligible once you've put it on a PCB, added a power supply and put it in some enclosure.

OTP is a real nuisance during  developing too, and that alone has dismissed them for me.
Both lack of a uart, and a very small family are other show stoppers.
It's just not worth learning the tools for that (for me at least).

I also wonder what other small uC's cost, once you go in the millions of parts and negotiate directly with the manufacturer. Even my toothbrush has an MPS430 in it, which is grossly overkil, but apparently it's cheap enough for the toothbrush manufacturer.

But I do find the question interesting. I'd also be curious about people actually using these things in an application where it is a good choice. But both the "price don't matter much for <10k units and the limited hardware would make it difficult to find such an application.

I'm guessing this mostly leaves the people who do it "just because they can". Just like the (quite a lot) people into retro computing and still designing Z80 computers.

[woofy]

I'm guessing this mostly leaves the people who do it "just because they can". Just like the (quite a lot) people into retro computing and still designing Z80 computers.

 I recently went on a nostalgia trip and designed a Z80 computer. Still have no intention of using Padauk chips though.   

[ebclr]

It's just the challenge, Nobody uses that  for real products, have much better alternatives who do more and simpler,  for example, the CH anything line that is also cheap and uses conventional western tools a good example is CH552 who has USB and do a lot more than any Padauk, with less than double the price, But still extremely cheap, and if you consider the cost of development, for sure is a much better alternative for problems that you may thing on padauk,  For the more complex problems is another history, but the reference was padauk,

[james_s]

I never understood the craze about the Padauk controllers.

Why would it even matter if a uC is 3ct or 30ct, for production runs of less than 10.000 or so.
Both prices are negligible once you've put it on a PCB, added a power supply and put it in some enclosure.

OTP is a real nuisance during  developing too, and that alone has dismissed them for me.
Both lack of a uart, and a very small family are other show stoppers.
It's just not worth learning the tools for that (for me at least).

I also wonder what other small uC's cost, once you go in the millions of parts and negotiate directly with the manufacturer. Even my toothbrush has an MPS430 in it, which is grossly overkil, but apparently it's cheap enough for the toothbrush manufacturer.

But I do find the question interesting. I'd also be curious about people actually using these things in an application where it is a good choice. But both the "price don't matter much for <10k units and the limited hardware would make it difficult to find such an application.

I'm guessing this mostly leaves the people who do it "just because they can". Just like the (quite a lot) people into retro computing and still designing Z80 computers.

It really doesn't. 10k is a very small production run in some products though, if you're making say 5M of something then a few pennies really adds up. That is assuming you can get the required large quantity of parts though.

[true]

I've made a couple devices with runs >100 units that have used these chips.

For people saying they're limited - sure. they are. For my uses, they've done exactly what I've needed and with fairly low power consumption.

Still have thousands of units on hand.

Haven't used the open source tools. I wrote assembly and even the weird Mini-C stuff and used the vendor ICE to debug.

If you don't know what you want to do, or can't deal with the limitations, then these aren't for you. There are ~$0.50 >100MHz RISC-V MCUs out there now. The cheap $0.25 range 8-bitters are slowly coming back in stock, like CH55x series.

I used them as they were interesting to me, and I like to cost optimize even when I don't need to. OTP and no chip markings was actually a positive factor in using the parts.

[westfw]

how anyone could ever think that a couple-cents MCU, this crippled, supported by a language that is not even C, and rather poorly documented to the point of requiring reverse engineering

You're kidding, right?

• They're not particularly more "crippled" than the 8bit PICs.
• The only thing requiring reverse engineering was the hardware development tools.  Which is about the same as any number of other last-generation microcontrollers.  It's only relatively recently that standardized and/or open source programming tools have become common.
• And their (free!) Tiny-C isn't so awful, either.  It's closer to being a "high level assembler" than many "languages", but ... that's sort-of nice, coming from a vendor.

I didn't get any in the first round, because I clearly didn't want a full reel, nor did I want to shell out $$$ for the HW tools.  Now, the supply chain and apparent obsolescence issues are more of a concern that the technical issues.

But if anyone is seriously ready to throw theirs out, I'll take them!

(I perceive that somehow the Chinese market is less ... sensitive to some types of risk ... than western markets.  Here, we'd say stuff like "We're not going to build a chip with a generations-old architecture that is really only good for blinking LEDs in toys; that's stupid!"  And THEY say "blinking LEDs are cool.  We'll make that chip and sell it to everyone who wants to make a $0.50 blinking toy!  And if no one buys it, we'll stop production and make something else."  I dunno how it works; I should probably actually read Bunny's book that I bought a while ago.)

[KaeTo]

I am sad to read, that so many have a bad opinion of this chip. I build a few of this reverve engeneered open souce programmers and the PFS154 became my new go to uC. I had so many projects in the past where I was hesitating to use an uC that costs 1-2€ just for a simple task (where the uC is overpowered for). Now I use the padauks nearly anywhere. For my chistmas lighting I needed 20 individual PWM outputs. I used 5 PFS154 with an serial data input. I build many lamps with capacative touch sernsors, consisting of an PFS154 (PMC150 also possible) an an 1MOhm resistor. I also build capacitive moist sensor with this uC.
I just love it. There are so many small tasks, where normaly a uC would be to much, but now I have an easy programmable 5ct solution.

[tim_]

I don't get all the hate. There is far more to the Padauk MCUs than them being "cheap" and "OTP".

There are zillions of PIC clones out there, but if you take a close look at the Padauk ISA, it is actually a very carefully crafted expansions of the original concept. In addition there are variants that offer simultaneous multithreading - a concept not used on many other devices.

The real problem is the availability, obviously. The interesting MTP types are almost impossible to buy at LCSC now, and if something is available it's a new and uncommon type. Also, the even more interesting variants with 2 FPPA have never even been offered.

There are plenty of other sub $0.10 cent MCUs like Nyquest, FMD, SHC etc. But all of them are just bland OTP PIC clones.

[TomS_]

I am sad to read, that so many have a bad opinion of this chip.

Don't be. Opinions are like a-holes, everyone's got one.

If it works for you in your application, and gets the job done, that is *all that matters*.

Who cares if it's not ARM based, or some other ISA (except the fanboys). It just needs to do what it needs to do, and it could be half as powerful as a 555 timer if that is all that was needed...

[brucehoult]

Where is the option "Didn't buy because I knew they'd be a huge pain in the ass to program and debug".

Yes, my feeling on them. For me, a more interesting  question is what is the next lowest cost microcontroller which has reprogram-able flash and a decent tool-chain.

There was this tweet recently:  https://twitter.com/Patrick_RISCV/status/1580384430996484101

48 MHz 32 bit RISC-V with 16k flash, 2k SRAM for "under $0.10.

Supposed to be on LCSC, but the only thing I see so far is a $13 eval board.

https://www.lcsc.com/product-detail/Microcontroller-Units-MCUs-MPUs-SOCs_WCH-Jiangsu-Qin-Heng-CH32V003F4P6-EVT-R0_C5187532.html

[SiliconWizard]

Just look at prices for the lower-end Nuvoton MCUs (Cortex-M0) on LCSC.

[bidrohini]

Never bought any.

[RazaHosyain]

ok