Sleep Mode breaks on cheap Pro-Mini? Testing ATMEGA328 - is it counterfeit?

[@phillkelley7700]

As usual, very interesting and informative. Gotta love Google's placement of the mid-roll ad. You say "twenty si" then the ad runs then you say "ixteen". Talk about the fine art of splitting an "i" in half. Have to say I'm getting real tired of mid-roll real fast. I've always let pre-roll play because I reckon the content maker deserves some revenue. Ditto post-roll (unless it's just a repeat). But I'm coming to the view that Dave Jones EEVBlog has the right idea - decamp to LBRY.

[@mr.bianchirider8126]

Phill Kelley Odd, I didn’t get any mid roll ad. Geez, I feel left out !

[@Kevindarrah]

oh man I know I hate the ads... haven't made a video in a while so will have to how you can turn the mid-roll ads off

[@phillkelley7700]

Kevin Darrah maybe see kzbin.info/www/bejne/iZLNfYCJdtCenq8 - might be helpful

[@zaprodk]

They are probably not counterfeits, but wafer rejects that is picked up from the trash bin, packaged and marked and sold for a profit since you didn't pay for the wafer.

[@Kevindarrah]

thanks - never heard of this kind of thing!

[@svea2005]

"B-selection" electronic components is something I have gotten from "a friend" at times. He works at a electronics repair company and from time to time they receive dodgy components from the manufacturer that won't reach the specs and still had passed the initial control. Normally they file a complaint and receive a refund. My thought is just that someone might be picking up those components already discarded from production and makes a few bucks out of them..

[@epiendless1128]

@@Kevindarrah That certainly happens (although I doubt it in this case). I had to debug a batch of faulty professional telecoms boards destined for the BT network, in which Infineon chips were failing. We sent the chips to Infineon, and they told us they came from a failed batch and should not have been on the market. We bet purchasing got them from the grey market.

[@user-dj1hy6zc6q]

Anyone viewing this in the future, it was proven to be counterfeit.

[@nowanda2nd]

Interesting video! Would be nice to do more comparisons on other characteristics such as stability at different supply voltages.

[@Datzfunk]

Just a thought.. Recently I was working on a project using a ATmega32u2 with a dc-dc converter on board. There was some tuning that needed to be done with the converter, and it made the 5v rail very noisy and prone to dips that would reset the controller. After awhile, I noticed that the board was consuming 10mA under really no load, and the chip was getting hot. Replacing the bootloader made no difference. The chip was somehow damaged and consuming a huge amount of power. Perhaps that is the case here? Were there other identical chips that exhibited the same behavior, or was it just that one?

[@wadefallin]

Kevin, this videos demonstrates what lengths you go to in order to provide the best performance from your products. Case in point, the front door alarm you provided me in September of 2019 (11 months ago) came charged to 4.1volts. Now, just over 11 months later of that front door being opened 1,475 times, the battery in the alarm has only dropped 0.3 volts! Thanks again for the great product and peace of mind it provides.

[@Kevindarrah]

whoa! thanks!

[@EmreKonca]

Nice video and very helpful, thanks Kevin.

What instrument do you use for current measurements?

[@nomadic_rider42]

328p has probably the best potential for counterfeiting. Nice video, Kevin.

[@ScientificSolutions]

I use dozens of the cheap Pro-Mini’s in complex assembly line test fixtures and, although I have not tested the sleep mode current, I have always been concerned about the ADC performance for the very same reason. The analog features/functions are the most difficult to get right and factory fallouts must occur often (and then where do they go?). However, the only thing I have noted as an inconsistency problem between the low cost devices is the settling time after switching the ADC mux to a different channel, this time can be considerably different between devices and has caused some problems for me until I bumped the delay time ( between switching mux channels and starting a new ADC conversion) up considerably; for all I know, the prime parts may also have this issue as well. Thanks for the heads up on this sleep current issue because I do use that feature in other Atmel applications but not on a low cost Pro-Mini yet.

[@Kevindarrah]

yea I could see that also being a problem... I'll think about testing other features as well. Like even a memory test

[@ScientificSolutions]

@@Kevindarrah : Good idea, a memory test (and others?) at the upper and lower operating voltages!

[@darkobul1]

I got few LGT8F328P to test and they where unmarked. Maybe someone is selling such kind of "pin compatible" chips as genuine atmel ones.

[@MaxintRD]

I've had some remarked chips. Some people go to lots of efforts to skim a few pennies on a chip. I guess pennies add up to big numbers when it's a full reel. My seller probably didn't even know he was selling counterfeit chips from a dodgy supplier.

[@SianaGearz]

If this should be investigated, i suppose next step would be to list up potential other chips that have been cleaned up and remarked, such as Leaf Green 328p and maybe different Atmel chip variants. Perhaps it's Atmega 328 without the P so it's non Picopower. I mean if they had a bunch of extra stock of less desirable variants of the genuine Atmega chip, it would make sense to manipulate their face to sell them quicker or for more. So you could look for spec differences unrelated to current draw to try to test for them. Clock drive circuit and fuses seem promising to look at.

[@zaprodk]

The signature for the non-p is different, so Avrdude would not program it. I wonder if the signature is hardcoded in the silicon, or if it can be changed with the right equipment ...

[@Kevindarrah]

yea, I'll see if I can check that!

[@kaypope1581]

If it’s a counterfeit, it’s still amazing. I can imagine a counterfeit of an ARM chip like a blue pill since it is a ubiquitous ARM architecture. But AVR is a proprietary so how can someone replicate entire AVR silicon?

[@AdityaMehendale]

This could be a cool 'razor' to pinpoint counterfeit chips. Nice!

[@luizcassettari]

Interesting, I wonder if the 'boot_signature_byte_get' has some random bytes. A simple example below: #include #ifndef SIGRD #define SIGRD 5 #endif void setup() { Serial.begin(115200); } void loop() { Serial.println(""); for (int i = 0x00; i < 0x40; i++ ) { if ((i % 0x10) == 0) Serial.println(); byte b = boot_signature_byte_get(i); if (b < 0x10) Serial.print("0"); Serial.print(b, HEX); Serial.print(" "); } Serial.println(""); delay(5000); } I use the 'boot_signature_byte_get' to simulate a Serial Number on atmega chips, some bytes have a serial behavior. Check the library for more technical detail: github.com/ricaun/ArduinoUniqueID/

[@Kevindarrah]

ah well that is interesting! I already have a part 2 planned!

[@x10iman]

Thanks for sharing. Would be great to do more investigation on this. It’s probably a bad chip.

[@jimstand]

He sent Kevin multiple boards. It isn't a single bad chip.

[@bardenegri21]

Could very well be rejected batches that tested wrong specifically for this issue, and instead of getting thrown out they somehow got re-marked and sold on the black market.

[@pedroramirez5905]

I had the same problem with a 328P-DIP one, I changed with a new one. But it took me a long time to realize that was a problem with the chip.

[@DesertVox]

Yes, I bought 10s of so-called Atmel DIP chips, and I'm pretty sure I had some orders that were fake ones. But, still, it's amazing they can fake these chips.

[@MaxintRD]

Some time ago I was sold some fake ATtiny44A SSOP chips. Their tops were filed off and remarked. These were chips with same pin count, but totally different type. Kevins chips could be remarked LGT8F328P chips. These are not fakes, but re-engineered Chinese chips with even additional features. However, if some culprit marked these as genuine Atmels, they became fakes when they received their fake markings...

[@raymondheath7668]

Yo got me confused. You proved the possible counterfeit chip theory from the pro mini, but you didn't go the one extra step and drop one of your known good chips on the pro mini board to eliminate any possible leakagein the pro mini board itself

[@glenndoiron9317]

Gonna assume that he correctly cleaned up the board, since he already replaced it with a 2020 IC and the current was the same. Another possibility is that the IC is a recovered / waste IC, which has not had a good prior life (or was taken off the board with a blowtorch).

[@MaxintRD]

Hi Kevin, nice video. I had some couterfeit ATtiny44A chips and made a video about them. At first look (from the top) they seemed the same, but they didn't work. On close inspection it was clear that the original markings were filed off and replaced by others. The side view clearly showed a lower, uneven profile that didn't have the same roundings as the originals. Perhaps you can also have a look at the chips from the side?

[@gregwmanning]

Interesting Video thanks. Is it worth the effort ti removing parts of your testing sleep code to see what section is contributing to current saving and what part is not. May give some clues to what is happening.

[@n8twj]

I agree with zapro_dk... they could be rejects or mislabeled parts accidentally pushed onto the secondary markets to make something back on production costs. A manufacturing engineer loaded the 328 build process, but the silkscreen operator loaded up the 328P version?

[@user-xn3ym4cr6h]

what's the device that you are measuring that nano amp current with ?

[@mr3748]

Good day Mr Kevin, I apologize in advance, this comment is not about the content of this particular video, but of another video of yours... Uhm a while ago you've posted a series of videos on how to control esp8266 from anywhere in the world, in Part 1 you've used an arduino connected to an esp8266 to control the brightness of a led. Also, in Part 1 you've mentioned that you've worked on industrial scale iot projects. I am working on circuitry which pulls data from waterpumps ( waterpressure, motor current etc.) to give farmers a way of keeping an eye on their waterpumps. I am also a mechanical engineer, so I am sorted with everything up to the point where the esp8266 must communicate with a remote server... Could you perhaps point me in a direction where I can find information on how to setup security for example, you know, things that would be used on an industrial type of project

[@sebastianhamburg2245]

hello, thanks for the video. For me it is very helpful because I am looking for a board that uses very little power. What kind of device did you use to measure? I'd be very interested in that. The device was definitely not cheap. Thank you for the information.

[@giuseppe6631]

maybe the chip is an atmega328 and not atmega328p ? its almost the same but the 328 version lack some feature... no BOD disable in software and no "picopower" technonlogy. the 328p cost more.

[@rodrimunoz9110]

Reaaally really good video Kevin. Just what I was looking for right now! because I'm having trouble with this issue as well. But I would love to know whats the software that you use to plot the current! and to can I get it?. Do you power the boards from the computer? Thanks in advance and keep it up! :)

[@Kevindarrah]

thanks - that's a sweet little device I picked up from Otii - called the arc

[@ed-jf3xh]

As soon as you popped up the pictures I thought the right hand chip was counterfeit. I have dealt with a lot of chips from ebay, the majority of which are counterfeit. In most cases, the logos, on close inspection, will appear to be printed with a dot matrix format. Dead giveaway. However, have you put one of your digikey chips on the blue board? Just curious if there are any other issues with the board. If you can't wake the chip up once in the deep sleep, what is the purpose of the mode?

[@Kevindarrah]

nah, haven't tested on the pro-min board, but it's probably fine. Yea, no real point of this mode, just a quick test... all sleep modes would have higher current though

[@ThisMicrophoneSoundsCheap]

It could be a LGT8F328P instead of an Atmega. At least 6 companies are now making clones of STM32F103C8T6, which are sometimes re-labeled as STM parts and sold on the pill boards.

[@AlbusUmbrus]

Did you read the fuses? Probably this one has BOD enabled. Your code to disable BOD before sleep may take longer than required 4 clock cycles, therefore might not work.

[@Kevindarrah]

yea tested with the stock fuse settings, then burned again identical to my known good chip.... still bad

[@nisstw]

Is that chip 328pB marked as 328p? Since some functions are different. Maybe you can test them by that!

[@x10iman]

I am going to try it on some boards.

[@marcrives279]

Can you remove chip's markings with acetone?

[@Kevindarrah]

it seems to be engraved actually

[@fench1234567]

It's OPEN SOURCE! There's no such thing as "counterfeit" arduino chips or boards! At most, you can call their engineering (and the rellevant source) faulty!

[@nodirbekchoriev1763]

does external interrupt enable increase sleep current for avr ics?

[@Techn0man1ac]

What exactly did you measure the currents (program and hardware)???

[@Kevindarrah]

Otii Arc

[@chahat.charvik]

It would be really helpful if you can find out whether IC is counterfeit or sold in the market which are not quality approved.

[@spacewolfjr]

I thought you were great in Titanic

[@Kevindarrah]

ha!

[@RickTizzy]

nice one, thanks! Made me curious as well as to how your chip performs on the pro mini board... or to put it differently, how the board performs with a proper chip?

[@Kevindarrah]

yea I think the board is perfectly fine... ha, well looking at the layout it is a mess... looks like a single trace zig-zagging everywhere for the power rail, but shouldn't affect power consumption

[@amita1974]

Did you try contacting the factory and see what they have to say about this? Maybe you will be able to get an explanation... I guess that they would be interested in getting the bad chip sample from you and investigate this issue.

[@tal.klinger]

Which tools (sw+hw) do you use for the current measurements?

[@nomadic_rider42]

Looks like Oti from Qoitech. www.digikey.nl/product-detail/en/qoitech-ab/OTII-ARC-001/1916-1000-ND/8275323

[@Kevindarrah]

yep, that's it!

[@abeleski]

Did you try and wipe off the marking to see if it comes off easily? The genuines should not be able to be rubbed off easily

[@Kevindarrah]

yea, but seems engraved

[@abeleski]

@@Kevindarrah Yeah you are right, so that would suggest a fake or bad copy that somebody is reselling I would say.

[@jellyfishjelly1941]

I got burned with I2C on the clone pro mini before. I was banging my head, switching breadboards, jumpers, soldering wires but couldn't get fast i2c working reliably... Now switching to board I deemed reliable is my first debugging step.

[@Kevindarrah]

ouch!

[@ch0rd2]

Blacktopped 328 masquerading as 328p? -edit, never mind :)

[@Kevindarrah]

actually I was thinking about that too... not just 328 but something else that spoofs the device signature

[@magnusterminus4728]

Could you share the code so viewers can do the same powertest?

[@Kevindarrah]

yea good idea: github.com/krdarrah/lowPowerATMEGAtest

[@b3ngtb]

Have you spoken to Microchip ? They must be intersted and. An probably find out. They probably want the maker community to buy their stuff.

[@beargun42]

I doubt they would care. They’d only care if these chips somehow get into the regular supply chain (for example, getting fake chips when ordered through Mouser or DigiKey).

[@Kevindarrah]

yea was thinking about leaving a post on the forums

[@yogeshitaliya473]

After long time

[@BersekViking]

Looks laser engraved. Typical of grey market chips. Could be a reused chip, torn off a board, sanded down and remarked.

[@Kevindarrah]

yea you're right, does appear to be engraved

[@marcrives279]

I will be cautious about that in the future. So far, I would say it's simply an issue on a single chip.

[@HQ_Nextpcb]

Hi, Kevin All of your projects were fantastic, I would like to sponsor your video, may I know your email for contact?