Fried pro mini 3.3v?, what not to do

INTRO:
For some time, I was wondering what's the problem to program pro mini 3.3v (Atmega328P 3.3V @ 8 MHz) with another board (Nano and Leonardo in my case) as ISP.

Initially the pro mini 3.3v worked properly out of the box, I can upload sketches through FTDI USB adaptor.
Problem comes when I tried ISP programming with another Arduino board, as i want to avoid a bootloader, so sketch starts at mini power up.

After trying programing through Leonardo/Nano as ISP, pro mini 3.3V is dead.
By dead I mean:

  • Error while burning bootloader. avrdude: Device signature = 0x000000 avrdude: Yikes! Invalid device signature.

  • And "unable to enter ISP programming mode" in Nick Gammon's scripts

I tried identical programming process with a 5V pro mini, and is perfectly running.

I get to this thread https://forum.arduino.cc/t/minipro-3-3v-uploading-problem/902480 that would mean I possibly fried the pro mini 3.3V

QUESTION 1:
How "fried" the board is at this condition? Is it just bricked (can I revive it through high voltage programming?) or is it definitively fried?

QUESTION 2:
Provided that a pro mini 3.3V is just a pro mini 5V with different regulator and oscillator, meaning that the core Atmega328P bears with 5 volt... can I next time (new boards) program it through ISP at 5 volt by connecting 5 volt from master board to VCC @ pro mini 3.3V , so the board is all powered at 5 volt and thus avoiding the deadly condition where "No IO pins should receive voltage over VCC+0.5, so 3.3+0.5=3.8V"
NB: It seems that this deadly condition happens if you connect 5V from master ISP (running at 5V) to RAW in slave (so slave will be running at 3.3V because RAW is connected to regulator) and at the same time ISP is connected to 5V from master ISP to slave through 11, 12, 13 pins

Thanks!

Q1/Q2 i have not heard about Atmega328P-AU 3.3V chip. but about that chip can work on 3.3V if frequency is lowered to 8MHz. So if the Mini board has 8MHz oscillator then it marked as 3.3V board. If you desolder chip from board and put it to programmer e.g. TL866A/CS, you will definitively know if chip dead.
if not dead, take bootloader from Ar.IDE for Mini Pro 3.3V 8MHz and upload it. verify fuses for this setting. after that you can solder chip back on the board.
UPD. wait a second. you don't need any bootloader at all. this is a Mini, it has no UART adapter to USB, so you can use entire flash space for your program. which will be uploaded with e.g. "ArduinoAsISP".

Yes, it's a pro mini with 3.3V regulator and 8MHz oscillator, I already indicated that in the first line.

I don't have a programmer nor I'll buy it by now, plus desolder the chip, test, resolder, etc, not worth, its much cheaper to buy a new pro mini... my Q1 was in direction to anyone having suffered the problem and having the answer already, so to know if chips are recoverable and consider the expense of a programmer.

Chips are not responding to ISP programing, as indicated... call it dead (maybe), in any case, if would not be in this state, I would not be doing neither Q1 nor Q2.

Not having USB doesn't mean you wouln't have or want a bootloader, that's completely independent of having an onboard USB interface.

And regarding uploading the sketch through ISP, is what the Q2 is about, how to do it with a 5V arduino board as master, without sending the pro mini 3.3V to trash.

You burned out the Pro Mini by connecting its 3.3V logic inputs to 5V logic outputs. I'm surprised that the 5V board survived the process!

Logic level shifters are required for communications between a 5V device and a 3.3V device. For example:

Use the ISP programmer to power the Pro Mini. ISP 5V to VCC, no connection to RAW input, only possible thing that would be damaged would be the voltage regulator, but if that were so then powering from an FTDI USB adapter would presumably also damage it. (Easy enough to test if the board is still getting the regulated 3.3V).

Thats a good one, checked and yes the regulator is alive, so seems that applying voltage above the regulated at the regulator output has not destroyed it in first instance.

But thats the point, the processor in a pro mini, whether it is "named" 3.3V or 5V is a ATmega328P, that is up to 6V tolerant, it is not a 3.3V device, so programming it with 5V at all pins, included power, shouldn't be a problem.
Again, the problem seems to be when you power it at one voltage, but then send inputs plus than 0.5V over this power voltage. That situation happens if its powered through RAW... I guess thats how I fried the processor.

Batch of new pro minis 8MHz are on the way, lets see.

If you look carefully at the ATmega328 data sheet, the absolute maximum voltage on any I/O pin is Vcc+0.5 V. If Vcc is 3.3V, then do the math to see why you fried the chip.

That rule holds for Vcc=0 (powered off) as well.

1 Like

I think we are saying the same, maybe from different prospective.
The chip is powered 3.3V when 5V is connected to RAW, and the regulator enters into action.
In this situation, you can not program the chip with 5V at MOSI, MISO or whatever other pin, or it will be dead (mine at least seems to be dead).
In that case, you are forced to employ 3.3V (+0.5) at any pin to avoid the deadly situation, thus being necessary a 3.3V arduino as ISP, or logic level converter, and thats a bit of a mess as usually the arduinos we have in the drawer are 5V

PDF schematic for the pro mini

But (and thats what I want to confirm) if you put 5V to VCC, then the chip is powered at 5V and shouldn't be a problem to program it introducing 5V at any pin. Does it have sense?

This is the complete table for the ratings, with the *notice and max VCC for the ATmega family.

Some others with this understanding:

Thats a very good puntualization indeed :thinking: Thanks :wink:

I tested a lot of configurations, and so, my main conclusion is:

It is perfectly OK to program "pro-mini 3.3V/8MHz" with 5V powered at VCC/GND as ISP peripheral with NO problem, just avoid power it through RAW

In other words, RAW pin is evil for programming purpose.


As discussed, 3.3V pro minis, has a voltage regulator that pick RAW and internally puts 3.3V in the ATmega328, that means all signals to it should be 3.3V as per explained in the thread.

To ensure only 3.3V signals all arround would be tricky if you are programming with a 5V arduino as ISP programmer, you can be tempted to connect power through RAW, but the ISP interface wiring straight between ISP "controller" and "peripheral". That will kill the pro mini you are trying to programm.


My idea in this post is to not employ specific stuff that is not already on everybodies drawer, and that in case you need to buy something, you can re-employ it for other things.


My arrangements for testing:

  1. Computer - (USB cable) - USB to UART (FTDI) - (UART 6 pin wire ribbon)- pro mini as ISP programmer -(ISP 6 pin wire PICO, POCI, CS, RESET, VCC, GND)- pro mini as peripheral

  2. Computer - (USB cable) - nano 5V as ISP programmer-(ISP 6 pin wire PICO, POCI, CS, RESET, VCC, GND) - pro mini as peripheral

I tested powering both 5V and 3.3V to both "pro mini 5V/16MHz" and "pro mini 3.3V/8MHz" and all possible combinations worked flawlesly.

If you are wondering, yes, even 5V/16MHz powered with just 3.3V worked flawlesly as ISP programmer

The only case I see where you mandatory need to power and have a programmer running at 3.3V is if you have the pro mini embeded in a system where other devices like sensors, runs at maximum 3.3V.

For this case, what I did is to buy a USB to UART (FTDI) adaptor with a 3.3V/5V jumper, so you can power the thing with USB, but send only 3.3V UART signals. Again, i buyed this for the shake of having a very usefull and universal tool like this in the shed, that I can employ thousand times later for other things, like Serial connection to any pro mini I want to embed in a project, or upload sketches by serial comunication instead of ISP.


Extra info, new naming convention for ISP communication:

controller = master
peripheral = slave

Reference to your post delete, there is no request to anybody to agree, thats a forum not a political congress

This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.