| FORUM

FEDEVEL
Platform forum

two pico-imx7d SOM modules died. I don't have RMA. technexion don't talk with me

neuberfran , 02-26-2020, 01:27 PM
Hi,
First of all, I spoke with Robert Feranec on linkedin and he authorized me to post
I have two pico-imx7d modules died. https://community.nxp.com/thread/517299 and https://community.nxp.com/thread/521956 or https://community.nxp.com/thread/520887

Now, I need help to start my SOM module (pico-imx7d) without baseboard. Using digital power supply, Where I need connect wires in SOM module? https://drive.google.com/file/d/1g0h...ew?usp=sharing (datasheet at the end of the document_)


Please, help me.
robertferanec , 02-27-2020, 12:38 PM
I do not know iMX7, but if it is similar to iMX6, only the thing what you may want to try is to load uboot through USB. However to try this, I would recommend a working module, so you know you are doing everything right and only then try it on the dead modules.

If modules would not boot up from USB, then you may need to replace CPU chip - and if only CPU is damaged, then it could help, however if other things are damaged you would need to debug the board.
neuberfran , 02-27-2020, 02:59 PM
Originally posted by robertferanec
I do not know iMX7, but if it is similar to iMX6, only the thing what you may want to try is to load uboot through USB. However to try this, I would recommend a working module, so you know you are doing everything right and only then try it on the dead modules.

If modules would not boot up from USB, then you may need to replace CPU chip - and if only CPU is damaged, then it could help, however if other things are damaged you would need to debug the board.

TKS

I already changed the SOC of one of the kits and this is the one I want to solve.(ReChip) I exchanged for another SOC of a module that I bought without wifi and it also spoiled (nor charged).


But it still hasn't resolved. The worst thing is that there is nothing hot on the SOM. In the other (which I almost fixed) there were two burned capacitors This link and its video helped me: https://www.nxp.com/docs/en/application-note/AN5094.pdf LOOK to SW1ANIN and C21 and C22) I changed and came back. And stopped. There is another burnt capacitor, in another region of SOM.

As I mentioned in this post, I am interested in connecting this SOM without the baseboard. I don't have an oscilloscope to measure the crystals. But I want to turn on the SOM without a baseboard. How much voltage and amperage can I play on the SOM and how do I connect the wires? How do I know where the defect is?
robertferanec , 02-28-2020, 06:58 AM
If you had to change components around PMIC, it means there can be much more damage than just CPU. It may be a lot of work and time to try to fix it (especially with limited resources).

If you still would like to try it, start from power supply. Find what PMIC is used on the board, find reference board from that PMIC and check out power sequencing. There is often only 1 pin what you need to toggle to switch on the whole power. However, if your power supply is not switching ON automatically - that chip may be damaged or some power rails may have short circuit.

To debug a board you really may need an oscilloscope, you can not do much without a scope.
neuberfran , 02-28-2020, 08:37 AM
Originally posted by robertferanec
If you had to change components around PMIC, it means there can be much more damage than just CPU. It may be a lot of work and time to try to fix it (especially with limited resources).

If you still would like to try it, start from power supply. Find what PMIC is used on the board, find reference board from that PMIC and check out power sequencing. There is often only 1 pin what you need to toggle to switch on the whole power. However, if your power supply is not switching ON automatically - that chip may be damaged or some power rails may have short circuit.

To debug a board you really may need an oscilloscope, you can not do much without a scope.

Just to make it clear. One of the SOM I changed 2 capacitors (it worked, and died again) I'm changing another capacitor of this SOM. In the other, I already changed the SOC, but I didn't change anything else. For that last module. There's nothing hot. I want to connect this last SOM mentioned in a digital power supply. If there is no change in consumption (amperage) then I know there is a short circuit.

In this module (and the other, too) I have access to the debug port (mini-USB V8). But I can't get into the SOC via UART or serial (putting the SOC in download mode). The Power-USB-C / OTG port works. But the SOC does not go up.

I think about using an oscilloscope (dainu ads5012, buy), last.

neuberfran , 03-01-2020, 06:16 AM
How can I connect digital power supply in rt8070 (dc/dc) of the pico-imx7d SOM module?



Can I solder a wire on the rt8070 Vin? Where I can sokder a ground of the power supply? How to the best way to do this?
robertferanec , 03-02-2020, 02:54 AM
One of the SOM I changed 2 capacitors (it worked, and died again)
- I am not sure if just trying and guessing what is wrong is the right approach. Firstly, there is something wrong how the boards are used - boards should not be dying. So this is the first thing what you need to investigate - why boards are dying. Then, as I explained in my previous post, once the board is damaged, there can be more problems and that may be a reason why even if you fix something, the board will dye again.

I would buy a new board and I would be extra careful in what I would do with it.
neuberfran , 03-02-2020, 07:46 AM
Can I solde wire in Vin of the U36 (RT8070 DC/DC Converter) or Not
?

I will Use 5.15V in digital power supply and test short circuit or current leak
robertferanec , 03-03-2020, 02:22 AM
It could be 5VIN, however you should follow design guide or datasheet for baseboard - that document should explain what powers you need to provide to the module.
neuberfran , 03-03-2020, 08:37 AM
Thanks for the help so far. Without this, the solution would be even more difficult.

I'm going to call imx7d-SOM-one the one that changed two capacitors, it worked for 3 hours and stopped again.

I will call imx7d-SOM-two the one that changed the main ic. It shows no sign of life. Not 19 degrees Celsius, nothing.


The problem now lies in the difference in measurement of the resistors. I need to find them (with your help) in the schematic (pdf in the link below) and confirm the problem.

As foto01 from the link below, on the resistors marked in yellow, on the imx7d-SOM-one. the values vary from 3.3v to 2.6v on one side and the other of the resistors. In the green part of 0.748v and the red part of 0.874v.

in the imx7d-SOM-two the resistors marked in yellow do not vary from side to side of the resistors, always giving 3.3v. BUT, both in the green part, as in the red part, they give 0.000v.

That is, I need to study the schematic. How do I find the aforementioned resistors in the file pico-pi-imx7-rev-b1-37-53-2-5.pdf which is in the link below?







neuberfran , 03-08-2020, 10:10 PM
Can you help-me more, please
robertferanec , 03-09-2020, 03:21 AM
@neuberfran, I explained my thoughts in my previous posts. I would recommend to find out why the boards are dying and then I would recommend to buy a new one. I am not saying it is not possible to fix a damaged board, it may be just very time consuming and there is a chance that the board may not work (or the board may not work reliably).
neuberfran , 04-07-2020, 09:57 PM
@@robertferanec +@robertferanec I bought oscilospe daniu 5012h. How to measure crystal? I tried today and got some results with 200ms and 5ms and got some results (without connecting the negative probe(1x), because when I connected that (negative) probe(1x) the value was zeroed
neuberfran , 04-08-2020, 08:16 AM
I last month, I bought new kit pico-imx7d android things too
robertferanec , 04-10-2020, 01:41 AM
I am not sure what crystal you are measuring, but take the crystal value e.g. 25MHz, put it into calculator as 25000000 and press 1/X. That will calculate the period e.g. for 25MHz the period is 0.00000004s = 40ns, so you would need to set your scope around that value.

As I said, I am not sure what crystal are you measuring, but 5ms may be too big and you still may see something but it will not be correct.
neuberfran , 04-11-2020, 01:14 PM
@@robertferanec +@robertferanec Here is a schematic photo of the main crystal I am measuring:

robertferanec , 04-12-2020, 12:59 AM
Please, did you make the calculation (about how to setup your scope) which I mentioned in my previous post?
neuberfran , 04-13-2020, 01:01 PM
@@robertferanec + @robertferanec In the figure this is marking 32.768 KHZ.........32khz egual 0,032mhz x 1.000.000 = 32.000 x 1/x = 0,000 031 25‬.....031 x 10 (because 031 stay in second box = 310µs (microseconds)

Neste vídeo demonstramos um correto diagnóstico em manutenção e reparos de placas de smartphone. Se você ainda não é inscrito já se inscreve que estaremos n...


neuberfran , 04-13-2020, 10:06 PM
@@robertferanec + @robertferanec I came to some conclusions by analyzing (regarding) crystals:

In my number response: #10 03-03-2020 06:37AM i talked about pico-imx7d-one and pico-imx7d-two. The one
have the crystals Y1 - 32K, Y2 - 24MZH e Y4 37MHZ ok. The two have Y1 OK. but Y2 and Y4 not OK.
robertferanec , 04-14-2020, 11:37 AM
Y2 - 24MZH e Y4 37MHZ ok. The two have Y1 OK. but Y2 and Y4 not OK.
- I just would like to add, please be sure, that your scope can measure 24MHz and 37MHz. Very cheap or slow scopes may not show the correct results.
neuberfran , 04-17-2020, 12:45 PM
@@robertferanec +@robertferanec

In your answer # 04(and #06) you said: "There is often only 1 pin what you need to toggle to switch on the whole power. However, if your power supply is not switching ON automatically"

How can I find this pin in pmic? datasheet mc32p3000a1

https://drive.google.com/file/d/1t0F...ew?usp=sharing
photos -> y3 cristal is on the back of the SOM module

I have a new board pico-imx7d now. For the time being, the only difference in measurements I found was in the two resistors in the highlighted photo. On the new (working) board they measure 0.000 and on the one I'm fixing, they measure 0.378v
neuberfran , 04-19-2020, 10:26 AM
@@robertferanec +@robertferanec
I know the issue. The main processor is heating up. I sent a load (4.8v more or less) in VDD_ARM_IN (which supports a maximum of 1.5v). It was during the PMIC measurements (pins 6, 7 and 8). Although the probe was thin, it closed short when the probe was between the pins and involuntarily sent extra power to the main processor.

I need your help in order to command the resolution of the problem. Does a good reballing solve? Or just a ReChip?

Please, look images below:

​​​​​​​


robertferanec , 04-21-2020, 10:48 AM
@neuberfran I am not sure how to help. I have already explained everything in my previous posts. Personally I would not be fixing that boards - you do not know what everything is wrong on them (e.g. it can be multiple damaged chips or components).
neuberfran , 04-21-2020, 11:15 PM
@robertferanec I'm already sure what the problem is with my boards (the last one spoiled while I measured it). Can you answer me about the reballing or reChip question
to finish?
I don't have any board working here. I sacrificed my last pico-inx7d (ic main) but I found the damaged one and last board. The ic main stay hot
neuberfran , 08-06-2020, 10:02 PM
@@robertferanec I solved two imx7d-pico with reChip/BGA

.
Use our interactive Discord forum to reply or ask new questions.
Discord invite
Discord forum link (after invitation)

Didn't find what you were looking for?