Привет, народ ,
Как я уже упоминал в заголовке, у меня есть RTX 3060 ti, у которого есть все напряжение, но оно не обнаружено, любые предложения, пожалуйста, спасибо.

биос 2 и 5 ноге есть шевеления?

Did you check PEX_RESET, LANE_0 +/-, REFCLK +/- ? Not the resistance but diode voltage drop?

Спасибо за быстрый ответ, но я не опытный ремонтник, а просто новичок, не могли бы вы объяснить более подробную информацию / изображения или схемы, еще раз спасибо.

А что вы имеете ввиду под "не обнаружено"? Через встройку смотрели? или материнка на посте 2A или 2B, останавливается. Чуть больше конкретики

Речь о падении напряжения на контактах A11, A13 и A14 на разъеме PCI-E видеокарты.

A11, A13, A14, B14, B15, B12
I think all signals should be checked from the first section to the notch and lane 0 section.

Не могли бы вы приложить фотографию или рисунок, показывающий pci-e графической карты, чтобы узнать, где находится A11, A13, A14, B14, B15, B12?
Пожалуйста, будьте терпеливы, я новичок в ремонте графических карт.
большое спасибо

https://en.wikipedia.org/wiki/PCI_Express
Scroll down...

PEX_RESET = 0L
REFCLK +/-=0,847v
LANE_0 +/-=0L

интерпретация, пожалуйста, для вышеуказанных значений, спасибо

LANE_0 is open, both pins
PEX_RESET open
REFCLK looks very good

All should read from 0.400 to 0.900V.
Check and compare LAN_0 to LANE_1, 2... to any random lane.
You should measure all to ground.

Open circuit is a typical failure of these pins directly connected to the GPU.
If this is true this card is dead and will never work again.

I have plenty of them, some with only PEX_RESET dead open.

Update :
I reversed the probe (red probe on shield black on mentioned pins ) readings changes :

PEX_RESET = 0,736v (A11) reading after reversing probes
REFCLK +/-=0,847v/0,847v (A13/A14)
LANE_0 +/-=0,395v/0,395v ( red probe on shield ) /// LANE_0 +/-=1,7v/1.7v ( black probe on shield ) ( B14/B15 )
LANE_1 +/-=0,395v/0,395v ( red probe on shield ) /// LANE_0 +/-=1,7v/1.7v ( black probe on shield )

the rest of LANES are same as LANE 0

PEX_RESET with Red on the shield
All others with Black on the shield. Test also Lane_1. Test other Lane if readings are the same as on the Lane_0 with red or black on the shield.
-----------------
OK
They are good So at least this is good.
On bottom Lane_0 input nocked out capacitors ? This is also lane 0 but INPUT with caps in series.

no knocked out caps, all the caps seems to be in their places .

i was checking LANES on another graphic cards , this is the last results for the LANES readings :

LANE_0 +/-=0,395v/0,395v ( red probe on shield ) /// LANE_0 +/-=1,7v/1.7v ( black probe on shield ) ( B14/B15 )
LANE_1 +/-=0,395v/0,395v ( red probe on shield ) /// LANE_1 +/-=1,7v/1.7v ( black probe on shield )

i tested all the DATA LANES and they gave the same readings

When the Lan_0 Outputs are similar to other random lanes it means they are good but they differ from card to card but must not be open or shorted.
Refclk is also important and it is always something like 0.8V maybe less maybe more but not open or shotred.
PEX_reset only with red to shield. In opposite it reads infinity.
If you have more dead cards how many of them have bad these pins?
Now when you have at least a good PCIe signals you can dig what else... It can be everything, from a resitor, via, trace, bios, ram to even GPU.

Now you must measure all voltage rails for shorts. It can be in resistance mode. PCIe +12, +3.3V, Molex secondary power, you have it on wikipedia where they are.

thanks a lot for your patient, just one more question "what's the thing which is preventing the gpu from being detected even if all the voltages are there ?

It can be anything. Do you have oscilloscope?

No, but I have multimeter with the option Hz .


Envoyé de mon iPhone en utilisant Tapatalk

So you can't do too much from now. You should check pulses and their shapes on inductors, look what is going on, on pins of the bios and many other things. You exhausted the multimeter capability.
Find even the cheapest digital 20MHz oscilloscope it will be better than nothing.