Nemo1985 wrote on 2024-08-20, 09:45:

I found the latest bios on vogons archive but I need to change the eprom since the original bios use an ST m27c256b-15f1.
What can I use to burn the latest bios version?
I have several spare W27C512, would they work? If not I could use a W27E257?

The latest BIOS for the 1542CF I know of is version 2.11, which is 32 kilobyte (equal to 256 kilobit). This will fit into the ST M27C256. If your BIOS chip is erasable, you have a UV eraser and a suitable programmer, you can re-use the original BIOS chip. On the other hand, if the original chip is a one-time programmable chip in a plastic package without an erasing window, a different chip is required. The W27E257 / W27C257 (nearly the same chip) is designed to be a drop-in replacement for 27c256 series chips, so that one can be used, no questions asked. THe W27C512 is a drop-in replacement for 27C512-style EPROMs, and uses pin 1 as address bit 15, which differs from the 27c256-style EPROMs, which use pin 1 as Vpp. For proper reading operation, the datasheet asks for Vpp to be connected to +5V. This means that a card designed for 27c256-stlye EPROMs pull address pin 15 high if you insert a 27c512-style EPROM (or a pin-compatible EEPROM like the W27c512). Having that pin reliably pulled high all the time means the chip will operate flawlessly, but only the second 32KB of that chip are getting used.

So you can use either the W27c512 or a W27E257/W27C257 as substitute. If you use the W27c512, make sure the upper half contains a the BIOS. It's considered best practice to just copy the 32KB image into both halves if you substitute a 32KB chip with a 64KB chip, which will obviously not hurt as well. Programming just the lower half will fail, though.

Furthermore, you should not mix different BIOS and microcode versions on the 1542 series controllers (or any other Adaptec controllers with multiple ROMs). The microcode (the program for the Z80 processor on the 1542CF) is stored in a separate 27c256 chip, so you should upgrade that one at the same time.

Thank you for the information.
The file I downloaded consist in 2 files, both 32kb, one called upper socket and the other lower socket. The original bios chips are ceramic so I suppose they are eprom but I lack the uv thing to erase them and I don't want to tear off the labels.
I have an eproom programmer the TL866II Plus, but it's a no go, I don't know if it's damaged or what but it can't write any of the chips I tried (both 256 and 512).
When I try to read the new eeprom I get this error:

APP Version : 12.67 Model : TL866II Plus […]
Show full quote

APP Version : 12.67 Model : TL866II Plus

Device : W27E257 @DIP28

Pins Detected Passed!

Check ID Error!

Checked ID is:0x FF FF

Wish to continue programming , please cancel [Check device ID] option!

So I uncheck the Devicd ID option, then I'm able to read the chip correctly.

But when I try to write I get this error:

APP Version : 12.67 Model : TL866II Plus […]
Show full quote

APP Version : 12.67 Model : TL866II Plus

Device : W27E257 @DIP28

Pins Detected Passed!

Erase ...Succeeded. Time : 0.203 S

Programming FLASH ...

ERROR! Address:0x000000 Buffer Data:0x55 Verify:0xFF

Programming ...Failed

That goes for every chip (256 or 512).

I tried the system self check and I get error to gnd pin number 14.

But then if I try to program a W29C020C it works like a charme.

Do you think the problem lies on the programmer or the chips (I bought them from ebay\aliexpress years ago) and as far as I remember they worked fine last time I checked them.

The 1542CF is the Flash Version of the 1542C, so there should be no need to remove any chips for an BIOS/microcode update.
The 1542A, B, and C have removable EPROMs that require erasing and new programming in an EPROM-burner device.
As written before one chip is BIOS while the other one is microcode. These two have to match in the version.

eisapc wrote on 2024-08-22, 14:20:

The 1542CF is the Flash Version of the 1542C, so there should be no need to remove any chips for an BIOS/microcode update.

No, not at all. The "F" in 1542CF is for "fast SCSI" (10 MB/s instead of 5MB/s), not for flash. None of the 1542 series is in-system upgradable, not even the lates one, the 1542CP. The P in 1542CP is for "Plug and Play", because the 1542CP supports the ISA Plug and Play protocol.

Well since the eproom burner seem problematic, until I can't solve that problem the update thing is on hold...

Nemo1985 wrote on 2024-08-20, 11:10:

Device : W27E257 @DIP28 […]
Show full quote

Device : W27E257 @DIP28

Pins Detected Passed!

Check ID Error!

Checked ID is:0x FF FF

This means your chip is most likely not an W27E257 or it is defective. As far as I know, the W27E257 and W27C257 are similar enough that the ID reading method of one chip should work on the other chip, but FF/FF sounds like ID reading does not work at all. Are you sure your chips are actual Winbond W27E257 or W27C257 chips?

Nemo1985 wrote on 2024-08-20, 11:10:

So I uncheck the Devicd ID option, then I'm able to read the chip correctly.

How do you determine you were able to read the chip correctly? Reading a ROM chip is a blind operation. You provide some address and control signals, and whatever you get on the data lines is treated as read result. There is no way for reading to fail in a way it causes an error, you might just get wrong data if the chip is defective or there are contact issues.

Nemo1985 wrote on 2024-08-20, 11:10:

Erase ...Succeeded. Time : 0.203 S

This still does not mean anything. "Erase success" just means the chip appears empty now. It might have been empty before, so it does not necessarily mean some kind of erasing actually happened right now. Furthermore, an erased EPROM just returns 0xFF on every read. If the EPROM does not respond to read cycles at all, the TL-866 would most likely also register the read result as 0xFF, and detect a "successfully erased" chip.

Nemo1985 wrote on 2024-08-20, 11:10:

ERROR! Address:0x000000 Buffer Data:0x55 Verify:0xFF

And this means that at the first time the TL-866 expected anything except 0xFF from the chip, the chip (still?) returned 0xFF.

Nemo1985 wrote on 2024-08-20, 11:10:

I tried the system self check and I get error to gnd pin number 14.

I don't know the details of the TL-866 self check method, because I don't know that programmer very well. Is the kind of check you tried to perform a check that should only verify the correct operation of the TL-866, or is it supposed to verify proper insertion of the EEPROM chip? In case of the former, that check is most likely supposed to be run without any chip inserted, and inserting an EEPROM may cause invalid results. On a 28-pin chip with standard pinout, pin 14 is supposed to be connected to GND by the programmer. If your TL-866 fails to connect that pin to ground, this explains why it does not operate properly with 28-pin chips...

Nemo1985 wrote on 2024-08-20, 11:10:

But then if I try to program a W29C020C it works like a charme.

...and this is a 32-pin chip.

Nemo1985 wrote on 2024-08-20, 11:10:

Do you think the problem lies on the programmer or the chips (I bought them from ebay\aliexpress years ago) and as far as I remember they worked fine last time I checked them.

I suggest you try to read (without any kind of ID check!) the original Adaptec ROMs, as they are known to work 28-pin chips. If you read them as entirely 0xFF, your programmer is likely defective. If you can read ASCII strings in the BIOS ROM, the TL-866 is likely OK, and the issues you are experiencing might be that the W27E257 chips you have are broken or counterfeit.

I also recommend you to check the EEPROMs you want to use for short circuits between pin 14 (GND) and pin 28 (Vcc). Depending on the construction of the TL-866, a defective chip that has a short circuit between these two pins might damage the circuit that provides GND to pin 14, although I hope the TL-866 has enough current limiting in place to prevent this from happening.

mkarcher wrote on 2024-08-22, 16:13:

[...]

Thank you for the complete answer.
The self check must be done without any eeprom installed (that's what I did).
I also tried to read the original eeprom.
I did some tests and yes the problem lies on 14 pin when "ground". I am also unable to read properly the original roms and it always claims there is a check id issue.

I tried to open the eproom reader but I can't find the issue... Since it's just a ground pin could it be such issue to read and write the 28 pin eeprom?

Anyway, I tried to reflash the firmware but obviously the issue is hardware based, therefore I opened it. The eeproom reader is composed of 2 modules, they put 2 pins to connect the two floors but they are just made to do not let people to separate it.

I took some pictures but I can't spot any obvious damage:
https://imgur.com/a/nRvCIdK

The problem is only present when the pin 14 is connected to ground, when it's tested for vcc or vpp everything works fine, though since the test stops at the first error (I think?) there may be more once this is solved.

As far as I remember the issue with pin 14 gnd was always present but I was able to read and write 32 pin chips and 28 pin chips without issues, so something has changed for sure...

I also have been able to find the schematics, which I can't understand but maybe someone can help.

Nemo1985 wrote on 2024-08-22, 16:27:

I am also unable to read properly the original roms and it always claims there is a check id issue.

If the ROM vendor / model you selected is just slightly different from the actual chip, you are supposed to get an ID check issue. That's why it is recommended to disable ID checking if you just want to read a "whatever kind of 27c256 this is" chip.

Nemo1985 wrote on 2024-08-22, 16:27:

I took some pictures but I can't spot any obvious damage:
https://imgur.com/a/nRvCIdK

As far as I remember the issue with pin 14 gnd was always present but I was able to read and write 32 pin chips and 28 pin chips without issues, so something has changed for sure...

The "pin 14 gnd" issue will prevent any 28-pin chip to be powered correctly, so I guess you misremember that 28-pin chips worked without issue. The schematics are clear enough for me: Inside the "ZIF socket" box, there is an explanation how the "GND Drivers", which are drawn as boxes labelled "Gxx" (with xx being the pin number) work: They consist of a transistor (Q4xx) that is supposed to short that pin to ground, if it receives a control current through the associated base resistor (R6xx). The transistor has the SMD marking code J3, which is an S9013 or S8050 SMD PNP transistor. The GND driver for pin 14, consisting of R614 and Q414 is driven by pin 5 of the serial-to-parallel converter (aka shift register) U24. On the last photo in your imgur album, U24 is clearly visible. The components of the GND drivers for pins 15 and 16 are right next to U24, the resistors are labelled appropriately as "R615" and "R616", although the transistors are not labelled Q415/Q416 as suggested by the schematics, but just as G15/G16. You also can easily identify G14 connected to pin 14 of the ZIF socket. The three resistors next to G14 are (from top to bottom) R612, R613 and R614 (without being labelled on the silk screen). So you should check the following:

• Do you have continuity between pin 5 of U24 and the "left" side of R614?
• Do you get around 2.2kOhm when measuring R614?
• Do you have continuity between the "right" wide of R614 and the "top left" pin of G14?
• In diode check mode on a multimeter: Do you get ~0.7V with the red lead at the "top left" pin (base) of G14 and the black lead at the "bottom left" pin (emitter) of G14?
• Also in diode check mode: Do you get ~0.7V with the red lead at the "top left" pin (base) of G14 and the black lead at the "center right" pin (collector) of G14?
• Do you get connectivity from the center "right pin" of G14 to pin 14 at the IC socket?

My guess it that G14 is broken, but I included the whole checklist, because the issue might as well be a bad solder joint. As GND for other pins works, I am confident that U24 is (mostly) OK, and unlikely to be the root cause of your issue, so the issue must be somewhere on the path from U24 to pin 14 of the ZIF socket. As the path from U24 to pin 14 is built from off-the-shelf discrete components, diagnosing and fixing the issue should be easy for someone with moderate electronics experience. Bad joints can be resoldered, broken traces can be substituted by bodge wires, and a blown transistor can be replaced. It's a bog standard transistor, so if you can get a jelly-bean transistor like a 2N2222 (US) or a BC848 (Europe) easier than you can get a S8050/S9013 transistor, they will do just as well as substitute. If it turns out that G14 actually is broken, you might also just swap G14 and G15. This will move the fault to pin 15, which is not required to be groundable for programming common EPROM chips.

mkarcher wrote on 2024-08-22, 21:09:

If the ROM vendor / model you selected is just slightly different from the actual chip, you are supposed to get an ID check issu […]
Show full quote

Nemo1985 wrote on 2024-08-22, 16:27:

I am also unable to read properly the original roms and it always claims there is a check id issue.

If the ROM vendor / model you selected is just slightly different from the actual chip, you are supposed to get an ID check issue. That's why it is recommended to disable ID checking if you just want to read a "whatever kind of 27c256 this is" chip.

Nemo1985 wrote on 2024-08-22, 16:27:

I took some pictures but I can't spot any obvious damage:
https://imgur.com/a/nRvCIdK

As far as I remember the issue with pin 14 gnd was always present but I was able to read and write 32 pin chips and 28 pin chips without issues, so something has changed for sure...

The "pin 14 gnd" issue will prevent any 28-pin chip to be powered correctly, so I guess you misremember that 28-pin chips worked without issue. The schematics are clear enough for me: Inside the "ZIF socket" box, there is an explanation how the "GND Drivers", which are drawn as boxes labelled "Gxx" (with xx being the pin number) work: They consist of a transistor (Q4xx) that is supposed to short that pin to ground, if it receives a control current through the associated base resistor (R6xx). The transistor has the SMD marking code J3, which is an S9013 or S8050 SMD PNP transistor. The GND driver for pin 14, consisting of R614 and Q414 is driven by pin 5 of the serial-to-parallel converter (aka shift register) U24. On the last photo in your imgur album, U24 is clearly visible. The components of the GND drivers for pins 15 and 16 are right next to U24, the resistors are labelled appropriately as "R615" and "R616", although the transistors are not labelled Q415/Q416 as suggested by the schematics, but just as G15/G16. You also can easily identify G14 connected to pin 14 of the ZIF socket. The three resistors next to G14 are (from top to bottom) R612, R613 and R614 (without being labelled on the silk screen). So you should check the following:

• Do you have continuity between pin 5 of U24 and the "left" side of R614?
• Do you get around 2.2kOhm when measuring R614?
• Do you have continuity between the "right" wide of R614 and the "top left" pin of G14?
• In diode check mode on a multimeter: Do you get ~0.7V with the red lead at the "top left" pin (base) of G14 and the black lead at the "bottom left" pin (emitter) of G14?
• Also in diode check mode: Do you get ~0.7V with the red lead at the "top left" pin (base) of G14 and the black lead at the "center right" pin (collector) of G14?
• Do you get connectivity from the center "right pin" of G14 to pin 14 at the IC socket?

My guess it that G14 is broken, but I included the whole checklist, because the issue might as well be a bad solder joint. As GND for other pins works, I am confident that U24 is (mostly) OK, and unlikely to be the root cause of your issue, so the issue must be somewhere on the path from U24 to pin 14 of the ZIF socket. As the path from U24 to pin 14 is built from off-the-shelf discrete components, diagnosing and fixing the issue should be easy for someone with moderate electronics experience. Bad joints can be resoldered, broken traces can be substituted by bodge wires, and a blown transistor can be replaced. It's a bog standard transistor, so if you can get a jelly-bean transistor like a 2N2222 (US) or a BC848 (Europe) easier than you can get a S8050/S9013 transistor, they will do just as well as substitute. If it turns out that G14 actually is broken, you might also just swap G14 and G15. This will move the fault to pin 15, which is not required to be groundable for programming common EPROM chips.

Thank you for the patience and the explanation. I'm sure the eeprom reader worked with 28 pins because I flashed some video bios in the past (S3 virge and SIS).
That being said I did the test you asked:
That's the actual orientation of the board:

• There is no continuity from pin 5 to the left side of r614 (there is on the right side though)
• When I measure from Pin 5 to left side I have 2,212kOhm, same when I measure from left to right side of r614
• No continuity between the R614 and any pin of g14 (the right has 2 pins, left 1)
• In diode check mode with black tip on top and red on bottom I get v2.590, while if I do it with G13 I get something around v0,7
• I get a value between v2.1 and v2.2
• I do get continuity between those two points

I think your hypothesis is right, g14 is broken.
I'm definitely not confident on desoldering such small components, but I may ask someone to do it. Do you suggest to swap G14 and G15, or should I get a new BC848 (I live in Europe)?
This one would work? https://www.ebay.com/itm/303984111153

Thank you so much for the troubleshooting.

Nemo1985 wrote on 2024-08-23, 01:10:

Thank you for the patience and the explanation. I'm sure the eeprom reader worked with 28 pins because I flashed some video bios […]
Show full quote

Thank you for the patience and the explanation. I'm sure the eeprom reader worked with 28 pins because I flashed some video bios in the past (S3 virge and SIS).
That being said I did the test you asked:
That's the actual orientation of the board:

The attachment photo_2024-08-23_02-54-09.jpg is no longer available

• There is no continuity from pin 5 to the left side of r614 (there is on the right side though)
• When I measure from Pin 5 to left side I have 2,212kOhm, same when I measure from left to right side of r614
• No continuity between the R614 and any pin of g14 (the right has 2 pins, left 1)
• In diode check mode with black tip on top and red on bottom I get v2.590, while if I do it with G13 I get something around v0,7
• I get a value between v2.1 and v2.2
• I do get continuity between those two points

I should have added another sentence about the orientation I used to describe "left", "right" and so on. They was specifically referring to the last image in your imgur album, which is 180 degrees turned to the "actual orientation" you are showing now. I will change terminology now and refer to the "actual orientation" image. So it is correct that continuity from pin 5 of U24 is to the right side of R614. R614 is at the top of the three resistors. If you get continuity from pin 5 to the bottom resistor, there is a short between R612 and R614. The resistor you probed is OK. If there is no continuity between R614 (the top resistor) and any pin of G14, there is a broken trace, or I reversed engineered the PCB wrongly. The left side of the resistors should be connected (via thin traces) to the bottom right pins of G14 (top resistor), G13 (center resistor) and G12 (bottom resistor).

The correct test points for G12, G13 and G14 are: red probe on the bottom right pin, and block probe on either other pin. You did get 0.7V on G13, showing that the test method works for the top right pin and you actually get around 0.7V. The defiation on G14 is indeed suspicious.

Nemo1985 wrote on 2024-08-23, 01:10:

I think your hypothesis is right, g14 is broken.

After looking at the schematics, it seems that G14 might break if you insert a broken 28-pin chip that has pin 28 and pin 14 shorted, or if you insert a 28-pin chip 180 degrees turned around.

Nemo1985 wrote on 2024-08-23, 01:10:

I'm definitely not confident on desoldering such small components, but I may ask someone to do it. Do you suggest to swap G14 and G15, or should I get a new BC848 (I live in Europe)?
This one would work? https://www.ebay.com/itm/303984111153

The idea of swapping the transistors was suggested a as temporary solution, in case you don't want to or can not get a replacement transistor. Also, you can swap first to try if it solves your issue, and postpone replacing the broken transistor until it is convenient. This is most notable if you could do the swap (or just remove G14 and move G15 to G14) yourself, as for a person experienced with SMD soldering, this should be completable in less than 30 minutes, so it's an easy thing to get started and verify that we are on the right track. On the other hand, as you need to ask someone else to do the soldering, I would recommend you get confident in understanding the issue(s) with your device, so all issues can be solved at once as a permanent solution. So if it turns out that G14 is the only problem (G12 and G13 behave the same way on the meter, but G14 is different, and all traces are OK, and then you get someone to swap G14, you can be quite confident you don't have to ask again for another minor fix. In your situation, arranging the work might be just as cumbersome as the work actually being performed, so it's best to get it right from the start.

If the someone you can ask to swap the transistor performs SMD soldering often, chances are that this person already bought a bag of 100 "standard NPN transistors" and can just use one of them, so you don't need to order at all. Current market price for these transistors is around 2ct/piece in small quantities, so the offer you cited has some markup over the pure component value, but the final price including national shipping and handling seems fair. And that's why people actually building stuff using these transistors often start at getting 100 of them at once, and factories buy them in reels of 2500 transistors. Furthermore, let's check the size: Those transistors exist in the "big" SOT-23 package and the "small" SOT-323 package. I tried to measure the dimensions of G14 by pixel counting and got inconclusive results. SOT-23 is around 2.9mm wide, with a pin distance (center-to-center) at the right side of 1.9mm. SOT-323 on the other hand is around 2.0mm wide with a pin distance of 1.3mm. If the transistor is SOT-23, the BC848B will fit. If the transistor is SOT-323, the BC848W would be the correct size.

mkarcher wrote on 2024-08-23, 07:25:

I should have added another sentence about the orientation I used to describe "left", "right" and so on. They was specifically r […]
Show full quote

Nemo1985 wrote on 2024-08-23, 01:10:

Thank you for the patience and the explanation. I'm sure the eeprom reader worked with 28 pins because I flashed some video bios […]
Show full quote

Thank you for the patience and the explanation. I'm sure the eeprom reader worked with 28 pins because I flashed some video bios in the past (S3 virge and SIS).
That being said I did the test you asked:
That's the actual orientation of the board:

The attachment photo_2024-08-23_02-54-09.jpg is no longer available

• There is no continuity from pin 5 to the left side of r614 (there is on the right side though)
• When I measure from Pin 5 to left side I have 2,212kOhm, same when I measure from left to right side of r614
• No continuity between the R614 and any pin of g14 (the right has 2 pins, left 1)
• In diode check mode with black tip on top and red on bottom I get v2.590, while if I do it with G13 I get something around v0,7
• I get a value between v2.1 and v2.2
• I do get continuity between those two points

I should have added another sentence about the orientation I used to describe "left", "right" and so on. They was specifically referring to the last image in your imgur album, which is 180 degrees turned to the "actual orientation" you are showing now. I will change terminology now and refer to the "actual orientation" image. So it is correct that continuity from pin 5 of U24 is to the right side of R614. R614 is at the top of the three resistors. If you get continuity from pin 5 to the bottom resistor, there is a short between R612 and R614. The resistor you probed is OK. If there is no continuity between R614 (the top resistor) and any pin of G14, there is a broken trace, or I reversed engineered the PCB wrongly. The left side of the resistors should be connected (via thin traces) to the bottom right pins of G14 (top resistor), G13 (center resistor) and G12 (bottom resistor).

The correct test points for G12, G13 and G14 are: red probe on the bottom right pin, and block probe on either other pin. You did get 0.7V on G13, showing that the test method works for the top right pin and you actually get around 0.7V. The defiation on G14 is indeed suspicious.

Nemo1985 wrote on 2024-08-23, 01:10:

I think your hypothesis is right, g14 is broken.

After looking at the schematics, it seems that G14 might break if you insert a broken 28-pin chip that has pin 28 and pin 14 shorted, or if you insert a 28-pin chip 180 degrees turned around.

Nemo1985 wrote on 2024-08-23, 01:10:

I'm definitely not confident on desoldering such small components, but I may ask someone to do it. Do you suggest to swap G14 and G15, or should I get a new BC848 (I live in Europe)?
This one would work? https://www.ebay.com/itm/303984111153

The idea of swapping the transistors was suggested a as temporary solution, in case you don't want to or can not get a replacement transistor. Also, you can swap first to try if it solves your issue, and postpone replacing the broken transistor until it is convenient. This is most notable if you could do the swap (or just remove G14 and move G15 to G14) yourself, as for a person experienced with SMD soldering, this should be completable in less than 30 minutes, so it's an easy thing to get started and verify that we are on the right track. On the other hand, as you need to ask someone else to do the soldering, I would recommend you get confident in understanding the issue(s) with your device, so all issues can be solved at once as a permanent solution. So if it turns out that G14 is the only problem (G12 and G13 behave the same way on the meter, but G14 is different, and all traces are OK, and then you get someone to swap G14, you can be quite confident you don't have to ask again for another minor fix. In your situation, arranging the work might be just as cumbersome as the work actually being performed, so it's best to get it right from the start.

If the someone you can ask to swap the transistor performs SMD soldering often, chances are that this person already bought a bag of 100 "standard NPN transistors" and can just use one of them, so you don't need to order at all. Current market price for these transistors is around 2ct/piece in small quantities, so the offer you cited has some markup over the pure component value, but the final price including national shipping and handling seems fair. And that's why people actually building stuff using these transistors often start at getting 100 of them at once, and factories buy them in reels of 2500 transistors. Furthermore, let's check the size: Those transistors exist in the "big" SOT-23 package and the "small" SOT-323 package. I tried to measure the dimensions of G14 by pixel counting and got inconclusive results. SOT-23 is around 2.9mm wide, with a pin distance (center-to-center) at the right side of 1.9mm. SOT-323 on the other hand is around 2.0mm wide with a pin distance of 1.3mm. If the transistor is SOT-23, the BC848B will fit. If the transistor is SOT-323, the BC848W would be the correct size.

Yes the picture on imgmur had the wrong rotation, I noticed when I answered your previous message, I'm sorry.
I do get continuity if I measure from pin 5 to the right side of the resistor (the trace), but I don't get continuity if I probe to the left side (so after the resistor), I think it should be ok that way?
I confirm that I get something like v0,7 from g12 and g13 while I get like v2.3 from g14 using the probes as you directed.

I definitely may have inserted the eeprom with wrong orientation or since I bought the eeproms from Aliexpress some may have been bad, either hipothesys are likely to have happened.

My experience as soldering is limited to very simple pcb and some capacitors from mb with mixed results. So yeah... I can ask to the guy who make repairs if he is willing to troubleshoot other than fix it.

Thank you very much for the help.

Nemo1985 wrote on 2024-08-23, 10:02:

I do get continuity if I measure from pin 5 to the right side of the resistor (the trace), but I don't get continuity if I probe to the left side (so after the resistor), I think it should be ok that way?

It is OK that there is no continuity from pin 5 to the left side of the resistor. That's the job of the resistor: It impedes the current flow strongly enough that the standard threshold for continuity is not reached. Usually, "continuity" in continuity testers is defined as "less than around 30 ohms", but that resistor imposes a resistance of 2200 ohms to that current.

On the other hand, the electric circuit through that resistor needs to be closed on both sides. We already established that the connection between pin 5 and R614 is OK, so no issue on that side. After the resistor, the current is supposed to flow into G14. If there is no continuity from the left side of the resistor to the bottom right pin of G14, the device is unable to work as designed. As there is a schematic that clearly shows the required connectivity, the missing connection can easily be added by a bodge wire if it turns out that the trace is damaged.

Another thing you should keep in mind, though: maybe your probe does not make good contact to the bottom right pin of G14 (possibly there is some oxidation or flux residue from factory soldering on it). This will both cause the voltage readings at G14 to be unexpectedly high (which they are), and show no connectivity from R614 to G14. You found that measuring R614 results in the correct value, so your probes do make good contact to R614. If the connection from R614 to G14 is OK, putting the probe at the left side of R614 should be the same as putting the probe at the bottom right pin of G14. So as a final test, I suggest to repeat the measurements of G14, but with the red probe connected to the left side of R614 instead of the bottom right pin of G14. If you get "OL" (or however your meter indicates "no continuity"), the trace between G14 and R614 is broken, which needs to be fixed as well. If you get 2.1V again, G14 is most likely broken, and the trace is OK. If you get 0.7V, the trace is OK, the transistor is OK, and just the measurement was giving bad results because your red probe didn't make good contact to the bottom right pin of G14.

Nemo1985 wrote on 2024-08-23, 10:02:

I can ask to the guy who make repairs if he is willing to troubleshoot other than fix it.

That's definitiely a good idea. With the schematic at hand, he should have no problems to exactly pinpoint the issue and perform the appropriate repair. It seems like the only components that might be at fault here are PCB traces, G14 or U24. All of these components are easy to diagnose and replace for experienced electronics service persons.

Sorry for the bump, the guy who I gave the stuff to fix it, did nothing in the end, so I will need to try by myself.
I don't have a calipers and I'm not sure if the transistor is SOT-23 or SOT-323, the problem is that the sot-23 are very cheap, while the sot-323 are much more expensive.
What can I do? I order the SOT-23 and if they won't fit I will need to find something else.

Nemo1985 wrote on 2024-10-08, 06:26:

What can I do? I order the SOT-23 and if they won't fit I will need to find something else.

I rechecked the image: You should be able to fit an SOT-23 transistor there. Even if it would be slightly too big, it would still cover part of the solder pads on the board. But it seems that SOT-23 is actually the correct size. Go forward with sourcing them.

mkarcher wrote on 2024-10-08, 20:56:

Nemo1985 wrote on 2024-10-08, 06:26:

What can I do? I order the SOT-23 and if they won't fit I will need to find something else.

I rechecked the image: You should be able to fit an SOT-23 transistor there. Even if it would be slightly too big, it would still cover part of the solder pads on the board. But it seems that SOT-23 is actually the correct size. Go forward with sourcing them.

Confirmed, I received the transistors, proceeded to desolder and solder (without any camera and cheap chinese equipment), I'm proud to say that the Xgecu is now back to life, changing the transistor also fixed the check id issue. I was able to flash the new bios.
Thank you very much for your help.

Nemo1985 wrote on 2024-10-23, 03:54:

Confirmed, I received the transistors, proceeded to desolder and solder (without any camera and cheap chinese equipment), I'm proud to say that the Xgecu is now back to life, changing the transistor also fixed the check id issue. I was able to flash the new bios.
Thank you very much for your help.

Thanks for your report. While you might argue that the Xgecu is not as robust as it could be (it would be better if you couldn't break transistors that easily), it is still very servicable: It seems to have a comprehensive hardware self-test that could pinpoint the failed transistor, it has these transistors as easily swappable components and there is a schematic available that helps troubleshooting. This is more than you get with average consumer equipment.

mkarcher wrote on 2024-10-23, 16:19:

Thanks for your report. While you might argue that the Xgecu is not as robust as it could be (it would be better if you couldn't break transistors that easily), it is still very servicable: It seems to have a comprehensive hardware self-test that could pinpoint the failed transistor, it has these transistors as easily swappable components and there is a schematic available that helps troubleshooting. This is more than you get with average consumer equipment.

I fully agree with you, I bought it for less than €40 5 years ago and it always served me well. Let's say they could put a safety thing to avoid this issue, I do not think that put something (a diode?) to avoid the burning of the transistor would have increased the cost so much. Do you know if they did on their latest product?

Nemo1985 wrote on 2024-10-24, 00:38:

I fully agree with you, I bought it for less than €40 5 years ago and it always served me well. Let's say they could put a safety thing to avoid this issue, I do not think that put something (a diode?) to avoid the burning of the transistor would have increased the cost so much. Do you know if they did on their latest product?

Protecting transistors that are meant to switch the power supply is not that easy without reducing the quality of the supplied power due to side effects of the current limiter. The only way to protect such transistors from getting damaged by excessive current drawn by the (E)EPROM is to make them handle higher currents than the supply can deliver. This would mean all 40 GND and all 40 supply voltage transistors need to be bigger - and thus won't fit the board as nicely as they do now, probably requiring a bigger and more expensive board as well. I don't expect newer device to have bigger transistors, but possibly they can limit the power supplied to the (E)EPROM and check for excessive current draw before applying the full operating power to the chip.