• Please review our updated Terms and Rules here

Troubleshooting a Socket 7 Motherboard

onesimus

Experienced Member
Joined
Jan 15, 2023
Messages
144
Location
Turtle Creek (Madison), West Virginia
Hey all, I just bought essentially a system without a case. It consists of a J-Mark J-656VXD motherboard with 96mb of ram and a Cyrix 6x86 processor. The system SEEMED to work before it was shipped, according to the buyer and his photos. It worked when I first started using it as well. What it would do is turn on and go through the bios screen, ram count, etc. sometimes but other times it would just give me a black screen. After this went on for a bit, it just stopped outputting video completely even with two different video cards. The jumper settings appear to be correct, I've taken out ram and tried all kinds of combinations to check that, I've tried to reset the CMOS, looked for broken solder joints, reseated the CPU, and various other things. I'm AM using an ATX to AT power supply adapter, I don't think that would have much to do with it, although I could be wrong somehow. I've tried rigging up a pc speaker by soldering one of those round ones to a connector, but I don't really know which pins would go on which wire so I haven't had any success with that.

I'm just kind of at a loss save for the CPU or motherboard going bad and unfortunately, I don't have any other socket 7 CPU's to test with it.

Do any of you guys have any ideas regarding what I should do? Do you think I could maybe get this to work? I have one of those diagnostics cards coming tomorrow, the kind that Adrian's Digital Basement uses, so hopefully that sheds some light on the situation. It's just disappointing because I've just been in a really tight spot lately and can't really afford a whole new board which may not even end up working if the problem is the CPU or something.

Any help would be GREATLY appreciated. Thanks and God bless! You guys have been really helpful to me.
 

Attachments

  • mobo2.jpg
    mobo2.jpg
    2.3 MB · Views: 12
PC speaker is almost always the outer 2 pins of a 4 pin connector. Polarity shouldn't matter.
POST card should give more information, but beep codes would really help.
Should be ok-ish for testing, but I'd get a bigger fan for that heatsink. Cyrix tended to run hot.
 
+
PC speaker is almost always the outer 2 pins of a 4 pin connector. Polarity shouldn't matter.
POST card should give more information, but beep codes would really help.
Should be ok-ish for testing, but I'd get a bigger fan for that heatsink. Cyrix tended to run hot.
Seems like there are no beep codes, left it on for a good bit without hearing anything. I will say this, I've found that a certain part of the board is getting strangely hot, the part is on a heatsink, so maybe this is normal, but if I touch it after 30 seconds of being on it's burning to the touch. I looked up the part number (D45H2A) and it says it's a PNP voltage amplifier, I'll post a picture.
 

Attachments

  • mobo.jpg
    mobo.jpg
    1.5 MB · Views: 3
If it's getting that hot, something could be shorted on a power line somewhere.
Are there tantalum caps on this board? Those could be suspect, also electrolytic caps. Perhaps those leaked and are shorting some other pins...

Have you taken the CPU out of the socket? RAM? Will it give you diagnostic beeps with either of those removed?
 
I don't believe there are any tantalum caps on the board, I don't see any of those rounded orange ones. I'm pretty new to things of this nature, so I could be wrong. The regular caps seem to look good though, not swolled in the least. Yeah I've reseatted everything that can be reseatted and rearranged the ram, tried with only one stick, etc. I tried to hook up a pc speaker but am getting no beeps.
 
Take a look in the link and check. If everything is ok, it's time to check the components alone, but this isn't possible, right?

I would suggest to remove every component, and try in this sequence:
PS and MB
PS, MB and CPU
PS, MB ,CPU and RAM.
PS, MB, CPU, RAM and VGA

Keep the RAM at minimum, look in the link possible configurations, if in doubt use different modules, you have some, you can work it out.

Remember to always keep the speaker on.
Also do not forget to visually inspect everything, even the back of the MB.


Keep posting the results and we will help you.
 
Take a look in the link and check. If everything is ok, it's time to check the components alone, but this isn't possible, right?

I would suggest to remove every component, and try in this sequence:
PS and MB
PS, MB and CPU
PS, MB ,CPU and RAM.
PS, MB, CPU, RAM and VGA

Keep the RAM at minimum, look in the link possible configurations, if in doubt use different modules, you have some, you can work it out.

Remember to always keep the speaker on.
Also do not forget to visually inspect everything, even the back of the MB.


Keep posting the results and we will help you.
Unfortunately, I have no other test bench, the only thing close would be a PS/2 model 70. Yeah I've had that link up while doing this. I wish I had the manual, but this is the closest thing to it I can find.

No dice on any of these attempts. Nothing seems to be coming out of the pc beeper ever. I have it hooked up like in the picture, so I assume it's right since most people are saying it goes on the two outer pins. Maybe I didn't do a good job at desoldering it from the other board and it's destroyed somehow? Maybe it's on the wrong pins? Maybe the board just isn't beeping at all. Unfortunately, according to my tracking number, my diagnostic card won't be here until tomorrow.

I just have no idea what else to try at this point, because I can't tell if it's the board or the cpu. I'll do another hardcore visual inspection and report back if there's anything of note.

Thanks very much for the time, by the way.
 

Attachments

  • pcspeaker.jpg
    pcspeaker.jpg
    1.4 MB · Views: 6
Do you have a multimeter? If so, check the main voltage lines on the board (eg, CPU and BIOS ROM have power?)

Try to reseat the BIOS, be careful if you're not experienced, if need help on this, just ask.
Try white eraser on the RAM pins.

I found the manual, do not think will be much of a help.
manual, jumpers, info
 
Last edited:
Do you have a multimeter? If so, check the main voltage lines on the board (eg, CPU and BIOS ROM have power?)

Try to reseat the BIOS, be careful if you're not experienced, if need help on this, just ask.
Try white eraser on the RAM pins.
I do have a multimeter, but I'm not sure how to know which lines are which. This is something I'd really love to learn how to do though and I've been wanting to for a while. I'll look into this and see what I can get accomplished. I will say that the ram pins seem pretty clean. I will reseat the bios though, that seems like a good idea.
 
Speaker looks to be wired correctly.

If you don't have it, here's the manual--
Looks like the voltage selects are set to 3.45 core and 3.52 i/o. Can you get a pic of the CPU without the heatsink?
 
Speaker looks to be wired correctly.

If you don't have it, here's the manual--
Looks like the voltage selects are set to 3.45 core and 3.52 i/o. Can you get a pic of the CPU without the heatsink?
Oh man thanks so much for the manual! That's awesome. Here is a pic of the CPU:
 

Attachments

  • cyrix.jpg
    cyrix.jpg
    2 MB · Views: 6
Misread the jumpers on the first pic. Looks like they're good.
Page 7 of the manual (9 of the PDF), JP11 should be set to 5-6, JP12 to 2-3.
Later socket 7 cpus jumped all over the place on voltage.
 
Yeah I believe all the jumpers are good. Not only have I meticulously checked the configurations but also there was the fact that it tried to work when I first got it. I'm not sure but the whole situation comes across to me like the psu fried something.
 
The VRM gets burning hot because it's a linear power regulator. Linear regulators burn the difference between the input and output as heat. The higher the voltage difference, the more heat that is generated. And if that regulator has an internal thermal cutout, it will cut power regulation if it exceeds its thermal junction temperature. You'll need to actively cool that regulator with a fan.

Cyrix CPUs of that era were known to have high power consumption and pushed the linear regulators used on motherboards of the time over the edge. I would recommend buying an AMD K6 or a Pentium and replacing that Cyrix.

I would also suggest buying a can of Deoxit Gold G5 and hosing down the contacts on the PCI edge connector on the video card, as well as the RAM. Old edge connectors tend to get a buildup of oxide that prevents proper connection to slots, which will cause no boot issues. Very common on old RAM and AGP/PCI/ISA addon cards.
 
The VRM gets burning hot because it's a linear power regulator. Linear regulators burn the difference between the input and output as heat. The higher the voltage difference, the more heat that is generated. And if that regulator has an internal thermal cutout, it will cut power regulation if it exceeds its thermal junction temperature. You'll need to actively cool that regulator with a fan.

Cyrix CPUs of that era were known to have high power consumption and pushed the linear regulators used on motherboards of the time over the edge. I would recommend buying an AMD K6 or a Pentium and replacing that Cyrix.

I would also suggest buying a can of Deoxit Gold G5 and hosing down the contacts on the PCI edge connector on the video card, as well as the RAM. Old edge connectors tend to get a buildup of oxide that prevents proper connection to slots, which will cause no boot issues. Very common on old RAM and AGP/PCI/ISA addon cards.
Yeah I do need to get some deoxit. Both PCI VGA cards work in another computer I have that's a pentium 4, so I know those are good at least. I'll try cleaning the ram pins a bit and report back if it goes well.

I also want to add to the thread that I can feel the CPU and the VGA card getting warm, so it does seem like they are getting power.
 
Last edited:
Don't know if it has anything to do with your problem but I have a HOT-591P and it admonishes you not to use SIMMs and DIMMs at the same times due to differing voltages. It looks like your board's manual "recommends" against it but it sounds like the previous owner likely had it this way for many years.
 
Don't know if it has anything to do with your problem but I have a HOT-591P and it admonishes you not to use SIMMs and DIMMs at the same times due to differing voltages. It looks like your board's manual "recommends" against it but it sounds like the previous owner likely had it this way for many years.
Yeah it comes off to me like it was someone's old "hot rod" machine or project machine. So that would be true even if they're different kinds of ram, for instance, like on this board the dimms are external cache and the simms are just regular ram. So is the manual recommending not to do this or to not mix the types in the regular ram slots. Might seem like an obvious question but I'm somewhat unfamiliar with boards with external cache and such.
 
The 168 pin DIMMs wouldn't be used as cache. In motherboards that tolerate both types of memory being installed at the same time, it most likely slows the system down. Synchronous DRAM is faster than asynchronous DRAM, which is what SIMMs are with EDO/FPM.

I would recommend removing either the DIMMs or SIMMs and keep one type of memory installed only. It'd reduce the electrical load on the memory controller, as well as reduce potential failures. With six modules, you have a much greater chance of having pins that aren't making complete contact and causing memory errors, as well as just the increased number of memory chips having more chance of being unreliable.
 
The 168 pin DIMMs wouldn't be used as cache. In motherboards that tolerate both types of memory being installed at the same time, it most likely slows the system down. Synchronous DRAM is faster than asynchronous DRAM, which is what SIMMs are with EDO/FPM.

I would recommend removing either the DIMMs or SIMMs and keep one type of memory installed only. It'd reduce the electrical load on the memory controller, as well as reduce potential failures. With six modules, you have a much greater chance of having pins that aren't making complete contact and causing memory errors, as well as just the increased number of memory chips having more chance of being unreliable.
Ahh, good to know. Thanks. Since that's the case I'll just take them out as it seems from what you're saying as there's no downside. Plus, it helps in the process of elimination at this point.
 
UPDATE:
I got the "PC Analyzer" diagnostic card. It's made the troubleshooting much more confusing at this point. Most of the time I get no code on the analyzer card, just two dashes. However, sometimes it WILL come on and I've gotten a whole host of different error codes, though there seems to be a pattern that many of these error codes involve the boards bios. When it does come on, it seemingly freezes after the ram count and I can't get into the bios menu or anything. The diag board has some LEDs on it which show the power rails, when I get no video and two dashes on the card, there is no light on for 3.3v but when it does put out video and codes, it does show a 3.3v rail.

Basically I have no idea at this point and I haven't accomplished much with the diag card besides creating a whole bunch of red herrings.
 
Back
Top