VOGONS


Can this Gravis Ultrasound PnP be repaired?

Topic actions

Reply 20 of 38, by Shponglefan

User metadata
Rank l33t
Rank
l33t
ales wrote on 2022-07-19, 20:34:

The card is dead cold - none of the ICs heat up. Is there any diagnostics that I can do to get an idea of where it's failing?

Have you performed basic continuity tracing? If not, I would start there, testing the circuit paths to ensure that electricity is going where it is supposed to, and that it isn't going somewhere it shouldn't (e.g. short).

You could also try supplying power to it and then testing to see if the expected voltages are making it through the circuit.

Pentium 4 Multi-OS Build
486 DX4-100 with 6 sound cards
486 DX-33 with 5 sound cards

Reply 21 of 38, by root42

User metadata
Rank l33t
Rank
l33t
weedeewee wrote on 2022-07-19, 20:51:

Just had another look at your photo and besides the rom being reversed (pretty sure that removing it will still allow the card to work if it was working to begin with, ofcourse without synthesizer music) , it looks like the pins of the 7905 are loose.
Fixing the 7905 solder joints likely will not solve your current problem

The fact that it is stopping your computer from booting, will imply that it is somehow messing some signals on the ISA bus up. this could be as simple as one of the logic/buffer ic's or worst case the interwave chip itself.

I would also start checking the 7905, also probing at least with a multimeter the voltages. If the 7905 has gone south it might have desoldered itself by heating up, even though it should have thermal protection. But sometimes weird things happen when components fail... As the 7905 is really cheap, it's worth simply replacing it (and the other VR).

YouTube and Bonus
80486DX@33 MHz, 16 MiB RAM, Tseng ET4000 1 MiB, SnarkBarker & GUSar Lite, PC MIDI Card+X2+SC55+MT32, OSSC

Reply 22 of 38, by ales

User metadata
Rank Newbie
Rank
Newbie

Ok a quick update.
I was wrong about the card being cold. I checked the VRs as suggested - they are fine and their output voltages are in spec.
I also checked all the ICs for input voltage. I found the datasheet for the InterWave and checked all VCC pins. Everything receives 5V as expected. Some random pins on the IW also seem to exhibit some activity on them.

I then decided to pull out the 16V8H from its socket. This resulted in the computer booting!

I don't have more time to continue today, but it looks like this chip is faulty in some way. I'll try to clean the contacts but possibly it needs reprogramming or replacement. Any idea what its role is on the card and how to diagnose it?

Thanks so much again for all the support. 😀

Reply 23 of 38, by weedeewee

User metadata
Rank l33t
Rank
l33t

Can you confirm the solder joints of the 7905 are good, as they look to be lousy on the photo you posted, which could cause intermittent audio output problems.

Have you removed or rotated the ROM chip, the IW78C21M1 chip that most of us think isn't oriented correctly ?

the 16V8, I assume, has something to do with address decoding logic. Likely someone from the gusar or argus ( @shock__ ) project could chime in on what that PAL on the original GusPnP is for and if someone has the code for programming a new one. heck, from looking at the board I can't tell what it is. It might just be there to support the IDE channel

Right to repair is fundamental. You own it, you're allowed to fix it.
How To Ask Questions The Smart Way
Do not ask Why !
https://www.vogonswiki.com/index.php/Serial_port

Reply 24 of 38, by ales

User metadata
Rank Newbie
Rank
Newbie

The joints on the 7908 as well as the 7805 (there's no 7905) seem good to me. Some cracks are apparent on the 7908 front side, but on the back of the card it looks smooth.
Here are some close-up photos. Apologies for the crappy quality from my phone's camera.

The attachment IMG_20220724_140532.jpg is no longer available
The attachment IMG_20220724_141131.jpg is no longer available

I haven't rotated the ROM yet. As that is going to be a more involved job, I'm delaying that until I run out of other things to try. I haven't yet ruled out the possibility that on this particular card it was mounted this way originally, given that there's at least one other card with this orientation of the chip.

Now I'm wondering about this PAL that's preventing the computer to boot. I see two possible explanations:
- the PAL is faulty and that's why the computer refuses to boot
- the PAL is ok, but without it, it's as if the card wasn't present at all and therefore the computer boots

I'll need to do some more research. Of course, if anyone is willing to share their knowledge, I'd very much appreciate it.

Reply 25 of 38, by Tiido

User metadata
Rank l33t
Rank
l33t

Running the card with inverted ROM is going to cause damage to the ROM and/or interwave, it absolutely shouldn't be powered in such state. You have to very least remove the ROM before any further tests.

PAL is likely for IDE only (Argus has none for example) and it is likely corrupt or has wrong content altogether. Whoever worked on this card likely didn't know that the PALs are programmable and have device specific info on them and use of some random one will alawys result in trouble of some sort.

T-04YBSC, a new YMF71x based sound card & Official VOGONS thread about it
Newly made 4MB 60ns 30pin SIMMs ~
mida sa loed ? nagunii aru ei saa 😜

Reply 26 of 38, by weedeewee

User metadata
Rank l33t
Rank
l33t

@ales
Since the computer boots with the card installed.
You can try running Unisound UNISOUND - Universal ISA PnP Sound Card Driver for DOS v0.81b
The download is at the bottom of the first post. The zip file contains a UNISOUND.TXT which has the usage info.

Right to repair is fundamental. You own it, you're allowed to fix it.
How To Ask Questions The Smart Way
Do not ask Why !
https://www.vogonswiki.com/index.php/Serial_port

Reply 27 of 38, by MJay99

User metadata
Rank Member
Rank
Member

@ales: Just out of curiosity, since I ran across this thread again today: did you try unisound and did it show anything about the ROM and did you try changing its orientation?

Reply 28 of 38, by ales

User metadata
Rank Newbie
Rank
Newbie

@Mjay99 - I did try unisound and it didn't find the card. But it might have been because the motherboard doesn't have a PnP BIOS (I don't know if that's a requirement for unisound). Also, I did not yet change the ROM orientation. I actually had to put this project on hold for a while. But in the meantime, I came across another card listed on eBay which had the reversed ROM and the seller claimed that it was working. So, I'm inclined to think that there might have been a series that was produced this way for whatever reason.

I'm still interested in finding the content of the PAL and try to reflash it.

Reply 29 of 38, by MJay99

User metadata
Rank Member
Rank
Member

I've used unisound on all sorts of boards, also without a PnP BIOS and it works well.

I've also put your card's picture and mine next to each other: the routing seems identical, just the ROM is reversed.
Since you're seeing more of those, there seems to have been some production issue then and you're the proud owner of one of those.

The card itself will work without the ROM (not sure it'll do that with a reversed IC) and most software will use their own samples or load the .pat - samples. If you want to verify the ROM being present and working it's e.g. by starting unisound (and checking its output) or playing a midi-file with e.g. modmxt or ultrasound's playmidi, etc.
So, I'm pretty sure the seller on Ebay never even used the ROM and simply deducted its condition from starting things like DOOM, etc. But there might be a datapoint that the card will run with it being reversed - but that's all mostly guessing now, of course.

As for the PAL content, usually you can't read them and either need to reverse-engineer its function or find some of the original developers 😉
https://www.techtravels.org/reverse-engineering-pals/

What I could and did test for you: Removing the PAL doesn't affect detection and regular usage in trackers or games - it works perfectly fine without. I didn't follow all traces, but from shock__'s reversed schematic and also the card I routed, there's no need for a PAL between the interwave and ISA bus if you're not using the IDE functionality. There's one 74*32 and that's present on the PnP as well as the reversed version.

I also put the PAL on a reader and tried reading its content - sadly and much es expected, to no avail. Still, as far as I see it, this shouldn't affect the usage as a soundcard.

Reply 30 of 38, by weedeewee

User metadata
Rank l33t
Rank
l33t
MJay99 wrote on 2022-11-09, 21:05:

Would you know if and does the guspnp works if the 93LC66/P has invalid/no data ?

Right to repair is fundamental. You own it, you're allowed to fix it.
How To Ask Questions The Smart Way
Do not ask Why !
https://www.vogonswiki.com/index.php/Serial_port

Reply 31 of 38, by MJay99

User metadata
Rank Member
Rank
Member

Without any data, it won't get detected. With corrupted data, anything might be possible.
But in any case, that's pretty easily fixable: pnpmap.exe from e.g. shock__'s thread can load it with a working config (just run it as 'pnpmap gravnocd.rom' (which will disable the IDE): Re: Newly made Gravis Ultrasound compatible cards ... what's the degree of interest?

Reply 32 of 38, by krivulak

User metadata
Rank Member
Rank
Member

From what I have read so far and what I can reply to - pulling the PALCE16V8H chip is basically equivalent to pulling the card out of the computer. The programmable array logic is dealing with addressing of the Interwave so without address the card is virtually non-existant for the computer hence why the computer starts booting after removing the chip.
The ROM is totally rotated because there is NO WAY it could be put two different ways on a board with the same PCB revision, same traces and same VIAs. Just look at the pinout of the chip. Rotating the chip would put address pins to data bus and vice versa, not speaking of the other stuff.

The attachment 2022-11-10 23_21_55-29F800 datasheet.png is no longer available

Yes, the original is custom chip IW78C21M1 but it turns out it is substitutable with off-shelf 29F800 FLASH chip. It is very probably fried now since it got 5V to the RDY pin. That is probably the reason the card was handled so rough since it probably never worked at all. The chip needs to be replaced no matter what.
The hungarian listing with the rotated chip is kinda fishy. Just look at the botched job of installing the additional 256x4 RAM chip. That's just AWFUL!
Thinking about how it might happen - despite the all-around belief, Gravis was pretty shitty maker so it totally could be a mistake from factory. OR there could be two revision of the chips, but that would be WAY beyond awful.

Just FIY - if you decide to give up on the card, I can take it off your hands... I love collecting broken stuff and then repairing it 😉
Last time I fixed GUS Classic 2.4 and before that GUS Classic 3.7. PnP is still missing in my collection though... But if you insist on fixing it yourself, be aware that this is going to be very tough job and sooner then later you'll need a working oscilloscope, better yet logic analyser and with all respect - since you ask questions like those you already did, you probably don't have any of these at your disposal...

P.S.: I know who you bought the card from and for how much. Know the guy personally and can tell you - if he sold it, it is thoroughly dead because he knows his stuff... Also there is a reasonable suspicion on who he got it from and an educated guess on where this guy got it from. It is from Austrian recycle center. I can tell why because me and the seller of the card both buy from the same person. There is VERY big chance he outbid me on the original listing 😁

Reply 33 of 38, by MJay99

User metadata
Rank Member
Rank
Member
krivulak wrote on 2022-11-10, 22:40:

From what I have read so far and what I can reply to - pulling the PALCE16V8H chip is basically equivalent to pulling the card out of the computer.

That's something I explicitly tested: pulling the PAL did nothing to the function of the Interwave as a soundcard. It worked perfectly fine without, it probably just disables the CDROM functionality (which I didn't bother to verify since I'm never using it anyway). The IC shouldn't be dead / shorted /damaged / containing wrong logic though - so, pulling it could be a good first step in trying to revive the card.

krivulak wrote on 2022-11-10, 22:40:

Rotating the chip would put address pins to data bus and vice versa, not speaking of the other stuff.

I definitely agree there and that's something that could even disable (but not necessarily destroy) the Interwave (it's the internal memory address and databus of the interwave, not the ISA-bus, with at least two of the pins also configuring the interwave on power-up), though it would need a much more in-depth look about what line gets connected where and the possible results of that. I remember having a small routing mistake at one point (VCC and therefore a logial high to a higher address line), which actually made the Interwave not start up at all, but also didn't damage anything.

krivulak wrote on 2022-11-10, 22:40:

It is very probably fried now since it got 5V to the RDY pin.

That's something I wouldn't outright expect to be deadly. But, to see if the interwave still works, the ROM should definitely be removed, the interwave tested without it and then put back in its correct orientation (or left off, since it's not used much also).

Basically and as far as I remember, to get it detected by unisound, (and not taking into account a damaged PCB or shorts in other places) mostly just the Interwave and 7432 need to be ok and there needs to be a 96C66 eeprom with proper contents (which can be done via the pnpmap tool).

Reply 34 of 38, by krivulak

User metadata
Rank Member
Rank
Member
MJay99 wrote on 2022-11-10, 23:25:

That's something I explicitly tested: pulling the PAL did nothing to the function of the Interwave as a soundcard.
...
I definitely agree there and that's something that could even disable (but not necessarily destroy) the Interwave (it's the internal memory address and databus of the interwave, not the ISA-bus, with at least two of the pins also configuring the interwave on power-up), though it would need a much more in-depth look about what line gets connected where and the possible results of that.

Well when you have the same card in your hands it is always easier. So far I repaired only GF1 based cards so my knowledge is derived from general repairs 😀

Reply 35 of 38, by ales

User metadata
Rank Newbie
Rank
Newbie

After may months I finally got around to follow up on this. Apologies for such long delay, but it just wasn't a priority.

TLDR; Good news: the card works! Bad news: the ROM does not.

MJay99 wrote on 2022-11-09, 21:05:
I've used unisound on all sorts of boards, also without a PnP BIOS and it works well. […]
Show full quote

I've used unisound on all sorts of boards, also without a PnP BIOS and it works well.

I've also put your card's picture and mine next to each other: the routing seems identical, just the ROM is reversed.
Since you're seeing more of those, there seems to have been some production issue then and you're the proud owner of one of those.

The card itself will work without the ROM (not sure it'll do that with a reversed IC) and most software will use their own samples or load the .pat - samples. If you want to verify the ROM being present and working it's e.g. by starting unisound (and checking its output) or playing a midi-file with e.g. modmxt or ultrasound's playmidi, etc.
So, I'm pretty sure the seller on Ebay never even used the ROM and simply deducted its condition from starting things like DOOM, etc. But there might be a datapoint that the card will run with it being reversed - but that's all mostly guessing now, of course.

As for the PAL content, usually you can't read them and either need to reverse-engineer its function or find some of the original developers 😉
https://www.techtravels.org/reverse-engineering-pals/

What I could and did test for you: Removing the PAL doesn't affect detection and regular usage in trackers or games - it works perfectly fine without. I didn't follow all traces, but from shock__'s reversed schematic and also the card I routed, there's no need for a PAL between the interwave and ISA bus if you're not using the IDE functionality. There's one 74*32 and that's present on the PnP as well as the reversed version.

I also put the PAL on a reader and tried reading its content - sadly and much es expected, to no avail. Still, as far as I see it, this shouldn't affect the usage as a soundcard.

Huge thanks to you MJay99 for the detailed guide. You were spot on in every single point.

The EEPROM was busted.
I ran UNISOUND with forced detection and it found the card but couldn't identify it.

The attachment IMG_20230529_002349.jpg is no longer available

Then I used PNPMAP to read the EEPROM contents and it was all just 0xFF from top to bottom.
I proceeded to overwrite the EEPROM which worked and finally the card was operational. I validated the new EEPROM contents.

The attachment IMG_20230529_003001.jpg is no longer available

UNISOUND detected the card, but couldn't find the ROM.

The attachment IMG_20230529_003220.jpg is no longer available

I tested playback in Scream Tracker and it works there great. But when I attempted to play a MIDI file with PLAYMIDI it hung and did nothing.
So just like you said - the InterWave is good, the ROM is not working, PAL is not needed for sound operation.

I guess now I can finally try to reverse the ROM and see if it somehow survived. If not, at least the main functionality of the card still works.

The mystery of why the ROM was reversed on multiple units remains. I guess as you said, it was some production issue. We could speculate that these were supposed to be scrapped but somehow got into the wild.

Anyway, I'll try to rotate the ROM at some point, not sure how soon, and will report back how it went. Hopefully I won't mess it up too much. 😀

Thanks again to everyone who helped, and to MJay99 especially for the sharing your expert knowledge.

Reply 36 of 38, by ales

User metadata
Rank Newbie
Rank
Newbie

Correction: the EEPROM wasn't busted, it was never written. Silly me managed to forget that it was one of the parts I had to replace in the first place. 😀

Reply 37 of 38, by MJay99

User metadata
Rank Member
Rank
Member

Now that's a great step forward, thank you for the update, ales! 😀

Always great to hear an ultrasound is back amongst the living!

Btw. I've had a quick look at the PCB and it seems a reversed ROM might not cause damage to the ROM - the VCC pin is actually a N/C if reversed (which makes VCC go to an N/C pin), so at least the ROM is not powered and might be safe:

The attachment VCC.jpg is no longer available

The GND pads on the other hand do connect two different address lines together (MA2 / RLA16), so they might step on each other when the ROM gets addressed (which could maybe have adverse effects on the Interwave or the latch, but that would need an even more in-depth look at it):

The attachment GND.jpg is no longer available

So, if you want to go ahead and swap / reverse the ROM and just in case it might be of any help (but much as so often, this is just my personal opinion and others might see things differently - therefore: caveat emptor 😉:

Details

I would see a few ways to go about it:

The first one would be to get some low melting solder with which the melting temperature of all pins could be lowered (this might be a good idea in any case) - the second best option would otherwise be to use regular lead based solder and going over all pins, as that might also help in removing the IC more easily.

Then there's two major ways in my opinion:

Either heat up the whole IC from above with a hot air station (personally, I even remove all nozzles from my cheap Chinesium one, to have the air spread out as far as possible), or do the same from the underside (which is going to be less strain on the IC, but more on the PCB).

In this case, I'd probably cover up everything around the ROM in two or three layers of tin foil and fix it with kapton tape (or put some metal box around the IC) and rather let the ROM experience a bit more strain than the PCB (which is something I've actually done already, too, since I had a card where the ROM wasn't working anymore, though it ended up being a defect on the Interwave side).

When doing this, though: be very conservative while trying to move the IC - best is not trying to lift it upwards, but carefully pushing it from the sides first - only if it moves freely, it can be picked up. Otherwise you might end up damaging or tearing off pads or traces, which is much worse than baking the ROM a little longer 😀 My personal recommendation with this valuable card: try a few similar sized ICs on some old and broken PCBs first - what needs most training in my opinion is the patience to not rush things 😀

There's also a way of using a cheap heating plate from Aliexpress and heating from the underside. This has the advantage of heating the solder joints evenly and avoiding most risk of damaging pads.
BUT: these plates usually have higher temperatures than PCBs are specified for - so if they're left too long, they might damage the soldermask (or let it change color) or even damage the PCB.
Personally, I've used them with success in many occasions, but mostly to remove parts from PCBs I don't care as much about (or which were defective / broken in the first place). Therefore, with this valuable card, I wouldn't recommend this route here.

And finally (even though I personally usually rather try than bail out): if this feels like too much risk, sticking to your gut instinct and letting someone with experience do it isn't wrong either - there's only this one chance not to damage the PCB 😁

This all said, removing or swapping such an SO package is at least one of the easier things, compared to e.g. the Interwave or even smaller pitch packages.

Reply 38 of 38, by ales

User metadata
Rank Newbie
Rank
Newbie

Wow, those are some amazing tips and tutorial. Thanks a lot! 😀

Even though this seems beyond my skill level, I think I'll still try to do it myself. It's sort of part of the excitement and if successful, the satisfaction will be that much greater. I promise that I'll practice and be careful. 😁