VOGONS


Gravis Ultrasound Classic problems

Topic actions

First post, by c0keb0ttle

User metadata
Rank Member
Rank
Member

Hi all!

So, a couple of days ago I picked up a really dirty Gravis Ultrasound Classic rev 3.4. The previous owner thought it was working, but since he couldn't check I bought it "as-is" cheaply.

When I try to install it I run into problems. I'm using the 4.11 installer, and when the setup utility scans for resources the system just hard hangs, and I have to reset it manually. It doesn't matter what I scan - IRQ, DMA, or even the "mem test", everything results in the same hard freeze.

I should add that the CONSCAN.EXE recognizes the card at the proper adress 240h, and the setup also detects it since it doesn't throw a fit about "missing card".

Anyone want to venture a guess; is it a dead card, bad memory, or just a good ole' resource conflict? It's a bit disconcerting that it hangs when I test the memory as well... Unfortunately I don't have any extra memory chips to swap and try with.

Attachments

  • Gravis2.jpg
    Filename
    Gravis2.jpg
    File size
    2.16 MiB
    Views
    2003 views
    File comment
    Gravis Ultrasound Classic 3.4
    File license
    Fair use/fair dealing exception
Last edited by c0keb0ttle on 2019-06-01, 20:50. Edited 1 time in total.

Reply 1 of 22, by Jepael

User metadata
Rank Oldbie
Rank
Oldbie

Could be a resource conflict, jumper it to 220h and try again.
Also push the four socketed chips ever so slightly if there is some oxidation in the contacts.

Reply 2 of 22, by elianda

User metadata
Rank l33t
Rank
l33t

While gussetup comes with the installer it is not suited for serious resource testing, as it has bugs e.g. DMA test always fails.

What you do is:
Jumper the Port on the card. (Fromt he pic it looks like your card is jumpered to 240)
On the command line do a SET ULTRASND=port,irq_play,irq_rec,dma_rec,dma_play
I usually jumper to 240 and use SET ULTRASND=240,7,7,7,7
Same IRQs and DMAs for playback and record limit the card to non-full-duplex operation, but full-duplex this is a very rare usage case on a GUS.
Now run ULTRINIT, it will set the IRQ and DMAs given in the ULTRASND variable.

Checking:
Run GUSDRIVE2 which shows the set resources as well as the revision and memory detected. It detects also no memory configuration (should be 256 kB in your case).
If it runs fine, run GUSDRAM for PIO RAM Test.
if this runs fine run XTC-PLAY with a small tracker music file (however > 64 kB).
If it hangs on sample load then DMA transfers to memory fail -> check memory
If it enters the player but freezes on the first notes then playback DMA fails -> change playback DMA
If it plays the first few notes and stops then playing the playback IRQ has a conflict -> change playback IRQ
You can change the resources just by dropping back to DOS, do a different SET ULTRASND=... and run ULTRINIT.
No need to reboot. (unless it froze the whole system)

Retronn.de - Vintage Hardware Gallery, Drivers, Guides, Videos. Now with file search
Youtube Channel
FTP Server - Driver Archive and more
DVI2PCIe alignment and 2D image quality measurement tool

Reply 3 of 22, by c0keb0ttle

User metadata
Rank Member
Rank
Member

Thanks for the responses.

Unfortunately it seems I can't manage to get past ULTRINIT. It shows the ULRINIT version (2.31) and after that just a blinking cursor. The keyboard is locked up, and the only thing I can do is a hard reset.

But I haven't tried every combination of port and irqs yet.

I also tried the procedure in another, completely different, system, and the results were the same. 🙁

Reply 4 of 22, by elianda

User metadata
Rank l33t
Rank
l33t

It does not depend on IRQs/DMAs when ULTRINIT hangs, it just uses the port to set the resources.

You may as well try GUSDRIVE2 to set resources as it implements the same functionality. Maybe it reports some error?

Retronn.de - Vintage Hardware Gallery, Drivers, Guides, Videos. Now with file search
Youtube Channel
FTP Server - Driver Archive and more
DVI2PCIe alignment and 2D image quality measurement tool

Reply 5 of 22, by c0keb0ttle

User metadata
Rank Member
Rank
Member
elianda wrote:

It does not depend on IRQs/DMAs when ULTRINIT hangs, it just uses the port to set the resources.

You may as well try GUSDRIVE2 to set resources as it implements the same functionality. Maybe it reports some error?

Unfortunately it doesn't. It just shows me which resources it will set (from the SET ULTRASND), and then hangs with a blinking cursor just like ULTRINIT. 🙁

Reply 6 of 22, by h-a-l-9000

User metadata
Rank DOSBox Author
Rank
DOSBox Author

This dirt doesn't look like the usual dust.
Maybe you should clean the board with water, soap and toothbrush. After that, dry at 100°C. Remove all socketed chips before doing so.

Another thing you can try is to swap the crystal.

1+1=10

Reply 7 of 22, by c0keb0ttle

User metadata
Rank Member
Rank
Member
h-a-l-9000 wrote:

This dirt doesn't look like the usual dust.
Maybe you should clean the board with water, soap and toothbrush. After that, dry at 100°C. Remove all socketed chips before doing so.

I removed all the chips, including the big GF1 mofo. SOme of the pins (or contacts or whatever they are called) of that one weren't exactly shiny. More like dull gray surface on them.

Cleaning the chip itself was pretty easy with rubbing alcohol. It looks pretty good now.

chip1.jpg
Filename
chip1.jpg
File size
85.96 KiB
Views
1885 views
File license
Fair use/fair dealing exception
Last edited by c0keb0ttle on 2019-06-01, 20:50. Edited 2 times in total.

Reply 8 of 22, by kaputnik

User metadata
Rank Oldbie
Rank
Oldbie
c0keb0ttle wrote:
I removed all the chips, including the big GF1 mofo. SOme of the pins (or contacts or whatever they are called) of that one were […]
Show full quote
h-a-l-9000 wrote:

This dirt doesn't look like the usual dust.
Maybe you should clean the board with water, soap and toothbrush. After that, dry at 100°C. Remove all socketed chips before doing so.

I removed all the chips, including the big GF1 mofo. SOme of the pins (or contacts or whatever they are called) of that one weren't exactly shiny. More like dull gray surface on them, so I guess it's corrosion?

Cleaning the chip itself was pretty easy with rubbing alcohol and a nail file. It looks pretty good now. The chip holder on the other hand is another matter. I can't use the nail file, and a stiff brush doesn't seen to really reach into the tiny recesses where the metal connectors are...

chip1.jpg

This side is fine!

chip2.jpg

Yuck!

chip3.jpg

Yuck again! Sigh...

A fiberglass brush will do the trick. If you don't feel like waiting weeks for delivery from some Ebay seller, Elfa got 2mm ones. If I remember correctly, I saw you mention where you live in that other forum we both frequent, not too far from one of their shops 😀

Edit: link.

Reply 9 of 22, by c0keb0ttle

User metadata
Rank Member
Rank
Member
kaputnik wrote:

A fiberglass brush will do the trick. If you don't feel like waiting weeks for delivery from some Ebay seller, Elfa got 2mm ones. If I remember correctly, I saw you mention where you live in that other forum we both frequent, not too far from one of their shops 😀

Edit: link.

Thanks for the reminder! I do actually have one of those at work. It's a bit bigger than 2mm, but I'll try that one first. 😁

Had completely forgotten about that...

Reply 10 of 22, by c0keb0ttle

User metadata
Rank Member
Rank
Member

So, I spent some time with a multimeter and this card today.

My skills with a multimeter is usually limited to measuring 9V DC adapters and the like, to see if they are alive, so I'm not sure I did things correctly today.

I tried measuring frequency on the crystal. Result was a big fat zero. Not sure I did it correctly.

Checked some of the capacitors: The big orange 470nF ones seemed to be alive and well, but when I tried it on the smaller 10uF ones, I didn't get any results at all - except one, which I think showed as 2uF.

So... I wish I knew if I measured things correctly, but oh well...

Reply 11 of 22, by keenmaster486

User metadata
Rank l33t
Rank
l33t

Seems like you'd need the caps out of the circuit to properly test them.

I really think your problems are due to that horrible dirt on the chip socket. Have you tried it after cleaning yet?

World's foremost 486 enjoyer.

Reply 12 of 22, by c0keb0ttle

User metadata
Rank Member
Rank
Member
keenmaster486 wrote:

Seems like you'd need the caps out of the circuit to properly test them.

I really think your problems are due to that horrible dirt on the chip socket. Have you tried it after cleaning yet?

I tried cleaning it with some contact cleaner spray and a fiber glass brush yesterday, but the brush was a bit too big to be effective. Haven't tried the card since then, so we shall see if it did any good at all.

Reply 13 of 22, by mightylaocrahcot

User metadata
Rank Newbie
Rank
Newbie

Did you have any progress with your GUS?

I have the exact same problem with the exact same revision 3.4 card, but didn't stumble over this thread until now. I also tried in different machines with only GUS and VGA and with replacing the memory chips, but no success.

I assumed that the GF1 chip was dead, since the only references I could find online of somebody actually resolving this issue was by replacing the GF1 with a chip from another card. So I gave up for the moment.

Probably a long shot, but... since we have the same odd revision, maybe there could be a specific version of the software that we need?

Reply 14 of 22, by c0keb0ttle

User metadata
Rank Member
Rank
Member

I found a broken trace on the circuit board itself. Not sure if that causes my problem. I have my eyes on another classic GUS, and if I get it I will try to swap chips.

Edit: The trace is fine.

Last edited by c0keb0ttle on 2019-06-01, 20:48. Edited 1 time in total.

Reply 15 of 22, by Batyra

User metadata
Rank Member
Rank
Member

I have the same problem with my GUS Calssic 3.7
My other GUS 2.4 and 3.73 are working OK and 3.7 just frezes when initializing... even power buton is not workig...
It happens only on 220 on any other like 240 card is not detected. When I start setup in ultrasnd directory every test freezes system again.

I've tried to change memory, to change GAL from other card... I have not tried to change main chip.

@c0keb0ttle - cleaning or changeing chip helped??

UPDATE----
swapping chip with other GUS classic helped... co it is a chipset problem...
Now I have to seek for groken classic for spare parts...

anyone can help?

Visit my website: http://www.collection.batyra.pl

Reply 17 of 22, by Batyra

User metadata
Rank Member
Rank
Member

Interesting if there is possibility to flash this chip again... I mean to take working one and just "copy it" to this not working... of course if it's not mechanical damaged. There're no signs of any damage, no scratch no rust or anything on it...

Visit my website: http://www.collection.batyra.pl

Reply 18 of 22, by c0keb0ttle

User metadata
Rank Member
Rank
Member
Batyra wrote:

There're no signs of any damage, no scratch no rust or anything on it...

That's what makes it so annoying. 🙁

Last edited by c0keb0ttle on 2019-06-01, 20:47. Edited 1 time in total.

Reply 19 of 22, by Frasco

User metadata
Rank Member
Rank
Member

It's clear now all you gotta do is change the GF1, but here we go...
(why not? It may take years for you to find one cheaply and get away with it).

My skills with a multimeter is usually limited to measuring 9V DC adapters and the like, to

see if they are alive, so I'm not sure I did things correctly today.

Warning: I know electronic like a roach knows quantum physics, but it won't destroy
your sound card for good (and even if she worked):
Take out the card from your system and measure resistance from every electrolytic capacitor
(Horse shoe symbol in your multimeter). You don't want to see 0 (zero) Ohm - it indicates
something went wrong with the card (you don't say 😀 😀 😀 Maybe the capacitor itself, diode, etc..)

Furthermore there is a simple charge test, if you can trust it - A multimeter has only 3v
from battery and a capacitor rate vary from 6 to 100v. Anyway, I don't think it's good for you to
know about this test. You have a capacimeter after all.

As I implied, I want to learn a little bit about practical electronic, cause theorical part
is out of this world. So yeah, correct me if I am wrong, but electrolytic capacitor in the
worst scenario can even do some damage to the components on a card.

Shorted capacitor.JPG
Filename
Shorted capacitor.JPG
File size
10.04 KiB
Views
1469 views
File comment
fig A= Capacitor compensates and stabilizes tension
fig B= I don't know what it does, but no tension is not a pretty picture.
File license
Fair use/fair dealing exception

My results so far doing the tests I mentioned:
Geforce 4 MX 4000 - No video signal, just blank screen. I removed one capacitor and it worked.
ESS AudioDrive 688 Eagle - No sound output. Removed two capacitors and it came to life. Lucky ?
I don't buy that...Although I had a few fails:

Sound Blaster 16 VIBRA16C - No sound output. A 470uf capacitor was presenting 10-20uf in the
capacimeter. I replaced it with a new one to no avail. Two 47uf capacitors were a little bulged
and I replaced both with 100uf rate. It didn't work either - Speaker out still muted.
Later I compared some photos on net and realized the problem was two missing jumpers on J5.
At last! The Speaker Out is working 🤣 but men, WHAT A NOISY SOUND CARD!
Holy mother of god 😠 Maybe the 100uf capacitor is no good replacement for 47uf.
The fact is I can hear a static noisy (radio AM/FM???) in almost every game when volume is high.
I must go to some thread talking about VIBRA16C and confirm CT2960 is a terrible thing or undone
the unnecessary recap operation as I'm beginning to suspect I cannot double the UF rating of capacitors.

Back to your problem, until you declare GUS as a dud. Can't you perform a continuity test on the broken
trace ? My experience tells me it is only finishing paint scratched, just the protector damaged and
rarely a broken continuity. Our fellows nailed it. Likely you just need to replace the GF1
(get another exemplar for better chances of diagnosing your defective one).
Here I am just venturing in the multimeter. I think it is great! Be patient with my caps ramblings.
Guys here recap everything. You can be sure capacitors aren't making your card working improperly.

And although I think "dirty" in the socket chip isn't making any difference *** You should see one of
my BIOS chips *** I would like to know more about that fiber glass brush. It's like what sponge does
for metals - shinning them ? Please, tell me I can restore those tiny metals from memory slots.
Are there others applications? Thanks for any insight.