VOGONS


First post, by pico1180

User metadata
Rank Member
Rank
Member

Greetings everyone,

Pretty simple "problem" here. If I have memory installed in my CT3600, when I go into DIAGNOSE.EXE the "AWE synthesized music" test plays back heavily choppy and distorted music. If I take the memory out and try again, it's perfect.

I've tried several types of memory and of course ran the memory tests on all configurations and no errors are returned.

The memory slots were replaced, but I am under the impression in issue with the replacement would be detected in the memory test portion of the diagnostics.

Any ideas why AWE synthesized music would be distorted with memory installed but work fine with memory not installed?

Reply 2 of 14, by Horun

User metadata
Rank l33t++
Rank
l33t++

Are you using matched simms ? In my CT3930 manual it states 70nS or faster and the picture shows 3 chip parity simms though 9 chip will work, also most non-parity 8 chip should work though the 2 chip non-parity are known to cause issues.
2 x 1Mb, 2 x 4Mb, 2x 16Mb possible RAM combinations, mixing ram sizes or types will not function proper. I tested 2 x 256K and it also worked....

Hate posting a reply and then have to edit it because it made no sense 😁 First computer was an IBM 3270 workstation with CGA monitor. Stuff: https://archive.org/details/@horun

Reply 3 of 14, by pico1180

User metadata
Rank Member
Rank
Member

Please see attached photo. These are the two pairs I used. Both created garbled results.

Attachments

Reply 4 of 14, by JayAlien

User metadata
Rank Member
Rank
Member

I have a ct3600 although it’s not currently in a pc. I know you can run 32mb although I think only around 28mb is addressable in that case. I cannot tell the size of the first simms but I’d be interested to know if the choppy sound is still present if the total <= 16mb. Possibly you’ve already tried that.

As horun states, it should be 70ns or faster, but have you tried exactly 70ns simms? Apologies if the first pair in the picture is, but I see 60ns on the second pair.

386sx25 SBP2 2M
P75(486) SB16 8M
P133 S3 Vir DX A64g 32M
P233MMX R128Pro A64 64M
Pii400 TNT2 Live! 128M
P3-1G V5 M80 256M
P3-1.4G R8500 A1 256M
A3200 9700Pro A2 512M
X6800 X850XT A2ZS 1G
E8600 X1950XTX Xfi 2G
QX9650 3870 Xfi 2G
i7-975 GTX570 Xfi 3GB

Reply 6 of 14, by BitWrangler

User metadata
Rank l33t++
Rank
l33t++

I ain't expert on these, but it looks like you might have managed to bracket it with too slow and too new.

Unicorn herding operations are proceeding, but all the totes of hens teeth and barrels of rocking horse poop give them plenty of hiding spots.

Reply 7 of 14, by pico1180

User metadata
Rank Member
Rank
Member

This is interesting:

https://dosdays.co.uk/topics/sb_awe32_64.php#SB32

I've always been under the impression the SB32 can handle 32MB. That link says it can only handle 8MB. I know my SB32 detects and tests out 32MB, but that doesn't necessarily mean it can us 32MB. I will do some more testing with other sticks and see what I come up with.

Reply 9 of 14, by Disruptor

User metadata
Rank Oldbie
Rank
Oldbie

It can handle 2x16 MB 30 pin SIMMs, but you just can access 28 out of the 32 MB. You can go from 512K over 2M and 8M to 28/32M.
However, the SIMM holders are not designed for swapping modules too often. Be careful.

Reply 11 of 14, by jesolo

User metadata
Rank l33t
Rank
l33t

Just to add to vstrakh's post.

The top pair are 1 MB simms each. Since there are 9 chips on each module, the one is for parity but, the access time for these chips are too slow in terms of the requirements for the Sound Blaster 32.
The bottom simms should work but, there might just be an incompatibility with the sound card.

Reply 12 of 14, by pico1180

User metadata
Rank Member
Rank
Member

I can confirm, the 32MB (2x16MB) pictured above does not work. Horrible chop. Sounds like the 1x2MB doesn't work because of insufficient speed. The below setup DOES work, but it sounds like absolute hot garbage. Bad distortion and cracking/garble. Though it isn't choppy so that is something.

I would describe it as distorted, garbled and phasing in and out with high to low modulation. Apposed to when I tried the 32MB setup where there was just blurbs of sound coming in with periods of silence at random intervals.

Attachments

Reply 13 of 14, by mkarcher

User metadata
Rank l33t
Rank
l33t
pico1180 wrote on 2023-04-07, 03:08:

I would describe it as distorted, garbled and phasing in and out with high to low modulation. Apposed to when I tried the 32MB setup where there was just blurbs of sound coming in with periods of silence at random intervals.

With all the issues you are getting, please make sure that all address pins on the new RAM sockets are properly soldered. A missing address bit may cause issues as you describe. The effect of a missing address bit can vary with RAM size. I used to think that the AWE32 synthesizer test in DIAGNOSE doesn't use the sample RAM at all, but that may be wrong, it could try to upload stuff from SYNTHGS.SBK and use GS banks if available. A simple RAM test that just fills RAM with a fixed pattern and then reads back that pattern won't detect addressing errors. This could explain why the RAM test passes and the synthesis still fails.

To verify the theory of addressing errors in the RAM: Please try any real-mode DOS MPU401 application (e.g. DOSMID) with AWEUTIL. As long as there isn't anything horribly wrong, having RAM installed should make no difference when you loaded SYNTHGM.SBK (refers to ROM only), but you are expected to get issues if you use RAM-based samples, e.g. from 2MBGMGS.SBK.

Reply 14 of 14, by Ydee

User metadata
Rank Oldbie
Rank
Oldbie

I tried 2x1MB SIMM (Goldstar with parity) on the card and no problem on the AWE test, that's all I can help, unfortunately.

Attachments

  • sb32.jpg
    Filename
    sb32.jpg
    File size
    653.12 KiB
    Views
    556 views
    File license
    Fair use/fair dealing exception