Trouble with Reply Vibra-16 MCA soundcard
Peterwendt
Got a Reply Vibra-16 MCA soundcard recently (ID:5138) and this is a pretty
tricky beast as it seems.
A) the Vibra-16 chipset was originally intended for ISA-cards only
B) the card requires a Low 8-bit and a High 16-bit DMA ... from which using the
high DMA either wont' work at all or crashes the machine
(DOS-based DIAGNOSTIC does not even play 16-bit testing WAV)
C) If you use the "map high DMA to low DMA" feature the card disqualifies for
Windows, since Win does not allow to set both DMAs to the same ressource.
So far I have managed to get the card going under OS/2 Warp4 on a 9595A with a
Pentium 120.
The CONFIG.SYS lines read
DEVICE=C:\MMOS2\SB16D2.SYS /C:1 /D:1 /H:1 /I:5 /A:220 /B:16 /M:330 /T:6
/N:SBAUD1$ /P
DEVICE=C:\MMOS2\AUDIOVDD.SYS SBAUD1$
The parameters are
C = card number
D = low DMA
H = high DMA
I = card IRQ
A = base I/O
B = DMA buffers (ommitting this defaults to B:8)
M = MPU-301 I/O adress for MIDI
T = card type (MCA = 6 I would say)
N = device sound queue
P = uh ... forgot about that
The Parameter /Q to "quiet output" the driver at startup is not used, because I
wanted to see the settings when OS/2 pulls up.
Anyone found a workaround to get this card going under Windows ?
Results so far:
- either crashes the machine with a MCA Bus Timeout (using IRQ 10 and 2
different DMAs)
- starts, but garbles the picture when playing the first tune after network
login (green vertical stripes in the top of the screen) then hangs.
- playing MIDI files, but no WAVs.
I used all drivers I could get: the ones from Win95 as well as the SB16 drivers
and the "specific" Vibra-16 from Creative. The latter ones are intended for ISA
only as it seems. Get strange error messages (empty error field with [OK]
button only etc.)
Very friendly greetings from Peter in Germany
http://members.aol.com/mcapage0/mcaindex.htm
Tim Clarke
Try using the SBP2D2 (SBPro [MCV] V2) driver (only one DMA, Type-4 [/T:4]
adapter, IIRC)?
Tim
Harald Pollack
an All:
Servus Peterwendt!
P> Got a Reply Vibra-16 MCA soundcard recently (ID:5138) and
P> this is a pretty tricky beast as it seems.
P> A) the Vibra-16 chipset was originally intended for
P> ISA-cards only
There are at least 2 differnt Vibra-16 Chipsets (16C and 16X)
P> B) the card requires a Low 8-bit and a High 16-bit DMA ...
will be the 16C chipset
P> from which using the high DMA either wont' work at all or crashes the
P> machine (DOS-based DIAGNOSTIC does not even play 16-bit testing WAV)
Do not mix up 16-bit DMA with 16-bit WAV :-)
The two DMA channels are used for playback and recording (one for each)
P> C) If you use the "map high DMA to low DMA" feature the card
P> disqualifies for Windows, since Win does not allow to set both DMAs to the
P> same ressource.
Why the same resource? Use 1 and 3 (both are low DMA).
P> So far I have managed to get the card going under OS/2 Warp4
P> on a 9595A with a Pentium 120.
P> The CONFIG.SYS lines read
P> DEVICE=C:\MMOS2\SB16D2.SYS /C:1 /D:1 /H:1 /I:5 /A:220 /B:16
P> /M:330 /T:6 /N:SBAUD1$ /P
You are use DMA channel 1 for both ???
P> P = uh ... forgot about that
ISA PnP (device driver will do 'isolation/configuration' according CONFIG.SYS
values)
P> Anyone found a workaround to get this card going under
P> Windows ?
Well, first of all, verify WHICH Vibra-16 chip you will have (is printed on the
chip)
Here the different setting for C and X (OS/2, but the same in Linux) for the
ISA PnP card
Rem =========================== SB 16 PnP Vibra16C =========================
DEVICE=E:\MMOS2\SB16D2.SYS /C:1 /D:1 /H:5 /I:10 /A:240 /M:330 /N:SBAUD1$ /P
Rem =========================== SB 16 PnP Vibra16X =========================
DEVICE=E:\MMOS2\SB16D2.SYS /C:1 /D:1 /H:3 /I:10 /A:240 /M:330 /N:SBAUD1$ /P
Only diferrence: 16C has both low and high DMA, 16X has twice low DMA (but two
different channels of course!)
Herzliche Gruesse, Harald
-+- Message created on Thursday January, 04 2001 13:22:37 MEZ
Peterwendt
All the time I see your postings I remember that I'd wanted to send you an AVM
B1 MCA "for evaluation" .... and forget it a moment later again. I grew old.
But the year just started :-)
>There are at least 2 differnt Vibra-16 Chipsets (16C and 16X)
>
> P> B) the card requires a Low 8-bit and a High 16-bit DMA ...
>
>will be the 16C chipset
Chip says:
VIBRA-16
(C) Creative Tech '94
CT2501-TBQ
Hmmm.
> P> from which using the high DMA either wont' work at all or crashes the
> P> machine (DOS-based DIAGNOSTIC does not even play 16-bit testing WAV)
>
>Do not mix up 16-bit DMA with 16-bit WAV :-)
Improper expression. What I *wanted* to say was, that the card (in Diags)
chokes on testing the 16-bit DMA already and plays none of the testing-sounds.
Interestingly they say "8 bit Test" .. and "16-bit Test" during that :-)
>The two DMA channels are used for playback and recording (one for each)
That's the often-discussed "Full Duplex Capability". Which the card has. And
which makes things difficult.
> P> C) If you use the "map high DMA to low DMA" feature the card
> P> disqualifies for Windows, since Win does not allow to set both DMAs to the
> P> same ressource.
>
>Why the same resource? Use 1 and 3 (both are low DMA).
I would have done that - if the driver settings support it. And if Windows
would accept it. Neither of that is true.
> P> DEVICE=C:\MMOS2\SB16D2.SYS /C:1 /D:1 /H:1 /I:5 /A:220 /B:16
> P> /M:330 /T:6 /N:SBAUD1$ /P
>
>You are use DMA channel 1 for both ???
Yep. A) it is supported in the ADF and B) demonstrates the superiority of MCA
machines running with OS/2. Funny, eh ?
>Here the different setting for C and X (OS/2, but the same in Linux) for the
>ISA PnP card
>
>Rem =========================== SB 16 PnP Vibra16C >=========================
>DEVICE=E:\MMOS2\SB16D2.SYS /C:1 /D:1 /H:5 /I:10 /A:240
>/M:330 N:SBAUD1$ /P
I hadn't got it going with two differerent DMA channels .... "MCA Bus Timeout".
Unfunny.
But Thanks. I'll try some more later.
Brad Parker
but I am running Win NT on those boxes :-)
Seems winblows 9x has problems with it, eh? I have heard from another
party that is having similar problems.
I am using the stock NT soundblaster drivers.
I'll check my resources when I get home tonite and post what I'm using,
but I am certainly using two different DMAs for full duplex operation.
Brad
--
Bradley M. Parker Systems Programmer
Center for Computer-Aided Design Computing/Simulation Services
The University of Iowa E-Mail: b_pa...@ccad.uiowa.edu
234 ERF http://www.ccad.uiowa.edu
Iowa City, IA 52242 Phone: (319) 335-5723
T. Sunagawa
If you use stock Win95 SB driver for your Reply Vibra 16, your machine will hang
with "MCA BUS timeout error" or video will be crashed showing odd stripes as
you described.
WIN95SND.EXE, Win95 driver for Reply Vibra 16, was once stored at neointeractive
web site but it's not available now.
Will send WIN95SND.EXE. It is a self extracting file, beta version made by
Creative Corp. for the card.
I'm sure that I once used the driver for my Reply card and one of my PS/55s.
--
Tatsuo
Ulrich Trunk
>
> Hi !
>
> Got a Reply Vibra-16 MCA soundcard recently (ID:5138) and this is a pretty
> tricky beast as it seems.
> A) the Vibra-16 chipset was originally intended for ISA-cards only
True. Usually Vibra-16 gives you troules in ISA systems too....
> B) the card requires a Low 8-bit and a High 16-bit DMA ... from which using the
> high DMA either wont' work at all or crashes the machine
The ISA ones have the same "feature". It is usually caused by a
bad/incompatible DMA controller there...
> (DOS-based DIAGNOSTIC does not even play 16-bit testing WAV)
> C) If you use the "map high DMA to low DMA" feature the card disqualifies for
> Windows, since Win does not allow to set both DMAs to the same ressource.
Yep. Two DMA channels are required to use the card in full duplex mode
(i.e. record and playback operation at the same time).
>
> So far I have managed to get the card going under OS/2 Warp4 on a 9595A with a
> Pentium 120.
Ahhhh! It might be incompatible with the 95A's MCA features. Did you try
the card in a "conventional" (i.e. "stage 2" MCA) PS/2?
AFAIK IBM claims that some MCA cards are incompatible with the
965A/Server500. The Reply Vibra-16 MCA might be one of them.
Another possible reason for the failure is DMA-sharing. No other divice
must use any of the Vibra's DMA channels.
>
> The CONFIG.SYS lines read
>
> DEVICE=C:\MMOS2\SB16D2.SYS /C:1 /D:1 /H:1 /I:5 /A:220 /B:16 /M:330 /T:6
/N:SBAUD1$ /P
<SNIP>
> P = uh ... forgot about that
P = initialize ISA Plug 'n Pray cards. You'd better remove that one.
Ulrich
--
Note that my email address has changed: Replace "ihep" woth "kip" to
reply
Carroll Bloyd
I'm sending you a copy of the Reply files, including Windows 95 driver, that
I received with my Reply Vibra-16 sound card. Perhaps that will help.
I've used my Reply also on a 95A with Warp 4, and IIRC on the 95A with
NT4--so I've never tried the Win95 driver.
Also note that there are two .adf files--one 5137, is for the Reply with FD
SCSI, and 5138 is for the non-SCSI card.
Good luck,
Carroll
"Peterwendt" <peter...@aol.com> wrote in message
news:20010104035417...@nso-fo.aol.com...
Peterwendt
>If you use stock Win95 SB driver for your Reply Vibra 16, your machine will
>hang with "MCA BUS timeout error" or video will be crashed showing odd stripes
>as you described.
Epp.. Yep. Precisely. Incredible. A true shock to me.
I *used* to see that when I stuff an old Soundblaster CT-5120 or 5130 (Pro MCV/
Pro2 MCV) in something faster than a 386DX-25 ... but ... haven't seen that for
years anymore.
>WIN95SND.EXE, Win95 driver for Reply Vibra 16, was once stored at
>neointeractive web site but it's not available now.
Yeah. Steve obviously sold off all his stocks and moved away without leaving a
new adress. Modern times. No moral.
>Will send WIN95SND.EXE. It is a self extracting file, beta version made by
>Creative Corp. for the card.
Got it already - thanks. Will try that out and let you know what comes out.
There are some more people suffering on Reply Soundcards (some with integrated
SCSI). That's why *I* jumped in here this time :-)
Will write you a longer direct letter later. Thanks again for now.
Hope the japanese New Year celebrations were fine for you this time too !
(Did you sit all lined-up in front of the TV with your pyjamas - as rumours say
?)
;-)
>I'm sure that I once used the driver for my Reply card and one of my PS/55s.
Hmmm. Basically the card is not too bad - even if "Creative" is some sort of
"Warning Sign" in the soundcards' world as is "ATI" in videocards ....
See you !
Peterwendt
>> A) the Vibra-16 chipset was originally intended for ISA-cards only
>
>True. Usually Vibra-16 gives you troules in ISA systems too....
I vaguely recall that IBM sold a Vibra-derived card for the earlier Aptivas /
PS1 machines ... and it was "less than good" too.
(Or was it an ISA Jazz16 ? Whatever. Crap.)
>The ISA ones have the same "feature". It is usually caused by a
>bad/incompatible DMA controller there...
DMAs tend to have a "schedule": once you miss them you cannot re-trigger them
or hold the request for inifity. That will surely shoot your system.
>Yep. Two DMA channels are required to use the card in full duplex mode
>(i.e. record and playback operation at the same time).
Maybe I could run "CoolEdit 2000" with the "ClickFix" plugin on my P-120 later
... once it runs at 120 I go up to 210 probably ...
:-)
(Great Stuff BTW. I have that CoolEdit / ClickFix on the PC-300GL to
accustically clean my audiorecordings from old analogues)
>> So far I have managed to get the card going under OS/2 Warp4 on a 9595A
>>with a Pentium 120.
>
>Ahhhh! It might be incompatible with the 95A's MCA features.
Aiigh ... now you'd drowned.
If it were incompatible with the hardware basics - how (or why) would it then
run fabulous and superb under OS/2 without going into TRAPs ? OS/2 is my
favourite hardware consistency test. If parts don't TRAP there they can be
judged as "in good working condition".
>Did you try the card in a "conventional" (i.e. "stage 2" MCA) PS/2?
You guessed, right ?
Mod. 70-A21 with 8MB system RAM, 486Now! upgrade.
With and without IBM "Tropic" TR adapter and "short" SCSI adapter without
cache.
No Go. Same results, bus hangs or green vertical stripes.
That was my first idea to be honest: maybe the "Big One" is a tad to fast for
that card.
Can be answered with "No. Not really."
>AFAIK IBM claims that some MCA cards are incompatible with
>the 95A/Server500. The Reply Vibra-16 MCA might be one of them.
That's a different story.
Old 16-bit cards *should* not be used (includes the original famous "Tropic" TR
cards - old designs prior to the very short one). Some cards cannot cope with
the SynchroStream Mode on these boards ... worser on the Server 720 by the way,
for which only a handful of cards have been announced anyway.
Early LAN-Streamer Adapter with faulty chipset, early Auto-LAN and those with
the early microcodes can crash a 95A platfrom. Same for the Etherstreamers.
There are / were ECAs running on these cards. Few may have survived.
The old "Tribble" SCSI is not compliant with the 95A. Was failure-prone in the
fast 80s already. Also ECA on earlier cards, but later must be suspected to be
at least far too slow.
Some Adaptec AHA1640 with downlevel microcode hold the bus for too long. Result
... Bus Timeout Error. Generally all busmaster cards running on 80C188
busmaster chips *might* be a potential cause for "mysterious" crashes. A timing
issue on the data-exchange between board and card IIRC.
Too long ago. Rumours from an IBM technical training.
>Another possible reason for the failure is DMA-sharing. No other divice
>must use any of the Vibra's DMA channels.
Wasn't on the 70 ... nontheless: did crash.
On the 95A I even went a step ahead and disabled all parallel ports (DMA 7 and
6).
Killed everything off from the machine and started with a bare board with only
memory, processor card, SCSI "Corvette" and XGA-2 installed. No net no other
gimmicks. No AVM B1 ISDN, which insists to sit on IRQ5 (if not otherwise
directed to move).
On my trusty old workhorse (9595-S30 / P60) I had to move that card to IRQ3
already for the ACPA, which will positively fail to work on an IQ other than 5
(have COM2 disabled here).
I confidentially hope that Tatsuos file will bring more positive results.
Will test out tomorrow.
And let you know.
Brad Parker
9595-3NT upped to a 90 Mhz Type 4 Pentium.
The 3NT is using the Reply combo SB/SCSI, the OMT is using the SB only
version.
> I am using the stock NT soundblaster drivers.
Maybe, maybe not - Creative Labs Sound Blaster 1.X Pro, 16
This may have been put on by SP 4, but I'm not sure.
> I'll check my resources when I get home tonite and post what I'm using,
> but I am certainly using two different DMAs for full duplex operation.
I/O 220
IRQ 5
DMA-L 1
DMA-H 6
MPU401 300
DMA Buffer size is 16 Kb
IIRC, I increased the buffer size to 16Kb, it might have been 4 or 8Kb
at setup, but experience has taught me to keep this larger than the
standard setting.
I'm running CoolEdit Pro on this thing. Yes, it works great-I've been
able to mix 16 tracks simultaneously, probably do more, but I'm not that
patient....
Other hardware in the OMT includes a Spock, AHA-1640, SMC8013, Dual
Serial Card, XGA-1, ATi GUP. 64 Megs RAM
3NT has a Spock, Passplay, AHA 1640, IBM-SVGA, Olicom 10/100. 144 megs
RAM. This thing has 7 drives in the RAID, a 540 meg boot drive and it's
the box with 64 megs of cache on the Passplay.
Brad
Art
I am one of the (other) guys having a heck of a time with this card running
Win95B on my 0QT 200Mhz. All the symptoms Peter is getting and more, (green
lines, hangs time-outs, etc.,).
I have the Reply SB Virba16 w/ Future Domain SCSI Extension.
Thank , Art
Could you send me WIN95SND.exe also?
"T. Sunagawa" <sa...@mbc.nifty.com> wrote in message
news:3A54B0F0...@mbc.nifty.com...
Art
Could you please send me what you have as well?
I have the Reply SB Virba16 w/ Future Domain SCSI Extension.
Thank , Art
"Carroll Bloyd" <cbl...@home.com> wrote in message
news:F8456.230869$U46.7...@news1.sttls1.wa.home.com...
Peterwendt
Some more funny -or unfunny- results:
Setting the Vibra16 to
IRQ5, Low DMA 1, High DMA 5 (or 6), IO 220, MPU 330
results in a "machine gun sound" when starting OS/2.
When I set the DMA buffers to 16 (B:16 parameter) the "noise interval" gets
longer.
(Hard to describe: "Ra-Ta-Ta" instead of "Ratatata" ....)
Same machine. It works perfectly well with using DMA 1,1 .... :-)
The only change was the High DMA setting. At least the system does not crash.
Brave OS/2 !
I'm going to play around with the settings - but while I wanted a "Dual Boot"
machine into OS/2 Warp4 and Win95C I need to find a setting that works for
*both*.
I leave the simple things to others ....
Something *very* funny in this respect:
I *physically* set the board in the machine configuration to IRQ = 5, DMA1 = 1,
DMA2 = 5 .... but the driver line in the CONFIG.SYS under OS/2 has the driver
set to /I:5 /D:1 /H:1 ... *And It Still Works* with no problems.
Okay ... given that works for OS/2 ... then I'm going to boot into Win95 and
try out the drivers I got from Tatsuo (and Carroll B. too about 2 hours later
!)
Harald Pollack
an All:
Servus Peter!
P> All the time I see your postings I remember that I'd wanted
P> to send you an AVM B1 MCA "for evaluation" .... and forget it a moment
P> later again. I grew old.
You 'grow' old, I 'am' old :-)
P> But the year just started :-)
Yes, but do not bother, I have enough projects to play with :-)
If you have time, than send me the card, but no need to do it NOW.
P> Interestingly they say "8 bit Test" .. and "16-bit Test"
P> during that :-)
Yes, this is misleading.
>>You are use DMA channel 1 for both ???
P> Yep. A) it is supported in the ADF and B) demonstrates the
P> superiority of MCA machines running with OS/2. Funny, eh ?
Yes, but I believe, that the sound driver itself cannot handle this DMA sharing
correct. Sharing DAM (in MCA) is not really a sharing, it is a 'multiplexing' in
fact (because you cannot transfer different data to/from differnt
source/destination with ONE physical DAM channel at the same time (you can
'share' an IRQ, but a DMA channel you can only multiplex)
P> I hadn't got it going with two differerent DMA channels ....
P> "MCA Bus Timeout". Unfunny.
Than - obviously - the second DMA (on the same channel as the first one) was
never recognized by driver (has used another as default, which is not connected
by hardware, so it times out).
Herzliche Gruesse, Harald
-+- Message created on Friday January, 05 2001 10:03:42 MEZ
T. Sunagawa
> Could you send me WIN95SND.exe also?
It's in your mail box.
--
Tatsuo
Peterwendt
Latest news:
The drivers that Tatsuo (and Carroll !) supplied worked so far.
Pitfalls: WinAmp and CoolEdit hat to be explicitely told to use "8 bit" output.
That might be a workaround - but you cannot playback 16-bit (stereo) files as
-for instance- the "Jungle" or "Utopia" audio-scheme in Windows.
Solution: once you'd installed the drivers (which are internally marked as
"Beta") you simply run the Win95 drivers update from Creative Labs for the
Soundblasters.
I found a copy of this driver pack in www.windrivers.com under "Audio" - file
name is SB95UP.ZIP, about 296KB.
Results so far: WinAmp plays, CoolEdit too. All audioschemes can be used for
the Win95 "audible events".
While typing this I'd sent out a copy of the driver update to Art.
Anyone interested could leave me a note and I send that stuff too.
Art
Still received bus time-outs and hangs.
Sooo,
Downloaded SBW9XUP.exe from here: Drivers were a bit newer (1998) versus
WIN95SND (1996).
http://www.soundblaster.com/drivers/welcome.asp?reg=0&os=win95&prod=sb_16&Se
lect=Get+Files&x=62&y=14#filelist
Was able to boot past Time-out errors and use MediaPlayer and RealPlayer.
Card settings are as Peter stated.
Still have to test Net2Phone and NetMeeting Video Calling (Full-Duplex
Apps.).
But first I want to reinstall them under the new drivers.
Ulrich Trunk
>
> Hi !
>
> Some more funny -or unfunny- results:
>
> Setting the Vibra16 to
> IRQ5, Low DMA 1, High DMA 5 (or 6), IO 220, MPU 330
> results in a "machine gun sound" when starting OS/2.
>
> When I set the DMA buffers to 16 (B:16 parameter) the "noise interval" gets
> longer.
> (Hard to describe: "Ra-Ta-Ta" instead of "Ratatata" ....)
>
> Same machine. It works perfectly well with using DMA 1,1 .... :-)
> The only change was the High DMA setting. At least the system does not crash.
> Brave OS/2 !
Hi Peter!
First I have to admitt that I neither have an MCA-Vibra16 soundcard, nor
a 95A, but you caught me with that already...
Leaving the problems with the Win95 driver aside, the above clearly
indicates a Hardware problem.
My two soundblaster (AWE32) equiped ISA boxes (6586 & a Clone) use
different channels for high and low DMA (/D:1 and /H:5) and your 95A
should do so either.
However, the SB16 is a somewhat strange beast. The codec actually
consists of TWO codecs: A SB pro (8bit) one, which uses the "low DMA"
channel and an additional 16-bit codec, which uses the "high DMA"
channel. Since it needs 16bit wide data, a 16bit DMA channel *SHOULD* be
chosen for this part.
The SB16's split personality enabled Creative Labs to write "full
duplex" drivers when it became an important buzz word for marketing.
Compared to codecs by Crystal Semi or ESS, the full duplex capabilities
of the SB16 are limited: One of the audio streams has to use the SB pro
part, i.e. is limited to 8bit resolution.
Drivers are a different issue:
While OS/2 drivers do what you tell 'em on the command line, the Win95
(and later) ones show the chaotic behaviour related to the "Plug 'n
Pray" syndrome and thus are totally inappropriate for hardware testing.
A golden Rule for Hardware Testing is:
"If it doesn't work with DOS, it won't work at all"
There are only two exceptions to this:
- There is no DOS device driver for that piece of hardware
- It's "parasitic" Hardware which uses an OS-specific API (Winmodems,
GUI-printers and similar crap).
OS/2 is a good tool for reliability testing; I have some memory modules
with "OS/2 tested" or so stickers on them.
Thus the DOS-based SB16 diagnosis tool should work ok before you
proceed.
In your earlier posting you state that it doesn't (won't play 16bit
audio). So it *seems* that you have a hardware problem with 16bit
resolution. Unfortunately you didn't check with the DOS diagnosis tool
on the 70.....
I also *belive* that in your case, OS/2 (or '95 with the drivers you got
from Tatsuo) will not record or playback any 16bit audio data (system
sounds are 8bit only!).
Seting highDMA and lowDMA to the same channel *seems* only to hide that
problem upon driver initialisation.
In your intrest, I hope I'm wrong....
Ulrich
BTW:
The card IBM sold for the Aptivas was a Jazz16 (by Mediavision) and
*probably* a SB16 for the PC300/700 line.
--
Please note that my email addres has changed: To reply, replace "ihep"
with "kip".
Peterwendt
>First I have to admitt that I neither have an MCA-Vibra16 soundcard, nor
>a 95A, but you caught me with that already...
Yeah - looks that way, doesn't it ? ;-)
You should visit eBay USA ... and have a good chance to get both: the 95A and a
Reply sound card. Saw some last time I visited there. One of the cards I saw
(5137) has the Future Domain SCSI chipset on it
(Hi Art R. how are things proceeding with yours ?)
>Leaving the problems with the Win95 driver aside, the above clearly
>indicates a Hardware problem.
You bet. If that's not a hardware problem I wouldn't know any better to show.
>My two soundblaster (AWE32) equiped ISA boxes (6586 & a Clone) use
>different channels for high and low DMA (/D:1 and /H:5) and your 95A
>should do so either.
It does. From the hardware it is set to 1 and 5 ... but OS/2 still believes it
were 1, 1 ... and since OS/2 doesn't complain ... who am I to *convince* it it
were wrong ?
I'm glad it works now.
>However, the SB16 is a somewhat strange beast.
Yep. Definitely. Ever had the "fun" to play with a SB Pro or Pro MCV ?
Aaak !
>The codec actually
>consists of TWO codecs: A SB pro (8bit) one, which uses the "low DMA"
>channel and an additional 16-bit codec, which uses the "high DMA"
>channel. Since it needs 16bit wide data, a 16bit DMA channel *SHOULD* be
>chosen for this part.
That for "staying compatible with keeping all the old crap".
The reason why modern processors still have the A20 gate ...
>Drivers are a different issue:
>While OS/2 drivers do what you tell 'em on the command line, the Win95
>(and later) ones show the chaotic behaviour related to the "Plug 'n
>Pray" syndrome and thus are totally inappropriate for hardware testing.
With that sort of PnP you often don't know which of the ressources are actually
in use ... I prefer hardware PnP like in the PS/2. Reading ADFs tells you where
you are - even on cards that have "no user selectable options" (like 8514A like
cards).
>A golden Rule for Hardware Testing is:
>"If it doesn't work with DOS, it won't work at all"
True. It worked on DOS ... mainly.
>There are only two exceptions to this:
>- There is no DOS device driver for that piece of hardware
Hmm. Okay.
>- It's "parasitic" Hardware which uses an OS-specific API >(Winmodems,
GUI-printers and similar crap).
Hahaha. Yeah.
>OS/2 is a good tool for reliability testing; I have some memory modules
>with "OS/2 tested" or so stickers on them.
My testbox in the company used to be a Mod. 90 for years running under OS/2.
The T2 platform accepts a wide variety of modules ... and once it passes a day
of a test under OS/2 it is good. Period.
>Thus the DOS-based SB16 diagnosis tool should work ok before you
>proceed.
>In your earlier posting you state that it doesn't (won't play 16bit
>audio). So it *seems* that you have a hardware problem with 16bit
>resolution. Unfortunately you didn't check with the DOS diagnosis tool
>on the 70.....
I did check on the 70. I started with DOS on it. I only didn't tell.
The results were basically the same: crash-boom-bang ... blue screen.
Wonder what that was ... maybe the *Diagnostic* itself is inappropriate for MCA
cards and machines. Most likely.
Once set the machine runs - apart from the above-mentioned OS/2 misbehaviour.
>I also *belive* that in your case, OS/2 (or '95 with the drivers you got
>from Tatsuo) will not record or playback any 16bit audio data (system
>sounds are 8bit only!).
1. I ran the updated SB drivers on the machine in W95 - and now those badly
distorted 16-bit stereo sounds that did not play prior *do* play now.
2. The base Winsounds are 8-bit .. but the "Audio Schemes" like Jungle, Utopia,
Robot etc. are partially stereo, most are 16-bit.
>Seting highDMA and lowDMA to the same channel *seems* only to hide that
>problem upon driver initialisation.
Physically I'm not using the same .. only for the old SB16D2.SYS in OS/2 it
appears as if I were using the same DMA for 8 and 16 bit. I suspect the driver
to be a little too narrow-minded ... :-)
>In your intrest, I hope I'm wrong....
Okay, Thanks. Most problems I have sorted out so far and next on the testlist
is WinNT 4. Then Linux. Maybe.
>BTW:
>The card IBM sold for the Aptivas was a Jazz16 (by Mediavision) and
>*probably* a SB16 for the PC300/700 line.
Yeah - something like that. I recalled the Jazz16. Fortunately the company I
work for did not sell Aptivas. The PC300 / 700 often came with the onboard
Audio. ESS ? Crystal Audio ? So many manufacturers ... so many names to keep in
mind ...
Whatfor ?
Art
Still no-go, trying to run audio software the utilizes two DMA
(Full-duplex). Cannot make a PC2Phone call of use NetMeeting with out system
hanging up. Peter, the settings are just as you suggested?
Strange that with my older 8 Bit SB PRo card with just one DMA I got these
to work, just could not talk when the other party was talking. Now with this
new Full-duplex card, nothing...
Thats progress for ya.. One step forward and two steps back...
I'll give Carroll's files a try but I think I'll be in the market for a
ChipChat and not completely sure that will solve my problems.
"Art" <arei...@cox.rr.com> wrote in message
news:9Tn56.5674$NL3.6...@typhoon.southeast.rr.com...
Art
SB95UP files don't help me run Full-duplex apps.. Possibly your additional
(Reply files) you received with your card will help. Heck, I've tried
everything else...
I am tending to believe my problem is more a timing problem with my 200Mhz
cpu 66 Mhz oscillator than confused DMA's.
You don't happen to have a spare ChipChat lying around do you?
"Carroll Bloyd" <cbl...@home.com> wrote in message
news:F8456.230869$U46.7...@news1.sttls1.wa.home.com...
Peterwendt
To add another detail here:
I can fully confirm the results posted here by Brad Parker about Vibra16 and
NT.
Once running Win NT 4.0 (Workstation in my case) you only need to open "Control
Panel" - "Multimedia" - "Audio Devices" - "Add ..." and pick the "Creative Labs
Soundblaster 1.X, Pro, 16" driver. Then set the values for I/O adresses, IRQ,
Low DMA, High DMA to appropriate values and it works. Still have the card set
to I/O 220, IRQ 5, DMA1 (8) and DMA5 (16) and I/O 330 for the Midi device. Same
setting as for OS/2 and Win95. Still don't know if the card is *really* in full
duplex mode.
The setup even survived a Service Pack 6a update with no problems (which is not
always given ...)
Main disadvantage here: the volume slides are very "inverse logarithmic" ....
the most part of the volume changes appear in the lower 5% of the way - after
1/3rd of the way the result is merely noticeable.
But I won't complain on that. The card works fine for the Midi and Wave part
and that's what I want.
Mike Ford
VIBRA-16
(C) Creative Tech '94
CT2501-TBQ
And after reading this thread I am wondering what the final consensus
was? Works with OS/2, NT, and Linux, but not W95?
Peterwendt
>And after reading this thread I am wondering what the final consensus
>was? Works with OS/2, NT, and Linux, but not W95?
Works with OS/2, Win95, Win NT 4 (server and workstation) but not under my SuSE
Linux 6.1 or 7 professional. Other Linuxes ... well ... you might try with
Debian (2.2 I guess ?) - it seems to be the most recommended for MCA-units.
Brad Parker
Works "mostly" with all of the above Operating Systems. If you try to
do anything beyond the basics, you will find that it is software
dependent.
Be prepared for resource hell though, takes two DMAs, 1 IRQ and some
memory slots. Add more resources if you got one with the Future Domain
SCSI host adapter on board....
All that said, I like mine. Works good for basic audio applications. If
you tire of it, drop me a line, I'll take it off your hands.
Art
I've gotten mine to work under 95B and NT for basic sound requirements with
good sound quality.
Anything more complicated requiring Full-Duplex is a No-Go:
- Under 95B, Green vertical bar across screen, requires reboot
- Under NT, Can run Full-Duplex apps but not in Full-Duplex mode
(i.e., NetMeeting, Net2Phone, etc.,)
I'm looking for a ChatChat. Hell I've tried everything else...
Peterwendt
Another bit of info to this old thread:
Under OS/2 Warp 4 I did'nt manage to get the Vibra-16 going in Slot 7 - which
is the 8514/A compatible AVE slot. No matter what I'd tried - setting the card
to only one DMA for Hi and Lo ... altered the CONFIG line various times.
After I relocated (and reconfigured) the card from Slot 7 to Slot 8 it worked
without any further fiddling with the drivers or CONFIG.SYS settings. I'm using
the already welknown "misconfiguration" under OS/2: Low DMA = 1, High DMA = 5
in the physical setup - but Low DMA = 1, High DMA = also 1 in the CONFIG.SYS.
Strange, eh ?
Those of you which tried -in vain- to get the card to run: test for another
slot before you finally give up.