Today I've been attempting to diagnose why my recent Turbo XT acquisition keeps corrupting the HDD when it's in turbo mode... even though the machine has managed to pass multiple different diagnostic tools' testing with flying colours and zero errors.

Investigations are ongoing!

BetaC wrote on 2020-09-20, 07:13:

I managed most of the cabling within my general purpose 98SE/Late-DOS/W2k system, freeing up some extra resources that can eventually find their way in to a Pentium/S7 build I want to eventually do. I can still work a little more to make it completely clean, but I'm good for now.
image0.jpg

Hi! Browsing the thread I noticed you haven't (yet) got the J12 and J13 jumpers shorted on your GA-6BXC despite using a Voodoo3 3000 AGP.
Just in case you didn't know: Those jumpers are intended for use with Voodoo3 cards. The original AGP VRM on many Gigabyte boards would often overheat when a Voodoo3 or Geforce card was used so they incorporated those jumpers on later revisions of affected boards to bypass the VRM and feed the slot 3.3V directly from the ATX PSU.
If you haven't done so already I recommend setting the jumpers so your board doesn't fry and potentially take the graphics card with it like it happened ever so often back in 1999.
Cheers! 😀

kdr wrote on 2020-09-26, 04:58:

Today I've been attempting to diagnose why my recent Turbo XT acquisition keeps corrupting the HDD when it's in turbo mode... even though the machine has managed to pass multiple different diagnostic tools' testing with flying colours and zero errors.

Investigations are ongoing!

ISA Bus overclock most likely?

liqmat wrote on 2020-09-25, 22:09:

Turbo -> wrote on 2020-09-25, 21:01:

liqmat wrote on 2020-09-25, 14:56:

My favorite Atari computer model and this one cleaned up nicely.

What kind of cable did you use to hook your Atari to TV?

A composite A/V cable.

Does the cable look like this? I'm asking because I'm getting ready to change some RAM's on the same unit for my friend, but before I attempt to do this, I would like to connect it to TV and see how it behaves.

appiah4 wrote on 2020-09-26, 08:00:

ISA Bus overclock most likely?

I measured the DMA controller's clock, it's 4Mhz when in turbo mode (and 4.77Mhz in normal mode) so seems within spec.

I guess it's time to start writing a minimal test program that can reproduce the corruption, and then trace through its execution to find out exactly what gets corrupted and how. 😀

Turbo -> wrote on 2020-09-26, 08:32:

liqmat wrote on 2020-09-25, 22:09:

Turbo -> wrote on 2020-09-25, 21:01:

What kind of cable did you use to hook your Atari to TV?

A composite A/V cable.

Does the cable look like this? I'm asking because I'm getting ready to change some RAM's on the same unit for my friend, but before I attempt to do this, I would like to connect it to TV and see how it behaves.

Yes sir. Purchased it off Ebay IIRC.

kdr wrote on 2020-09-26, 10:05:

appiah4 wrote on 2020-09-26, 08:00:

ISA Bus overclock most likely?

I measured the DMA controller's clock, it's 4Mhz when in turbo mode (and 4.77Mhz in normal mode) so seems within spec.

I guess it's time to start writing a minimal test program that can reproduce the corruption, and then trace through its execution to find out exactly what gets corrupted and how. 😀

Honestly would just try a different/newer MFM controller. Most problems I have run into with MFM tech is a failing cable or controller. Those full metal jacket Seagate MFM drives usually hold up pretty well in my experience. I mean, yeah, I have to tag some bad sectors, but overall they just keep on spinning. Unfortunately I am not near my gear right now otherwise I would send you a couple of newer model MFM controllers for testing.

Whoops, just realized you are in NZ. That would not be cost effective. 🤣

liqmat wrote on 2020-09-26, 14:02:

Honestly would just try a different/newer MFM controller. Most problems I have run into with MFM tech is a failing cable or controller. Those full metal jacket Seagate MFM drives usually hold up pretty well in my experience. I mean, yeah, I have to tag some bad sectors, but overall they just keep on spinning. Unfortunately I am not near my gear right now otherwise I would send you a couple of newer model MFM controllers for testing.

Whoops, just realized you are in NZ. That would not be cost effective. 🤣

Well the problem is far more sinister than just a bad controller/drive or an overclocked ISA bus...

SpeedStor and SpinRite say that my ST-238R is in perfect condition, not a single bad sector (!!!) or seek error (!). [Yes, while running in turbo mode.] But any access via the DOS/BIOS disk routines provokes corruption on both read and write when in turbo mode. Even worse, it affects the floppy drive too!

This is such an interesting bug that I'm determined to get to the bottom of it! Surely must be some hardware glitch, after all. At least debugging that is a realistic option on a turbo XT clone motherboard (everything is in DIP packages on a plain old double-sided PCB with <10Mhz signals). I'd hate to have to track down a hardware fault on my 486 VLB motherboard, for instance... 😁

kdr wrote on 2020-09-26, 22:21:

Well the problem is far more sinister than just a bad controller/drive or an overclocked ISA bus... […]
Show full quote

liqmat wrote on 2020-09-26, 14:02:

Honestly would just try a different/newer MFM controller. Most problems I have run into with MFM tech is a failing cable or controller. Those full metal jacket Seagate MFM drives usually hold up pretty well in my experience. I mean, yeah, I have to tag some bad sectors, but overall they just keep on spinning. Unfortunately I am not near my gear right now otherwise I would send you a couple of newer model MFM controllers for testing.

Whoops, just realized you are in NZ. That would not be cost effective. 🤣

Well the problem is far more sinister than just a bad controller/drive or an overclocked ISA bus...

SpeedStor and SpinRite say that my ST-238R is in perfect condition, not a single bad sector (!!!) or seek error (!). [Yes, while running in turbo mode.] But any access via the DOS/BIOS disk routines provokes corruption on both read and write when in turbo mode. Even worse, it affects the floppy drive too!

This is such an interesting bug that I'm determined to get to the bottom of it! Surely must be some hardware glitch, after all. At least debugging that is a realistic option on a turbo XT clone motherboard (everything is in DIP packages on a plain old double-sided PCB with <10Mhz signals). I'd hate to have to track down a hardware fault on my 486 VLB motherboard, for instance... 😁

Interesting you bring up the floppy. I had a friend who was having oddball MFM HDD issues. Sometimes it would boot, sometimes not. Sometimes it would have bad sectors, sometimes not. He unplugged the floppy drive and all the MFM HDD issues disappeared. Have you tried removing the floppy from the mix and trying it again in turbo? Worth a shot.

Doornkaat wrote on 2020-09-26, 06:14:

Hi! Browsing the thread I noticed you haven't (yet) got the J12 and J13 jumpers shorted on your GA-6BXC despite using a Voodoo3 […]
Show full quote

BetaC wrote on 2020-09-20, 07:13:

I managed most of the cabling within my general purpose 98SE/Late-DOS/W2k system, freeing up some extra resources that can eventually find their way in to a Pentium/S7 build I want to eventually do. I can still work a little more to make it completely clean, but I'm good for now.
image0.jpg

Hi! Browsing the thread I noticed you haven't (yet) got the J12 and J13 jumpers shorted on your GA-6BXC despite using a Voodoo3 3000 AGP.
Just in case you didn't know: Those jumpers are intended for use with Voodoo3 cards. The original AGP VRM on many Gigabyte boards would often overheat when a Voodoo3 or Geforce card was used so they incorporated those jumpers on later revisions of affected boards to bypass the VRM and feed the slot 3.3V directly from the ATX PSU.
If you haven't done so already I recommend setting the jumpers so your board doesn't fry and potentially take the graphics card with it like it happened ever so often back in 1999.
Cheers! 😀

Thank you, I was completely unaware.

kdr wrote on 2020-09-26, 04:58:

Today I've been attempting to diagnose why my recent Turbo XT acquisition keeps corrupting the HDD when it's in turbo mode... even though the machine has managed to pass multiple different diagnostic tools' testing with flying colours and zero errors.

Investigations are ongoing!

Wow, is that motherboard white ? Can you post a good picture of the motherboard ?
Also: if you did not format that MFM drive with that controller you could have odd issues if formatted with a different but same make/model controller... just a thought.

Picking away at what I call the Retro Console/Audio corner of the rec room (what my wife calls "the room full of your s**t"). Again, not 100% retro, but the retro consoles, turntable, component stereo are all here, and most of it has been either thrift/yard sale finds, family/friends, or mine from childhood.

LHN91 wrote on 2020-09-27, 03:55:

Picking away at what I call the Retro Console/Audio corner of the rec room (what my wife calls "the room full of your s**t"). Again, not 100% retro, but the retro consoles, turntable, component stereo are all here, and most of it has been either thrift/yard sale finds, family/friends, or mine from childhood.

IMG_20200926_233240.jpg

Badly missing a retro PC 😜

Iv been at this hobby, casually. For maybe a year and a half now. ...and it finally hit me, just how bad Sound Blaster 16s are with the notorious DMA Clicking/Popping bug.

I sware. I haven't noticed it before. Maybe in one or two games, but now I'm noticing it like..everywhere. It's driving me insane. Everywhere that's not offering a native SB16 option anyway.

The minute I go slightly audiophile on my main modern PC, the minute I start noticing quirks with my retro PCs 8(

LHN91 wrote on 2020-09-27, 03:55:

Picking away at what I call the Retro Console/Audio corner of the rec room (what my wife calls "the room full of your s**t"). Again, not 100% retro, but the retro consoles, turntable, component stereo are all here, and most of it has been either thrift/yard sale finds, family/friends, or mine from childhood.

IMG_20200926_233240.jpg

Is that Toto?

https://www.youtube.com/watch?v=6v6eqXGU2VE

LHN91 wrote on 2020-09-27, 03:55:

what my wife calls "the room full of your s**t"

Ah, the famous "Manc Cave"!

SodaSuccubus wrote on 2020-09-27, 05:27:

Iv been at this hobby, casually. For maybe a year and a half now. ...and it finally hit me, just how bad Sound Blaster 16s are with the notorious DMA Clicking/Popping bug.

I sware. I haven't noticed it before. Maybe in one or two games, but now I'm noticing it like..everywhere. It's driving me insane. Everywhere that's not offering a native SB16 option anyway.

The minute I go slightly audiophile on my main modern PC, the minute I start noticing quirks with my retro PCs 8(

You will not refer to it as a bug, dear sir, but rather a feature. 😀 Some even find it charming. They even made a cereal after it.

Horun wrote on 2020-09-27, 03:31:

Wow, is that motherboard white ? Can you post a good picture of the motherboard ?
Also: if you did not format that MFM drive with that controller you could have odd issues if formatted with a different but same make/model controller... just a thought.

MFM drive had a fresh low-level format, and as mentioned the issue seems to also affect the floppy. I can reproduce it with two different floppy controllers as well. I'm going to write a small test utility to exercise the floppy (far easier to debug because the controller is so simple) and see if I can find a particular bit pattern in the error bytes.

And yes - it is a pure white soldermask on this motherboard! It looks kind of like a NuXT except manufactured in the '80s. Here's a better photo of the entire board.

kdr wrote on 2020-09-27, 22:05:

Horun wrote on 2020-09-27, 03:31:

Wow, is that motherboard white ? Can you post a good picture of the motherboard ?
Also: if you did not format that MFM drive with that controller you could have odd issues if formatted with a different but same make/model controller... just a thought.

MFM drive had a fresh low-level format, and as mentioned the issue seems to also affect the floppy. I can reproduce it with two different floppy controllers as well. I'm going to write a small test utility to exercise the floppy (far easier to debug because the controller is so simple) and see if I can find a particular bit pattern in the error bytes.

And yes - it is a pure white soldermask on this motherboard! It looks kind of like a NuXT except manufactured in the '80s. Here's a better photo of the entire board.

That Is a suprisingly clean and modern look for a board made in the 80s. Well done to the folks in the factory that day!

Great find 😀

SodaSuccubus wrote on 2020-09-27, 22:30:

kdr wrote on 2020-09-27, 22:05:

Horun wrote on 2020-09-27, 03:31:

Wow, is that motherboard white ? Can you post a good picture of the motherboard ?
Also: if you did not format that MFM drive with that controller you could have odd issues if formatted with a different but same make/model controller... just a thought.

MFM drive had a fresh low-level format, and as mentioned the issue seems to also affect the floppy. I can reproduce it with two different floppy controllers as well. I'm going to write a small test utility to exercise the floppy (far easier to debug because the controller is so simple) and see if I can find a particular bit pattern in the error bytes.

And yes - it is a pure white soldermask on this motherboard! It looks kind of like a NuXT except manufactured in the '80s. Here's a better photo of the entire board.

That Is a suprisingly clean and modern look for a board made in the 80s. Well done to the folks in the factory that day!

Great find 😀

Yes is a great looking board !