STe - Bad DMA Chip

Troubles with your machine? Just want to speak about the latest improvements? This is the place!

Moderators: Mug UK, Zorro 2, Greenious, spiny, Moderator Team

Post Reply
alanh
Hardware Guru
Hardware Guru
Posts: 1630
Joined: Mon Jul 24, 2006 9:01 pm
Location: North Wales, UK

Re: STe - Bad DMA Chip

Post by alanh »

While there are multiple solutions to the 'Bad DMA" phenomenon, it should be said that there are differing opinions on what is right and what is wrong.

Either way, we should be able to agree to disagree and move on and not just keep repeating what's already been said.

For those wishing to investigate further and produce alternative solutions should be encouraged.

The world would be pretty boring if we all just accepted only one solution to any problem.

If other folks have alternative solutions to try then I'd be happy to hear them.
Falcon CT60, Falcon CT63 x2, TT x3, MegaST x2, MegaSTE x2, STFM x2, STE x2, STacy, STBook, (Dead) Hades 060, Milan 060, T40.
aag
Atari maniac
Atari maniac
Posts: 77
Joined: Wed May 09, 2007 9:20 pm
Location: West Yorkshire, UK
Contact:

Re: STe - Bad DMA Chip

Post by aag »

This user reported using emutos loaded into ram addressed issues with acsi2stm on ste

https://atari-forum.com/viewtopic.php?p=421465#p421465

(Again mentions "reducing noise on bus")
czietz
Hardware Guru
Hardware Guru
Posts: 2451
Joined: Tue May 24, 2016 6:47 pm

Re: STe - Bad DMA Chip

Post by czietz »

alanh wrote: Tue Apr 04, 2023 9:37 pm For those wishing to investigate further and produce alternative solutions should be encouraged.

The world would be pretty boring if we all just accepted only one solution to any problem.

If other folks have alternative solutions to try then I'd be happy to hear them.
Very wise words! Thank you.

It would indeed be boring and sad at the same time, if one opinion was always to win simply because it's the "loudest" opinion. That would stifle further investigation and, thus, the progress for everyone here.
joska
Hardware Guru
Hardware Guru
Posts: 5693
Joined: Tue Oct 30, 2007 2:55 pm
Location: Florø, Norway
Contact:

Re: STe - Bad DMA Chip

Post by joska »

dlfrsilver wrote: Tue Apr 04, 2023 11:00 am Joska, some people got problem when installing HC CPU, and other people not. As said earlier, it's not a single CPU problem, not a single DMA problem.
This is not really an explanation, is it? You claimed that the HC CPU solved the issue because it's "less noisy". If that's the case - how can *less noise* cause problems?

I don't buy the "noise" hypothesis. To me it's more likely that it's a timing issue, as explained by mpattonm:
mpattonm wrote: Mon Apr 03, 2023 1:32 pm CMOS logic is not less prone to noise, nor it it "less noisy" as you claim. It is evaluating different voltage levels as logical low/high than TTL technology. Meaning on same (slanted) signal, it changes output in different moment compared to TTL.
The slight difference in timing caused by the different logic levels in CMOS and TTL is IMO a much more plausible explanation for the behaviour we see when replacing the TTL CPU with a CMOS CPU. Both in the cases where it compensates for the "bad DMA" issues and in the cases where it *causes* DMA issues.
Jo Even

VanillaMiNT - Falcon060 - Milan060 - Falcon040 - MIST - Mega STE - Mega ST - STM - STE - Amiga 600 - Sharp MZ700 - MSX - Amstrad CPC - C64
12MHz
Retro freak
Retro freak
Posts: 12
Joined: Fri Sep 11, 2020 2:49 pm

Re: STe - Bad DMA Chip

Post by 12MHz »

joska wrote: Wed Apr 05, 2023 11:09 am I don't buy the "noise" hypothesis. To me it's more likely that it's a timing issue, as explained by mpattonm:
mpattonm wrote: Mon Apr 03, 2023 1:32 pm CMOS logic is not less prone to noise, nor it it "less noisy" as you claim. It is evaluating different voltage levels as logical low/high than TTL technology. Meaning on same (slanted) signal, it changes output in different moment compared to TTL.
The slight difference in timing caused by the different logic levels in CMOS and TTL is IMO a much more plausible explanation for the behaviour we see when replacing the TTL CPU with a CMOS CPU. Both in the cases where it compensates for the "bad DMA" issues and in the cases where it *causes* DMA issues.
I also don't agree with the "noise" hypothesis.
It's more a question of asserting and taking back the signals /AS (adress strobe) and the both data strobes (/LDS and /UDS) at the mainboard. There are some differences between NMOS and CMOS CPUs for these signals. We know, that the whole issue "bad dma" depends at little differences between assert and delinguish signals. These small differences can already decide whether the DMA access works or not.

Exxos's efforts on the topic of "DMA" are worthy of respect. But the remedial measures unfortunately only combat the effects and not the cause. The exchange of the NMOS CPU to a CMOS CPU is certainly one of the best measures. But it does not solve the problem at the root cause, unfortunately. The documentation on the GSTMCU provides some clues as to the context.
http://www.chzsoft.de/asic-web/
The drawings ST-4082 gives more information about the signals /BR and /BG which will assert after addressing xFF8604 and xFF8606.
But more about that should rather be reported by Christian. He deserves the honor.

Best regards
Robert
Robbiet9999
Atarian
Atarian
Posts: 7
Joined: Mon Jan 02, 2023 8:08 pm

Re: STe - Bad DMA Chip

Post by Robbiet9999 »

dlfrsilver wrote: Sun Apr 02, 2023 4:34 pm
Robbiet9999 wrote: Sun Apr 02, 2023 4:20 pm I'm still have got nowhere it would seem I have to download a firmware upgrade but don't have anyway of getting it to the US. As my pc is windows 11 and laptop is 10

I've not had much luck with the US . Its only a month old
Can you make a picture of your STE motherboard ? thanks :)

EDIT : Is your STE serviced ? Have you changed the caps on the PSU and motherboard ? What DMA chip do you have and what is the manufacturing date ?
I have a new psu to go in a tos upgrade to 2.06 and have bought a real hard drive which works so probably is the US so will email Retro lemon. Will take a photo while doing upgrades
Robbiet9999
Atarian
Atarian
Posts: 7
Joined: Mon Jan 02, 2023 8:08 pm

Re: STe - Bad DMA Chip

Post by Robbiet9999 »

US sent back to Retro Lemon for the software to sorted
User avatar
mpattonm
Hardware Guru
Hardware Guru
Posts: 841
Joined: Mon Oct 21, 2002 8:52 am
Location: Czech republic
Contact:

Re: STe - Bad DMA Chip

Post by mpattonm »

On a side note, this has just been posted on FB "Atari ST and STe users" group:
2023-04-13 104054.png
the rest of the post is also interesting, but irrelevant to this topic. Hopefully this will help to clear the waters.
You do not have the required permissions to view the files attached to this post.
User avatar
dlfrsilver
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 2199
Joined: Mon Jan 31, 2005 1:41 am

Re: STe - Bad DMA Chip

Post by dlfrsilver »

mpattonm wrote: Thu Apr 13, 2023 8:44 am On a side note, this has just been posted on FB "Atari ST and STe users" group:
2023-04-13 104054.png
the rest of the post is also interesting, but irrelevant to this topic. Hopefully this will help to clear the waters.
I saw that too. He is lucky that Atari fixed his machine for him. Problem : he doesn't know what they did on the motherboard.

A picture of the motherboard could maybe help understanding what modification they applied.
Now SPS France representative since the 19th of June 2014. Proud to be an SPS member !
User avatar
mpattonm
Hardware Guru
Hardware Guru
Posts: 841
Joined: Mon Oct 21, 2002 8:52 am
Location: Czech republic
Contact:

Re: STe - Bad DMA Chip

Post by mpattonm »

dlfrsilver wrote: Thu Apr 13, 2023 11:03 pm I saw that too. He is lucky that Atari fixed his machine for him. Problem : he doesn't know what they did on the motherboard.
Good. Then do not speak for OP, ask him if you are in doubt. Judging from wording OP used, my feeling is he is quite familiar with the issue. One thing is for sure though, I am not going to spend any more time on this ridiculous discussion.
User avatar
dlfrsilver
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 2199
Joined: Mon Jan 31, 2005 1:41 am

Re: STe - Bad DMA Chip

Post by dlfrsilver »

mpattonm wrote: Fri Apr 14, 2023 8:25 am
dlfrsilver wrote: Thu Apr 13, 2023 11:03 pm I saw that too. He is lucky that Atari fixed his machine for him. Problem : he doesn't know what they did on the motherboard.
Good. Then do not speak for OP, ask him if you are in doubt. Judging from wording OP used, my feeling is he is quite familiar with the issue. One thing is for sure though, I am not going to spend any more time on this ridiculous discussion.
You're free, do as you want.
Now SPS France representative since the 19th of June 2014. Proud to be an SPS member !
czietz
Hardware Guru
Hardware Guru
Posts: 2451
Joined: Tue May 24, 2016 6:47 pm

Re: STe - Bad DMA Chip

Post by czietz »

I like to add a link to my investigation into the "bad DMA" phenomenon to this thread:
https://www.chzsoft.de/site/hardware/ne ... stigation/

Among other things, it outlines a potential software fix, that has been successfully tested on multiple STEs, so far.

Thank you to everyone who was involved in this, be it by testing, reviewing the contents, or contributing measurement results!
evil
Captain Atari
Captain Atari
Posts: 271
Joined: Sun Nov 12, 2006 8:03 pm
Location: Devpac

Re: STe - Bad DMA Chip

Post by evil »

czietz wrote: Sat Apr 22, 2023 9:47 am I like to add a link to my investigation into the "bad DMA" phenomenon to this thread:
https://www.chzsoft.de/site/hardware/ne ... stigation/
Among other things, it outlines a potential software fix, that has been successfully tested on multiple STEs, so far.
You're a legend Christian, thanks a bunch for this!
czietz
Hardware Guru
Hardware Guru
Posts: 2451
Joined: Tue May 24, 2016 6:47 pm

Re: STe - Bad DMA Chip

Post by czietz »

evil wrote: Sat Apr 22, 2023 10:10 am You're a legend Christian, thanks a bunch for this!
Thank you. :D

As noted in my article, the software workaround should be further tested. In particular, I'm interested in doing some tests with a C025913-20 DMA IC; the model only used in the very first STs (and never in the STE). But these chips are hard to find. If someone here wants to donate or at least lend me a C025913-20, please contact me via PM.
User avatar
RealLarry
Captain Atari
Captain Atari
Posts: 261
Joined: Sat Jan 31, 2015 12:05 pm
Location: San Junipero

Re: STe - Bad DMA Chip

Post by RealLarry »

czietz wrote: Sat Apr 22, 2023 9:47 am I like to add a link to my investigation into the "bad DMA" phenomenon to this thread:
https://www.chzsoft.de/site/hardware/ne ... stigation/

Among other things, it outlines a potential software fix, that has been successfully tested on multiple STEs, so far.

Thank you to everyone who was involved in this, be it by testing, reviewing the contents, or contributing measurement results!
Hum...reads more exciting than a detective story! :cheers:
On the other side of the screen, it all looks so easy.
User avatar
Cyprian
10 GOTO 10
10 GOTO 10
Posts: 3087
Joined: Fri Oct 04, 2002 11:23 am
Location: Warsaw, Poland

Re: STe - Bad DMA Chip

Post by Cyprian »

@czietz well done!
Thanks for the analysis and that article.
Lynx I / Mega ST 1 / 7800 / Portfolio / Lynx II / Jaguar / TT030 / Mega STe / 800 XL / 1040 STe / Falcon030 / 65 XE / 520 STm / SM124 / SC1435
DDD HDD / AT Speed C16 / TF536 / SDrive / PAK68/3 / Lynx Multi Card / LDW Super 2000 / XCA12 / SkunkBoard / CosmosEx / SatanDisk / UltraSatan / USB Floppy Drive Emulator / Eiffel / SIO2PC / Crazy Dots / PAM Net
Hatari / Steem SSE / Aranym / Saint
http://260ste.atari.org
User avatar
dhedberg
Atari God
Atari God
Posts: 1371
Joined: Mon Aug 30, 2010 8:36 am
Contact:

Re: STe - Bad DMA Chip

Post by dhedberg »

Impressive work and dedication! After all these years we finally understand the problem. Thanks to you Christian, and yes... you're a legend now!
Daniel, New Beat - http://newbeat.atari.org.
Like demos? Have a look at our new Falcon030 demo It's that time of the year again, or click here to feel the JOY.
User avatar
npomarede
Atari God
Atari God
Posts: 1496
Joined: Sat Dec 01, 2007 7:38 pm
Location: France

Re: STe - Bad DMA Chip

Post by npomarede »

That's what I call a proper analysis ! Facts and only reproducible facts, specific software test case to exhibit and possibly circumvent the issue.
All in all, great work :thumbs:
User avatar
dlfrsilver
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 2199
Joined: Mon Jan 31, 2005 1:41 am

Re: STe - Bad DMA Chip

Post by dlfrsilver »

czietz wrote: Sat Apr 22, 2023 10:16 am
evil wrote: Sat Apr 22, 2023 10:10 am You're a legend Christian, thanks a bunch for this!
Thank you. :D

As noted in my article, the software workaround should be further tested. In particular, I'm interested in doing some tests with a C025913-20 DMA IC; the model only used in the very first STs (and never in the STE). But these chips are hard to find. If someone here wants to donate or at least lend me a C025913-20, please contact me via PM.
You did an excellent job. However, i have a question after reading what you wrote :

-why replacing only the DMA chip by a "good DMA chip" is not enough to fix the problem ?

(This is what i saw on my machines.....)
Now SPS France representative since the 19th of June 2014. Proud to be an SPS member !
User avatar
Cyprian
10 GOTO 10
10 GOTO 10
Posts: 3087
Joined: Fri Oct 04, 2002 11:23 am
Location: Warsaw, Poland

Re: STe - Bad DMA Chip

Post by Cyprian »

czietz wrote: Sat Apr 22, 2023 9:47 am I like to add a link to my investigation into the "bad DMA" phenomenon to this thread:
https://www.chzsoft.de/site/hardware/ne ... stigation/

Among other things, it outlines a potential software fix, that has been successfully tested on multiple STEs, so far.

Thank you to everyone who was involved in this, be it by testing, reviewing the contents, or contributing measurement results!
I have a two questions:
1) does 'BAD DMA' has an impact onto FDD?
2) What should be corrected (I mean 16bit writes to the DMA registers) the TOS in ROM, HDD driver or both?
Lynx I / Mega ST 1 / 7800 / Portfolio / Lynx II / Jaguar / TT030 / Mega STe / 800 XL / 1040 STe / Falcon030 / 65 XE / 520 STm / SM124 / SC1435
DDD HDD / AT Speed C16 / TF536 / SDrive / PAK68/3 / Lynx Multi Card / LDW Super 2000 / XCA12 / SkunkBoard / CosmosEx / SatanDisk / UltraSatan / USB Floppy Drive Emulator / Eiffel / SIO2PC / Crazy Dots / PAM Net
Hatari / Steem SSE / Aranym / Saint
http://260ste.atari.org
czietz
Hardware Guru
Hardware Guru
Posts: 2451
Joined: Tue May 24, 2016 6:47 pm

Re: STe - Bad DMA Chip

Post by czietz »

Cyprian wrote: Sat Apr 22, 2023 12:02 pm 1) does 'BAD DMA' has an impact onto FDD?
No, because with floppy disk access there are no long-word writes to begin with.
2) What should be corrected (I mean 16bit writes to the DMA registers) the TOS in ROM, HDD driver or both?
As it is writing to the disk that corrupts the data, the change needs to done to the hard disk driver. (TOS 2.06 has a DMAwrite routine in ROM, that would require changing, too; but it's almost never used.)

Then again: Before advocating any changes to drivers or TOS, we still need to figure out the reason for Atari's recommendation of using MOVE.L.
User avatar
Cyprian
10 GOTO 10
10 GOTO 10
Posts: 3087
Joined: Fri Oct 04, 2002 11:23 am
Location: Warsaw, Poland

Re: STe - Bad DMA Chip

Post by Cyprian »

@czietz thanks for clarification
Lynx I / Mega ST 1 / 7800 / Portfolio / Lynx II / Jaguar / TT030 / Mega STe / 800 XL / 1040 STe / Falcon030 / 65 XE / 520 STm / SM124 / SC1435
DDD HDD / AT Speed C16 / TF536 / SDrive / PAK68/3 / Lynx Multi Card / LDW Super 2000 / XCA12 / SkunkBoard / CosmosEx / SatanDisk / UltraSatan / USB Floppy Drive Emulator / Eiffel / SIO2PC / Crazy Dots / PAM Net
Hatari / Steem SSE / Aranym / Saint
http://260ste.atari.org
czietz
Hardware Guru
Hardware Guru
Posts: 2451
Joined: Tue May 24, 2016 6:47 pm

Re: STe - Bad DMA Chip

Post by czietz »

dlfrsilver wrote: Sat Apr 22, 2023 11:31 am -why replacing only the DMA chip by a "good DMA chip" is not enough to fix the problem ?
(This is what i saw on my machines.....)
I can imagine that you're observing an entirely unrelated problem that also manifests itself in data corruption. For example, I've seen an Ultrasatan that repeatedly corrupted the data on the SD card regardless of the machine. (Also on STFs and a TT.)

I'd be interested in getting a logic analyzer trace from a supposedly affected machine with a C398739 DMA, containing at least the same signals as in this trace: https://www.chzsoft.de/site/site/assets ... ansfer.png. That should confirm that the issue is an unrelated one.
mikro
Hardware Guru
Hardware Guru
Posts: 4115
Joined: Sat Sep 10, 2005 11:11 am
Location: Kosice, Slovakia
Contact:

Re: STe - Bad DMA Chip

Post by mikro »

Excellent work, Christian.

I find it quite ironic that the explanation *really* is a bad DMA IC, despite all of the mocking and sarcastic comments (see Exxos' Phoenix PCB ;-)).
User avatar
MiggyMog
Atari Super Hero
Atari Super Hero
Posts: 986
Joined: Sun Oct 30, 2005 4:43 pm
Location: Scotland

Re: STe - Bad DMA Chip

Post by MiggyMog »

Great work Christian, nice to know there is a software work around.
('< o o o o |''| STM,2xSTFM,2xSTE+HD,C-Lab Falcon MK2+HD,Satandisk,Ultrasatandisk,Ethernat.
Post Reply

Return to “Hardware”