BAD DMA - Myth or fact ?

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

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

User avatar
exxos
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 4933
Joined: Fri Mar 28, 2003 8:36 pm
Location: England
Contact:

BAD DMA - Myth or fact ?

Postby exxos » Fri Sep 11, 2015 10:52 am

I have started this thread so people can post everything they know about "Bad DMA" and why they think it is bad and experiences , in particular when "you" changed the DMA only to still have the same problems as before. This subject keeps coming up over and over. So lets have it out on this thread.

My results in a short-hand version are this:
I have several STE's all with the -38 DMA and work perfectly well with UltraSatan. I also tried the "new" DMA the 001 and it operated exactly the same.

The main issue I have had With UltraSatan (relating to what looks like DMA problems) is a "bad PSU". As I have suggested countless times, the DMA circuit is the first circuit to fail when the PSU is bad. If the PSU has bad capacitors (these are 20+ years old now!) then things will fail. I had exactly this issue while testing some things out for someone in another thread. Everything worked perfectly well for about a hour, then it rapidly got worse until I could no longer access the C: drive. A "updated" PSU and it ran stable for several hours after that. All this is with the -38 (old) DMA.

I am going to call the "BAD DMA" a myth until I actually see some machine or DMA which is actually faulty in some way. I have been asking for years for someone to come forward with a "bad DMA" chip but not 1 person has sent one. In fact I pretty much hear mixed views about the "good DMA" (001) about solving problems. This is understandable. IMHO the "good DMA" (001) probably has better noise immunity than the older DMA (-38). So while this may solve some peoples problems, once the PSU ages a little more, be it in a month or years time, it will start to fail again. So the "good DMA" to me is in no way a cure, but a "stop gap measure".

It is also possible (although I am not sure if it is even verifiable) that early STE's which are supposed to have the "Bad DMA" that the PSU was in those machines was bad to start with. I have done a lot of work on testing various PSU's over the past couple of years (http://www.exxoshost.co.uk/atari/last/psu/index.htm ) and found that various types have different regulation rates. In fact I wouldn't class any PSU of having particularly good regulation even after changing the capacitors. Capacitors years ago were not as good as they are today by far, So its not a far stretch to speculate that some PSU's were pretty poor even when manufactured. If such "poor PSU's" were produced on the early STE's with the older -38 DMA then this could may well explain the problem. Then Atari probably changed manufacture of the DMA chip and magically these DMA problems went away. Which could may well be the birth of the "BAD DMA MYTH". As mentioned before, the only time I had DMA problems was with a bad PSU. So its reasonable to assume it could be true.

Also there is a lot of issues with bad floppy drives appearing. This caught me out for some time. But seems to be also some odd DMA related problem. Information about this is posted here http://www.exxoshost.co.uk/atari/last/DMAfix/index.htm#2015 But these floppy issues seem to be something which has started to appear recently, which I am still investigating.

I think it would help is forum members post their experiences of "DMA problems" in this thread so we can build up a better "picture" of what is going on with these machines. Though in conclusion, for me at least, I have never had any "bad DMA" problems relating to the older -38 DMA in the STE. If someone thinks they have a bad DMA, get a updated PSU from me, or the re-cap kit and give that a try before spending big bucks on a "New DMA".

If anyone has some documentation as to where Atari said about this "Bad DMA chip" then please post it here. This "bad DMA chip" myth must have started somewhere. Though until someone can provide me with a "buggy DMA" I class this case as closed, that the "bad DMA" is a myth. Of course the -38 DMA had been used in the STFM for many years. So unless there was some bad batch of DMA chips then there should be no reason why the -38 shouldn't work in the STE's. Works perfectly in all my STE's.
4MB STFM 1.44 FD- VELOCE+ 020 STE - Falcon 030 CT60 - Atari 2600 - Atari 7800 - Gigafile - SD Floppy Emulator - PeST - various clutter

http://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
http://www.exxoshost.co.uk/atari/last/storenew/ - All my hardware mods for sale - Please help support by making a purchase.
http://ataristeven.exxoshost.co.uk/Steem.htm Latest Steem Emulator

User avatar
troed
Atari God
Atari God
Posts: 1424
Joined: Mon Apr 30, 2012 6:20 pm
Location: Sweden

Re: BAD DMA - Myth or fact ?

Postby troed » Fri Sep 11, 2015 11:28 am

FWIW

I have an STE that had the "bad" DMA revision. This STE has had its PSU switched out for a picoPSU. I'm positive it can't get much better than that.

I haven't had any issues with any device connected to ACSI - I very briefly at one point in time connected a Megafile and rescued data. Then I waited, due to all the horror stories, to use ACSI for anything else until I had swapped out the "bad" DMA for a good one.

However. I had a lot of issues with my HxC floppy emu. Every now and then, and it seemed to be related to how often I wrote to an image, the HxC would suddenly claim that the disk was corrupt. Inserting the SD-card into my Mac, duplicating the image, and then putting it back into the HxC often worked fine. Other times I just created a new image and worked with that instead.

All those problems disappeared when I switched out the "bad" DMA for the good one.

/Troed

User avatar
exxos
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 4933
Joined: Fri Mar 28, 2003 8:36 pm
Location: England
Contact:

Re: BAD DMA - Myth or fact ?

Postby exxos » Fri Sep 11, 2015 12:07 pm

Thanks or the info. There are 2 problems with the DMA circuit, hard drive and floppy. Strange thing is, I have issues with the floppy and not the hard drive. The -38 (bad dma ) I had to put pull ups on it to solve the floppy problems. Swapping to "good DMA" also solved the floppy problem.

I think the floppy is a transient supply problem. As each time the motor turns on, or the heads move, the 5V dips and you can see the white border on the monitor go dim each time the floppy heads move. Mix that into regulation problems, and it probably explains why the pull ups are needed on the -38 DMA. Overall it still suggests to me that the new DMA is slightly less susceptible to noise problems, in particular the 5V rail. It is possible as the chips have aged, the specs have drifted out enough to now cause problems. At least with the floppy drive.

I have got a good capacitor across the floppy drive 5V rail on the connector itself and see get the dimming problem on the video. I may try a small adapter and use the 12V rail and a 7805 regulator to power the floppy. That in itself could solve a lot of problems. Of course this is getting off topic of this thread a little as there is another thread dedicated to the floppy problems :) http://www.atari-forum.com/viewtopic.php?f=15&t=28449

I never got a PicoPSU to work, I went though a few of them, gave up in the end. Would be interesting to know what the regulation is like on them (if you have a scope) and also report back if the white border dims while the floppy heads are moving.
4MB STFM 1.44 FD- VELOCE+ 020 STE - Falcon 030 CT60 - Atari 2600 - Atari 7800 - Gigafile - SD Floppy Emulator - PeST - various clutter

http://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
http://www.exxoshost.co.uk/atari/last/storenew/ - All my hardware mods for sale - Please help support by making a purchase.
http://ataristeven.exxoshost.co.uk/Steem.htm Latest Steem Emulator

User avatar
troed
Atari God
Atari God
Posts: 1424
Joined: Mon Apr 30, 2012 6:20 pm
Location: Sweden

Re: BAD DMA - Myth or fact ?

Postby troed » Fri Sep 11, 2015 12:25 pm

exxos wrote:I never got a PicoPSU to work, I went though a few of them, gave up in the end. Would be interesting to know what the regulation is like on them (if you have a scope) and also report back if the white border dims while the floppy heads are moving.


No dimming at all - it's rock solid. Unfortunately I don't have a scope though.

/Troed

ijor
Hardware Guru
Hardware Guru
Posts: 3462
Joined: Sat May 29, 2004 7:52 pm
Contact:

Re: BAD DMA - Myth or fact ?

Postby ijor » Wed Sep 23, 2015 12:38 pm

IMHO the DMA chip is, by far, the worse design on the ST chipset. Because the ACSI bus is asynchronous (no clock on the bus), the chip seems designed with the goal to be as much independent on the ST clock as possible. The chip then uses a lot of async and ripple techniques. At the day it wasn't probably unheard. But today we know these techniques are calling for trouble.

AtariZoll
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 2978
Joined: Mon Feb 20, 2012 4:42 pm
Contact:

Re: BAD DMA - Myth or fact ?

Postby AtariZoll » Wed Sep 23, 2015 2:12 pm

I never seen or had chance to test "bad DMA" chip. but really don't see any reason to not believe to people who had problems, and solved it by chip replacement.

ijor wrote:IMHO the DMA chip is, by far, the worse design on the ST chipset. Because the ACSI bus is asynchronous (no clock on the bus), the chip seems designed with the goal to be as much independent on the ST clock as possible. The chip then uses a lot of async and ripple techniques. At the day it wasn't probably unheard. But today we know these techniques are calling for trouble.


Ijor, what you saying explains not why it was OK in ST machines, and suddenly became troublesome in STE. Only difference what I see is added bus driver chip in STE. Does it cause problems is what should be tested - with problematic DMA chip and with some old ST DMA,and with good STE DMA chip.
I can not do it, because I have no bad DMA. I mean by simple removing bus driver and shortcut I with O .
I don't think that ACSI bus needs clock. There is no clock on SCSI bus too.
DMA chip IMHO is very good design (talking about ST DMA), very fast for it's time - peak speed is 2 MB/sec for all versions - I tested in diverse ST, STE, TT - same in all. Some literature mentioning less.
Of course, this days, when machines are very old there are diverse problems with it. Probably mostly power related.
Famous Schrodinger's cat hypothetical experiment says that cat is dead or alive until we open box and see condition of poor animal, which deserved better logic. Cat is always in some certain state - regardless from is observer able or not to see what the state is.

User avatar
exxos
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 4933
Joined: Fri Mar 28, 2003 8:36 pm
Location: England
Contact:

Re: BAD DMA - Myth or fact ?

Postby exxos » Wed Sep 23, 2015 9:38 pm

So far I have had no issues with STE and -38 or 001A DMA. But TTL buffers generally float HI anyway, so they could be acting as a pull up which makes the -38 working on some machines. It could be that some other buffer chip was used in STE's which fail with DMA, different manufacture of chip for example.

Really regardless of machine, the -38 has this "bug" , it needs stable PSU for starters, same with 001A DMA. The 001A DMA only seems to work as loading the databus seems to extend the pulse width of the signals. So the more load on the databus pins, the "slower" the databus changes state. The -38 changes voltage, so no stable logic HI or LO. That is why it fails. As to why this problem only show up now, who knows.

So here is a little story to tax your brain...

I saw odd faults in audio hardware in the past. When I worked brefly as a audio engineer for a company, they had some power amplifiers which would work for 6-12months, then the right channel would fail. Same fault on ever amplifier which the company brought. nobody could figure out why. other engineers showed me they change the 8 mosfets in the channel, and on turn on, it would explode and catch fire. so nobody was able to fix them as they did not know why that happened.

What I found was a protection diode was in the circuit backwards, the silk screen print was actually wrong, so it was manufactured wrong. So on power up, this diode was pushing over 50volts into the mosfet gate which had a limit of 20volts I think. So when they changed the mosfet, 50volts into 20volts junction, and boom! I turned the diode around and solved the problem.

Point here, why did these amplifiers work for some months before the channel failed ? If they mosfets were changed (exactly same number and brand as original) they would explode instantly. Clearly it should have been impossible for these amplifiers to work for several months before failing, but they did.

<rant>
Incidentally I got no thanks for solving this problem ( and some other similar issues) and saving the company huge sums of cash. All I got was the argument with the manager as I spent 3 hours on something when apparently I shouldn't have spent that long on something. Same crap different day, also reason why I no longer work for any companies :)
</rant>
4MB STFM 1.44 FD- VELOCE+ 020 STE - Falcon 030 CT60 - Atari 2600 - Atari 7800 - Gigafile - SD Floppy Emulator - PeST - various clutter

http://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
http://www.exxoshost.co.uk/atari/last/storenew/ - All my hardware mods for sale - Please help support by making a purchase.
http://ataristeven.exxoshost.co.uk/Steem.htm Latest Steem Emulator

ijor
Hardware Guru
Hardware Guru
Posts: 3462
Joined: Sat May 29, 2004 7:52 pm
Contact:

Re: BAD DMA - Myth or fact ?

Postby ijor » Thu Sep 24, 2015 2:56 am

AtariZoll wrote:Ijor, what you saying explains not why it was OK in ST machines, and suddenly became troublesome in STE.


Of course it doesn't explain that. My post wasn't meant to explain any specific issue, obviously.

I don't think that ACSI bus needs clock.


I didn't say ACSI needs a clock, did I? I actually didn't say that anything in ACSI is inherently bad. I was only talking about the internal design of the chip.

DMA chip IMHO is very good design


Have you seen the internal logic of the chip? Because that's what I'm talking about.

Of course, this days, when machines are very old there are diverse problems with it. Probably mostly power related.


Agree 100%. And btw, it might be just coincidence, but ripple techniques are very sensitives to power rail issues.

User avatar
troed
Atari God
Atari God
Posts: 1424
Joined: Mon Apr 30, 2012 6:20 pm
Location: Sweden

Re: BAD DMA - Myth or fact ?

Postby troed » Thu Sep 24, 2015 7:32 am

I think it might be a red herring to focus on power issues as an explanation. My floppy-DMA problems with the -38 DMA were _after_ having switched out the internal PSU for an 80W picoPSU.

(Sorry for bringing this up over and over but I think it's important)

/Troed

User avatar
exxos
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 4933
Joined: Fri Mar 28, 2003 8:36 pm
Location: England
Contact:

Re: BAD DMA - Myth or fact ?

Postby exxos » Thu Sep 24, 2015 7:59 am

troed wrote:I think it might be a red herring to focus on power issues as an explanation. My floppy-DMA problems with the -38 DMA were _after_ having switched out the internal PSU for an 80W picoPSU.

(Sorry for bringing this up over and over but I think it's important)

/Troed


Input is always welcome :)

I documented power issues *somewhere* probably on here and on my site, that I had a STFM which after about a hour started to seriously have hard drive issues to the point where I couldn't even access the C: partition. After I swapped out the PSU for a re-capped one, the problems went away. But this still leaves the "floating data bus" problems which are not always there, but 90% of the time they are. This is why I suggest people re-cap their PSU for starters as the DMA circuit fails (also) because of poor regulation. Really nothing is going to work when the 5V rail is bouncing between 4-5volts.

As for those PicoPSU, I never got one working at all on my falcon. I have in line to design a new PSU for the ST/STE, but with a million other projects on the go.. its not high priority.
4MB STFM 1.44 FD- VELOCE+ 020 STE - Falcon 030 CT60 - Atari 2600 - Atari 7800 - Gigafile - SD Floppy Emulator - PeST - various clutter

http://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
http://www.exxoshost.co.uk/atari/last/storenew/ - All my hardware mods for sale - Please help support by making a purchase.
http://ataristeven.exxoshost.co.uk/Steem.htm Latest Steem Emulator

AtariZoll
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 2978
Joined: Mon Feb 20, 2012 4:42 pm
Contact:

Re: BAD DMA - Myth or fact ?

Postby AtariZoll » Thu Sep 24, 2015 8:17 am

ijor wrote:I didn't say ACSI needs a clock, did I? I actually didn't say that anything in ACSI is inherently bad. I was only talking about the internal design of the chip.
...
Have you seen the internal logic of the chip? Because that's what I'm talking about.
...

OK. Let's say just that you were not exactly clear about what is bad. We would like to see that internal logic. Until that, I can give here only my experiences. Internal logic design could help in possible development of CPLD clone.
Famous Schrodinger's cat hypothetical experiment says that cat is dead or alive until we open box and see condition of poor animal, which deserved better logic. Cat is always in some certain state - regardless from is observer able or not to see what the state is.

ChrisTG
Atari nerd
Atari nerd
Posts: 44
Joined: Sat Apr 30, 2011 2:57 pm

Re: BAD DMA - Myth or fact ?

Postby ChrisTG » Sat Sep 26, 2015 9:06 am

I did some measuring myself using a cheap USB-based logic analyzer yesterday. I can confirm the observations of exxos: The DMA-chips seem to have some fault in their design resulting in intermediate undefined (or tristate) signals on the data bus. If you study the rest of the schematic, you will see that such a behaviour makes absolutely no sense. It looks to me as if the whole curcuit of the DMA only works at all because of the capacity of the circuit pathes. These small "caps" act like a buffer keeping the signal on the recent level for a very short period of time. Maybe this "buffering" is enough delay for the rest of the chips to acknowledge the data correctly.
Now this awkward circuit design in conjuction with an aged power supply or any other worn electronic part in the whole schematic may then result in problems with the floppy and/or the ACSI bus.
From this point of view it absolutely makes sense that a pull-down or pull-up helps to allay the trouble, but it is a complete mystery to me why you need a different approach for -001 and -38.
I believe that some kind of buffering that acts like flipflops by keeping the signal on the last "valid" state would be the best solution here, though I don't yet have no idea how to build such a circuit.

User avatar
exxos
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 4933
Joined: Fri Mar 28, 2003 8:36 pm
Location: England
Contact:

Re: BAD DMA - Myth or fact ?

Postby exxos » Sat Sep 26, 2015 9:12 am

Thanks for the confirmation.

I did posted later that pull ups work ok on the 001. I did a re-test on that. I might have been trying my CF drive at the time, not the floppy drive, the CF drive is iffy anyway, so the pull down's on the 001A wasn't a fair test.

I did post similar information on my site also :) http://www.exxoshost.co.uk/atari/last/DMAfix/index.htm#2015 I need to tidy the page a bit, but will probably remove the pull down comments if its going cause confusion.

I don't think any "IC" solution is needed. Pull ups solve the problem, the IC solution would still need pull ups anyway. The DMA seems to have good pull down current, so 1K pull ups do not tax the DMA in anyway way. 1K is more than enough to drive everything else on the bus.

It is possible the DMA outputs are "open collector" types. I think that is the best way to look at it.
4MB STFM 1.44 FD- VELOCE+ 020 STE - Falcon 030 CT60 - Atari 2600 - Atari 7800 - Gigafile - SD Floppy Emulator - PeST - various clutter

http://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
http://www.exxoshost.co.uk/atari/last/storenew/ - All my hardware mods for sale - Please help support by making a purchase.
http://ataristeven.exxoshost.co.uk/Steem.htm Latest Steem Emulator

ChrisTG
Atari nerd
Atari nerd
Posts: 44
Joined: Sat Apr 30, 2011 2:57 pm

Re: BAD DMA - Myth or fact ?

Postby ChrisTG » Sat Sep 26, 2015 9:47 am

Hm, I still believe that some circuit that keeps the signals on the last valid state would make sense as it would also eleminate all that inexplicable timing issues that we had to fight against in the past when building new hardware for the ACSI bus. IIRC Jookie had a lot of trouble with the timing being too fast for the bus when he created the first revision of the Satandisk.

ChrisTG
Atari nerd
Atari nerd
Posts: 44
Joined: Sat Apr 30, 2011 2:57 pm

Re: BAD DMA - Myth or fact ?

Postby ChrisTG » Sat Sep 26, 2015 11:01 am

Btw, the "1v pulses" are crosstalk from a nearby line while the particular line is in "floating" state. So just another indication for a bad circuit design concerning the DMA stuff.

Edit: The signals generated by the GLUE seem to have the same strange "floating" states, though they already placed a pull-up on the RDY-line (with a clear understanding I think). Oh boy, we just opened pandora's box! 8-O
Last edited by ChrisTG on Sat Sep 26, 2015 11:30 am, edited 1 time in total.

User avatar
exxos
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 4933
Joined: Fri Mar 28, 2003 8:36 pm
Location: England
Contact:

Re: BAD DMA - Myth or fact ?

Postby exxos » Sat Sep 26, 2015 11:14 am

ChrisTG wrote:Btw, the "1v pulses" are crosstalk from a nearby line while the particular line is in "floating" state. So just another indication for a bad circuit design concerning the DMA stuff.


Yep , I really wish the internals of the ST chips were published, it would make life so much easier.
4MB STFM 1.44 FD- VELOCE+ 020 STE - Falcon 030 CT60 - Atari 2600 - Atari 7800 - Gigafile - SD Floppy Emulator - PeST - various clutter

http://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
http://www.exxoshost.co.uk/atari/last/storenew/ - All my hardware mods for sale - Please help support by making a purchase.
http://ataristeven.exxoshost.co.uk/Steem.htm Latest Steem Emulator

ChrisTG
Atari nerd
Atari nerd
Posts: 44
Joined: Sat Apr 30, 2011 2:57 pm

Re: BAD DMA - Myth or fact ?

Postby ChrisTG » Sat Sep 26, 2015 11:43 am

exxos wrote:Yep , I really wish the internals of the ST chips were published, it would make life so much easier.

Yes indeed, it would a lot. :roll:

I'm thinking about building a small pcb with a simple bus-holder circuit based of two inverters and a resistor per data-line, like described here: https://en.wikipedia.org/wiki/Bus-holder
What do you think?

User avatar
exxos
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 4933
Joined: Fri Mar 28, 2003 8:36 pm
Location: England
Contact:

Re: BAD DMA - Myth or fact ?

Postby exxos » Sat Sep 26, 2015 11:52 am

Possible that could work. Though you will have the same problem that I had with the ABT IC, even with a cmos input, it can't drive a logic HI, so a circuit like that would latch probably the wrong value, you would still need the pull ups on the data bus.

It is probably less of a problem on the STE as the bus is loaded with TTL chips which generally float HI anyway.
4MB STFM 1.44 FD- VELOCE+ 020 STE - Falcon 030 CT60 - Atari 2600 - Atari 7800 - Gigafile - SD Floppy Emulator - PeST - various clutter

http://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
http://www.exxoshost.co.uk/atari/last/storenew/ - All my hardware mods for sale - Please help support by making a purchase.
http://ataristeven.exxoshost.co.uk/Steem.htm Latest Steem Emulator

ChrisTG
Atari nerd
Atari nerd
Posts: 44
Joined: Sat Apr 30, 2011 2:57 pm

Re: BAD DMA - Myth or fact ?

Postby ChrisTG » Sat Sep 26, 2015 12:33 pm

I think I will try my approach with 3x 74ACT14 and 8x 1k resistors. Using inverting schmitt-triggers and 1k resistors looks like a good plan to me.
Ordering parts now, we'll see how it comes out when it is built and running. ;-)

ijor
Hardware Guru
Hardware Guru
Posts: 3462
Joined: Sat May 29, 2004 7:52 pm
Contact:

Re: BAD DMA - Myth or fact ?

Postby ijor » Sat Sep 26, 2015 1:06 pm

ChrisTG wrote:
exxos wrote:Yep , I really wish the internals of the ST chips were published, it would make life so much easier.

Yes indeed, it would a lot. :roll:


I did partial reverse engineering of most of the chipset internals. Already published internal schematics of the whole atari 8-bit chipset a few years ago. For the ST, I was waiting to finish the job before publishing. Unfortunately, I never had the time. And since then I was away from Atari for a few years.

Guess I should find the time to perform a minimal clean up and publish what I have at the state it is, never mind it is not complete. Sad to hear that we still don't get original Atari work (not reverse engineered). I am aware at least about two people that have something original, not sure exactly what ( schematics, layout, or what).

User avatar
exxos
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 4933
Joined: Fri Mar 28, 2003 8:36 pm
Location: England
Contact:

Re: BAD DMA - Myth or fact ?

Postby exxos » Sat Sep 26, 2015 9:03 pm

ijor wrote:I did partial reverse engineering of most of the chipset internals. Already published internal schematics of the whole atari 8-bit chipset a few years ago. For the ST, I was waiting to finish the job before publishing. Unfortunately, I never had the time. And since then I was away from Atari for a few years.

Guess I should find the time to perform a minimal clean up and publish what I have at the state it is, never mind it is not complete. Sad to hear that we still don't get original Atari work (not reverse engineered). I am aware at least about two people that have something original, not sure exactly what ( schematics, layout, or what).


Would be interesting to see nonetheless :)
4MB STFM 1.44 FD- VELOCE+ 020 STE - Falcon 030 CT60 - Atari 2600 - Atari 7800 - Gigafile - SD Floppy Emulator - PeST - various clutter

http://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
http://www.exxoshost.co.uk/atari/last/storenew/ - All my hardware mods for sale - Please help support by making a purchase.
http://ataristeven.exxoshost.co.uk/Steem.htm Latest Steem Emulator

User avatar
Cyprian
Atari God
Atari God
Posts: 1504
Joined: Fri Oct 04, 2002 11:23 am
Location: Warsaw, Poland

Re: BAD DMA - Myth or fact ?

Postby Cyprian » Sat Sep 26, 2015 10:28 pm

ijor wrote:I did partial reverse engineering of most of the chipset internals.

wow, how did you do that?
Jaugar / TT030 / Mega STe / 800 XL / 1040 STe / Falcon030 / 65 XE / 520 STm / SM124 / SC1435
SDrive / PAK68/3 / CosmosEx / SatanDisk / UltraSatan / USB Floppy Drive Emulator / Eiffel / SIO2PC / Crazy Dots / PAM Net
Hatari / Steem SSE / Aranym / Saint
http://260ste.appspot.com/

ijor
Hardware Guru
Hardware Guru
Posts: 3462
Joined: Sat May 29, 2004 7:52 pm
Contact:

Re: BAD DMA - Myth or fact ?

Postby ijor » Mon Sep 28, 2015 8:45 pm

Cyprian wrote:
ijor wrote:I did partial reverse engineering of most of the chipset internals.

wow, how did you do that?


Google for decap, chip reverse engineering.

SteveBagley
Captain Atari
Captain Atari
Posts: 165
Joined: Mon Jan 21, 2013 9:31 am

Re: BAD DMA - Myth or fact ?

Postby SteveBagley » Tue Sep 29, 2015 9:36 pm

exxos wrote:If anyone has some documentation as to where Atari said about this "Bad DMA chip" then please post it here. This "bad DMA chip" myth must have started somewhere. Though until someone can provide me with a "buggy DMA" I class this case as closed, that the "bad DMA" is a myth. Of course the -38 DMA had been used in the STFM for many years. So unless there was some bad batch of DMA chips then there should be no reason why the -38 shouldn't work in the STE's. Works perfectly in all my STE's.


Certainly, Atari were quoted in the press at the time attributing the fault to the DMA chip (see the attached bit from ST Format Issue 19). I suffered with a bad DMA in my STe and had to have it replaced in 1996 when we got a HDD (you can read the details in an 'Atari Computing' article I wrote) but before it was fine with floppies (that machine got absolutely hammered back then :)). Soon after getting the HDD, I started to get intermittent data corruption…

After I'd had the chip replaced, the machine (and hard drive) were fine for the next two years or so before I moved to a PC.

Steve

atari-st-format-issue-019-137.jpg
You do not have the required permissions to view the files attached to this post.

User avatar
exxos
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 4933
Joined: Fri Mar 28, 2003 8:36 pm
Location: England
Contact:

Re: BAD DMA - Myth or fact ?

Postby exxos » Tue Sep 29, 2015 10:10 pm

SteveBagley wrote:
Certainly, Atari were quoted in the press at the time attributing the fault to the DMA chip (see the attached bit from ST Format Issue 19). I suffered with a bad DMA in my STe and had to have it replaced in 1996 when we got a HDD (you can read the details in an 'Atari Computing' article I wrote) but before it was fine with floppies (that machine got absolutely hammered back then :)). Soon after getting the HDD, I started to get intermittent data corruption…

After I'd had the chip replaced, the machine (and hard drive) were fine for the next two years or so before I moved to a PC.


Thanks for posting that Steve, I will probably put that file on my site :)

I don't really get why it would only be a issue on the STE when the same DMA was used in the STFM. I have a -38 in my STE and it works fine, it doesn';t work in my STFM's though. I suspect the STE buffers act as pull ups so it works better, but probably a few other issues involved elsewhere also.

Do you know which AC issue it was in ?
4MB STFM 1.44 FD- VELOCE+ 020 STE - Falcon 030 CT60 - Atari 2600 - Atari 7800 - Gigafile - SD Floppy Emulator - PeST - various clutter

http://www.exxoshost.co.uk/atari/ All my hardware guides - mods - games - STOS
http://www.exxoshost.co.uk/atari/last/storenew/ - All my hardware mods for sale - Please help support by making a purchase.
http://ataristeven.exxoshost.co.uk/Steem.htm Latest Steem Emulator


Social Media

     

Return to “Hardware”

Who is online

Users browsing this forum: No registered users and 9 guests