DMA ISSUE ON H4

Post by **exxos** » Mon Dec 23, 2019 11:56 am

btl wrote: ↑Mon Dec 23, 2019 9:58 am I'm sorry, i don't understand the problem & solution here.

You really would have to read the whole thread.. But in short, the DMA is sort of multiplexed between the HDD and FDD. Those two buffers control the switching. However I mistakenly assumed that HDCS was a inversion of FDCS , where czietz pointed out that it is only active on commands to the hard drive. So with the original H4 circuit, the hard drive would be recognised, but you would not be able to transfer any actual data to or from it. The floppy drive would however work perfectly fine. So to basically solve this problem, we remove the chip and jumper link over it. The inverter comes into play for inverting FDCS and re-adding the buffer back in.

I later did a fix PCB for this as shown below, but I do not think anyone has actually added this in yet.

https://www.exxosforum.co.uk/forum/viewt ... 721#p30306

PaulJ · Post by **PaulJ** » Mon Dec 23, 2019 2:38 pm

I went back and tested the ASCI port last night since it appears to be a 100% with the ultra satan. To my surprise it fails 100% of the time on the 1st pass. Error codes are meaningless and say to replace DMA, or GLUE. So I went back and tested the ASCI on the REMAKE board and the mega 4 I've been using as a development machine and they both pass.

Since it appears to be working fine I will revisit this issue after I finish the second H4. Just a heads up there's something a little different from a stock ASCI port.

Post by **exxos** » Mon Dec 23, 2019 2:49 pm

There is not much difference other than the added buffers... Currently as long as the hard drive buffer is jumpered over it will work.. And of course if you have a long cable then without the buffering it may malfunction.

PaulJ · Post by **PaulJ** » Mon Dec 23, 2019 2:54 pm

exxos wrote: ↑Mon Dec 23, 2019 2:49 pm There is not much difference other than the added buffers... Currently as long as the hard drive buffer is jumpered over it will work.. And of course if you have a long cable then without the buffering it may malfunction.

The Atari test board comes with a short ASCI cable. Very strange it gets the hard failure and all others all pass. Not at all concerned but will look into it after I finish the second H4. See what that ASCI port on that board tests like.

Post by **exxos** » Mon Dec 23, 2019 3:07 pm

PaulJ wrote: ↑Mon Dec 23, 2019 2:54 pm The Atari test board comes with a short ASCI cable. Very strange it gets the hard failure and all others all pass. Not at all concerned but will look into it after I finish the second H4. See what that ASCI port on that board tests like.

I've never had the test cable, but have you tried it on a actual STE to see if it works there ?

PaulJ · Post by **PaulJ** » Mon Dec 23, 2019 3:13 pm

exxos wrote: ↑Mon Dec 23, 2019 3:07 pm
PaulJ wrote: ↑Mon Dec 23, 2019 2:54 pm The Atari test board comes with a short ASCI cable. Very strange it gets the hard failure and all others all pass. Not at all concerned but will look into it after I finish the second H4. See what that ASCI port on that board tests like.
I've never had the test cable, but have you tried it on a actual STE to see if it works there ?

Never even seen a STE. Not many here in the US. I do have a MSTE and it works fine there, and on a Mega ST, and ST... and the Remake.

Post by **exxos** » Mon Dec 23, 2019 3:22 pm

PaulJ wrote: ↑Mon Dec 23, 2019 3:13 pm Never even seen a STE. Not many here in the US. I do have a MSTE and it works fine there, and on a Mega ST, and ST... and the Remake.

If the remake works then the H4 should, its the same circuit pretty much. With IC6 jumpered, its DMA IC goes direct to the DMA port like it does on the remake. Assume its a bad connection on the socket or iffy solder joint on it.

PaulJ · Post by **PaulJ** » Mon Dec 23, 2019 3:43 pm

exxos wrote: ↑Mon Dec 23, 2019 3:22 pm
PaulJ wrote: ↑Mon Dec 23, 2019 3:13 pm Never even seen a STE. Not many here in the US. I do have a MSTE and it works fine there, and on a Mega ST, and ST... and the Remake.
If the remake works then the H4 should, its the same circuit pretty much. With IC6 jumpered, its DMA IC goes direct to the DMA port like it does on the remake. Assume its a bad connection on the socket or iffy solder joint on it.

Well as you've most likely observed I'm the king of assumptions and I have a lot of injuries to show for all those assumptions. But the 19 pin ASCI connector on this H4 is reclaimed from a donor board and reused. The ultra satan connector does have pins that fit tighter than the cable on the test board. So a faulty connection is possible. If it was faulty solder I would "assume" the ultra would also be bad. Notice I didn't hasitate to assume.

So since everything that counts is working I'm going to wait for H4 #2 which will have a new 19 pin connector and see if it also fails the ASCI test.

Post by **exxos** » Mon Dec 23, 2019 3:54 pm

@PaulJ Your initial post seems a bit confusing, but to clarify, ultrasatan works, but the diagnostic cartridge doesn't ? I was confusing over it and reading that neither worked..

PaulJ · Post by **PaulJ** » Mon Dec 23, 2019 4:25 pm

exxos wrote: ↑Mon Dec 23, 2019 3:54 pm @PaulJ Your initial post seems a bit confusing, but to clarify, ultrasatan works, but the diagnostic cartridge doesn't ? I was confusing over it and reading that neither worked..

You are correct. The Ultra Satan works fine. The Atari test board fails while is passes on every other Atari machine I have. I'm going to complete the second H4 and see what happens there. If the second H4 also fails the DMA test I will look into it.

DMA ISSUE ON H4

Re: DMA ISSUE ON H4

Re: DMA ISSUE ON H4

Re: DMA ISSUE ON H4

Re: DMA ISSUE ON H4

Re: DMA ISSUE ON H4

Re: DMA ISSUE ON H4

Re: DMA ISSUE ON H4

Re: DMA ISSUE ON H4

Re: DMA ISSUE ON H4

Re: DMA ISSUE ON H4