exxos's Atari, Amiga & retro forum

sporniket wrote: 18 Apr 2022 10:36 If I understand things correctly, with the "1772 pull-up fix" and using a drive with internal pull-ups as well, the overall pull-up will be stronger -because one would ends up with 2 pull-ups in parallel for each line-.

@DoG , very likely a typo indeed, looking at the pictures I think the resistors on the pictures of the fix, I believe it is [grey/black/black/brown] i.e. 8K (because decoding in reverse direction would give [red/brown/black/black] i.e. 210.

I think @DoG was referring to the SIL resistor network on the underside of the board where stween says:

"I've thrown in a 1k resistor network. I've done the same as far as I can tell from what I see in other photos."

This should be a 10K SIL and not 1K according to the mandatory fixes page.

@stween Can you confirm if the SIL resistor pack on the underside is 1K or 10K?

sporniket wrote: 18 Apr 2022 10:36 If I understand things correctly, with the "1772 pull-up fix" and using a drive with internal pull-ups as well, the overall pull-up will be stronger -because one would ends up with 2 pull-ups in parallel for each line-.

@DoG , very likely a typo indeed, looking at the pictures I think the resistors on the pictures of the fix, I believe it is [grey/black/black/brown] i.e. 8K (because decoding in reverse direction would give [red/brown/black/black] i.e. 210.
...

I'm a bit lost too I think. These pull ups you describe here is after the 7406 and before the floppy cable. But the fix that he was referring to was done on the data lines between WD1772 and DMA? Different side of the WD1772.

My experience with a similar issue on one of my STFM was that in the end is was a weak DMA chip.

I did try changing the address and data bus pull-ups but that just made things worse. I tried the WD1772 fixes and that made no difference.

Swapped out the DMA chip and the problem went away.

With the weak DMA chip the floppy wouldn't read at all if the UltraSATAN was plugged in (though it partly worked with other ACSI devices, such as an ICD AdSCSI host adapter).

First things first!

stween wrote: 17 Apr 2022 16:56 Also I opted to try the DMA pull-up fix mentioned up-thread. On the bottom of the board, I've thrown in a 1k resistor network. I've done the same as far as I can tell from what I see in other photos.

1 - As said, 1K is way too strong, it should be 10K (or may be 4.7K, if you want) ! If this is not a typo, please replace those resistors before doing anything else.

2 - Can you replace the TOS chips with a later version of TOS, even TOS 1.02. This will probably not fix the problem, but because of a bug that TOS 1.0 has, the behavior might change and will give us more clues about the problem. Conceivable, even when not likely, your TOS chips might be corrupt. But this is not the main reason I recommend to try a later TOS.

2 - Again, it might be a good idea to test a different DMA chip.

3 - What pullups do you have at the CPU buses?

ijor wrote: 18 Apr 2022 12:49 1 - As said, 1K is way too strong, it should be 10K (or may be 4.7K, if you want) ! If this is not a typo, please replace those resistors before doing anything else.

How are you coming to that conclusion ?

Taken from the 7406 DS..


40mA output drive. 1K is only 5mA.

Taken from Atari's own schematic they are using 1K on the 1772 directly. So where's the problem ?

exxos wrote: 18 Apr 2022 13:11

ijor wrote: 18 Apr 2022 12:49 1 - As said, 1K is way too strong, it should be 10K (or may be 4.7K, if you want) ! If this is not a typo, please replace those resistors before doing anything else.

How are you coming to that conclusion ?
Taken from the 7406 DS..

Exxos, we are talking about the pull-ups on the DMA bus. Not about the pull-ups on the 7406 output.

Taken from Atari's own schematic they are using 1K on the 1772 directly. So where's the problem ?

But these pull-ups are at the the 1772 inputs pins, not at the output pins. Anyway, again, I was talking about pull-ups on the DMA data bus that would involve the DMA chip (and the FDC).

ijor wrote: 18 Apr 2022 13:18 Exxos, we are talking about the pull-ups on the DMA bus. Not about the pull-ups on the 7406 output.

I was reading the posts where Dog was talking about the 1772 pullup so assumed the posts after were still relating to the 1772 as it mentioned 1K. I did not realise you had gone back to the DMA resistors. My bad sorry!

But where are you reading that the DMA resistors are 1K ? AFAIk I never mentioned that. Indeed that post would need correcting if that is the case.

You are are indeed correct that the "DMA pullups" need to be 10K. The 1772 pullups should be 1K. Using any different values on any fixes , is effectively performing an experiment where the results are unknown.

exxos wrote: 18 Apr 2022 13:28 I was reading the posts where Dog was talking about the 1772 pullup so assumed the posts after were still relating to the 1772 as it mentioned 1K. I did not realise you had gone back to the DMA resistors. My bad sorry!

No worries. There are lots of pull-ups fixes involved here, and we are all getting a little confused :)

But where are you reading that the DMA resistors are 1K ? AFAIk I never mentioned that. Indeed that post would need correcting if that is the case.

No, it wasn't your mistake. It is the OP that, apparently, installed 1K pull-ups at the DMA bus by mistake (if it wasn't a typo when he posted).

rubber_jonnie wrote: 18 Apr 2022 10:35 Just to recap, have you tried another DMA yet

It's one thing I'd like to try, but I don't have a spare here without removing one from an STE :shock:

Also, have you got an oscilloscope, and if so have you scoped the signals at the WD1772 during diag cart tests and then at the FDD to see if they match?

Also something I don't have; I'm at the max of what my tools can do for me! Might be my reason to finally pick one up though.

DoG wrote: 18 Apr 2022 11:04

sporniket wrote: 18 Apr 2022 10:36 If I understand things correctly, with the "1772 pull-up fix" and using a drive with internal pull-ups as well, the overall pull-up will be stronger -because one would ends up with 2 pull-ups in parallel for each line-.

@DoG , very likely a typo indeed, looking at the pictures I think the resistors on the pictures of the fix, I believe it is [grey/black/black/brown] i.e. 8K (because decoding in reverse direction would give [red/brown/black/black] i.e. 210.
...

I'm a bit lost too I think. These pull ups you describe here is after the 7406 and before the floppy cable. But the fix that he was referring to was done on the data lines between WD1772 and DMA? Different side of the WD1772.

rubber_jonnie wrote: 18 Apr 2022 10:45 This should be a 10K SIL and not 1K according to the mandatory fixes page.

@stween Can you confirm if the SIL resistor pack on the underside is 1K or 10K?

Yeah I haven't done anything to the 7406.

For the 1K I looked at the page on the bad DMA investigation and based on the text at the top about a 1K SIL, I opted for that. (Now I see that this post says 10K, so I can find time later in the week to pick that back out and try again!)