ICD host adapter with extremelly strong pull-up !?

General discussions or ideas about hardware.
ijor
Posts: 428
Joined: Fri Nov 30, 2018 8:45 pm

ICD host adapter with extremelly strong pull-up !?

Post by ijor »

According to a recent reverse engineering of the ICD Advantage-Plus host adapter by Tori, one signal has a so extremelly strong pull-up that seems hard to understand.

This a link to the atari-forum post where he posted the schematics. https://www.atari-forum.com/viewtopic.p ... 12#p433212
There is a 100 OHM pull-up at the DRQ signal at the ACSI output connector (for daisy chaining another ACSI device or adapter). The DRQ signal at the Atari motherboard has an 1K pull-up. This is also the "official" value according to Atari documentation. Tori double checked and confirmed the presence of the 100 ohm pull-up value.

I am wondering how it could work with a pull-up so strong. The required sinking current (50ma !) would exceed the capabilities even for high current parts like the 7406 (specified for up to 40ma). Let alone that most ACSI devices don't use a high current part like the 7406. Even the ICD host adapter itself uses a standard GAL to drive that signal.

So not only I am wondering how it could work reliably, because in the best case it probably would be difficult to pull the signal low enough for a good noise margin. Couldn't this even damage the connected device?

Edit: Added the link
http://github.com/ijor/fx68k 68000 cycle exact FPGA core
FX CAST Cycle Accurate Atari ST core
http://pasti.fxatari.com
User avatar
exxos
Site Admin
Site Admin
Posts: 23496
Joined: Wed Aug 16, 2017 11:19 pm
Location: UK
Contact:

Re: ICD host adapter with extremelly strong pull-up !?

Post by exxos »

Could someone please upload the schematic as it's not viewable on AF to guests it seems.

Indeed 50mA is a lot even for more modern logic chips. I think it needs confirming by someone else that the value is 100R. It also should be measured on a scope the signal to see what the signal is switching between. I've seen odd things where undershoot on signals is the "logic low" when the actual driven low could be over 1volt.
https://www.exxosforum.co.uk/atari/ All my hardware guides - mods - games - STOS
https://www.exxosforum.co.uk/atari/store2/ - All my hardware mods for sale - Please help support by making a purchase.
viewtopic.php?f=17&t=1585 Have you done the Mandatory Fixes ?
Just because a lot of people agree on something, doesn't make it a fact. ~exxos ~
People should find solutions to problems, not find problems with solutions.
ijor
Posts: 428
Joined: Fri Nov 30, 2018 8:45 pm

Re: ICD host adapter with extremelly strong pull-up !?

Post by ijor »

exxos wrote: Wed Jun 29, 2022 6:44 pm Could someone please upload the schematic as it's not viewable on AF to guests it seems.
I forgot about that. Here it is:
Attachments
ICD Advantage Plus.pdf
(58.85 KiB) Downloaded 39 times
http://github.com/ijor/fx68k 68000 cycle exact FPGA core
FX CAST Cycle Accurate Atari ST core
http://pasti.fxatari.com
User avatar
stephen_usher
Posts: 5580
Joined: Mon Nov 13, 2017 7:19 pm
Location: Oxford, UK.
Contact:

Re: ICD host adapter with extremelly strong pull-up !?

Post by stephen_usher »

I wonder if anyone ever actually tried plugging anything into the OUT port.

It was probably supposed to be 1K ohm but someone got the decimal place wrong in the specs. :-)
Intro retro computers since before they were retro...
ZX81->Spectrum->Memotech MTX->Sinclair QL->520STM->BBC Micro->TT030->PCs & Sun Workstations.
Added code to the MiNT kernel (still there the last time I checked) + put together MiNTOS.
Collection now with added Macs, Amigas, Suns and Acorns.
User avatar
exxos
Site Admin
Site Admin
Posts: 23496
Joined: Wed Aug 16, 2017 11:19 pm
Location: UK
Contact:

Re: ICD host adapter with extremelly strong pull-up !?

Post by exxos »

stephen_usher wrote: Wed Jun 29, 2022 7:24 pm It was probably supposed to be 1K ohm but someone got the decimal place wrong in the specs. :-)
Ah so not the ST end being talked about then. Yeah if it's additional devices then it's possible it's simply wrong. It's not like many could afford one drive back in the day nevermind 2.
https://www.exxosforum.co.uk/atari/ All my hardware guides - mods - games - STOS
https://www.exxosforum.co.uk/atari/store2/ - All my hardware mods for sale - Please help support by making a purchase.
viewtopic.php?f=17&t=1585 Have you done the Mandatory Fixes ?
Just because a lot of people agree on something, doesn't make it a fact. ~exxos ~
People should find solutions to problems, not find problems with solutions.
Steve
Posts: 2570
Joined: Fri Sep 15, 2017 11:49 am

Re: ICD host adapter with extremelly strong pull-up !?

Post by Steve »

stephen_usher wrote: Wed Jun 29, 2022 7:24 pm I wonder if anyone ever actually tried plugging anything into the OUT port.

It was probably supposed to be 1K ohm but someone got the decimal place wrong in the specs. :-)
I concur with this assessment of the situation... it is similar to the whole 'bad dma' debacle. People forget that back when Atari was alive and kicking, their target audience was the low-cost consumer segment, the average home user didn't have a hard drive and most couldn't afford to buy one. So you can imagine that Atari didn't get many complaints about their ACSI issues compared to the amount of customers who had ST's in their homes. You can imagine how stuff like this got missed.
ijor
Posts: 428
Joined: Fri Nov 30, 2018 8:45 pm

Re: ICD host adapter with extremelly strong pull-up !?

Post by ijor »

stephen_usher wrote: Wed Jun 29, 2022 7:24 pm I wonder if anyone ever actually tried plugging anything into the OUT port.
It was probably supposed to be 1K ohm but someone got the decimal place wrong in the specs. :-)
Well, conceivable, but I doubt it. We are talking about the most popular manufacturer of ACSI host adapters, by far. And this is not their first generation of products. Furthermore, they advertised that their host adapters were specially designed for daisy chaining, supporting the two different ways of connecting two devices, in parallel, with an Y cable, or with daisy chaining at the OUT port. Hard to believe they didn't test that.

I actually connected a second device to my ICD host adapter, but that was an earlier generation, it might be different.
http://github.com/ijor/fx68k 68000 cycle exact FPGA core
FX CAST Cycle Accurate Atari ST core
http://pasti.fxatari.com
User avatar
exxos
Site Admin
Site Admin
Posts: 23496
Joined: Wed Aug 16, 2017 11:19 pm
Location: UK
Contact:

Re: ICD host adapter with extremelly strong pull-up !?

Post by exxos »

I only had a quick look at the schematic. It looks like pin19 are inputs to the GAL which doesn't make sense as pin19 on the ST is a input. Anyway, if ICD had other hardware then I'd have to assume they had chips capable of driving a 100R pull up.

In fact the 74F07 datasheet quotes 64mA output drive,128mA max. Not looked into it anymore as to tired now. Though maybe ICD used such chips in other hard drive interfaces. Or maybe just doubled up on LS gates in parallel.

Of course I've never seen any ICD stuff. So no idea what hardware they all had.
https://www.exxosforum.co.uk/atari/ All my hardware guides - mods - games - STOS
https://www.exxosforum.co.uk/atari/store2/ - All my hardware mods for sale - Please help support by making a purchase.
viewtopic.php?f=17&t=1585 Have you done the Mandatory Fixes ?
Just because a lot of people agree on something, doesn't make it a fact. ~exxos ~
People should find solutions to problems, not find problems with solutions.
ijor
Posts: 428
Joined: Fri Nov 30, 2018 8:45 pm

Re: ICD host adapter with extremelly strong pull-up !?

Post by ijor »

exxos wrote: Wed Jun 29, 2022 11:58 pm I only had a quick look at the schematic. It looks like pin19 are inputs to the GAL which doesn't make sense as pin19 on the ST is a input.
This is because the ICD board implements buffered daisy chain. The outputs of the daisy chained host adapter are inputs to this board.
Anyway, if ICD had other hardware then I'd have to assume they had chips capable of driving a 100R pull up.
ICD can't control what kind of device you'll daisy chain. As a matter of fact, it would be very unusual to daisy chain another ICD board. Normally what you'll connect to the output port might be an Atari hard disk, or more probably something that is not a hard disk at all, like a laser printer. So there is no way to tell which kind of chip would drive this signal.

In the meantime Czietz confirmed that his ICD board has indeed a 100 ohm pull-up. Furthermore, he commented that he actually daisy chained an Atari SH204 for years without problems! But according to schematics the SH204 uses 7406 drivers. The 7406, even when still out of specs in theory, probably can sink this strong pull-up. But other boards use other chips with much less sinking power than the 7406. Seems that typically most later boards use a PAL.

Btw, I think I understand why ICD used a very strong pull-up. I forgot that ICD has an (exclusive) buffered daisy chaining scheme. If you look at other boards, like the SH205, all the ACSI signals are connected in parallel to the output port. A daisy chained device in such case would drive directly the computer ACSI port. But these ICD boards buffer all the signals between the daisy chained device and the computer. Hope I'm explaining myself clearly enough ...

Buffering means delays, of course. So I am guess that the strong pull-up is to make the raising edge of the signal as fast as possible, otherwise the delay would be too long.
http://github.com/ijor/fx68k 68000 cycle exact FPGA core
FX CAST Cycle Accurate Atari ST core
http://pasti.fxatari.com
User avatar
exxos
Site Admin
Site Admin
Posts: 23496
Joined: Wed Aug 16, 2017 11:19 pm
Location: UK
Contact:

Re: ICD host adapter with extremelly strong pull-up !?

Post by exxos »

The GAL data sheets mentions short circuit current of 130mA. Possible it could affect timings a little as well. But I guess it would be a easy test for someone to hookup a GAL and code a open collector inverter or something and experiment with various load resistors around 100R and observe the high and low levels.

There could be peak power available but it's hardly mentioned. Some chips can give higher current pulses for a short time if it's not repetitive. Also mA ratings are guarantees. I had a long talk to maxim once about such things. Where they may well aim for 50mA for example, but in real world production it's likely more like half that value. So all chips could drive 25mA for example without problem. But it's out of guarantees.

I assume there is no series resistance in any of the cables as well. Might be a few ohms inherent in cables themselves but likely not enough to make any difference unless there is deliberate resistors added somewhere.

Also could be a factor that there may be a small voltage drop when a IC switches to zero. Also are the PSUs really delivering 5volts exactly. When a SCSI drive is running it could may well drop the 5V even lower. A few voltage drops here and there can add up. Of course it's only going to be a few mA, but it could be enough to to allow more ics to drive the signal, rated or otherwise.

But again, trying things on breadboard with chips will give a bit more information .
https://www.exxosforum.co.uk/atari/ All my hardware guides - mods - games - STOS
https://www.exxosforum.co.uk/atari/store2/ - All my hardware mods for sale - Please help support by making a purchase.
viewtopic.php?f=17&t=1585 Have you done the Mandatory Fixes ?
Just because a lot of people agree on something, doesn't make it a fact. ~exxos ~
People should find solutions to problems, not find problems with solutions.
Post Reply

Return to “HARDWARE DISCUSSIONS”