jmacz Mac IIfx Project

[Chuckdubuque]

Those of us with 3d printer’s would love that model too.

 

[Chuckdubuque]

Ran into a problem today. Booting is working perfectly after the fixes detailed earlier. But post boot (after reaching the Finder), I was getting random crashes and lock ups. As I wrote down the pattern (when it happens), it seems it's only during disk IO. I'm using a ZuluSCSI Mini connected via the external SCSI port, no internal SCSI device.

I tried swapping memory, video cards, etc to rule them all out and none had an effect on the issue.

I then removed the ZuluSCSI Mini and connected a ZuluSCSI 1.1 via the internal SCSI connector and used the same exact disk image I was using on the external ZuluSCSI Mini. No problems. Kept trying to repro but could not, even slamming it with disk operations.

The ZuluSCSI Mini has permanently enabled termination, but not sure how this relates to the need for a special IIfx terminator.

I then tried providing auxiliary power to the ZuluSCSI Mini (via USB) and that seems to work and stabilized things. For the internal ZuluSCSI, I had used aux power as well.

EDIT: spoke too soon, still getting the issue even with aux power. Just with external devices though.

Did you try formatting the SD card with SD Card Formatter. That has solved a lot of these problems for me.

 

[jmacz]

Those of us with 3d printer’s would love that model too.

Of course. I was planning to post it here publicly.

Did you try formatting the SD card with SD Card Formatter. That has solved a lot of these problems for me.

What is SD Card Formatter? The card was formatted Fat32. And was working fine in all other Macs except this one. But as I mentioned more recently, I can no longer repro, which doesn’t mean it’s fixe, but rather I will have to debug when it shows up again.

 

[Chuckdubuque]

Of course. I was planning to post it here publicly.



What is SD Card Formatter? The card was formatted Fat32. And was working fine in all other Macs except this one. But as I mentioned more recently, I can no longer repro, which doesn’t mean it’s fixe, but rather I will have to debug when it shows up again.

Cool!

Sdcard formatter is a tool that is supposed to better align a raw format to the peculiarities of an SD card. https://www.sdcard.org/downloads/formatter/

 

[jmacz]

Here's the model for the modified cover for the MeowToast. Let me know if it works for you.

 

[macuserman]

I can share the model when I get home tonight. Or are you saying you don’t have access to a 3D printer even if I share the model?

I don't have a 3d printer, but I realized the macbatt is really not that expensive so I just ordered a bunch more of those instead, I'll give the meowtoast away to someone or something not worth going to the extra trouble. I do appreciate it though!

 

[micheledipaola]

Finished putting the IIfx back together today.

I have dual CR2032 batteries installed, using two MeowToast's and a custom cover I made that holds the MeowToast better. Without the little support, the MeowToast keeps moving around so this locks it in.

View attachment 61438

I tried with a similar button battery holder, the one keeping batteries horizontally /parallel to the logic board, and my IIfx would not power up, while oldstyle batteries worked. I thought it was because button batteries in fact provide a little less power, but your experience seems to prove me wrong... so now I wonder how I could make that work. It doesnt seems to move, so maybe you have some advices to share? Thanks in advance

 

[jmacz]

For my IIfx, it's still on the stock power supply (I have not yet converted it to ATX and don't plan to until it fails -- and I'm told these IIfx power supplies are much more robust than the Mac II power supplies). I'm using two CR2032 batteries which provide 3V each. The stock 1/2 AA batteries provide 3.6V each so they do provide more power. But my IIfx starts just fine with the CR2032 batteries.

I just went and checked what voltage I'm getting from the two CR2032s. With the power unplugged to the IIfx power supply, I placed my multimeter probes on pin 15 on the PSU<->Motherboard connector and ground (one of the other ground pins) and then pressed the power button to see how much is being sent to the PFW pin 15 to signal the power supply to start. I saw around 4.3V to 4.4V. So it's not using all of the power the CR2032s have so I would think it should be fine (and on my system, it works fine).

I'm curious what you're seeing on pin 15 if you do the same experiment with the normal 1/2 AA batteries and your CR2032s.

 

[jmacz]

BTW, I was just checking Guide to Macintosh Family Hardware from Apple (second edition which includes the IIfx) and found this in the power supply section for the II/IIx/IIfx:



And here's Figure 6-7:



So it looks like Apple's expecting only 3V per battery and note that based on the text blurb above, only 3V is actually necessary to signal to the power supply to turn on. So two CR2032 coin batteries should clearly be enough.

Again, I think it would be good to see what you get measuring the voltage on pin 15 (PFW) when you press the power button and you have the coin batteries in place. That might help at least figure out the problem. If you have the proper voltage there, perhaps there's something going on with the power supply. If there's not enough voltage there, something must be going on with the power-on circuit.

Here's the pin reference for the PSU connector in case you need it:

 

[jmacz]

I have a gremlin in my IIfx.

I have had a SuperMac Spectrum 24 PDQ in my IIfx for a while now driving a NEC 52v. Also had a SuperMac Spectrum 24 III in the same IIfx driving a second NEC 52v. It was working great with dual 1024x768 screens at 24bit color.

Today I removed both video cards to replace the fan on the power supply since the stock one is like a jet engine. I replaced it with a Noctua and it’s so much quieter now.

After putting the power supply back in, I only put the SuperMac Spectrum 24 PDQ in. That’s when I started getting some weird behavior.

The NEC 52v (running at 24bit 1024x768 at 75Hz) started flickering whenever I move my mouse. The flickering wasn‘t the video going black but actually the sync on the 52v being lost. If I move the mouse quickly, it loses sync long enough to cause the monitor to show the ”no signal” message on the display. If I don’t touch the mouse, display is good. I tried the second NEC 52v, same issue, so it’s not the monitor. And this same video card had been working fine before in my IIfx. Disabling extensions did not help either.

I then took the video card out and replaced it with a SuperMac Spectrum 24 IV (had not been used in this Mac). Everything started working again.

I then placed the Spectrum 24 PDQ into a different computer, and the card with the same display (NEC 52v) works fine with no problems.

Not sure what mouse movement would cause an issue? Tried with a different mouse, same problem. Tried the second ADB port, same problem.

Note this computer had that problem earlier (and still continues to have this issue) where external ZuluSCSI mini experiences random IO errors if it’s powered by SCSI bus power. But works fine if I provide external power to the ZuluSCSI mini.

Feels like this one is going to be a pain to debug.

 

[bigmessowires]

This sounds like an EMI issue where ADB activity is coupling to the PDQ's sync outputs and creating enough noise to confuse the monitor. The III may have been physically acting as an EMI shield, and when you removed it the problem appeared. Can you try moving the PDQ to a different slot? Can you reproduce something similar if you press and hold a keyboard key? Or it might not be ADB directly causing the EMI problem, but CPU activity related to updating the mouse cursor. I once had a computer that made and audible squealing noise from the speaker whenever the mouse was moved.

 

[jmacz]

The issue does happen with keypresses as well.

I tried moving the PDQ around to other slots. Same issues, doesn't get better. Move the PDQ into my Quadra 700, the card works fine. Move a S24 IV into the IIfx, it works fine. Just the PDQ in the IIfx is having issues. Also tried moving the S24 III back into the same slot "shielding" the PDQ and still happening at this point. That doesn't help either.

Found something though... I pulled the IIfx power supply (it's completely stock except the Noctua fan I put in yesterday). And I replaced it with my ATX converted power supply from my Mac II. No issues. PDQ works great. And so far, it looks like the ZuluSCSI mini without additional external power is also working fine but haven't played with it long enough.

Possible something is up with the stock power supply.

 

[3lectr1cPPC]

Sounds like it's running a little weak. Has it been recapped?

 

[jmacz]

Sounds like it's running a little weak. Has it been recapped?

Nope, I had left it alone. Looks like it's time to recap or convert to an ATX.

 

[volvo242gt]

Looks like it. Had the same Spectrum 24PDQ in my previous IIfx, which was running an old Astec Mac II power supply. Card worked fine in that computer.

-J

 

[jmacz]

Been hammering the system with disk IO for a few hours now (using the ATX PSU from my Mac II) and cannot repro the external zuluscsi mini issue. So hopefully all of my remaining issues with my IIfx come down to the power supply.

I'm just going to convert the PSU to ATX.

I had used the Coolmax I-400 ATX PSU on my Mac II. I had thought it only had 0.8A on the -12V rail but looking up the specs on Coolmax's website, it seems to indeed have 1A on the -12V rail which is great. So comparing specs:

Voltage	Stock PSU	Coolmax I-400 PSU
5V	18A	21A
12V	2.5A	25A
-12V	1.0A	1.0A

The PSU is only $25. So only a bit more than getting new caps + mouser's shipping fee. And I don't have to worry about something else going wrong since it's a new power supply.

 

[zigzagjoe]

Been hammering the system with disk IO for a few hours now (using the ATX PSU from my Mac II) and cannot repro the external zuluscsi mini issue. So hopefully all of my remaining issues with my IIfx come down to the power supply.

I'm just going to convert the PSU to ATX.

I had used the Coolmax I-400 ATX PSU on my Mac II. I had thought it only had 0.8A on the -12V rail but looking up the specs on Coolmax's website, it seems to indeed have 1A on the -12V rail which is great. So comparing specs:

Voltage	Stock PSU	Coolmax I-400 PSU
5V	18A	21A
12V	2.5A	25A
-12V	1.0A	1.0A

The PSU is only $25. So only a bit more than getting new caps + mouser's shipping fee. And I don't have to worry about something else going wrong since it's a new power supply.

I'm going to be real with you: if you care about the machine I wouldn't recommend using that unit, or other cheap PSUs.

Cheap PSUs are typically worse at regulation (ripple), and machines of this vintage are more sensitive to that than modern machines (as they re-regulate most rails anyways). Additionally, that PSU is likely a simple group regulated type which means the voltage output on individual rails are roughly tracked together corresponding to the load on 12v rail. Thus, they rely on a significant 12v load. These machines do not use 12v much, so it is likely that 5v will go out of spec.

Additionally, there's the more typical quality concerns: I wouldn't put that in anything I would care about if it died and took the machine with it. On a rareish vintage machine? Honestly, I'd trust the original PSU uncapped more.

 

[jmacz]

Additionally, that PSU is likely a simple group regulated type which means the voltage output on individual rails are roughly tracked together corresponding to the load on 12v rail. Thus, they rely on a significant 12v load.

Thanks. I had not considered this.

I have a few PSUs that continue to misbehave even after recapping thus wanted to avoid the hassle of debugging further. I will try recapping the IIfx power supply but need to find a replacement ATX unit for my Mac II. I guess I am looking for a PSU with independent regulation or DC to DC, with 1A on the -12V rail.

Do you know if the Seasonic SS or SSPs that folks are using as replacements (IIci and SE/30) are not group regulated? Can’t tell from the specs. eg. Seasonic SS250-SU or SSP-300SUB.

 

[Bolle]

Do you know if the Seasonic SS or SSPs that folks are using as replacements (IIci and SE/30) are not group regulated? Can’t tell from the specs. eg. Seasonic SS250-SU or SSP-300SUB.

They are group regulated as well. For the most part the SS250 in my testbench SE/30 is fine but the 5V line sags a little bit below 5V with PDS card and accelerator in place.
It does get problematic for me with 2 PDS cards and an accelerator to the point that the machine stays in reset on power-on.

While I wouldn’t say I trust an old Sony power supply any more than a cheaper modern one, the old one can at least be set up for a nice and healthy voltage on the 5V line (with the 12V getting a bit out of spec, but I that hasn’t caused any issues for me so far)

 

[zigzagjoe]

They are group regulated as well. For the most part the SS250 in my testbench SE/30 is fine but the 5V line sags a little bit below 5V with PDS card and accelerator in place.
It does get problematic for me with 2 PDS cards and an accelerator to the point that the machine stays in reset on power-on.

While I wouldn’t say I trust an old Sony power supply any more than a cheaper modern one, the old one can at least be set up for a nice and healthy voltage on the 5V line (with the 12V getting a bit out of spec, but I that hasn’t caused any issues for me so far)

Yeah, the adjustable 5v line is big. In my SE/40, with 2 PDS cards + 040 accelerator and a HDD, it sits at a depressing ~4.8v at the LB with a SSP-300SUB. This is somehow fine, but if I introduce an extension cable for testbed work it gets real sensitive to how well the molex connectors are making contact. In a worst case scenario, it drops down to 4.6 - 4.5v with the extension.

SSP-300SUG seems to have independent rails, but it is not fanless at low loads and you'd need to stick your own 7912 regulator in there to have a -12v rail with enough power.

Another option to consider is SFX PSUs - more variety of quality PSUs available in that space while having a convenient form-factor. Either way though, I'd not skimp on the PSU: I saw too many cheap ones die and take components with them in another life when I was building PCs for a living.