Turbo Freezer Stopped Working

[+Larry]

I have two TF units -- one for XL and one for XE.  I haven't used them for a couple of years, but I got them out yesterday to test.  The XE module has stopped working.  The XL unit works fine with either the PBI or ECI adapter.  Pressing the freeze switch give a solid black screen on my XE.  If I turn on the Cart Emu switch, that menu comes right up. The Old OS appears to work. I looked at the PCBs in the freezer, and looks like everything is soldered.

 

Any thoughts to help pinpoint what is going on?  I do have a system disk, so I can attempt to re-flash that portion of the system.

 

Edit: I reflashed the freezer software (banks 120-127) from my system disk, and now pressing the freeze switch results in nothing happening at all. (So no black screen -- if I'm in Dos, the screen doesn't change at all.)

I also noticed that in attempting to make a new system disk (from the working XL module), the disk drive (or image) does not format the disk -- it attempts to boot it giving a boot error.

 

-Larry

[Mclaneinc]

Probably something you have already done, giving its connector a good clean..

[+Larry]

Thanks, Paul.  I did clean the ECI and adapter and the pins on the module.  Unfortunately, no effect.

[Mclaneinc]

You should start getting some more technical replies about now, best of luck there...

 

Paul..

[HiassofT]

Can you try flashing the Freezer to (latest version) stock firmware with the production tool? https://www.horus.com/~hias/tmp/prodtool-200726.zip

 

This will wipe the whole flash memory and also perform some basic Freezer RAM tests.

 

Just set the flash write and ramdisk switch to "on", all other switches to "off" position and boot the ATR. If something's not quite right then you should get some error messages on screen - a screenshot of that could help narrowing down if there's some issue with flash or RAM.

 

so long,

 

Hias

[+Larry]

Thanks, Hias.  I will try that tomorrow.

[+Larry]

Hi Hias-

 

I performed the update with no error messages, but I still get a black screen when I press the freezer switch. The Cart Emu still will come up, and the Oldrunner OS still works. Since I used it for the update, the ramdisk seems fine.

 

Erase OK

Program OK

Verify OK

Then a normal exit from the program. 

 

That is my kind of update tool -- amazingly easy to use!

 

-Larry

[+Larry]

Update:

The first XE that I tried was stock 128K, no mods at all.  My other Turbo Freezer module comes right up with no issues on that XE.  Further, I've had this XE in service off-and-on for several years, and the "bad" module definitely worked with this XE in the past. (And the "good" TF module works with it right now.)

 

But I thought that I should try the "bad" module on another XE, so I got out another stock XE.  It is without any upgrades, but it has been 100% socketed (so not quite "stock").

 

No cleaning of the ECI, no nothing -- I plugged it in and booted the system and pressed the freezer switch -- the freezer menu comes right up.  Switched to the previous XE -- still a black screen.  ???  So something must be slightly different between the XE's, but something also must be different between the two modules.  So for now, I'll just leave the "good" XE in service, but I am puzzled by this one!

 

-Larry

[HiassofT]

Very puzzling. I had expected to see issues from the production tool as well - not sure why flash and RAM access seem to be fine but freezing doesn't work on this one XE model. I can't remember ever seeing such an issue combination.

 

A shot in the dark: if you have a scope it might be interesting to probe the PHI2 and the shortened PHI2 signals (the latter is on pin 13 of the HCT123), both on the working freezer and the non-working one when plugged into the problematic XE.

 

The falling edge of the shortened PHI2 should be about 50ns before the falling edge of PHI2.

 

so long,

 

Hias

Edited August 4, 2020 by HiassofT

[_The Doctor__]

marginal XE cpu?

marginal XE support chips ( the two in between the LSI's on the right hand side with the factory wires going to them)

[+Larry]

I tested the "bad" XE with several add-ons -- no problem with cartridges, IDE+2, or high speed SIO.  Everything worked as you would expect, but maybe this has turned out to be something relatively easy to fix.  I noticed that there was a slight boot-up color difference between the two XE's. I also noticed that the power switch on the "bad" XE didn't feel quite right when it was powered on -- kind of "sticky." So I sprayed some contact cleaner on the switch plunger and worked it in (several times).  Booted it up again and the screen color was the correct shade of blue.  Tried it several times and was OK each time. Plugged in the Turbo Freezer and booted the computer.  Pressed the TF switch and the freezer menu came right up!  If this "fix" only turns out to be temporary, I do have some new power switches, so I can replace it for hopefully a permanent fix. 

[HiassofT]

Ah, bad/insufficient power makes perfect sense, that's about the number one cause of most odd/unexplainable/intermittent issues.

 

so long,

 

Hias

[Mclaneinc]

Hi Larry, just glad that it's not a bad board...Glad it's working in some form as it should..

 

Paul...

 

 

[+Larry]

Thanks, Paul!  I'm still curious why originally one module worked and one didn't, but at least the Turbo Freezers both still work.  I later got out a 3rd XE, and it also worked with both TF units. 

[_The Doctor__]

slightly more power draw of one over the other or components that are slightly more tolerant of noise than another.... bottom line... power conditions matter and you solved it.

This is more common than you think... and often the last thing though of.

[HiassofT]

Simple, non-technical explanation: without a proper power supply anything can happen and it's pure luck if something works or not.

 

A bit more technical/detailed: glitches in the power rail may lead to shifts in the thresholds of the input-stage of circuits. Or internal circuitry in ICs, like flip-flops, may loose their state.

 

Glitches on the power rail also usually also affect the output stage of ICs and produce glitchy output. Input stages of other circuits, even if their power source is stable, react differently to those glitches as the exact switching point between low and high isn't specified but dependent on manufacturing processes and variations in silicon etc.

 

So, what possibly could have happened is that a short glitch on the power rail caused a glitch on the PHI2 signal which was "noticed" on one CPLD (your problematic freezer) and lead to double-clocking of the Freezer state machine, but not on the other CPLD. A glitch on the reset input at the wrong time could have caused similar fatal effects.

 

There are a lot of other effects that can happen due to short or longer power glitches and one common thing of these is that they are nasty and can be hard to diagnose from the software/logic side.

 

So, in case of odd issues (no matter if it's on a PC, an old Atari, a router or some other system) it's always a good idea to check all external and internal power rails with a scope. Switch-mode power supplies are the norm nowadays so a standard DMM won't do a job and may show a stable 5V or 3.3V supply. But a scope will reveal if there's ripple/glitches on the rail, eg due to aging capacitors.

 

so long,

 

Hias