X68000 Compact - Stand-by power-on non linear signal

[wild_potato]

Hi!

I've got a X68000 Compact (not mine but a friend').
PSU died. We didn't try to fix it, we found a new and standard slim-PSU which could fit exactly with a few mod (FSP-180).
Easy thing with what is in the wiki
I would have liked to give contribution to the Wiki with a full example of replacement for the Compact
...but I don't know how, I run into troubles. The standby mode is not working properly now.

I tried the PSU on a breadboard and connect all 5V/12V/-12V and system was starting at first.
I worked on the inverter and the power-off delay circuit days after, based on a NE555 and the 74LS04.
I wanted to try the new circuit with the 5Vsb / PwrOn / Gnd wire connected, but my PSU fan started to turn a little every second. I didn't understand at first but something has changed, and I must have damaged something.

The standby mode (i.e. power button in release state) is not working properly anymore.
IO diagnostic: PwrOn output is unstable and oscillate every second.
Curve is like a discharging capacitor or something?


Concretely what did I do to get in this situation? Could have happened anytime, I must have misconnected 5V to something, but I don't have any clue
So... how to find what's wrong now without schematics? Hard mission I know.

Fortunately the standby mode is not a big part of the Compact board.
I mesured the nearly exact same voltage and frequentcy on the !ALARM pin of the RF5C15 RTC chip. Which is why the issue is 1Hz I guess.


Of course, do measurement on a big circuit is always not representative of what is wrong inside as you know.
I mesured this kind of signal in some other test points and the power button for instance.


The only troubleshoot repair page which could help me is here
https://gamesx.com/wiki/doku.php?id=x68000:soft_power_switch_troubleshooting_and_repair
But I don't know where to search which component is reponsible of this. A cap? The RF5C15?

Does someone have encountered one day this kind of signal instead of a linear one?
I imagine so the PwrOn signal should be constant and near from linear 4-5V instead of this unstable signal (as it is inversed then).
Anything would be helpful.

Many thanks to the community

[caius]

Why did you use also an NE555?I always successfully used only an inverter (74LS04 or 74HC04) following this guide:

https://nfggames.com/nfg/x68repair/

My advice is to try this circuit along with another PSU, also a simple ATX one just for testing.If you still have issue then you will have narrowed the fault on the Compact motherboard so you can try to troubleshoot it.IN particular check if all the conditions for the POWER ON output are satisified.You can launch the X68k Power Test utility and analyze the results.It could be a bad IC (74LS08 or 74LS244) like it happened to my X68000 ACE HD.

[wild_potato]

Thank you Calus,
I went wrong with the 555 I imagined something complex by "Also, you need a couple of small parts to engage a timed poweroff that X68k PSUs employ" (in my mind timed ment delayed...). So I though the final signal transition had to be delayed but of course this is useless.

Anyway I removed this part definitively and I still have the weird signal problem and I can't start the Compact 
Depending the power button
- state power off: PwrOn signal is 0V with weird 5V peaks every second like on my previous screens. With inverted signal, this makes the PSU start / stop every second.
- state power on: PwrOn signal is continuous 5V. The reversed signal is 0V, PSU is on.

Should be a dead IC definitely :/
So I continued to inspect around the RF5C15.
- pin VCC (18) is 4.10V
- pin OSCOUT (17) is alternating (2V)
- pin OSCIN (16) is alternating (1.4V)
- !ALARM output is measured (in the circuit) as the weird 0V with 4.10V peaks every second
- other pins are near to 0v when in sleep mode

The RF5C15 documention can be downloaded here. What is weird with my measures:
- "The ALARM outputs alarm signal and 16Hz and 1Hz clock pulses. This pin is an open drain output."
An open drain output means I should have Gnd or Vcc I guess
- OSCOUT is documented to work with a 32.768kHz quartz, I measure 50Hz, why? 

So I cut the pin 15 !ALARM pin to see the behaviour


-- switch power off state, pin 15 cut off --
!ALARM is oscillating ~20mV values
I measure 4.9V after pin 15
PwrOn 14.5mV => 4.53V after the 74LS04 !not, and PSU is always off, good!

-- switch power on state, pin 15 cut off --
171mV direct after pin 15
PwrOn 1.25V => 40mV after the 74LS04 !not, and PSU is always on, good!

Now I cut off !ALARM I can see what really happen, and I'm not sure to understand
Why does my RF5C15 !ALARM output this? Is it normal?


Ahah, so now my Compact can start and stop my PSU properly, but I imagine this won't work with !alarm cut off
(I did not try with a monitor, I will soon but I'm not confident if anything will happen...)

For now, I'm not sure the problem comes from the RTC because I don't really understand how it works 
I already ordered for a new RF5C15 RICOH chip in case I really burn it, but I'm still thinking for now what is next step

[wild_potato]

Ahah I don't understand anything but I've got a signal 


So to sumarize:

• I had a weird oscilating peak signal when in stand-by mode, and the x68k did not start when power button pushed neither
• I cut the !ALARM pin of the RTC
• And it "works" (seems to for now)

I'll let my friend restore his system via floppy + scsi
and we'll test if anything is wrong for now.

Question I'm asking now:
- is the RTC dammaged? I don't see what it could be from my measurements.
- What is the !ALARM for? I would have said it was necessary for on/off system signal but it seems to properly shut down
- if I need to replace it, how will I do without risking to melt the plastic connector on the left 


Anyway, let's flash and I will let you know

[caius]

The ALARM signal from the RTC is one of the conditions for generating the POWER ON.You can see on attached image how this signal is ANDed by a 74LS08 along with other conditions.Now the machine powers ON because you simply changed the logics by removing the ALARM condition, now you must figure out why.For me the culprit is the 74LS08 itself.

[wild_potato]

Finally, I finished to include the new PSU, do tests ; all were working fine so I reclipped together.

Caius, you must be right. I think the ALARM signal is normal. The 20mV noise I measured on the clock is insignificant. This must interfer with the logical door directly connected to it. What I did to make my x68K compact work : nothing ;

My friend tested his machine - setup to boot on an SCSI controller, play game.
All worked fine (for this use case), so I did not consider to make any hardware searches.

So maybe that:
- The ALARM pin must be useful, but not for my usecase,
- The ALARM signal is not useful at all for the X68k compact (because it is a remaster of previous x68k releases)

If never someone unfortunate had the same behavior as me, just use your cutting tool
I'm a bit ashamed of the solution, but all is working fine, that's all I want. Many thanks Caius.
I will post an example of my PSU integration on my x68k Compact