Saturday, December 24, 2011

JX-8P +5V supply goes crazy

The +5V looks OK most of the time...

... but sometimes... it looks like this:

Nice +5V voltage! No wonder it crashes...
What the .. !?

Ok, before I found this out I replaced the 7805 on the PSU as well as all the electrolytics on the PSU board. I knew this thing was crashing and I was hoping it was the 7805 that was busted somehow or maybe the electrolytics had gone bad. After replacing them I found out the JX was still crashing. Disappointing...

So, tonight I decided to look at it again. I figured that since it's resetting, as if it was just turned on, it might be something with the +5V supplying the CPUs. I had checked the +5V before and it was flat and nice. No noise or anything (at least not worse than what my shitty USB PicoScope introduced).

Hmm.. So what is the problem?
Is the +5V failing due to the 7805 thermal protection or short-circuit protection kicking in?
I'm not sure, but since it could happen soon after power on, or a very long time after power on, I think it's not really related to the temperature of the regulator. Maybe something is shorting out, but, what would that be? Doesn't make sense.

I figured that if I measure the +5V supply (hooked the probe on to the out-pin of the 7805) and set the scope to do a single sweep when the +5V drops below +4V (negative flank) I will wait for the JX to reset/crash and see if the scope has caught anything, and it sure did! It's the picture you saw above. You can see it making two short dips down to -7V ! Huh?

I repeated the test a few times. Interestingly, when I saw this it would not always cause the JX to reset.
Out of the 6 times I saw something happen on the scope it only reset 3 times. So, I assume the capacitors across the PCBs kept the thing alive half of the times.

It took anywhere from a couple of minutes and up (to an hour in one case) before I got anything that triggered the scope. Sporadic error. Very annoying.

Here are a couple of other results I got.
This did NOT cause a reset.

This didn't cause a reset either!

Two guesses

First guess
I sent a picture of only the first scope screenshot to a friend. His comment was something like:
"That doesn't look right, even for broken regulator. Negative voltage? It looks like you're switching a motor on or off, as if there is an inductive load."

... which got me thinking... There is a VFD display on this thing... They require higher voltage than +5V. (Around +50V or so?)
I have a faint memory of that there's some sort of oscillator and a small transformer somewhere... Inductive load... That would be one way of producing the higher voltage needed by the VFD.
Is there one?
If there is, I wonder if it is powered by the +5V line...
Yup.. Found it.

IC1 is display controller. The display itself is above it.
I'm not sure how that TR1/TR2/C3-circuit works but I assume that it could be an oscillator of some sort.
So.. Could it be that a fault with the display causes some spikes to travel backwards through the transformer and onto the +5V, causing those horrible spikes?
On some of the scope pictures it does look like a "sine wave" (although ugly) rather than just spikes. Perhaps those sine waves are coming from this transformer somehow, leaking back into the +5V supply and causing all the problems?

Second guess
I read somewhere that the frequency usually used for VFD transformer applications was above the audible range to reduce interference. Sounds reasonable. I thought I'd check what frequency those sine waves were.

Let's see... I'll check the time between the peaks.
First picture..looks like.. 20ms.. Second picture...looks like 20ms...  Third pic... 20ms...

Uh oh.. This is 50Hz... I live in Europe. We have 50Hz in the mains voltage here.

Is there a problem with a rectifier in the PSU!? I put my finger on the rectifier bridge used for the 7805. It was HOT! Almost burnt my finger. But does that mean it's damaged or just that something is drawing a lot of current, causing the rectifier to heat up?

It's getting late. I have to go to bed. It's Christmas Eve tomorrow...

Grrr... :)

To be continued...


  1. Seems like a diode in the rectifier bridge is loose from the chip and sometimes makes no contact. Did you replace it yet?

  2. If there was no contact there'd be no voltage at all I suppose, or at least a considerable drop if the capacitors did their part. I think the bridge occasionally conducts in reverse direection. Didn't get the replacement bridges yet but I will make a new post when I have some results :)