Switching PSU Theory of Operation / Repair Tips & Techniques

From: Douglas Quebbeman <dhquebbeman_at_theestopinalgroup.com>
Date: Tue Jul 25 07:38:12 2000

> > So, I'd like to eventually repair it. It's a monster, and
> > potentially lethal. From poking around with a voltmeter,
> > I can see that it chops the AC to DC and doubles the voltage
> > to about 280 volts, one positive supply, one negative supply.
> Ok, so it's a switch mode power supply that produces +280v and -280v
(wrtg) rails right?
> Is this used as a rail to feed low voltage supplies, or are they generated
in the main supply.
> Or do you REALLY mean that it produces a postive and a negative with 280v
across them?
> That is what I would expect to see.

No, if I measure across the positive and negative, I'm seeing nearly 600vdc.

These HT supplies appear to feed the switching section of the PSU. Here's a
quote from a former prime engineer (who's been as helpful as his memory

: These things are much more complicated than that. Functionally the Line is

: directly connected to a special rectifier configuration called a voltage
: doubler. It produces rougly 280 volts DC. There is no isolation
: which makes these things potentially lethal. The 280V DC is input to
: switcher, which runs the 280V DC through the torroidal transformers to
: produce the other voltages. The Big Bridge is to convert the output of the

: switch from high frequency (probably about 25Khz) AC to DC.



Tony's aware, and has replied with useful information. Even with a dozen
people helping, I'm not expecting to get this repaired over night. If I
manage to get it repaired sometime next year, that would be cool.
> A circuit diagram would be useful here, but psu's aren't THAT
> complicated really. Often the type of fault will indicate where the
> problem lies.

I can probably sketch the HT portion out, but this PSU includes two rather
complex looking boards (although they're only 2-layer) with custom PALs and
the like.

One symptom I hope leads me to finding the problem is that the PSU has
four LEDs onit, 3 green, 1 yellow, the green are status indicators for
+5, +12, & -12; the yellow LED is an indicator for the AC. Although the
AC feed to the PSU is fine, the yellow LED glows dimly, not brightly, as
it once did. The green LEDs are unlit.

> I take it the 280v rails you mention are still present in it's non
> functional condition?


> That suggests that the mains rectifier is still functional, they don't
> 'chop' the ac mains.

Uh, being an analog idiot, forgive me if this is a stupid remark, but
I thought "chopping" the AC with a rectifier yielded DC; when I measured
the ~300v, I was on the DC scale, and I didn't see the needle vibrate
like I have in the past when I stupidly tried to measure AC in DC mode.

Right now, I'm stuck using this old Radio Shack analog VM; my Fluke is
in need of a new LCD display. :-(

> Switch mode supplies generally rectify and filter the mains,
> and use the resultant 250-350vdc to drive the rest of it.
> If you have no low voltage stuff running, and no blown
> fuses, it could be as simple as a startup resistor being
> open circuit. Look for a high value (~300k - 1Meg) resistor
> from the + side of the dc rail to the electronics of the switch
> mode supply. These go open circuit at switch on fairly often,
> and the supply doesn't get the initial starting pulse it needs.

If it matters, the room in which I keep the Prime was hotter than
I usually let it get when I have the Prime running (I don't have
central A/C, so I have a window unit in an adjacent bedroom and
usually use a fan to direct the cool air into the room with the
Prime while it's up. That night was hot and I had the air directed
instead to my bedroom, so I was running the Prime warmer than I had
been, but not, I didn't think, out of spec (site prep guide says 86 degF
top ambient temp). So, keeping in mind someone else said that dust
is the enemy of a power supply, and being able to tell you that the
PSU was indeed dusty inside (the only thing I didn't pull and clean
when the system was delivered), and I operated it kinda warm, if
the resistor is an item likely to die if it can't dissipate enough
heat, then I'll look for that.

> ****WARNING****
> These supplies are filtered by one or more largish reservoir
> capacitors that can and do store a potentially dangerous amount of
> power, they are DANGEROUS even when the unit is off and unplugged until
they are
> discharged. They may or may not have a bleed resistor across them. I
> strongly recommend that you check with a meter after about a half hour
> with it switched off to see if they have discharged before you stick
> fingers in there or you might get a nasty surprise.

I've been measuring across the caps before I touch anything.

> If you identify a likely resistor, you will find that you get ambiguous
> readings unless you remove it from the circuit. Open circuit startup
> resistors are a common fault in switch mode supplies.

It's gonna be a bit of a chore to disassemble the PSU, it's in a form factor
that allows it to plug into the backplane; separating the supply PCB from
aluminum carrier will be a bitch, and I don't have a lab bench anymore.
I miss my lab bench!
> > possible for devices that vary in design not only from
> > manufacturer to manufacturer, but wven model to model?
> Never fixed TV sets have you. :^) Emphatically yes.
> Hope this helps.

Thanks Geoff!

Received on Tue Jul 25 2000 - 07:38:12 BST

This archive was generated by hypermail 2.3.0 : Fri Oct 10 2014 - 23:32:58 BST