HP 16700 PSU in 230V region woes

[FrodeM]

Have anyone here experienced problems with the power-supply of the HP 16700 logic analyzer blowing up when running in 230V regions? I am in particular thinking of the Celestica 7000 variety, which does claim to handle 230V just fine but I have had two blow on me so far.

The first time it was the switcher for the standby power source. An all-in-one TOP227Y switch mode regulator in a single TO-220 package. This blew quite violently, taking out a whole trace on the PCB, some resistors in its path, as well as D7.

For the second one , it was a small 8-pin chip, probably part of the main regulator. Fortunately, not as violently and this far I only found D7 as the only other part that has blown.

Is it just me, or does this PSU have really slim margins?

Some notes on the design. I plan to fix the second one with parts from the first one, and I'm in the process of drawing up a schematics as a consequence. I now see that the main board of the PSU uses a charge-pump type of design, instead of switching a transformer. That explains why D7 is very prone to damage: it's the main boost converter diode. All current driving the unit passes through it, and with 230V mains it is natural that the flyback suges it needs to pass will peak harder.

Does anyone else have similar experience with these supplies?

[alan.bain]

I've got a dead one here also in 230V-land.  R30 (the 10 ohm series resistor in aux power has exploded). Looks like the TOP223Y has also had a bad day and possibly also U1 (UC3854) as resistance to earth between pins 15 and 1 is about 35 ohms.  Do you have a service manual? 

[FrodeM]

I've got a dead one here also in 230V-land.  R30 (the 10 ohm series resistor in aux power has exploded). Looks like the TOP223Y has also had a bad day and possibly also U1 (UC3854) as resistance to earth between pins 15 and 1 is about 35 ohms.  Do you have a service manual?

No, but I did draw up some schematics of the aux supply a few years ago. Those are available on Bitsavers:

https://bitsavers.org/test_equipment/hp/167xx/Celestica_7000_Aux_Power_Regulator_Schematics.png

The 10 ohm resistor is a natural thing to go when the TOP227 goes, since it's the only thing between the high voltage reservoir and ground when that happens.

[alan.bain]

Thanks for sharing that very useful schematic.  I'm a bit surprised by the 12V marking though as the UC3584 needs more than that - thought somewhere > 15V.

I'll replace the obvious failures and see what goes bang next!

[FrodeM]

Thanks for sharing that very useful schematic.  I'm a bit surprised by the 12V marking though as the UC3584 needs more than that - thought somewhere > 15V.

I'll replace the obvious failures and see what goes bang next!

That 12V doesn't go to the UC3584, that has it's own supply on the lower half of the regulator board. The Aux-supply is completely standalone and is in the PSU only used to drive fans, as well as the U6 "Aux Power Good" and the U5 "On/Standby" optocoupler enabling/disabling the bank of secondary regulators providing the main supply for the analyzer.

The problem here, using an inductive charge-pump design for filling the high-voltage side, is that any sort of short to ground after the absolutely huge inductor is not immediately going to trip the fuses before it. What's worse, if something trips short and then open (for example, by blowing a PCB trace), you can get additional voltage-spikes over the inductor, potentially causing more carnage for anything connected to the high-voltage bus. There should really have been a clamp in there on the HV bus to prevent spikes accidentally exceeding maximum ratings of components, but from this far into the reverse-engineering/schematics project I can tell you there are none.

In my own case, I ultimately ended up getting one of the older PSUs made early in the life-cycle of the 16700 series. This was made by a different company, Autec Power Systems (Model MFC75-WEDDGG), so it hopefully has a better design internally.

[alan.bain]

I think I may have a slightly different version.  The secondary feed here is definitely to U1 - at least removing U1 removed the low Z to ground seen at the 390uF capacitor C18 (in honestly my first guess was a leaky capacitor) and continuity checks C18 +ve is connected to pin 15 of U1.

Seems logical really - usually the reason for an aux supply is to run the controller for the next stage.

[nctnico]

I've got a dead one here also in 230V-land.  R30 (the 10 ohm series resistor in aux power has exploded). Looks like the TOP223Y has also had a bad day and possibly also U1 (UC3854) as resistance to earth between pins 15 and 1 is about 35 ohms.  Do you have a service manual?

No, but I did draw up some schematics of the aux supply a few years ago. Those are available on Bitsavers:

https://bitsavers.org/test_equipment/hp/167xx/Celestica_7000_Aux_Power_Regulator_Schematics.png

Thanks for sharing. However, are you sure D6 is the right way around? It has been a long time ago I used a Topswitch device but I remember the datasheet being adament to clamp the voltage on the switching node. With D6 the way it is (assuming it isn't a zener diode), that won't happen which could cause the margins to be slim if this is a universal voltage design.

[FrodeM]

Thanks for sharing. However, are you sure D6 is the right way around? It has been a long time ago I used a Topswitch device but I remember the datasheet being adament to clamp the voltage on the switching node. With D6 the way it is (assuming it isn't a zener diode), that won't happen which could cause the margins to be slim if this is a universal voltage design.

I'll double-check later today. Maybe that's one of the reasons this failure-mode seems to be reasonably frequent with these?

[alan.bain]

On the one I have the anode of D6 is connected to the TOPswitch D, cathode to C23 & D8 in parallel which is the opposite to the way drawn, but agrees with the TOP223Y datasheet.

Alan

[FrodeM]

On the one I have the anode of D6 is connected to the TOPswitch D, cathode to C23 & D8 in parallel which is the opposite to the way drawn, but agrees with the TOP223Y datasheet.

Alan

Ok, on close inspection of where the trace goes, I see I have drawn it wrong. It must have measured as low enough resistance through the inductor for the multimeter to call continuity when I did the initial work on this.