PSU capacitor choice

Those caps are to suppress spikes arising from rectifier-diode reverse-recovery-time current snap-off, and the transformer's leakage inductance. See AoE x-Chapters, 9x.6, for discussion and scope traces. The values aren't critical, and are related to diode current and transformer L_sub-ell.

--
 Thanks, 
    - Win

I'm most grateful, your Majesty I should mention the top cap in that diagram failed short circuit, blowing the fuse immediately preceding it. I had a spare cap of exactly the same value and type in my spares bin - and it was 600VDC rated as against the original's 400 so I'm guessing it shouldn't happen again. Fortunately, there's so much real estate available in these old boat anchors you can slot in physically larger replacements that would never fit into the footprint available in today's products. You can't beat vintage electronics.

Pure swank. A lot of us don't have forelocks anymore.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

It was a virtual forelock, Phil. I'm in the same boat in reality. :-D That 8566B of yours - have you ever had the not uncommon "YTO unlocked" error appear on the screen?

On one unit, yes. It went away after a few minutes' warm-up and hasn't returned.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

Well, don't worry about it. It's generally fixable and there's a fair bit of info on the web about this issue.

That is a bad idea. Once the spikes are generated, they go everywhere and are hard to filter.

Best to stop the noise before it is generated. 0.1uF caps right across the diodes.

Given the huge filter caps on the other side of the diodes, the two connection approaches are largely equivalent. Longtime practice by HP engineers and others seems to be caps to ground. Dunno if they had a good reason. Maybe clamping high dV/dt at that node. Both methods are very effective.

--
 Thanks, 
    - Win

Output caps are irrelevant. The diodes are fed by transformers, which are inductors and have high impedance to spikes. Caps connected directly across the diodes prevent the snap-off by slowing the discharge current.

Output caps? All of HP's diodes were at the transformer-diode node. They're input caps.

--
 Thanks, 
    - Win

There's a papah:

I meant the large caps fed by the diodes. We can call these filter caps.

Connecting the diode bypass caps to ground puts these caps in series with the diodes and bypass caps. The large caps have internal resistance which reduces the effectivness of the diode bypass caps.

The object is to slow the current transition through zero and prevent the spikes.

Recall that step recovery diodes require a sharp transition to generate harmonics. If the transition is slow, there is no step recovery and no harmonics.

------------------------

** You are joking ?

** The cap values are proportional to the current flow in each case.

Their purpose is to suppress RF noise that might affect the analyser.

You see them is analogue TV sets, scopes and radio gear, but not much else.

Same nobody else answered your ACTUAL question.

..... Phil

..... Phil

A couple of power supplies shown in _ARRL Handbook_s from back in the day when dissipative regulators reigned supreme, use 0.01?F capacitors as snubbers across each individual diode in a bridge. One 2,000V beast uses a 1000 PIV, 3A diode, a 0.01?F, 1kV capacitor, and a 390k?, 1W resister connected in parallel as a basic building block. The blocks are then serially stacked six high in each leg of the bridge. The capacitors suppress transients while the resistors equalize.

Thank you,

--
Don Kuenz KB7RPU 
There was a young lady named Bright Whose speed was far faster than light; 
She set out one day In a relative way And returned on the previous night.

(I think) Win's approach was approximating an optimum R value for appropriate damping to practically eliminate the effect to avoid having a hard-C load directly on the transformer (which has its own issues... see below).

The diode recovery delay induces the effect, but the spike originates from the secondary leakage inductance so it makes sense to put the snubber on the transformer secondary.. when the current abruptly stops the resulting spike has somewhere to go. When the caps are placed across the diodes, then the current spike has to travel through the filter cap then through another cap (if a bridge) to complete the path. But yeah, caps across the diodes works fine too. Best to use ceramic caps if not including series R, they can handle the current and extra voltage from high frequency line transients.

Some experiences I've had with snubbers or lack thereof...

Rivera tube amps usually include a film-type capacitor across the transformer HV secondary. It almost always blows, sometimes blowing itself off the PCB and "fixing" itself. I always remove this cap and replace it with 4 0.01uF 500V ceramic caps placed across the rectifier diodes. If I bother, never noticed any noise if the caps are omitted but after reading about the "spike" I now just add the caps.

I suspect that it's less stress on the caps when placed across the diodes (the voltage doesn't reverse) than when placed across the secondary. But more likely ceramic caps just take the heat better than film caps, especially when there's no series R.

Old Fender tube amps (center-tapped secondary) have no snubber at all but never noticed any noise issues, even when souped up with super-high-gain overdrive preamps. Next time I'm at the shop working on one (when the shop finally reopens! ugh..) will hang a probe on the secondary to see if there's a spike. Could be high-voltage transformers have enough winding capacitance to naturally absorb the spike.