PSU Design

Yes, they're zeners and they're using them as Vrefs. If I'd have posted a clearer diagram it might have been evident, but you have to work with what you can find online sometimes, if there's no better scans available from original hard-copy you can access yourself.

John I guess I started design in Aerospace at the right time in 1975. and it was probably my draftsman who said it stood for CURRENT RECTIFIER which a common descriptor. Although I read most of the Mil-Specs and STD HBBK's from our vast library at Bristol Aerospace, I can't recall seeing it there. It was nice to have any test equipment you needed.

Here is the TEK7904 mainframe service manual.

I guess you youngsters don't remember point-contact rectifiers, selenium rectifiers, vacuum tube rectifiers, copper oxide rectifiers, tungar rectifiers, mercury rectifiers, ignitron rectifiers, vibrator rectifiers, synchronous rectifiers, foxhole (rusty razor blade) rectifiers, galena rectifiers, or electrolytic rectifiers.

I actually saw a tungar rectifier in operation when I was a kid. In a gas station.

There were germanium, silicon, and gallium arsenide rectifiers in WWII. I *don't* remember WWII.

I had to download the .png file and print it out before I could find them.

Even then I could only find two - one at the top of the diagram close to Q850, where VR851 43V doesn't seem to do anything sensible at all and VR890 9V about halfway down the page next to the dual transistor Q866 which looks equally daft. Both look more like parts put in to stop voltage excursion during power up than anything intended to set up a predictable output voltage.

There may be zener diodes in there, but they don't seem to be being used as voltage references. Back in the 1970's I used to put together circuits vaguely like that,and I would have been in deep trouble if any of my schematics had been that unintelligible.

They may be zeners but neither VRR851 43V orVR8900 9V looks as it is is being used to set up a reference voltage.

I'm a skosh younger than you, but I've built things with selenium, tube, point contact, synchronous, and galena rectifiers. I have a couple of dozen point contact diodes in a Lista cabinet bin (1N34A, 1N38B, and

Of course the galena was just for fun, and never actually did very much. The seleniums and tubes were salvaged from old TVs when I was a boy. (It was a good thing for me that frying seleniums smell so horrible--I didn't realize that the fumes were dangerous.)

There are still things that tubes are unbeatable at. My last tube circuit used an 811A to control a grid in an electrostatic drift chamber--it needed to switch some hundreds of volts, and then go completely away, i.e. have very low leakage and capacitance when off.

(Of course, that was in about 1990.) ;)

My tube-type GDOs are some of my favorite instruments.

Cheers

Phil Hobbs

Your synchronous rectifiers were probably mosfets, but I was thinking of the motor-driven mechanical contact types.

(Somebody built a DRAM with leaded capacitors on a spinning disk.)

One cool device was a DPDT vibrator, sometimes used in car radios. One set of contacts pushed 12 volts into a transformer primary and the other set rectified the secondary. Old tube car radios had some interesting technology. I used to get them free from junkyards.

(I once needed some vibrators to help parts in tubes slide down low-slope feeders on a pick-and-place line, for a friend who had an assembly shop. I found a suitable gadget at Good Vibrations, a ladies', umm, accessory store. That was fun.)

Some of the Keithley electrometers used a special electrometer tube. Mine uses an early vintage mosfet.

VR890 seems to be the ref for -50, and the other supplies track that.

That's a really component-rich design. It was probably much-tinkered.

I think that was John D. Atanasoff (of Atanasoff-Berry Computer fame).

Yup. As a teenager, I had an old T19 Mk III tank transceiver that (iirc) had a companion vibrator supply. Interesting gizmo--I never did get it to work. I remember a weird E1148 VHF superregen oscillator tube with two caps on it.

I have one of each: a 405 and a 610C like yours. The 405 works on the

If you can find one that isn't gate-protected, these days. I have a reel of 2N7002E's because NXP discontinued them.

Cheers

Phil Hobbs

John, I do remember my 1st pickle-jar AM radio with Xtal earbud and pick+xtal diode my Dad made in 1962. I used a along wire antenna to the tree outside my bedroom window. I also remember a summer job climbing 6 story scaffolding with other with back-back vacuum cleaners to remove dust from the ceiling at the Dorsey HVDC Station where English Electric installed huge Mercury diode valves. It wasn't until I started testing 5 MVA transformers for partial discharge (PD from 20 kV up to 200 kV) that I discovered the harmful effects of of ceiling and wqll dust with high E + H fields . Even the paint on Faraday cages got electrified asa we stood outside the cage, the hair on the back of our necks stood out.

Don't forget Mercury Amalgam diode detectors in your teeth can detect AM radio signals under certain conditions.

In the end they all become polarized Current Rectifiers. CR1 , CR2 but D is more relevant.

My 610 doesn't work in fast mode, just goes bonkers. Sometimes a measurement takes half an hour to settle and I have to walk around my office holding a ground wire. Does yours work in fast mode?

Generally, but not on the three bottom ranges (10, 30, and 100 fA FS). The 10 fA scale is apparently just there to decorate the front panel--it's too unstable to be useful even in slow mode.

Possibly hooking up a LabJack to the analog output and averaging awhile might fix it, but if I needed to do that I'd certainly wire up something dead bug style in a cast-aluminum stomp box instead of using the Keithley.

Sometime in my declining years I might retrofit my old 602, which is sort of like a more primitive 610C that runs off a zillion dry cells.

I'd probably pick a 100-pf polystyrene cap out of my drawer, wire it up as a current integrator with something like an LMC6001 op amp, and connect it straight to the input. with some very low leakage reset switch.

Then I'd do a F/V for higher currents, and differentiate the ramp for lower ones.

Cheers

Phil Hobbs

One way to measure low current is to let it dump into a cap for an hour or a week, and then measure the voltage on the cap. That will work to aA's.

Not exactly. VR890 seems to be setting up a -9.0V reference for the dual transistor Q886 to sense, and R880,R881 and R882 divide down the -50V rail to produce a matching -9.0V but R883 (to 0V) and R884 to -53V suck out a bit of current through CR883 which complicates the situation.

If that's a design, I'd hate to see an improvisation.

And not well-thought out.

CR883 would normally be reverse biased, it is there to cope with the remote sense line getting disconnected.

The schematic is not drawn for great clarity but I am not critical of the design. Looks competant me and has interlocks. The use of higher voltage rails as current sources to the diff amps make it look more confusing than it is.

piglet

I thought of that after I'd made my post, which was a bit embarrassing. European circuit design has adopted a number of conventions that make that kind of mistake less likely.

There are couple of PNP transistors used as current sources on the schematic, but they aren't drawn in a way that makes this clear.

There is a rule of thumb that says if you can't draw your circuit in a way that makes it clear to other people what it is doing, you probably don't have a clear idea of it's operation yourself, but American circuit diagrams from that period were pretty much uninformly horrible.

I meant the three dots on the base lead.

Yandex rained down some spare decimal points?

The schematic on TekWiki is clearer and does not show those dots. Q835 et al are the graticule lamps dimmer.

piglet

Oh, I thought it might be a ferrite bead or something.

Was that an old scope with an external edge-lit plastic graticule? Or the later internal one?