FLUKE 332A pre-regulator cycle dropping

[Vollumscope511]

Howdy All-
I'm new here-- this is my first post.

I have a very clean Fluke 332A DC standard. I've owned it for at least twenty years. It looks like it's been updated (way before I acquired it) because many of the boards have 1984 copyright dates-- and device date codes, but the transformers are stamped with a 1966 date code.

I want to put my ol' side kick to use but....

The unit's pre-regulator is on the fritz.  It drops every other mains cycle-- that is the standard's voltage transformer should be energized every 8.333ms (60Hz US power main) for a period of the mains cycle proportional to the voltage drop across the pass transistors.  But the pre-regulator is "ticking" off at 16.67ms...

The UJT oscillator works. It responds to the voltage drop across the pass transistors, there's drive out of the pulse transformer. The driver circuitry seems to work right but when it comes to driving the pass transistor every 8.33ms-- there's base drive every 16.67ms.

I'm stumped.

The diode bridge around the switching transistor is OK...

I performed a pile of other inconclusive tests but I'm not going to drone on...

Anybody have experience with the pre-regulator in this beast?
BTW-- it's a relatively common pre-regulator set up...


Any thoughts, help, musings hints greatly appreciated.

[Roman oh]

I can’t immediately find my notes on the pre-regulator, but over the years I have made a number of mods to my 332 mainly with a view to reducing its power consumption and hence warm- up and post-turn-on stabilization time. A replacement pre-reg was part of that, so I have fiddled in this area. 

IIRC, the UJT circuit associated with the Series Pass element fires a stream of pulses through T1 to the pre regulator, which is synced to the mains by +V applied through R18 to the quasi-SCR circuit made up of Q6 and Q7. +V is (in your case) 120Hz positive half sine wave coming from CR1/CR3 driven in turn from a nominal 10V AC transformer winding. If one of those diodes is U/S, that would explain your symptoms.
Have you checked these, and the +V waveform? BUT BE CAREFUL, because the pre-reg circuit is connected to the mains, so unless you have some isolated form of working, don’t try to probe +V directly with any un-isolated instruments. But test the diodes in the normal way.
Roman

[Vollumscope511]

Howdy Roman--

Thanks for replying to my query about the Fluke 332A pre-reg SNAFU I'm having.
I've got more experience in electronic design & repair than the average bear but this 332A pre-reg issue has me stumped.

I'd like to avoid a massive redesign of the pre-reg circuitry-- altho it's tempting.  Over the production cycle of the 332x and like voltage standards, Fluke made a few minor changes to the pre-reg circuit(s)-- very likely improvements having to do with dynamic range and regulation stability.

The basic pre-reg circuitry seems to be fully functional.  I can energize the 332A's DC regulator with an external supply providing the pass transistor energy and "steer" the pre-reg UJT pulser/drive all the way to the switching transistor. Of course the AC mains to the voltage power trans is disconnected (that's the transformer that provides the output energy of the 332A and not the trans/supply that runs the 332A).

It looks like the UJT and associated circuitry is firing at the 60Hz mains zero X-ings and a "turn-off" time delay varies in proportion of the voltage drop across the pass transistors. 

The circuitry on the "other side" of the pulse transformer seems to be functioning as well.  I can't say correctly.

In my 332A the original pre-reg switching transistor was replaced. I can not find any doc on that original transistor. My guess is it was custom made for Fluke NPN silicon device with somewhat higher hFE and maybe 250-300V Vceo / Vceo(sus) than what was commonly found back then (even today).   I know it was made by Westinghouse but the device bears a non-JEDEC number. Search-engine look-up of the device number provides nothing useful.

I am very suspicions that transistor static hFE is at the heart of my problem. It could be that a lack of base drive keeps the transistor from saturating or the transistor pops out of saturation as the AC mains voltage rises and hence the wattage demand increases. The replacement transistor is rated at an hFE of 7 at 4A Ic. The way Fluke has the switching transistor driver rigged, it looks like they depended on over an ampere of base current to the switching transistor.

I replaced the present switching transistor with a device I found in a junk box-- and found some data on-- it had safe Vceo and Ic capabilities. Also rated hFE of 35 and 5A Ic-- and low and behold it worked but failed (shorted) at the 332A's 1000V range-- I suspect second breakdown.

I replaced that transistor with a Darlington pair device. A big assed brute TO-3 switching transistor BUT-35.  I used my transistor curve tracer to establish base drive requirements vis a vis likely transistor IC.  With some changes to the 332A's pre-reg driver load resistor(s)-- for a mere 50ma base drive to the Darlington-- low and behold it works! But... but..  pre-regulator dynamic range is severely compromised.  The drive to the new Darlington transistor will stop if I raise the mains voltage (to the switching transistor) to the nominal 120V.  It will also grind to a halt if I draw too little wattage from the 332A's output terminals..   transistor looks like it's saturating as it doesn't heat at all.

One last thing before I close-- that pre-regulator switching trans is straddled in a diode bridge so that the alternating current flows through the switching trans in one direction. But as the mains current flow alternates the transformer primary looks like it appears in the collector circuit and then on the alternate 60Hz cycle in the emitter circuit.  The base drive return to the pre-regulator common is through the diode bridge.  I'm wondering if this affects the switching transistor biasing?  I haven't come up with a safe probing approach. I'm using battery powered and thus isolated 'scope 'n' meters. Also running the pre-regulator thru an isolated variac.

All for now.. and thanks for replying to me..

Regards,

Robert (aka Vollumscope)

[Dr. Frank]

I also had an old 332B/AF with a defective pre regulator board.
The power transistors Q1, Q2 were dead, unobtainium.
What I also found out, that nearly all of the medium power transistors Q3-Q9 showed very low hFE, below 10, or so, as if they all deteriorated over time.
If I remember correctly, these sit in sockets, so it's easy to check and replace if necessary.
Frank

[Vollumscope511]

Thank you Dr Frank for replying to my plea for help with my Fluke 332A pre-regulator..

I suspect that somehow the mains switching transistor's (Q107) emitter voltage is rising above it's base voltage during transformer primary current draw in that cycle dropping issue. While the base drive and DC from Q107 return directly back to the pre-regulator power supply, somehow voltage drop across the regulated power transformer primary is also lifting the emitter voltage above ground (although there is no common circuit between Q107 base and mains/regulated transformer primary)-- and therefore above the base voltage. This back-biases the switching transistor B-E junction and cuts off the transistor (Q107).  The driver transistor for Q107 is also a saturated switch and its base drive current is part of Q107's base drive current. So when Q107 becomes reverse biased, the driver transistor no longer conducts because its base driver current stops flowing.

This is a guess. I have to probe the circuit to verify if my guess is plausible.

See my other reply post to Roman Oh.

Again,
Thank you for your help.

Regards

Robert, aka Vollumscope

. 

[Roman oh]

Hey Robert,
Your problem drove me back to thinking about this. I would love to be able to respond to your individual points by quoting them like some posters do, but can’t see how to do that. So I’ll just talk about what I think I know, hoping it may help you a bit….

Firstly, I can’t find a schematic of this part of the 332A. The manuals I have been able to find seem to have all the schematics except the series pass assembly and pre-regulator. I have found an early 332B schematic which is quite different from my 332B/AF, but the component ID scheme is quite different.

Assuming that the design philosophy is the same, the way the pre-regulator works is as follows:
The pre-reg works at a nominal 100V AC (85V in the B/AF) derived from a tap on the line side of the main power transformer (hence the reason why probing with an unisolated set-up is hazardous). The control is via a single control transistor in a conventional diode bridge.
At every mains zero crossing, the control transistor is turned on with each mains transition  via a rectified but unsmoothed voltage sourced, in the A version, from a bridge rectifier CR1201-1204 (this schematic from the A I do have). In my book this is labeled “-V”.
The control transistor switches ON when -V goes nominally to zero (ie, mains transition). There is a bunch of stuff between the pulse transformer and the control transistor which then locks it ON until a pulse hits it from the pulse transformer driven by the UJT (or excessive current in Q107 as sensed by a small resistor in its emitter…although I don’t know if this is present in the A). The UJT is a conventional oscillator which responds to the voltage across the last series pass transistor (I’ll call it Q8 as per the B, but the A may be different) in a chain of 8, the first 7 of which do nothing more than absorb the bulk of the up-to-1200V or so coming from the HV transformer fed by our control transistor, which is the one I think you call Q107. (If the control loop is working correctly, then in the steady state these first 7 transistors are simply saturated with almost zero voltage drop. They only see action during dramatic range or load changes, and then only briefly.)
Now, the UJT circuit is such that it starts running when the voltage across Q8 is around 40V, and when that voltage gets to about 55V or so, it’s running flat out well beyond 20kHz.
(In the B/AF, there is extra circuitry here to prevent the UJT running for a very brief period at the mains transition, but that circuitry is absent in other B schematics I have seen, and probably in the A as well.)
So, with each mains transition, Q107 gets turned on and stays on until Q8 calls “enough!”. But if Q8 already has enough (such as high input voltage or light load, which I think is where you were seeing issues) Q107 could well be being turned off immediately, which may look like loss of drive that you are seeing. I think you mentioned that you were running some parts of this independently of the overall control loop, so perhaps it will help to watch the voltage across your Q8 as you monitor your drive to Q107.

Apologies if you already know all this…..


Roman

[Dr. Frank]

Hello, there are many complete manuals available, covering the different versions also. But search for 335 B and 335D. The only difference is the differential null/voltmeter

[Roman oh]

Dr Frank,
Yes, there are many manuals on line. But there are also a lot of subtle differences in the circuitry as it evolved. I have downloaded many variations of 332A, B, D and B/AF manuals (commercial and MIL) from Bama, k04bb, fluke, electrotanya, and a bunch of random edu and other sources. Naturally, the  Fluke guys evolved the design over many years, and the manuals sort of kept pace, I guess, although my particular unit is still different in some small details from any version of the manual, and its change notes, than I have, in areas that are clearly original. And I still cannot find one that contains the particular schematic of the pre-regulator that Robert is referring to.
But I am always looking…I have been studying, adapting, improving(?), modifying my unit over many years, initially with the aim of reducing its power consumption and hence stabilization time, with some success. So my search for manuals and info is continuous.

And from watching your posts over the years, I’m sure there are still some questions I have about the design that you probably know the answers to…
Roman

[Vollumscope511]

Hello Roman & Dr Frank.
Once again thanks for replying to my posted Fluke 332A question(s).

The complete Fluke 332A manual & schematic is on BAMA-- it's labeled 332A_sm.pdf...

I understand how the pre-reg system works.  It is a very crude voltage regulator as there's no independent reference standard that I can see-- there is a zener stack that is used to compare the voltage drop across the pass transistor line-up.  I also don't know how Fluke attempted to implement regulation hysteresis. As Roman mentioned it simply works by charging a capacitor through what looks like a quasi constant current source-- a zener two 220K resistors and then another zener and back to the pass transistor floating low voltage supply.

The cap charges to the a UJT's peak point which is a tiny fraction of the claimed 50V pass transistor drop and then the UJT is forward biased and dumps the energy in the cap through it's Emitter-Base1 junction and in turn pushes the current thru the pulse transformer. primary.  The pulse signals the pre-reg system to "turn off"..  we've gone thru that already.

My 332A was updated by somebody-- maybe Fluke? maybe a very advanced hobbyist or maybe a test equipment repair depot. The 332A in question is not military or ever was owned by the US military.  For example, the original 332A used incandescent lights as the front panel annunciators-- voltage setting decimal points, standby indicator, current limit-- on my 332A those have been very elegantly upgraded to LEDs.  The circuit board suite has also been upgraded-- such as the chopper board. It uses the MOSFET chopper.

One thing that was changed is the mains switching transistor-- which according to the Fluke manual (I have the original manual) bears some sort of a non-JEDEC part number.  I can't find any info on that transistor other than what the manual says. 

Whoever updated my 332A also likely changed that transistor.  It's a GE part # to which I have the data sheet.  It's a plausible replacement for the service instended-- but I have no clue as to what the static hFE of the original part was.  The replacement part has a typical hFE for a high voltage BJT device approx 7.

Given the fact that the UJT based regulator seemed to function as expected when I bypassed the 1KV AC mains transformer supply with a DC bench supply leads me to believe that the circuitry is functioning correctly or close to correctly.

I wondered if the original Fluke switching transistor had a high BETA (static hFE)  and thus required less base drive per collector current than typical high voltage power NPN transistors of the era.  I replaced that transistor with a suitable Darlington transistor and made some component changes to the driver transistor-- higher collector load resistor (47 ohms from 6 ohms).  This works but still cycle drops and the relay still buzzes loudly depending on current demand.  The new darlington switching transistor saturates on (seen by oscilloscope observation).

So now my guess is that the UJT voltage sensing circuitry is detecting no need to enable the transformer primary in 8.33mS (half cycle zero crossing time).  And hence the cycle dropping observation.  But that is not what the original design was to do.

The UJT oscillator is synchronized to the 60hz line zero crossings. The pre-regulator drive circuit is also synched to the 60hz zero crossings.  If for some reason the UJT oscillator failed-- the pre-regulator could get out of control.


 Here are a couple of observactions of my 332A--

1) The power relay that energizes the HV power transformer/pre-regulator buzzes loudly-- and does so in step with the current demand and cycle dropping for the pre-regulator.  The more current I draw from the 332A the louder the relay buzzes. Maybe all 332As do this.

2) My 332A creates RFI-- once again maybe they all do this.

3) The voltage drop across the DC pass transistors is higher than Fluke claims it to be-- mine is approx 80-90V depending on 332A current load instead of the 50V...

That's all I've got for now...

Developing...

Thanks again for your time.

Regards,

Rob aka "Vollumscope511"