AVO MKIV calibration: heater voltage *way* too high

[ianbatty311]

Hi, Folks - greetings from Down Under

I have an AVO MKIV on the bench for a fellow HRSA club member.

I initially set the mains voltage selector so that the meter needle sits at the correct place on the scale.

The anode voltage (measured with a DC voltmeter at the top socket panel) reads about half the selector voltage (i.e., 200 V selected gives pretty well 100 V on the DC meter).

My problem is that the actual heater voltage is *way* too high, even under load.

No load, I get 0.83 V (0.625 selected), 1.56 V (1.25 selected), and so on, until 6.56 V (5 V selected), and 7.7 V (6.3 selected).

Plugging in a 6J5 (octal triode, 0.3 A heater as most receiving valves use) only drops the actual voltage for the "6.3" setting to around 7.5 V.

I know you can 'over-run' heaters for 'a while', but the biggest problem is that the excess emission gives a wildly exaggerated value for mutual conductance (up to 1.5 times what it should be).

The AVO is original and has not been messed with apart from the power cord having been removed (neatly), which I've replaced.

The RV6 calibration pot is at the extreme end of its travel, giving maximum indication on the AC adjust setting, so changing the pot's setting *reduces* the meter indication.

I did suspect the electrolytic cap across the meter and paralleled a new one in with no effect.

So, where do I go from here?

As the voltages are pretty well derived from transformer secondaries, I'm lost for an explanation of how the selected anode voltages seem pretty well on spec, but the heater voltages are as much as 25% (!) high on the 5 V position and almost as bad on the 6.3 V.

I did check with an EL34 in the socket (1.5 A heater), as I've read that AVO over-designed the heater supply to account for transformer drop under load, but even the EL34 ended up with some 7.2 V (!) on the heater, with the potential (again) for misleadingly high gm readings.

Any thoughts?

Cheers,

Ian Batty, M.Ed.

[Dekatron]

What AC mains voltage do you have and at what position is the fuse positioned and what position does the voltage selector on the front panel have?

[Radio Wrangler]

To close up a couple of loop-holes:

What are you using to measure AC voltage?

How good is your mains waveform?

Have you measured your mains voltage, and what tap is selected for it?

David

[Craig Sawyers]

From the AVO IV service manual 5V should read 5.5V off load with limits of 5.3V min and 5.7V max.

The heater transformer has a lot of taps, but is otherwise very straightforward

So that you read 6.56V off load suggests that all is not well in the camp. Any evidence it has been modified in some way?

[ianbatty311]

Hi, folks,

Thanks for your replies.

Dekatron - I have the fuse at the 210-230 V position, but need my a Variac set to 240 V AC output (!) to get the mains voltage correct according to the meter (it's around '90' on the 0-100 scale) - I also have the mains adjust switch maximum clockwise to get this setting.

This gives around 110 V DC at the A1-C connection, and heater volts of 7.3 V AC.

Running the mains down to about 225 V gives me 100 V DC at the A1-C connection (correct according to the service notes), and puts the mains indication at around '83' on the 0-100 scale (the correct setting is '90' on the scale). This still gives around 6.8 V with a 300 mA heater valve in the test socket.

So, I can set the anode voltage for 200 V and get the correct indication (100 V DC) by jiggling the mains input, but the measured heater voltage is still way too high.

Radio Wrangler - my mains supply is domestic AC power from the local grid - I'll check the waveform and let you know.

Craig Sawyers - yes, your 5 V calibration (5.5 no load, max 5.7 min 5.3) agrees with the AVO service data (which I have), and I have re-checked the actual heater voltages on the AVO and they are still wrong.

A general question - given that the two voltage/supply transformers (T2, T3) are fed by *tappings* on the heater transformer (the "200 V" and the middle of the "set mains" tapping set), is it advisable to trace out where they connect and try different connections?

Not something I like to do on pro equipment, but is it an option?

And a real puzzle - I have the original circuit, and a later one (1965 build), but NEITHER shows a complete circuit path for the meter in the 'mains test' position. I *have* re-checked, but am I missing somethine?

Cheers,

Ian.

[Dekatron]

What happens if you set the fuse in the 240V position and use your variac and set it to 240V, with the voltage control knob on the front set to its middle position?

What heater voltages do you get now?

Forget about setting the meter to the SET AC line for now, just measure the heater voltages with this setup.

With this setup you should have approximately correct heater voltages and by turning the voltage control knob on the front you should be able to get 200VAC across the primary windings on T2 (grid voltage transformer) as well as across the primary of T3 (Anode/Screen & Backing-off transformer).

Which position do you need to rotate the knob to get 200VAC?

What heater voltages do you get in this position?

If the needle isn't possible to adjust to the SET AC line with this setup with the fuse in its 240V position while supplying the tester with 240V, assuming you have correct heater, anode/screen and grid-voltages with 200V across the primaries of T2 & T3, you will have to check R19 and possibly perform the calibration steps with RV4 & RV6.

[ianbatty311]

Hi, Dekatron,

Thanks for the hint - I'll try it today and get back to this forum.

As well, the electrode selectors are *really* hard to move.

Is there any way of freeing them up?

Cheers,

Ian.

[Dekatron]

I am no expert on what to use to freeing up the rotary switches - but do check them before turning them so that there is no problem with the contacts or the wafers.

[Craig Sawyers]

Originally Posted by ianbatty311

Hi, Dekatron,

Thanks for the hint - I'll try it today and get back to this forum.

As well, the electrode selectors are *really* hard to move.

Is there any way of freeing them up?

Cheers,

Ian.

You mean the thumbwheel switches on the top? Alas there is no shortcut. They need to be carefully desoldered, dismantled, cleaned at a part level, contact grease and reassemble.

Take photos and copious notes. There will be shim washers in there, the position of which you will need to note.

If done two of these for my AVO valve testers. It is fiddly, but you end up with something you don't lose a fingernail moving!

Craig

[ianbatty311]

Hi, folks,

So I set up for:

Fuse in 230~250 V socket,

240 V AC input,

Adjusted the mains adjust switch for 200 V at T2 primary.

1. I found I could adjust the 'set mains' (rotary) switch for a range of approx 188 to 207 V AC at T2 primary.

2. I set for 200 V at T2 pri.

3. Looking at the meter, I got an indication of '80' on the 0~100 scale (that is, the mains setting was 'too low' according to the meter).

4. The correct setting for mains voltage (the sine symbol above the rectangle at top right of all scales) is '90' on the 0~100 scale.

5. I set for 200 V at T2 pri, and got heater volts (6.3 selected) of 6.8 V no load, 6.7 V with a 300 mA load, and 6.4 V with a 1.5 A (EL34) load.

As I have mentioned before, the RV6 calibration pot is fully advanced, and it is still not bringing up the needle from '80' to '90' - which it should do, as all circuit voltages are correct with the needle at '80'.

I cannot see where to make a further sensitivity adjustment - there must be some fixed-vale resistor in series with the meter on the 'set mains' position.

The AVO uses a 30-something microamp meter, and I have RV6 at minimum resistance, so there must still be a large series resistor somewhere that (if adjusted) would allow me to back off the position of RV6 so that it has some leeway for adjustment.

My circuit schematic shows an OPEN CIRCUIT when the meter is on the 'set mains' position. This must be a drawing error, as the meter is indicating on the 'set mains' position.



Help please.

1. Can anyone tell me how the meter connects to the circuit in the 'set mains' position. As I have said, my CIRCUIT does not show a connection to the left-hand end of the meter/capacitor/diodes assembly. Does anyone have a corrected schematic?

2. Do I have a low-sensitivity meter? I checked for FSD and got (IIRC) about 37 microamps. Should the instrument have enough leeway/some adjustment that will compensate if this sensitivity is too low?

Thanks in advance,

Ian.

[Dekatron]

Good, now that it seems that you can get the correct voltages you should start by adjusting your meter as it has 37uA FSD. You can adjust the circuit as it has an FSD below 39uA, but you can also put some small neodymium magnets on the old magnet to help it reach a better (lower) FSD current.

Have you read the large document on circuit comparisons that I wrote a few years back? If not you can find it here: https://www.vintage-radio.net/forum/...ad.php?t=86262 as it might give you an understanding of the circuit involved in the calibration (even though I mainly discuss the CT160 the same holds for the Mk IV and other AVO VCMs with minor circuit differences). In Appendix 5.2 you can read about how to adjust the shunt resistor to accomodate for a meter with higher FSD, however some of these adjustments need changing or shorting of the internal swamp resistor which then throws off the temperature compensating effect - so adding neodymium magnets or having the original magnet re-magnetized is a better option in my mind as it keeps the original swamp resistor unmodified.

You might also have an older revision where RV6 is taken from the 125V tap on the Anode/Screen transformer, later models had it moved to the 150V tap for easier calibration. I can recommend that you move it to the 150V tap if it sits on the 125V tap now.

It might also be R19, the "calibration resistors" that have changed their (it is usually a pair of resistors in series) resistances over the years. Do check the resistance of R19 and see if you can correct it as necessary and do the check for SET AC again.

[ianbatty311]

Hi to all who chipped in.

My original issue was an AVO MK IV that, with the mains selector set to put the meter at the mains cal point ('90' on the 0-100 scale), the heater voltage on the 6.3 V winding was well over 7 volts (around 7.5 open circuit).

So the 'let's get into a flap about this and run about for a bit' response was it had to be a dud meter.

I measured the meter, and found an FSD of 33.5 microamps.

Since that was within expectations, *AND* since (i) it was not my AVO, and (ii) I really, really don't want to open up a seriously delicate non-replaceable meter 'just in case', I started looking elsewhere.

I was puzzled that, with the fuse/selector in the '230~250' position, I had to crank my mains variac up to almost 250 V and switch the mains selector fully clockwise to get the mains cal close to correct (it was *under* reading).

The manufacturer's circuit is one that should have seen the bod doing the drawings given a good talking to, or (if a competent artist), protesting that AVO needed to come up with circuits that were not begging for the annotation 'throw six to start'.

Seriously, I have looked at both circuits, the one in the manual which really should never have been published and that helpfully supplied by JACMUSIC, and I

can

not

trace

the complete meter circuit when in the Mains Cal mode.

I took advice that I should check the voltage at the primary of the Grid Bias supply (which also feeds the cal cct), and adjust the AVO to give 200 V at that point.

Doing so (i) put the heater voltage in range: 6J5 (300 mA) 6.9 V, 6V6 (0.45 A) 6.8 V, 6AG7 (0.6 A), 6.6 V, 6DQ6 (1.2 A) 6.4 V, EL34 (1.5 A) 6.3 V, and put the meter needle pretty much at '80' on the scale.

Using these settings, I made up a Calibration Valve (why don't the instructions say you have to *parallel* the two sections?), and tested the AVO.

It agreed pretty well with the cal valve, as did two CT160s.

So, back to Martin's tip about the calibration resistor/s (thanks, Martin).

I looked back at JACMUSIC's circuit, and the illustrations in the parts list and discovered calibration resistor R19 (not easy to find, AVO have put just about all minor components on the switches, using blank wafers as tagstrips).

It measured high - a bit over 3 megohms, so I tried paralleling it with series-ed up 10 meg resistors.

That (using 2 x 10 megs in series) brought the pointer pretty much onto '90', the correct scale point.

So this AVO was basically working OK, but the calibration circuit was out of spec.

I recommend that, if you have an AVO that won't mains cal,
(i) check the meter first anyway, so you can be confident it's OK, then
(ii) check the heater voltage with the selector on 6.3 V (see above for loaded voltages, and also the table on Page 10 of the service manual), and
(iii) if the heater voltages are correct, *double-check* for 200 V at T2 primary and - if all that makes sense - take up Martin's tip that the calibration circuit is likely at fault.

You may wonder at my fussiness over heater voltage. Using an ordinary 12AU7, I found (for grid bias of –8.5 V) anode current 4.7 mA and gm of 1.8 mA/V at heater 6.3 V, but this increased to 5.7 mA, 2.05 mA/V at 6.9 V, and fell to 3.7 mA,1.5 mA/V with a heater at 6.7 V.

If you are matching valves (for output stages, say), you may assume that, so long as they are both tested at the same heater voltage, your match is accurate.

I've not tried to do this at, say, excess or insufficient heater voltage, but I reckon you need to test under working conditions.

That said, for *accurate* gm readings on lower-current valves (12AU/AX7, 6J5, EF86, etc) you may need to adjust for correct heater voltage according to a meter connected across the heater pins at the top socket.

This *will* put the mains cal meter a bit below '90' - I've not tested the instrument's accuracy under those conditions.

Thanks again,

Ian - Historical Radio Society of Australia (at hrsa1.com)

p.s. the gripe about the diagram probably sounds like a Grumpy Old Man. It's really a Grumpy Old Draftsman With Forty Years On The Job.



Ian.

[Craig Sawyers]

I agree totally about the circuit schematic. The cryptic diagram is in common with all AVO's valve testers. The CT160's schematic is just as bad or worse.

By contrast, Tektronix and HP classic era products are clear and easy to understand.

I'm happy though that you seem to have the MkIV sorted!

Craig

[ianbatty311]

Hmm....

I'm seriously thinking of doing one (or two) articles on the MK IV and the CT160.

Maybe I'll just redraw the schematics - I *did* redraw the Philips EL3302 Cassette Recorder (Australia's Silicon Chip magazine, July 2018) for one of my articles, so how much harder can a manky old valve tester be?

...or is that a question one should never ask?

:-(

Ian.

[Dekatron]

More articles on the AVOs are always welcome but before you spend your time on doing so you could check this forum for the ones I've written on these testers, one is mentioned above which describes the inner workings of the CT160, which also compares it to the Mk III and Mk IV and there is also a special mention of the VCM163 too (here's the link again: https://www.vintage-radio.net/forum/...ad.php?t=86262). At the end of that document there are some references to other threads here on the forum like why AVO split the circuit in the CT160A for insulation resistance where the "calibration" resistor is discussed.

You can download a lot of AVO documents from here: https://frank.pocnet.net/instruments/AVO/index.html and here: www.nvhr.nl (you have to search for them under "Schema's" in the left menu and enter AVO in the "Merk" field)

There are no 100% correct schematics on any of the AVO VCMs that AVO have published in their documentation to my knowledge, the two best ones is an old one from 1960, the one Jacmusic has scanned, (you can see the Grid Volts design best in this one, but there are some omissions and also the Leakage Switch is hard to understand) and a later one from 1966 (you can see the Leakage Switch best in this one, but here the Grid Volts schematic is wrong). I corrected the one from 1966 a long time ago and I think I put it up here on the forum but can't find it right now.

There has been a few discussions on putting a Sticky Thread here on the forum with information on how to properly do a check and calibration of AVO testers but unfortunately nothing has come from them.

[Dekatron]

Originally Posted by ianbatty311

Hi to all who chipped in.

.......... portion deleted .........

I measured the meter, and found an FSD of 33.5 microamps.

.......... portion deleted .........

Thanks again,

Ian - Historical Radio Society of Australia (at hrsa1.com)

p.s. the gripe about the diagram probably sounds like a Grumpy Old Man. It's really a Grumpy Old Draftsman With Forty Years On The Job.



Ian.

I'm glad to hear that you got it working!

However since the meter has an FSD of some 33.5uA and you adjusted the "calibration" resistor to get to the SET AC line at 90% you now have 30.15uA at that line and you will then have 24.8uA at the 1mA/V line and these two points should really be 27uA at 90% and 22.2uA at the 1mA/V line - this means that your readings will be some 10% off.

For these AVO VCMs to work properly you have to start by making sure that the meter has an FSD of 30uA precisely at 3250 Ohm internal resistance, if not you'll have to either re-magnetize it or adjust it with shunt and/or series resistors to get those values, otherwise all of the measurements will be off, it is unfortunately not enough to adjust the calibration resistor. The circuit relies on the meter having exactly those specifications, otherwise all measurements will be off.

[Station X]

Not forgetting the magnetic shunt inside the meter.

[ianbatty311]

Hi, Martin,

Thank you for the comments - how do they address the fact that I found the calibration resistor to be high, and that correcting it brought the indications back to correct?

And on the article, I will certainly look at what you have already written. I would be writing for the Australian Electronics magazine Silicon Chip, or for our own HRSA Radio Waves.

I'm aiming for the popular readership, and will certainly (as I always do) give a list of references for further reading, for fact-checking, and for getting alterative perspectives. You might like to check a sample of my work on:

https://www.siliconchip.com.au

and do a search for the article on the EL3302 (June 2018).

Hi, Station X,

About the magnetic shunt - yes, I would probably have gone to that depth if it were my own meter, or if I had a spare.

Given the delicacy of the AVO meter, I was not going to mess with equipment that I would need to hand back in operational condition, even if - as Martin suggests - it would give a low-but-known indication in use.

Cheers,

Ian.

[Dekatron]

It is quite common with these old testers that the resistors have gone high. I once spoke to a gentleman who worked for Welvyn resistors about all of the resistors used since the AVO Mk III up to VCM163 and he told me that they were never intended for this long usage. He knew that most of them didn't have a sealed outer coating so they corroded over time due to the moisture in the air. They also become more temperature sensitive over time and that also changes for the worse over time he told me, this is especially affecting the black painted resistors. Only the green ceramic power resistors with a clear or green coating were intended for longer use as they have a sealed coating.

Lets compare the AVO valve tester to a normal AVO 8. If you find out that your AVO 8 suddenly reads wrong on one or more ranges you will have to check if the AVO 8 still reads correct on a few ranges as a first step, if it does its moving coil meter is probably within tolerance and you will then have to find the resistors involved in the circuit that results in incorrect readings. However if the moving coil meter is out of tolerance all ranges will show an error (some might still have resistors that needs to be replaced if they are more out of tolerance compared to other ranges).

However when it comes to an AVO VCM it is not as easy as with an AVO 8 to perform similar checks, but you still have the service manual with checking and calibration instructions. If you follow them step by step you can usually figure out if it is the meter or the range selecting resistors that are out of tolerance. As with the AVO 8 you must know that you have a working moving coil instrument in your AVO VCM to be able to use it to check the readings against when you perform the calibration procedure. You can replace the moving coil meter with a 3250 Ohm resistor with a DMM in series (you will possibly have to adjust the total value for the internal resistance of the DMM) and read the current flowing there, it should then be 27uA at the SET AC line at 90% and 22.2uA at the 1mA/V line. These currents will only flow if you have the correct secondary voltages and the correct value of the calibration resistor as per the calibration instructions - you can of course reach the same currents for other voltages and values of the calibration resistance, but then your measurements on the other ranges during check/calibration will be off.

The important thing with the calibration resistors is not only to act as a calibration means, they are also used in the isolation tests as they supply the isolation measurement with a known current - the isolation test is a resistance measurment which is shown on the outer scale. So if you don't have the correct current flowing in the circuit with the calibration resistors, nor the correct FSD of the meter, your measurement will be off on these tests too.

If you read the circuit comparison document I refer to above you'll find the explanation for the "fudge" factor of 0.52 as it has been called by a few people. It has been written as 0.51, 0.515, 0.52 and 0.525 over the years in different instruction/service manuals by AVO. It took me quite a few years until I understood where it came from as I couldn't find any maths equations that would explain it.

I've included a corrected circuit diagram of the AVO Mk IV that I made many years ago, almost 10 years ago I see now, where I traced the circuit from one of my working AVO Mk IV VCMs, there might still be errors but I traced it as well as I could, when I couldn't get correct readings when I wanted to understand the inner workings of the AVO VCMs. It and the corrected circuit diagrams for the AVO CT160, CT160A and VCM163 were the base for the circuit comparison document I wrote.

When it comes to the magnetic shunt it works but mostly has very little effect on the movement if the magnets are weaker than intended, it is mainly used to shunt away excess magnetism and will seldom work much if the magnets have become weaker, then it works best if you remove it completely - I was told this by a gentleman who serviced them for Herts Meters Ltd when I asked about it. However if you use a few neodymium magnets to increase the magnet strength you can sometimes use it to decrease the flux enough to reach 30uA FSD.

A friend of mine, Euan MacKenzie, in Australia wrote a few articles on AVO VCMs in some magazines where he and I collaborated. I am looking forward to your contribution to the many AVO articles and documents out there! Feel free to send any questions you have to me, I'll try to answer them as best as I can. You are welcome to use references to my material.

[ianbatty311]

Hello, Martin

Thank you for the revised schematic - it is much clearer than what I have in the service manual, and it does appear to clarify the Mains Cal position.

I was puzzled by the SE switch (why not S1, S2, etc, :-(, but IIUC, on leakage it tests from the selected element to *all others*.

Am I correct in this?

I will go look at what you have written, also.

Cheers,

Ian.