Questions on check and calibration of AVO Mk III and IV testers

[vintage port]

Re. circuit, I'll.have to dig it out but it used a simple, ancient, very low bandwidth instrumentation IC from Mouser I suspect. It sits on the rear of the meter, screwed onto the two terminals as you indicate. It must take a few volts from a filament winding, and a simple bridge rectifier. The setup also allows you to offset the zero point as well, which is not (to my mind) remotely significant. As an aside, meter linearity isn't a biggie to me either, the main readings are, or can be, taken at near FSD. A few percent won't change any decisions made about valve suitability.

[TomAW-60]

Thanks VP for lining out the set-up for the opamp fix! I’ll spend some time reading a bit on opamps to get more understanding of how they function (if you happen find a circuit drawing that would be highly appreciated!)

Taking of drawings - I was struggling to get this current splitter set-up correct for measuring the meter, and seem to totally have missed/ignored the very nice drawing presented in post no 9… apologize for this… (I could have dropped my post no 17…)

I fully agree with your comment about the measuring of tubes - in practice we probably will not need these very exact numbers for Ia and Gm. An emission tester saying good/bad would probably suffice. I presume it’s about the potential higher precision these AVO VCM testers can provide if calibrated correctly - then again balanced versus work put in. I think of this (a bit ambitious) Mk 3-4 project of mine as a way to learn more, perhaps help me to service my point to point wired tube amplifiers on my own in the future. Sound wise, tubes rule!

[TomAW-60]

Switching to my Mk 3 now:

I ran the meter test on my Mk3 as well (no 3910, type3). I’ve been very lucky with the meter on this one, meter showing 99/100 and very close to 3250 ohms. Did the 90 degree test, both ways - and almost no deviation of the pointer position The previous owner seems to have been Swedish, taking good care - it’s written “Rør Ej” = “Do not touch” in big letters on the front. Maybe that’s the trick for these AVO testers!

Then I tried to test an EC86 valve (used in my DDC). And then the Ia measure fainted. Suddenly no Ia reading at test position. I then checked the voltage over the Ia link on top, which gave 0VDC. Opening up and looking inside I spotted one electrolytic cap - if I understand the schematics correctly it must be C1(?) - looking very old (see picture), so measured and got OL. I guess this explains what happened? From the part list it’s specified as “8uFd Electrolytic”. Would this best be replaced with a metal polyprop of the same value?

(I checked the meter once more, and it was ok).

[vintage port]

Check the resistance of the switches. These can go high but are rarely irredeemable, they just need cleaning.

[VT FUSE]

I fitted the diodes and capacitor to my mk3 & mk4 VCM’s, did before and after comparison using a known good valve and could not see anything affected on the scale by any more than a needle width after fitted ( my notes) at the time I did no further measurements but it was definitely worth it to damp the rate of needle traverse. I did not try to emulate a scenario that shot the needle fsd however.
I used polypropylene film caps in both my testers but if I ever chose to increase value to 100uF it would be axial Tantalum or smd type.

Mike.

[TomAW-60]

Thanks for advice/comments! My Mk3 8uF cap measures 21nF - I thus conclude I should replace this cap (and hopefully be back in operation?) - and also do the same in my Mk4 (also fitted with an electrolytic cap, however looks like a replacement, age unknown). In the mk4 there is also a C2 cap (rated 0.01 uF +/- 20%). Perhaps avoid electrolytic type here as well?

The switches feel ok (I can try to measure them, but will have to find out how to do it). One related observation though: When measuring an E88CC before the “incident”, setting Vg (neg grid v) of -1.2v (by the way, the last 3 digits of AVO code “...510” should be 410 for E88CC/ECC88!), the scale is not very precise and just a small touch on the knob will have some impact on the Ia reading. Any smart way of knowing if -1.2v is set accurately (and not 1.1/1.3) - some form of side v check?

As for cleaning - I see some do worn against using some of the cleaner fluid types as it will just make roller/switches become even more sticky in the long run - any safe ones? (isopropanol?)

Thanks for suggesting a 100uF cap type to look for in the meter protection circuit! With reference to the parts list (attached) - is there any general rule as to what V and resistor W rating (and perhaps brand) to go for (Mk3 and 4)? I also see that prices vary quite a bit…

The OA210 diodes listed - are there more modern types to look for?

[vintage port]

Originally Posted by TomAW-60

The switches feel ok (I can try to measure them, but will have to find out how to do it). One related observation though: When measuring an E88CC before the “incident”, setting Vg (neg grid v) of -1.2v (by the way, the last 3 digits of AVO code “...510” should be 410 for E88CC/ECC88!), the scale is not very precise and just a small touch on the knob will have some impact on the Ia reading. Any smart way of knowing if -1.2v is set accurately (and not 1.1/1.3) - some form of side v check?

As for cleaning - I see some do worn against using some of the cleaner fluid types as it will just make roller/switches become even more sticky in the long run - any safe ones? (isopropanol?)

The OA210 diodes listed - are there more modern types to look for?

A lot of questions in here. The switches are robust but IMHO cleaning involves disassembly and thorough cleaning of the mech, a bit boring but...

The grid volts IS a bit sensitive but not unduly so. Some pots develop dead spots and measurements then become jumpy. This is "a bad thing" because I have read that the pots are difficult to replace. This is hearsay however.

The AVO manual is wrong!! Err, yup, quite regularly. Sometimes a lot more than that. For rare or precious valves, cross reference with a datasheet first.

I haven't read all the thread, are the 0A210 in the protection circuit for the meter or used as rectifiers? At a guess, if the former, 1N5820, 21, or 22 might be better?

[vintage port]

As for negative grid volts, to check this, from memory, you quite specifically need an AVOmeter 8 MK4. The grid volts are pulsed and derived from a sine wave. You therefore need an averaged volts reading, not RMS, and allegedly only that AVO does this, others may know better? Either way, a regular DMM is near useless for this task, maybe a modern oscilloscope might do the maths for you?

Luckily that AVOmeter is not difficult to pick up in decent condition.

[TomAW-60]

Thanks. When reading some of the documentation provided in a thread by Martin (Measurement definitions), there seems to be a fixed relationship between mean DC volt and RMS volt for a correct sine wave? Would “RMS=1.11 Mean” hold true? (meaning I can convert the specified AVO mean values to RMS values - thus can use a DMM for the measurements when taking such a conversion into account?)

[TomAW-60]

Yes, I’m afraid the list of (perhaps basic) questions seem to add up on me as I try to gain insight. As I will have to replace the C1 caps:

For a replacement of the C1 cap (which I read can be 8-10uF) - will this cap see max 222V RMS? (and not the 444V at the max Ia voltage setting, ref the "half wave rectifying method"?) Or is peak V the important figure? Will a 250V cap suffice?

[Chrispy57]

Re. Post #29, unless your waveform is actually a pure sinewave then those relationships do not apply, sorry. Is the waveform you are trying to measure with your DVM a pure sinewave? I suspect not, since the Avo valve testers use a halfwave rectified derivative of the mains sinewave in some areas of their circuitry.
There are quite detailed explanations of this in posts on here by Avo experts, that you should be able to find with the search facility.

[TomAW-60]

@Chrispy57 Thanks for clarifying - I will then have to go look for an AVO 8 meter (hopefully one that function ok)

[Dekatron]

You can use a TRUE RMS DMM for al measurements, as for instance the Fluke 87V - however you can't use an old Beckman 3030 RMS DMM as that one is NOT a TRUE RMS meter and it will not give correct readings.

You do not need an AVO 8 or an AVO Electronic Test Meter nor the EA113 for any of these measurements, it's just what they had at hand as DMM's weren't really invented at that time - there were some DMM's early on but they were prohibitively expensive compared to an AVO 8.

If you don't believe me just hook up an AVO 8 in parallel with a Fluke 87V and compare the measurements.

[vintage port]

Does it not come down to the crest factor of the DMM, Martin? I am soon out of my depth but what I understood was that the half wave rectified, pulsed grid voltage was further rectified by the physical mechanism of the AVOmeter. Happy to be enlightened, I most certainly need it!

[Craig Sawyers]

Originally Posted by TomAW-60

Switching to my Mk 3 now:

“8uFd Electrolytic”. Would this best be replaced with a metal polyprop of the same value?

(I checked the meter once more, and it was ok).

That capacitor is only in circuit for testing power rectifiers. The idea being that in a valved rectifier circuit in practice, it is loaded by 8uF smoothing.

It is a red herring for testing anything other than rectifiers.

Craig

[TomAW-60]

Thanks, that’s helpful! Yes, in the schematics I now see that C1 is connected to D2. I jumped to a conclusion here… should replace that cap anyway, but no explanation to the “incident” then.

[vintage port]

Originally Posted by vintage port

Does it not come down to the crest factor of the DMM, Martin? I am soon out of my depth but what I understood was that the half wave rectified, pulsed grid voltage was further rectified by the physical mechanism of the AVOmeter. Happy to be enlightened, I most certainly need it!

To TomAW and Martin; my apologies. I dug out my calibration folder from way back to find that the negative grid volts is a simple half-wave rectified sine and not particularly pulsed at all as I was incorrectly remembering. So any reasonable modern DMM should suffice, no Avometer needed. I've ignorantly been causing trouble for you both!

As a punishment, thinking to assess readings between my DMM and my scope I discovered that the rectifying diode for the grid voltage has now futzed out in my Mk4. Hey ho, another job on the list! Tom, you're not alone on the re-calibration journey, I have re-joined it

[TomAW-60]

All good, VP! Nice that you confirm that we can use a DMM for calibration (that is a True RMS!) - and good to be in company on Mk4 work

But this made me think - what is actually displayed on an AVO 8 meter? If it’s mean values, we should multiply the values with factor 1.11 to get to RMS values. However, it seems like what is displayed on the AVO meter is actually RMS values (mean values x 1.11) - I asked ChatGPT about it and got (and hope it's correct!):

“How the AVO 8 uses this
The AVO 8 (and other classic analog multimeters) actually measures the mean of the rectified voltage or current, but its scale is calibrated in RMS volts or amps under the assumption that the waveform is a pure sine wave. So, internally:
- The movement measures a DC proportional to the mean of the rectified AC.
- The scale then multiplies that mean value by 1.11 to display “RMS” volts or amps.
This works perfectly only for sine waves.

What happens for non-sinusoidal waveforms
If the waveform isn’t a sine wave (e.g. square, triangle, distorted AC), the ratio between RMS and mean changes. For example:
Vrms/Vmean
Sine 1.11
Square 1.00
Triangle 1.155
So, if you measure a square wave with an AVO 8, it will over-read by 11%, since it assumes the waveform is sinusoidal when it’s not.

“True RMS” meters
A True RMS meter instead measures the true power-equivalent value by electronically squaring, averaging, and taking the square root of the waveform.
It gives the actual RMS, regardless of shape or distortion.

[vintage port]

I've never looked in an AVO 8 but assume it rectifies the waveform with a couple of diodes and that's that. I read somewhere that it uses the sheer physical inertia of the meter mechanism to smooth out the ripple and display an 'accurate' reading, without visible vibration. I don't know but assume it was made with 50 Hz in mind as a lower end frequency?

[Dekatron]

It has helped me a lot to think like this:

A modern instrument should of course work as well as an old one, both have their limitations too of course but if you read the instructions and follow them you will get correct readings within those limitations - however understanding when those limits apply is up to the user to do, this is usually where things go wrong in my experience. Like RMS and TRUE RMS confuses people as they are not the same and different DMM's for instance always don't tell you which they measure (usually cheaper and older DMMS's measure RMS and more modern and more expensive measure TRUE RMS)

Something similar goes to the inner workings of a tube/valve when it is being tested - the tube/valve doesn't know anything about its surroundings, it just obeys physics laws and when being presented with identical (after conversion) test voltages no matter if they are DC, short pulses or an AC voltage it will work in the same way if those voltages correlates between themselves as the physics laws does between those voltages - so translations between Peak, RMS, Mean, DC & AC holds true.

I've included a PDF which shows the correlations.

Now, don't forget that the calibration factor that AVO mentions as being 0.525 is the scale factor on the Grid-volts control and has nothing to do with any AC/DC/RMS or other voltages to do at all! Avo has written this calibration factor as 0.515, 0.52 and 0.525 in different documents and that has confused people over the years trying to wrangle with some calculation to get that factor to fit. I did too until I suddenly realised that it was just the scaling factor on the Grid-volts control when I plotted the values in Excel. I have described it in the huge AVO document I wrote in this thread: https://www.vintage-radio.net/forum/showthread.php?t=86262 in chapter 2.3. AVO also did a nice design for the mA/V control to get a non-linear mA/V control to get higher resolution at lower values.