AVO MKIV VCM, anode voltage change

[vahakuja]

Recently I was testing four UL41(NOS) tubes with my AVO MKIV VCM. The test setting was as instructed on AVO Valve data manual (edition 15th): Vh 45V; -Vg1 9V; Va 175V; Vg2 175V  Ia 54mA and gm 9,5mA/V.
I got the following test results: Tube1= Ia 56mA, gm 8,0mA/V. Tube2= 66mA, 5,0mA/V, Tube3= 80mA, 4,4mA/V. Tube4= 80mA, 6,2mA/V. It was strange that the anode current did outgrow so much over nominal values and as well gm was clearly lower especially on tubes 3 &4. Then repeating tests of tubes 3&4 and using Vg2 as 150V gave “normal” values ie. abt. 55mA and 9mA/V.
This led me to check what shall happen to the anode voltage, when the anode current increases. By changing -Vg1 to achieve several current values I measured voltages between anode-cathode and sreengrid -cathode top connectors. Ia 5mA) Va 87Vdc/Vg2 87,4Vdc. Ia 10mA) 87V/87V. Ia 15mA) 85V/87V. Ia 30mA) 83,5V/86V. Ia 55mA) 80V/85V. Ia 80mA) 77V/84V. So actual anode voltage has decreased abt. 20volts.
Is it normal in AVO MkIV VCM, that this kind voltage drop occurred and what might be the reason for it.?
Later by re-examining other newer Valve data manuals (editions 18th and 23rd) I did notice that test values for Va was 200V and Vg2 175V.

I’ll happy to get some directive what to do if any.

Regards Ake

[Hartley118]

You may be encountering the rather common problem of oscillation which can happen in the AVO VCMs and gives unusual results, particularly abnormally high anode current. Mine tends to do it with 807s.

Even at the best of times, DC voltage measurements tend to be puzzling because of the rectified AC waveform used by the VCM. Add RF oscillation to that and it's anyone's guess what you'll read. There are various cures suggested, including ferrite beads on the internal wiring. I haven't messed with mine however and just assume that a valve that's got enough emission to oscillate is probably healthy in other respects.

Martin

[Craig Sawyers]

I think the IV (and III) has ferrite beads at periodic intervals around the valve wiring loops already. Perhaps adding grid stopper resistors and/or screen grid capacitors to cathode (ie ground) as well might be effective in stopping certain valves from hooting.

They would have to be at the roller selectors of course, so might not be as effective as in a real circuit where they would be as close as possible to the valve.

Craig

[Dekatron]

It is possible that you have high contact resistance in any of the switches or contacts involved in the measurement loop which then results in different voltage drops depending in how much current the valve draws. Turn of the tester and turn all rotary switches, potentiometers and toggle switches a few times and also the roller selector on the top to see if the problem goes away. There is a chance that the contacts on the cut out relay need cleaning to get rid of the contact resistance there, if there is any charring/blackening of the contacts from previous cut outs.

Oscillation as mentioned above might also affect the measurements.

There are ferrite beads to stop oscillations in the wire loops on the Mk IV but you can still experience oscillations.

Try to start with a more negative grid voltage and slowly turn it to the value in the data book and see how the measurements goes. But first start with trying to get rid of any contact resistance.

/Martin

[vahakuja]

Thank you for comments and directions to solve the voltage drop problem. I shall to make closer study of anode voltage trail between rectifier diode MR5 and A sockets. MR5 has been replaced early with new Si version. Cleaning of all proper contact might be worthwhile task to do.
Probably the anode and sreengrid voltages are clean and there isn’t any oscillation. Look at attached picture.

The tube I used are (Nos?) and it is uncertain are those according the data values.

Regards
Ake

[Dekatron]

I spotted that you are using a Beckman RMS3030 meter!

I had two of those, bought them new a very long time ago, also had an RMS3020 if I remember the number correctly.

When I started to check the voltages on the AVO series of VCMs many years ago I could never get a correct reading and never succeeded with calibrating them. After speaking to some knowledgeable people I found out that the RMS3030 doesn't work well at all on half wave rectified voltages, only on full sine wave voltages - it just can't make the correct measurement due to how it is designed internally. My RMS3030 meters showed something between 3% and 12% incorrect values on half wave rectified voltages!

I wrote about it here on the forum, so search for Beckman and RMS and you will find the threads.

You really need a True RMS reading DMM which can handle the half wave rectified voltages, or an old analog meter that works, otherwise you won't get correct readings - not all modern DMMs will make these measurements correct either!

I use Fluke 87V for my measurements, I bought that meter as I asked Fluke if they will make correct measurements and they told me that they would as long as the signal you measure is within some levels on the range you use and also if the crest factor is within some values depending on if you measure full or half wave rectified voltages.

So I can only recommend you to use another meter for these measurements.

[Radio Wrangler]

I should point out that the metering in the AVO valve testers isn't true RMS, it's just a plain DC d'Arsonval moving coil instrument which has been scaled to read whatever AVO wanted, but the pointer responds to the rectified signal just as a non-RMS meter would.

Any multimeter is either scaled to give the true RMS value of a sinewave, while responding to the peak of the rectified voltage. This scaling is OK only when an entire sinewave is presented. On a halfwave recified sine it will be scaled wrongly.

Trying to get a meter scaled to read comparably to what an AVO valve tester is scaled to read on the waveforms present in these machines is difficult. A proper true RMS meter should read the true RMS value of any waveform (within limits) but that's not what the AVO reads on its internal meter.

Difficult!

David

[ms660]

It says in the manual I'm looking at to use an ordinary analog meter to check the anode DC voltage, eg: Avo 7 or Avo 8.

Lawrence.

[Dekatron]

You can use a digital meter - BUT, and that is a big BUT - you need to know that it works correctly under the conditions that you will use it. Unfortunately a lot of digital meters don't work correctly with other waveforms than pure sine wave AC or pure DC.

My own measurements started out exactly this way - the wrong way with a digital meter not up to the job - with a Beckman RMS3030 meter that had served me well many many years but then only under pure sine wave AC or pure DC conditions. I had sold off all of my old analog meters a long time ago and unfortunately didn't have any to compare the readings to as I trusted the DMM. I wish I had kept just one analog meter as that would have saved me many many hours of pulling my hair in frustration about not being able to calibrate my AVO VCM MkIV.

When I asked a friend about the strange readings I got he asked me to compare to a good analog meter, I bought an AVO 8 Mk7 and found that it showed the correct readings that I was expecting and that the Beckman RMS3030 didn't.

Then I started to look into what DMM could give me the same correct readings and the only way I could find that out, with my limited knowledge on this subject, was to ask the manufacturers what their DMMs would show on the waveforms that I was interested in measuring compared to what an AVO 8 MK7 would show.

I finally bought two used, but still in calibration, Fluke 87V as I liked their handy size and good reputation and I could only afford second hand ones but most importantly since the Fluke support personnel told me that they would do the job correctly, they even showed me what to expect and how to calculate some tolerance figures from their datasheets. They do the job properly under the limits explained by Fluke, there are limits to the voltage/current you can measure and also on the crest factor of the signal you measure. For crest factor, RMS and MEAN voltages explained I can recommend this webpage/PDF-file: Bull-DMMglossary-E.pdf as it has it all summed up nicely with simple explanations.

My recommendation for an analog meter is to use an AVO 8 Mk7 (or any earlier model up to the job) that you know give correct readings or for a digital meter a Fluke 87V as I know that both of these will work in calibrating an AVO VCM properly. AVO often states that you should use meters that have been calibrated at some voltage, that is to ensure that you get the correct readings on important calibration points as well as a way to compare your meter against a known standard to see that it is still in calibration. With moderns meters that is usually not a big problem, as long as you use a meter as discussed here that works under these conditions, or use one that you know is within its tolerances.

If you are computer savvy enough you can use LTSpice and simulate the internal design of an AVO VCM to see that there are no magic waveforms inside an AVO VCM, just pure sine wave voltages and half wave rectified voltages that are combined with some resistors passing through a meter and the valve under test. LTSpice will show you the curve forms, how they add up, and also the Mean (Average) and RMS voltages and currents for all measurements. LTSpice and Microsoft Excel have been my main tools in understanding how the AVO VCMs work. The AVO VCMs works due to how physics work in our world and how that is expressed in mathematical equations shown in AVOs patents and not due to some magic formulas - however magic would have been a lot nicer in my opinion!

Just as I started out with meters that weren't up to the job many other people have done the same "mistake" without knowing about the problems you can get into. I didn't know enough of how a DMM worked compared to an analog meter and getting to this point where I am today took me many many years of reading books and asking on forums and asking manufacturers. Like me, other people asking questions, has probably reached an end where searching a forum isn't enough as they can't find the "right" answer among all questions and answers on a forum so I started a new thread and asked and usually I was referred to a thread with the answer or it was kindly explained once more to me on the forum or in a private message.

Please also have a look at this thread: AVO CT160's & MK3's & 4's VCM's on some more AVO discussions relating to this topic and the value of starting a sticky thread with advice on how to calibrate an AVO VCM. Such a thread would have to evolve over time as different people will add their knowledge to the thread and someone might have found good short cuts or additions to the knowledge expressed in the thread. Having two ways to check the measurements and calibration is always good, just look at this: Measurements of the gravitational constant using two independent methods.

[vahakuja]

I do happen to have several old analog and digital voltage meters in my radio hoppy workshop. To meet my curiosity, I took six of those and measured VCM anode voltage on connector A1. UL41 was again as the test tube and settings: Va175V, Vg2 150V, -Vg10,5V Ia abt. 40mA. Pls look at the test results from attached picture’s table. Those “voltage” meters are not calibrated after hard industrial use. 100V range was selected (if available) and measurements were made both DC and AC selections.

Beckman gives different result when AC selection was done. Two years old, low cost Chinese Aneng, value of 15 euros, seems to show similar results as few others in different price class!

Regards Ake

[Dekatron]

Do you get the same results on all tubes, even the ones where the anode voltage varied?

Do you have an oscilloscope that you can use to check what the anode and grid voltage looks like when you are testing these UL41s?

[vahakuja]

Oscilloscope pictures of anode and g1 voltages taken from UL41 and EL84 valves are attached. Both valves are working abt. 50mA anode current. EL84 anode set voltage was 250V and measured dc-voltage at 123V (Ia ~0mA) and 116V (Ia=47mA). Anode dc- voltage drop is thus 7volt. UL41 settings as earlier.
As known the effect of small anode voltage change to anode current of pentode is not significant but change of screen grid voltage Vg2 is.

Because the conditions of UL4 (4) valves I have is uncertain it will better to test those first with pure Dc-voltages. This hopefully eliminates the question why test results deviate so much comparing to AVO valve data manual.
Also, when having better time, I’ll try to identify where loaded anode voltage drop happens between Va -voltage rectifier and A1 connector.

Thank you for your comments.
Ake

[Craig Sawyers]

I think those voltage waveforms are just about right. You need to read and understand AVO's 1948 patent 606707

[David Simpson]

Great photos Ake. This is the road to understanding how these AVO VCM's work which I've been pushing for some time now. Referring to photo 2 - the scoped continuous peak grid waveforms(approx. -11.6V) might well produce a DC value on a DMM of -7.3V, but remember that only one peak -ve grid waveform per sinewave period is in anti-phase with the Anode's peak +ve waveform(i.e. only every other grid waveform, not every one)(thus controlling the valve's Ia). So actually only has a DC value of approx. -11.6/2 whilst the valve is conducting for 10mS. In reality, for calibration purposes, AVO in their Manuals don't explain this.
Cleverly, AVO can quote a particular Ia for a valve in their Valve Data Manual, and yes - the valve can attain that but just for a very short period at the waveform crest, 50 times a second. But in fact, the DC (RMS) values of voltages & Ia are only ever approx. a half of the peak (AC) values. AVO have saved weight & expense by leaving out bulky chokes & electrolytics, and their VCM's, when calibrated, can often get within a few % of pure DC testers, or ultra-modern digital circuitry testers. Even though they're now vintage items.

Regards, David