AVO VCM - mains quality and calibration

[fastflyer99]

I've not written much on this forum, so hello to everyone I've certainly read a lot though, in particular, about the Avo Mk IV VCM. I was lucky enough to be offered one by a chap whose father used to work at Saunders Roe here on the Isle of Wight, which was rescued from disposal a good many years ago. It's in excellent condition, and in the months I've owned it, I've had the meter serviced by the Herts Meter Company, and I've changed out of tolerance components, of which there were relatively few. A leaky electrolytic capacitor was the worst that I encountered during the work...

The biggest issue was actually outside the VCM; the mains quality in my workshop, as I discovered, is quite poor. Mains voltage fluctuations are quite marked, and the waveform is not great, which doesn't make for the best qualitative measurements on an Avo VCM, dependent as it is on a decent supply. Whilst Avo tried to compensate for mains variation with a switchable tapped mains transformer, I found it to be hard work trying to keep the needle on the mains calibration mark during testing.

To get around this, I decided to try using a pure sine wave supply to provide a clean and stable mains source. I worked out that a 150W source would do the job, and I managed to find a pure sine wave inverter eBay. I built a 24VDC stabilised power supply for it, and tried it with the VCM. Avo, in the service manual, caution against regulated supplies primarily (as I recall) because of harmonics potentially affecting the DC mean voltages, so I checked the VCM with several different valves both on the inverter and on the standard mains supply. The only difference I observed was that the inverter supply was totally stable and made life very easy; the "mains" voltage doesn't fluctuate at all now, no matter what valve is tested, or if the DC supply varies. I have a couple of NOS CV491 calibration valves, the characteristics of which I plotted on a calibrator based on the one in the Avo service manual, and the readings on the VCM are pretty much spot on. Incidentally, I also built a simple solid state regulated HT supply for the calibrator, which makes it much easier to plot an accurate slope for the calibration valve

Anyway, I'd be interested to hear any comments about this. I don't profess to be an electronics guru by any means, having returned to electronics three years ago after a break of many years since being a radio comms field engineer (I used to work for Storno), and these days I run a workshop repairing live music equipment. More than anything though, I'd like to thank all the contributors on this forum who have written so much useful information about Avo VCMs which helped immensely in restoring my Mk IV. In particular, the work that Dekatron has carried out has been invaluable - I'm very grateful for to him for that. I will likewise do my best to contribute in terms of VCMs, wonderful pieces of kit that they are, and when I can find time, in regard to other pieces of equipment where I have experience...

Mark

[Dekatron]

Thank's for the kind words!

It would be very kind of you to share what make and model number the inverter you bought has.

I have been using AVR's, Automatic Voltage Regulators, professionally made regulators consisting of a variac that is controlled by a small motor that is driven by a voltage sensing circuit to regulate the output voltage and also motor-generators which produce a pure sinewave output voltage. I've also tested a few mains filters used to "condition" and remove interference that we use at work with power line communications and they helped some. I've also tried a few mains voltage isolation transformers with magnetic and capacitive stabilization circuits and they worked well too, not generating any harmonics that upset the measurements. But so far I have not tested any modern inverter.

Since the AVO VCM's (except the VCM163) rely on the 50Hz signal (50-400Hz) they work quite well as long as the input mains power has a nice and clean signal looking like a sinewave. When the sinewave gets distorted you sometimes get incorrect readings, this will be both because of incorrect voltages but also because of harmonics which the valve picks up and amplifies which then affect the reading. You will also get interference from switching power supplies and power inverters that use low frequencies which spill over on the mains power lines. These two last also affect the VCM163!

All of this makes it harder to calibrate an AVO VCM without proper calibrated valves that you check the readings with. You must of course be sure that all components are within their tolerances (meter too) and also what readings you can expect on each and every one of the measuring ranges, and that all switches are clean so you don't get problems with contact resistance. Each range has to be checked with a diode and appropriate resistor (potentiometer) in series, to simulate a valve, before you make any attempts at calibrating the VCM, only then is it any idea to insert a calibrated valve and compare the results.

Checking only with one valve at one point on the valve curve doesn't really show much so you should have several valves with different characteristics that you have plotted to check against to be sure that you get correct readings on several grid voltage settings for a full calibration run.

I use my RoeTest valve tester to "make" calibrated valves, apart from using new valves that work as they should. I also save every valve I come across that has developed a problem like gassy and leaky valves, valves with poor measurements and other defects as those are also important to check with to see that I get the same readings on those as with new valves.

One thing that the AVO VCM's are much better at testing than modern computerized valve testers are is with valves that oscillate easily - the AVO valve panel construction is very very good at stopping oscillations. Part of this problem is of course with the wiring of modern computerized valve testers but also with the designs that really don't have any protection for what happens when a valve starts to oscillate which then affects the whole voltage regulating circuits getting them to swing like mad - I've cooked a few valves that way until I had a chance to realise what was happening, then I modified the RoeTest and also always use a grid stopper resistor with valves that are known to oscillate easily, and always with pentodes! I still have some modifications that I want to do to the RoeTest but I never seem to get time enough to do them.

[Bazz4CQJ]

I have been using a 'ferro-resonant' stabilising transformer made by Cetronics http://www.cetronicpower.com/product...e-stabilisers/ to power my VCM163. I have to admit that I don't fully understand the physics behind this device, and I suspect that they are expensive to buy, but this one was "donated" to me.

As for calibration valves, some time ago I took pair of brand new ECL80's and obtained their characteristics "by hand" and use those as "standards". The idea of using ECL80's was presented on a website somewhere, these being very inexpensive valves, and each bottle actually contains 2 valves to offers two test points.

[fastflyer99]

The inverter I'm using is a Sterling Power Products Pro Power S 150W unit. I paid about £12 for it - the company (charlesterling) advertises on eBay, and sells off remanufactured inverters. There's one on there now: I think the 150W versions are cheap because they are less useable than the higher power versions. Another benefit is that if you overload the inverter, it will shut down, offering another layer of protection for the Avo VCM, as I inadvertently discovered when I was working on it recently...

As for calibration valves, I've so far plotted the CV491s between 10 - 20mA (Vg ca. 7V - 8V, triodes paralleled), and an ECC83 at lower current ratings, with the results correlating well when tested on the Mk 4. I think the idea of using something like an ECL80 is a very good one, and I want to plot some valves with higher anode currents and varying gms as well. To this end, I've also fitted an octal base on my calibrator for future work.

Very interesting to read about your methods of providing a clean stable mains source though (and bazz4cqj) - I'll post any updates to my experimentation as I go.

[Bazz4CQJ]

Another valve which I'm adding to the my standards list is the PCL86, which is another cheapie (unloved old TV valve) and runs higher Ia. My 163 had quite a few faults on it when I first got it, mostly the usual ones which are described in the sticky on this sub-forum. I think that I have resolved most of them, but it is useful to be able to plug these standard valves in to check all is still OK.

[glowinganode]

I'd recommend the use of a motor driven variac type stabiliser, which is what I use with my CT160.
I picked up an "Autostac" unit quite cheaply which works really well. Saves continual rechecking and setting of the mains tapping adjuster.
I also sucessfully used a small UPS which can be picked up quite cheaply, killing the mains supply whilst testing so the CT160 runs off the internal battery and inverter.
I find it also helps to carry out testing late at night when the mains voltage tends to be more stable.
Rob.

[Radio Wrangler]

The ferroresonant CVT also fixes the mains waveform, which is quite important, and a motor driven variac can't do that.

David

[Bazz4CQJ]

Yes, the ferroresonant CVT is a wondrous device, though it does have the mass of a small planet

[Radio Wrangler]

Not all CVTs give a sine output, some are closer to a squarewave with rounded corners and this throws out the accuracy of peak-reading voltmeters with 'RMS' scales.

Small planet? What's that in AR88s?

David

[glowinganode]

Originally Posted by Radio Wrangler

The ferroresonant CVT also fixes the mains waveform, which is quite important, and a motor driven variac can't do that.

Granted, but how important is the waveform in the operation of the CT160 et al?
Rob.

[Radio Wrangler]

Well, it's using a DC moving coil meter to measure rectified mains, and that usually makes assumptions about peak to mean to RMS ratios. Even if if the bridge arrangement cancels it all out, the waveshape only has to affect the 'Set∼" operation and it will throw out the chosen bias voltages and the heater voltage.

David

[n_r_muir]

Hi,

In the AVO patent all of the calculations are based on a pure sine wave, if you have a distorted wave form, the harmonic content will render the readings invalid, in as much that a "reference valve" that gives a particular set of results when tested with a sine wave mains input will give different readings when tested on the same machine fed from a distorted waveform.

If you just want a basic indication that the valve under test is working more or less OK
then it will probably be adequate.

[glowinganode]

Originally Posted by n_r_muir

If you just want a basic indication that the valve under test is working more or less OK then it will probably be adequate.

This is essentially all you can expect from these testers.
Sure the theory says you need a pure sine wave, but the real world conditions are far from theoretical.
You have an under-rated transformer on the verge of saturation with significant internal resistance, connected to a highly non-linear device. The rectifiers are somewhere between mean and square law detectors, the meter is probably the only linear component (if it hasn't gone o/c!)
Don't get me wrong, I really admire these testers but these are at best a compromise.
Rob.

[Radio Wrangler]

There are many compromises in the bench-top valve testers, and mains waveform would be just one more. The problem is that a cult has grown up believing them absolutely. And they don't believe anyone who says otherwise.

I use a valve tester very rarely. When I do, all I want to know is 'Will it work when I plug it in my radio'.

If I felt a need to match high power valves, I'd make a hefty DC tester with regulated supplies and the capability of running the devices to their actual operating conditions. It would make sense to build a curve tracer if I was going to go to that effort.

As you said, Rob, the transformers are under-rated and a number of people have managed to burn them out.

David

[glowinganode]

Originally Posted by Radio Wrangler

And they don't believe anyone who says otherwise.

AVOphools?

Rob.

[David Simpson]

Mark, heaps & heaps of past AVO VCM info, CT160's in particular, can be accessed via the Forum's "Search" facility.
The VCMs only attempt to pass current through a valve(Ia) during the 10mS of the +ve half of a mains 50Hz sinewave. ie 50 pulses a second. But out of that 10mS pulse, the valve is only working at full efficiency when close to the top of the peak. Therefore, a waveform which is not fully sinusidal will further increase the margin of error under which these AVO VCMs work.
Also, please remember, CT160wise in particular, - once the valve's voltages have been selected - when testing - the only thing that controls Ia significantly is the -ve Vg Pot.
The coarse & fine "Ia" controls only "balance" the meter's bridge network.
I've attached a "Very" simplistic diagram of how a CT160 works under AC & 1/2wave rect. conditions.
Always remember, safety-wise, - the full whack of "mains derived" HT is present on the meter's terminals. Should you be delving inside.

Regards, David

[fastflyer99]

Interesting to read the comments. Today, after replacing a couple more resistors which were marginal in value on my VCM, I carefully re-checked the calibration of the Avo, and then measured my test valves again. The two CV491s, which have been carefully plotted on my calibrator (which has regulated supplies), both give almost identical readings on the VCM compared to the calibrator. Likewise, an ECC83 and an ECC81. Clearly, these are only three different types of valves, and all triode valves from the same family, but will (when both triodes are paralleled inside each valve) draw up to 20mA or so and have been tested with grid voltages anywhere between 1 and 9V, depending on which valve was on test. All readings have been well within Avo's tolerances, both gm and Ia - in fact extremely close to the calibrator readings across the range. I also have an EL34 which I have checked on my calibrator in the same conditions that the VCM uses. Here, I observed Ia at 80mA, and gm at 10 mA/V. On the Avo, I observed 74mA Ia and a gm of 9.8 mA/V - a little out on plate current, but still acceptable. The screen and anode are actually at the same potential for this test (250V), which does cause Ia to sag somewhat across R36, the transformer probably struggling a bit with the screen current as well.

So far though, I think the results are very encouraging. It is true that a test of a valve in any of the conditions it is to be used, is the only true test, but so far I'm seeing strong correlation between the results of my voltage regulated calibrator. I'm looking to improve the calibrator to try a greater variety of valves, and I will publish the results on here when that is done. I think the VCM does have some shortcomings, maybe in the internal resistance of its transformers and its inability to measure some low impedance triodes, and probably doesn't warrant the super elevated status that some people confer on it, but I have to say that the results I'm seeing so far are impressive - especially bearing in mind the age of the technology in use. It's a wonderfully elegant design those guys in Avo came up with.

[David Simpson]

Mark, you're spot-on, with your conclusions regarding the shortcomings of mains derived VCM's, and your decision to enhance your calibrator.
If you already have a MK4, which has heaps of empty space inside the framework, why not consider fitting a 9way C/O switch to divert the valvebase panel connections to your modified V/Reg DC standardizing calibrator. The switch, and a 10way Jones Plug & Socket connection could be fitted on one of the removable side panels.
It would mean that you'd greatly benefit from the versatility of a multitude of valvebases & the code setting thumbwheel switch.
I was extremely lucky in aquiring a duff MK2 with a pristine VB Panel as a donor for my LeyMarcAVO, which is just a glorified version of the DC standardizing rig which AVO feature in the first few pages of their VCM Manuals.
Right enough, mains derived VCM's offer a good enough indication of a valve's "Replace"(Red) - marginal(White) - "Good"(Green) condition. And some enable you to plot a reasonable Ia/Vg graph. But - what a difference using a fully regulated DC versatile tester makes, I can asure you ! And, it places the valve in the DC enviroment under which its meant to work in all radio/electronic equipment. No fannying about - worrying wheather the "Set ac" has change when your wife puts the Sunday roast on, or your neighbour fires up some hefty diy machinery.
Another advantage of using an AVO VB panel is the ability to externally monitor Ia via the Anode links, and any of the valve's electrode voltages via the adjacent wee sockets.

Regards, David

[Radio Wrangler]

The AVO valve base panel is also well salted with anti-oscillation beads etc.

When you're testing a valve whether at 50Hz or DC, you're biassing it up into its highest gain (and therefore friskiest) state.

David

[David Simpson]

Two Davids in total agreement. With a pukka gen AVO VB, AVO have already designed the inter-base wiring layout specifically to avoid any interactions or parisitic oscillations. And, as David says, the beads are also there for that purpose.
Mark, if you were to consider using your MK4's VB Panel, and also consider my suggestion about mounting a multi-contact switch & Jones Plug - if you have any reservations about compromising the MK4's value - just fit an alternative side panel made out of paxolin, laminate or a gash sheet of aluminium.
Hey, just thought - you Mark,could call your new apparatus - The "Mark's4-AVO" DC VCM.

Regards, David