The VCM163 Oscillator Unit

[Bazz4CQJ]

For quite a while, the idea of replacing the oscillator unit in the 163 with some entirely “new design” has been on my mind. More recently, I’ve been wondering about “re-manufacturing” the standard board (4 x 5 inch), i.e. making a brand new PCB and populating it with good quality new components, with only the transformer and the thermistors being carried over.

The 2N2926’s would be replaced by BC574’s; this issue has been discussed on this forum in detail. For the electrolytics, a decent Japanese make (Rubycon or Panasonic) from a major UK supplier would be used, rated at 25VDC. Axial component were used originally, but the PCB could be made to accommodate radial components as well. I think pots from Bourns would be used for VR1 & 2.The 6 pins which provide connections would be replaced by two 3-way PCB mount connector blocks.

If anyone has any other ideas about other “refinements” to an updated board, please feel free to make suggestions. I’m only thinking about making a couple of “DIY” PCB’s (in my garage); but if anyone is interested in having one, let me know. I'm currently trying to find somebody to cut up a 'few' 4 x 5 boards for me. If I actually do this project, a follow on would be the same thing for the amp unit, which, of course, also uses a 4 x 5 inch board.

B

[Radio Wrangler]

Amplitude control is quite difficult in any oscillator for decent precision, and its accuracy affects the Gm measurement accuracy.

So I'd make a square wave oscillator using CMOS to define the swing from a well-regulated 5v supply. Than I'd use an active filter to convert it to a sine. Quality will be good, accuracy good and it'll settle quickly. No ****dy thermistors

Freq can be tweaked to match the tuned metering transformer.

David

[Bazz4CQJ]

As you know, the question of how best to deal with these ageing boards persists. I first had problems in form of the electrolytics needing replacement about 4 or 5 years ago, then failure of the transformer last year, and Martin and others have had to deal with the 2N2926's doing funny things.

I think the original circuits were "adequate for purpose" for most requirements when everything was working well. As you will recall, when David ran his "round robin" exercise with the "standard valves" he had characterised by hand, the 163 owners (and this was a while ago) generally saw pretty good results. Additionally, in my own tests using valves I had standardised by hand, my 163 gave very acceptable results.

I doubt that I have the skills to develop new circuits along the lines you quite rightly suggest. "Re-manufacture" has the advantage of being easy and will keep the purists happy if I were ever to decide to sell the thing.

B

[Radio Wrangler]

If I wanted to soup-up a valve tester, I certainly wouldn't pick a VCM163 as the starting point, simply because of the whomped-up prices due to cult status. Applies to pretty much all AVO VCMs. Some people managed to pick then up at reasonable prices before the cult moved in, some even rescued VCMs destined for scrap. But for anyone starting from scratch, the financial step is ot sensible. Add in the fragile meter business and it just gets worse.

They are nice, collectable machines in their own right, And I'd rather like to have one again, but my Taylor 45D is perfectly adequate when I want to check that a valve is OK.

As an instrument designer, I'm more swayed towards DC supplies along the lines taken by Swordholder and David Simpson, though I'd probably make it a curve tracer as a computer peripheral. Wouldn't it be good to call up a library of datasheet graphs and have a machine that plots the characteristics of your valve superimposed on them?

Towards such a venture, the expensive AVO would be reduced to simply a donor of a heater transformer with switches, and a collection of valve bases with their switching. So it doesn't make sense to use one as a starting point. Even if you already had one, your accountant would tell you to cash it in and start from scratch.

So I think I've come full circle, and AVO VCMs ought to be repaired and kept going just as they are. Certainly any modifications to skirt the need for unobtainable parts need to be internal and invisible to the outside world, for else the instrument's glamour will be lost.

David

[Radio Wrangler]

Originally Posted by Bazz4CQJ

I doubt that I have the skills to develop new circuits along the lines you quite rightly suggest. "Re-manufacture" has the advantage of being easy and will keep the purists happy if I were ever to decide to sell the thing.

No one was born with those skills. All those with them acquired them along the way. Other people didn't because they were involved in other fields.

You could set out to acquire them. Yes, you'd be starting later than some people did, but also earlier than others. What is this forum for if it's not for passing along skills and knowledge which would otherwise be lost? There are plenty of folk hereabouts to provide guidance, suggestions and hand-holding as needed, so it's a comfortable place to do some learning in. No exam, no timescale, unlimited syllabus, no formality.

With each new thing we learn, we open up doors to doing things we couldn't do before, and doors to learning further things. There is no end, so the fun and entertainment are infinite.

An audio oscillator for a VCM163 is something that can be breadboarded on a piece of raw copperclad board, with a bench power supply and an oscilloscope. It would make a great starting point and would have an immediate application.

David

[Bazz4CQJ]

Many thanks for your encouragement. Developing new circuitry would certainly be a challenge, but I think it would take more time and the 163 has already taken quite a lot. I have a list of projects, started but not completed, and it would be good to get back on to some of those. I think that might be the deciding factor.

B

[Craig Sawyers]

Regarding valve curve tracers:

First there was the Sophia from Audiomatica http://www.jacmusic.com/Tube-testers...ofia-Index.htm . Long defunct. I actually looked at acquiring the design at one stage. But they wanted £30k, including a box of bits. Things like the codec were obsolete even fifteen years ago. I don't think so matey.

Then there is the exceptionally expensive Amplitrex https://amplitrex.com/ , but there is a somewhat unhappy review here http://www.jacmusic.com/Tube-testers...rex/index.html

There is also the Roetest, with critical parts available to build as a kit http://www.roehrentest.de/EnglishInfo.html

And the excellent utracer https://www.dos4ever.com/uTracer3/uTracer3_pag0.html , a Euro 300 highly developed and good specification kit which has the parts and circuit board. Box, valve bases etc are down to personal choice.

Craig

[Craig Sawyers]

This seems a pretty comprehensive comparison of computer based valve testers

http://www.jacmusic.com/Tube-testers...PARE-INDEX.htm

Seems a new addition to this site.

Craig

[Bazz4CQJ]

I think that for the very limited number of valves that I'm interested in testing these days, I'm content to take 5 or 6 measurements on the 163 at different Vg settings and plot the results in Excel.

Whenever I've done that in past, that little exercise has come up a value for Gm which is pleasingly close (within a few percent) to the value the Gm meter tells me using the 163 in the normal 'automatic' mode. But of course, that it what it should do .

Obviously, owning one of the modern curve-tracers would be very nice.

B

[Radio Wrangler]

It's an interesting balancing act between the need for some sort of valve testing instrument to help out when needed to either fix some set you're working on, or to sort out the ones worth keeping from a valve collection and the liking for having a nice instrument just as an end in itself.

The need is really just for something fairly basic, while the desire for an end in itself favours something special and almost jewel-like. Knowing where to stop is something the human race in general, isn't very good at.

From the point of view of design, you really want datasheets with statistical bounds on the variation of parameters, both as shipped and as aged over a reasonable life. You can't measure this from just one or two samples. But it is what you need to know to design something which can be manufactured and will just work unless some component is out of spec.

Even it you're only intending building one, then making it dedicated to the measured characteristics of one individual valve can extract the most performance, but at terrible cost if you need to replace that valve.

So whether you're fixing something or building something your, need comes down to just something to say whether a valve is in spec or not... but better if it indicates probable life expectancy in terms of parameters which age. Beyond that the data sheet rules.

David

[Bazz4CQJ]

Well, I think it was a German who described the 163 as "The Rolls Royce of valve testers" in the BVWS mag (?). Of course, RR is now owned by the Germans. I think my view is that a fully working 163 is all I need; all that has been / is lacking now is reliability. I think most people on this forum are happy with the capabilities of the Avo instruments, but I realise there are others who might like something better (possibly the audiophile community?).

B

[joebog1]

The link Craig posted re: comparison of computer based valve testers is excellent!!.

Looking at the prices asked, one can see why I am building my own.
As I have mentioned in the past, all I am really doing is making a bunch of power supplies that have very accurate metering. So all manual and no curve tracing. It will have an oscillator for modulating the grid to synthesise real signals though. I have been collecting the bits and pieces for some time and just about have everything needed. My budget is somewhat limited so it takes time to accumulate all the stuff thats needed.

Joe

[Bazz4CQJ]

I think I have decided to proceed with building a "replica" of the original 163 oscillator, changing the transistors to BC547. Retaining the original circuit rather than developing something new, will help retain confidence that old and new measurements are the same. It will also save me a lot of time.

However, thinking about the amp board, as along as the input transformer / filter is retained, the amplification there could be anything. On my 163, I'm already using a "bolt-on" op-amp to provide the drive for the Gm meter, which probably has shorted turns. The use of op-amps to correct for meter deficiencies in the Avo testers is a fairly well-established practice. So, "updating" the amp board and making it capable of driving "any old meter" would be a benefit. Perhaps it could all be done using a single IC?

B

[Diabolical Artificer]

There's a simple low distortion wein bridge oscillator circuit that uses a TL072 opamp that may fit the bill. It is a bit boingy as David says when first powered up but it settles quickly, it can drive a panel meter but may need an emmiter follower adding on. Power supply is +15v 0v -15v but a little PSU can be knocked up using those little PCB tfmr's Rapid sells.The lamp is a 28v automotive lamp, I have some if you need one. Lastly some values may need changing, I drew this from memory, I may have got the odd value of a resistor wrong. Adjust the 500r (or 470r) preset for a stable sinewave. The 100k pot could be subbed for a preset as this can do 20v RMS.

Andy.

[Bazz4CQJ]

Thanks for the circuit design, which I'll put on file. I have a 2-transistor Wien oscillator, which seems to work very well and I also spent a while looking at this XR2206 module https://www.ebay.co.uk/itm/224526656555.

However, the key question is, "what's the objective here" and it is simply to restore reliability to an old piece of test gear. The two PCB's in the 163 start off my needing the electrolytics changing, but then there's also been questions crop up about the transformers and the transistors. The quality of the boards themselves is not great so re-working them is an issue.

Having thought about this a lot, I think the design of the oscillator is adequate for purpose. In 2022, the reliability is an issue. One solution is to simply build a brand new "clone" board, re-using just the output transformer and the thermistors.

In the case of the amp board, that too has a design which was adequate for purpose, so long has you can go and buy replacements for the 50uA meter which it has to drive. But that's no longer possible and those meters can be a problem (mine certainly is).

So, I think at this stage, the most productive use of time in making my 163 more reliable is to produce a new amp capable of coping with either damaged or alternative meters. In fact, some time ago I built an AF millivoltmeter using a single LF353 IC and I think that could be the core for a new amp board.

The PCB material for this project arrived this morning, and by mid-afternoon, I was "cutting metal" .

B