Mullard HSVT 7600/4

[David Simpson]

Folks, a vintage radio pal asked me to look at his HSVT yesterday. He's had it a while but not used it, and wanted it checking out. I'll be honest & admit knowing sfa about HSVT's, but managed to work my way through the Mains/Reject/HT checking procedures.
Lot of fannying about, sometimes the crt's dot would just oscillate up & down, sometimes stayed steady. Couldn't get the dot to within 3mm on the HT check - more like 5mm.
In anticipation, a couple of days back, I'd tested & standardized & Gm graphed an NOS Mullard EF37A on my DC Valve Tester. Perfect valve.
The HSVT's Card 63 was in good nick, but Test6 -Emission - plonked itself firmly in the Red zone! Then tried several other of my pal's own Mullard spare valves & their cards - No emission.
So, bollox, have fetched it home. Removed several hwt of outer casing & started from basics. Also found the missing wee label off the coarse mains selector. DMM'd my own mains voltage, and made coarse & fine adjustments whilst also monitoring V5's(EL37) pins 2 & 7 - Heaters. Got a spot-on 6.3V. Also tested V2 & 3's(EF91's) Heaters - 6.3V also.
Now, I managed to deduct that all those sec windings on the mains T/F are on the outer side. But do their circuitry references(S1 to S9) or their circuitry connections(1C, 2C, 3C,4C, etc.) appear written in white ink on the transformer ? No they do not. Just numbers. Jolly confusing.
So, now popped in my super duper EF37A, and Card 63, clipped my brand new Newlec DMM to the valve's pins 2 & 3 - Pulled the lever( & noted that aprox1/3 movement activated the micro switch & thus applied the heater voltage to the valve under test,), and observed 8.5V being supplied - not 6.3V ! In fact, like AVO VCM's - all the valve holders have their pins wired in parallel. So checked several.
Yes, I've checked a number of HSVT threads via the "Search" facility, but so far haven't seen reports of faults concerning significant differences between the windings which supply the tester's own valve's heaters,(which obviously need to be constantly spot-on), and the windings for valves under test.
As the 6.5V Pilot Lamp is straight across S7 I expected to measure 6.5V-ish, but only got 3.6V.
Well, that's as far as I've got.
Any help or advice most appreciated by this HSVT novice.

Regards, David

[frankmcvey]

Hi, David,

There are a couple of wee gotchas with the HSVT wiring that are quite hard to spot from the E7600/4's wiring diagram, the way it's drawn. Better to look at the diagram for the military version when it comes to what pin on the valveholder is connected to what pin on the gate switch. (the CTA20 diagram).

Everything is nicely parallelled at the top row of the gate switch - you'll see it more clearly if you undo the 2 x 2BA s that secure the front of the valveholder panel and hinge the panel back for access - but the wiring from the top row of the gate switch to the actual valveholders ain't necessarily in parallel. So Pin 2 on the IO socket actually comes from contact A7 on the fixed gate contact; however this same contact goes to pin 4 of the B9G and pin 7 of the B7, for example. So you can't assume that that the holders themselves are actually in parallel.

Also, the pilot light is in series with a resistor, R22 (10 Ohm) so you won't be seeing the full S7 winding voltage of 6.3V if you measure at the lamp.

Also you mention that you checked the EF37A's heaters between pins 2 and 3 - was this a typo? This should of course, be pins 2 and 7 again.

Another good snippet is to read the manual for the CTA20 - it gives a good description of how the self-test cards set up the machine, with nice little circuit diagrams for each set-up, and the component numbers and values are (mostly!!!) the same as for the Mk4.

As always, the waxy caps on these machines are usually iffy, so I'd check on them as well as the low-stab resistors around the indicator chain. The mains card circuit ( a simple bridge) for the CTA20 is a good start.

Hope this helps a bit.

Cheers,

Frank

[David Simpson]

Many thanks Frank for all the helpful info.
Yep, my typing error regards the EF37A's heater pins. You mentioned the Manual for the military version CTA20. Would that be a maintenance/calib. Manual rather than the sub-technical Operating Manual just written for the end-user? Is it easily downloadable, perhaps from VMARS ? I haven't checked yet.
Right, me being an AVO enthusiast, just assumed(wrongly) that all valve holder pins were in parallel, numerically. I'm glad you've put me right. I'll delve further. Oh yes, the circuit diagram which I was given with the tester - definitely shows LP1 wired directly parallel to S7's terminals 2B & 1B. No 10 ohm dropper.
Off to work in the garden now. Such a lovely day. Will update later.

Regards, David

[frankmcvey]

Aye, the CTA20 manual is a real-deal test-equipment bay AP, as you'd expect and be well familiar with. You'll find it on George's, the Mullard Test Card man's, website at the bottom of this page, along with the correct circuits for all the Mullard HSVTs. If only all manuals were that comprehensive...

Shouldn't take a man of your experience long to crack it with that info!

FWIW, I've never seen any heater problems on the HSVTs I've worked on. The transformers are quite vulnerable to damp problems, but that's usually down to bad storage, and your mate doesn't sound like the sort of chap who'd do that!

I forgot to mention as well that, if the machine hasn't been used for some time, a good clean of the valve sockets, the main test selector switch, the push-buttons contacts and the gate switch might pay dividends. The gate switch comes apart for cleaning a LOT easier than you'd expect, and you'd be surprised sometimes by the amount of FOD that finds its way down the card slot into the gate switch.

[David Simpson]

Frank, thanks again for the gen. I'll certainly download George's copy of the AP. Right enough, we were spoilt for choice of AP's & Maker's Manuals in the RAF T/Eq. Bays I've worked in(Cosford & 103MU, Akrotiri). Sadly, from Mullard's point of view, we only ever had AVO CT160's to service & maintain, valve tester - wise. I can vaguely remember an HSVT coming through 103, for assessment. But it was rejected, mainly for it's susceptibility to FOD & damp ingress. Mind you, some ex Pongo or Matelo wags might suggest that RAF guys should be good at removing deposits from slots.
So my fondness & familiarity with AVO VCM's just comes from those far off days. Not any AVO orientated snobbishness against Mullards. Mind you, although I've only had this heavy beast for a couple of days, I'm already enamoured with its quality & robustness.
The previous chap who had it before my local pal had done a great job in replacing heaps of sus capacitors & resistors. However, I'll certainly follow your advice regarding cleaning. That gate switch assembly certainly frightened me, I must admit.
Yep, I've un-hinged the valve holder panel, & Bobs your uncle, 6.3V - wise on the IO's pins 2 & 7.

Regards, David

[frankmcvey]

Yep, the gate switch, on the Mk4, at least, is a doddle. It's only held together by the geometry of the various parts, no tools needed.

The hole in the card lever actuating link just slots on to the pin on the crank, so with the lever in the "Off" position, just slide the link on the RHS of the machine off the crank. Note the position of the crank wrt to the machine body - it usually sits at around 10 o'clock IIRC.

Then have a look at the rear of the machine, where the crankshaft runs through the two cams that cause the gate to slide in and out. There's a red line on each cam, so make a note of their orientation. Then it's just a question of withdrawing the crankshaft (from the RHS). You'll usually see a red spot on the underside of the crankshaft to assist you with orienting it again when you refit it.

The cams will now be free, so remove them from their guides, and the whole gate moving contact assembly should slide off rearwards on its 4 metal dowels. There's plenty of slack on the cable loom to get it completely away from the machine, although you might have to fiddle it past where the loom runs below the gate switch on the lower rear of the machine. Once the moving contact assembly is clear, you'll see a plastic mechanical interlock on the lower left as you face the rear of the machine. This is what prevents the two contacts behind it from applying power to the internals if someone tries to close the gate switch without a card in the slot. The plastic shutter on this gizmo isn't as robust as the rest of the machine, so take care in that area.

Check the contact springs are good on both sides of the gate while it's apart, as well as cleaning the brass knife-edges of the contacts. I use a very fine triangular Swiss file. It's probably not wise to use too much solvent cleaner on the gate. The 260 springs are untreated steel, so you don't want to wash the protective grease off them (if there's any left after all these years!)

As they say in the Haynes Manuals, re-assembly is the reverse of the stripdown procedure!

Like you, I had my reservations about these machines at first but they quickly won me over. Not much good for matching your KT66's, although after a bit of practice, you do develop a feel for the relative "goodness" of same-type valves. However, if you've a bread tray full of valves to work your way through, they're hard to beat; they don't seem to give false positives or false negatives and as you say, their build quality is outstanding. Also, there's no expensive meter to wrap around the stops if you get a switch wrong - the little green spot just flies off the screen...

Cheers,

Frank

[David Simpson]

Once again Frank, thanks for the info. Will keep me straight, I'm sure.
The VCM/Tester enthusiast in me suggests that, if I owned one, I'd be tempted to fit a nice little cheapo edge mc meter, & with added rectification would calibrate it to read one of the 6.3V windings - as a permanent ref to a correct mains supply. Bit like the "Set AC" facility in AVO VCM's. I agree that the CRT method of display must be a much safer & cheaper method than those horrendously expensive 30uA AVO meters.
I've drawn a sketch of all the sec winding's numbering & wiring. And voltages observed.Not sure if Mullard kept the same wiring colouring code throughout their range of HSVT's.

Regards, David

[David Simpson]

Folks, still no answer to an ident of the transformer tags ? I've identified half of them, but it would save some fannying about.
I've attached a copy of the standardized graphs of the EF37A I'm using in the HSVT. The HSVT still gives the valve 218V Va, 104V Vs, and -2V Vg, and connecting an ammeter in the IO pin 3 connection to the matrix switch contact shows a healthy 4mA Ia. But still the HSVT puts the "Emission" Test 6 Dot firmly in the Yellow zone ! V's 2,3,4,5 & 6 have all been either tested "good" on my Mk3 or replaced with NOS. I've repeated the Mains, Reject & HT Card tests many times - Spot-on.
Mutch head scratching by myself, and Frank McVey, who has been a great help.

Regards, David

[David Simpson]

Quite a coincidence that Stevehertz is now also experiencing Test'6' Emission problems. Hopefully, we'll both get a result soon.
I've un-hinged the gate assembly & cleaned all the contacts with a wee bronze wire brush. Also, gently went over all the contact recesses with a fine triangular file. Greased up all the mechanical linkages & bushes, etc., but didn't spray on any contact cleaner. Aren't I a good boy ?
Repeated all the Mains, Reject & HT Tests again - all spot-on. Tried my Standardized EF37A again - still the emission dot is firmly in the amber zone!
I've even tried PJL's R41 voltage tests, and get within 5%. Next, tried a thumping big EL34 - at last got about a third of the way into the Green zone, but the dot doubled in size & became fuzzy. Also the valve became seriously HOT, even though it was only in for a couple of minutes. Will investigate further. Oh yes, I've been back over 6 years of HSVT thread posts, via the 'Search' facility, and can't find any reference to similar Test '6' Emission problems.
I've trawled through both the AP & the Op. Manual, and note that neither the MOD nor Mullard, mention anywhere recommending using a Standardized Valve. Surely, out of their vast range, one could have been considered which would be expected to fall in the middle-ish of the green zone. Perhaps using an EF91 or EL37, which HSVT owners might hold as spares. What do the Forum's HSVT enthusiasts suggest ?

Regards, David

[ukcol]

I think the reason there is no documentation on using a "standard" valve to check the Mullard HSVT is do do with the fact that it is a "go, no go" tester rather than a measuring instrument. The AVO testers are, on the other hand, a measuring instruments and a standard valve is appropriate to check calibration.

Don't get me wrong, I'm not criticising the HSVT, I own one myself and I am a big fan. I thing you only need an AVO if you are matching valves or selling valves and want to quote their mA/V.

[frankmcvey]

Hi, David,

I've repeated the EF37A tests on a second HSVT, a Mk 4 this time. NOS valve, the spot sits 1mm above the dotted line during the emission test.

With Card 63 in, the gate closed and no valve in the valveholder, and Sw pos 6 selected, I'm seeing Va= 206V, Vs= 114V and Vcg = -0.979V. If you remember, we calculated that the grid voltage should be around - 0.9V from the scaling resistors that are selected when Card 63 is used.

The voltage across R41 was -0.78V, so pretty close to PJL's figure of -0.72V at the dotted line.

The fact that you're getting a good result from the Mains, Reject and HT cards would seem to indicate that the diff amp and the rest of the indicator section is OK.

What voltage are you seeing across R41 when you're testing the EF37A, and where exactly is the spot in relation to the line at the bottom of the green section?

Cheers,

Frank

[David Simpson]

Aye, Colin,
I realise that an HSVT is just a "Go, No Go" simple tester. I'm just looking at it for a local vintage radio pal. I've no intention of using it for accurate measurements. But surely, that EF37A of mine should put the dot somewhere in the Green zone. Middle at least, seeing as its a jolly good NOS one.
Surely, over the last 60 years, either Mullard themselves, or keen radio/TV shop owners, or more recent HSVT vintage enthusiasts, must have discovered a particular valve which exhibits 'middle green' readings, if its New or NOS. Whether it be a an EF, EL, ECC, or whatever. If someone can suggest one, then I'll tabulate & graph it on my AVO Mk3 & my DC tester. Then try again. I'll even willingly post it off to a keen Forum HSVT enthusiast or two to assess.
As I said in my first & third thread posts, I'm just an HSVT novice, and I'm not trying to demean HSVT's just because of my keen past interests in AVO's. They are a well built, sturdy, and a simple to use piece of test equipment. Long may they survive in the vintage radio fraternity. However, something within this pal's HSVT is putting a jolly good standardized EF37A( and some other spare NOS EF37A's of mine) firmly in the amber zone.

Regards, David

[David Simpson]

Frank, got out of sync with answers. Re the R41 readings - I got -0.39V at line 'B', ie. top amber/bottom green. Dotted line - I was -0.79 = pretty close to yours. However, I still think that your -ve Vg should be close to -2V, mine is. I just cant get a 'Green Zone' reading for my EF37A.
However, back to the 2nd hand EL34 I tried earlier, more tests, :-

AVO Mk3, Va = 250V, Vs = 250V, Vg = -13.5V, Ia = 70mA(should be 75mA), mA/V = 9(should be 11), lower green zone reading.

HSVT, Va = 281V(253V when warmed up), Vs = 257V, Vg = -13.6V, Ia = 90mA(should be 75mA), lower green zone reading(between A & B).

So this HSVT likes EL34's, and dislikes EF37A's, coloured zone-wise.
I've taken a picture of the IO pin3 link I made for reading Ia. Plus I've scanned the Cards I have for EF37A(63) & EL34(709), just in case someone spots something wrong with them.

Regards, David

[David Simpson]

Right Frank & folks, once again - popped the EF37A back into the HSVT. Connected my new Newlec DMM across R41 and the following ensued :-

Test 1. Heater continuity, Dot half way between lines B & A, and DMM = -0.617V
Test 2. Electrode Ins., Dot bitty higher, & = -0.617V

Test 3. H/C Ins., Dot spot-on dotted line A, & = -0.803V

Test 4. El. Ins., Same as Test 2.

Test 5. I grid. 2/3 up between B & , & = 0V.

Test 6. Emission. Just above line C in Amber zone, & = -0.303V

Note. By the way, during those tests, adjusting the 'Fine' Mains SW ACW or CW altered R41's voltage quite significantly, but not the Dot.

Further note. Can only get close to dotted line A separately - by using the HT Card. Actually 1mm above. Observed R41's voltage as -0.788V.

Tests on Diff. Amp :- Whilst EF37A on Test 6.
Anode/Screen Voltages - - V2 = 356V, V3 = 400V

Vg - - V2 = 215V, and Dot shoots to the very top of the screen. V3 = 210V & Dot shoots to the very bottom of the screen. ( Due to loading of DMM ?).

Joint Cathode connection = 221V

Fit else can I dae ?

This HSVT journey is certainly exercising my brain cells. Thank you, Mullard.

Regards, David

[MotorBikeLes]

I know nothing about these Mullard machines, but is it possible that the fault lies in the spot deflection circuit, NOT the combined Va, Ia and Vg.
I shan't attempt to think about it, but maybe somebody else will. (When all possibilities are exhausted, try the impossible).
Les

[David Simpson]

HSVT Controvosy, I hope not:-
At the back of the AP there is Part3 Table2. Seems to be the last definitive test, if everything else above it passes. But hey - what a lot of fannying about delving inside the chassis, from one side or the other, or underneath, to get to the various elusive circuitry test points. The weight of the beast doesn't help either.
Solution - take all the test points to one tag strip, and also read PJL's R41 voltages on the same tag strip. Taken up a fair bit of time, soldering, looming, lacing, etc. but boy - heluva lot easier to take measurements.
See test sheet - everything looks tickety-boo.
Fit else can I dae ?
Oh yes, I raised a query regarding Mullard recommending a standardized valve, a few posts back. However, (see red highlight) - they just say " VALVE TEST. Using known good valves with their appropriate test cards". Which is just what I've done. It certainly doesn't say " Don't use an EF37A, cos it'll fanny you about"!

Regards, David

[frankmcvey]

Hi, David,

Sorry for the absence - when Grandson visits, I don't get near the computer!

OK. With your EF37A as the VUT, you're seeing -0.3 across R41, with the spot just above Line C. This is about the deflection you'd expect for that voltage, so that would seem to prove that the indicator section is telling you the truth. This would seem to be borne out by the fact that when you feed R41 with -0.78V during the HT test, (and during the EL37A tests for H/K insulation @ -0.8V) you achieve the correct spot deflection to the dotted line.

As far as I can see, in the indicator section, there is no further scaling applied to any test results after R41 (other than during the grid current test, when R33 is shorted out), nothing else to muck up your test results. So it would seem to me that you have a perfectly serviceable indicator. It's giving the wrong indication because it's getting the wrong test result from R41.

Let me do a little more figuring and I'll come back to you.

When you're in despair, it's always worth bearing in mind the old truism "If the right components are present, serviceable and correctly connected and configured, and you give it the right power supplies, it MUST work". Well, it isn't working, so at least one of those statements isn't true!

NB - I like what you've done regarding the test point wiring modification. I've done several wee temporary lash-ups in the past to get test results more conveniently, but nothing so well engineered or permanent as your mod. Very nice workmanship. It would certainly have helped all of us if Mr Mullard had done the same at the outset!

Cheers,

Frank

[frankmcvey]

Hi, David,

Thinking firstly about our grid voltages to the valve under tests, where we have a discrepancy. I'm measuring -0.982V at the grid of the EF37A under test; you're measuring around -2V.

I'm calculating that the grid voltage for the EF37A should be -0.999V for the Vg for the EF37A; PJLs HSVT Calculator reckons that it should be -1.98V.

Looking more closely at your Card 63 image in an earlier post, it differs from mine; mine is the original Card 63, yours is Card 63 Revision A. The main difference in in Row I. Row I contacts happen to select the bank of resistors that control the grid bias voltages fed out to the valve under test.

The original card has holes I3, I4 and I5 punched; these connect resistors R14 (31.2K), R13 (62.5K) and R12 (125K) in parallel, giving a total resistance of 17840.805 Ohms.

The Revision A card has holes I4, I5 and I6 punched; these connect resistors R13(62.5K), R12 (125K) and R11 (250K) in parallel, giving a total resistance of 35714.285 Ohms.

The machine then uses this combined resistance in a potential divider consisting of R18(1.5 Megohms) plus the combined resistance, strapped in series between the -85VDC reference and Earth.

17840.805/(17840.805+1500000) = 0.11754; multiplying by -85V = -0.999V, which agrees with my calculations and my observed measured grid voltage.

35714.285/35714.285+1500000) = 0.0233; multiplying by -85V = 1.976V, which agrees with PJL's calculation and David's observed measured grid voltage.

To be fair to PJL, he did make the caveat that he didn't know what revision card 63 his HSVT calculator program was using, and I should have paid more attention!

Well, all of this explains why we're seeing different grid voltages, but it doesn't explain why you're still getting a deflection in the amber. Card 63 resembles Card 63 Revision A in every respect, other than the grid voltage selections mentioned above.

I've got to go and do some other stuff now, but we'll have a go again later.

Cheers,

Frank

[David Simpson]

Thanks Frank for sorting out the Card 63 discrepancy. Even Mullard's own EF37A Ia/Vg Graph puts the Grid at -2V at the 4mA point. I wonder if there are any other cards that throw up this discrepancy for particular valves.
Also, more thanks for your kind words regarding my modification. One of my RAF Instructor jobs was teaching looming & lacing. Wish I could get my hands on the original black lacing. I've white, white waxed & fine black cotton. Although I'm rather rusty.
I've made a further mod today - two terminal sockets behind the flap, which monitor V2/V3's 6.3V Heater supply. A click one way or other of the Fine Mains Switch quickly keeps the monitoring DMM spot on 6.3V. This corresponds with the setting up cards, and enables one to adjust the RED Pot, with confidence, spot on. Tested this at 5.30PM when our rural mains fluctuates a fair bit as ovens etc. are switched on for dinnertime. Was dead easy to adjust the fine mains with confidence.
Talking of V2/V3's Heater supply - Why does Secondary 6 winding have a centre tap connected to the cathodes(T/F connections 5C,6C(C/T),7C) ? Also, by the way, I've at last worked out all the secondary connections. So will post a diagram soon.

Regards, David

[frankmcvey]

Hi, David,

For lacing cord, you might try Roy at Electrojumble (about halfway down this page). I quite agree with you - nothing else looks just right to my eyes!

The mains monitor is another worthwhile mod as well. I wasn't too sure about your thoughts on using an edge-meter, but this is unobtrusive and in just the right place. I think I may well blag your idea. Any particular reason (other than safety!) for using the 6.3V winding to monitor the mains, rather than monitor it directly?

I don't really know about the CT on secondary 6. Perhaps it's to tie the heater voltage to the cathode voltage, so that there's less PD between the heater and cathode, putting less stress on the H-K insulation. Or might it be a hum-cancelling arrangement?

Anyway, back to our problem.

More food for thought. I found a Card 63 Revision A in another set of cards, so I fired it up with the same NOS EF37A.

I'm now seeing -1.947 VDC on the grid, slightly lower than you, but same ballpark now. The spot, as you'd expect, is much lower than with the old Card 63, but it's still about 3mm into the green. A second NOS EF37A is 6mm into the green and a old EF37A which had previously read as amber is now showing in the red.

I'll start doing a few more measurements of the voltages around the machine using Card 63 Rev A.

Meanwhile, now that we've established that you are indeed getting the correct grid voltage applied, it looks as if the resistor bank connected to Row I on the gate, is OK, and those gate contacts ore OK as well. So it might be worthwhile checking the anode load for the EF37A. You should be able to do that with power off, Card 63 Rev A in the slot, Test Selector to position 1 and the gate switch closed and card hole G10 blanked off by a piece of tape.

The contacts made wrt the anode load selection are G1, G2, H3 and H4, respectively selecting R69(8.2K), R70(4.3K) R73 (560R) and R74(300R) in parallel, making a total anode load resistance of 182R. So if you do a resistance check between moving gate contact E10 and any moving gate contact in the H row between H 1 and 9 you should see 182R.

Using your EF37A Ia/Vg curve from your DC tester, -2V on the grid should give us about 4.2mA Ia.

Working this through, the voltage which should appear across the anode load should be IR, = 4.2 mA x 182.7 = 0.767V, which according to PJL's voltages for R41 should put you almost smack on the dotted line...

Cheers,

Frank