Heathkit AV-2

[Tubetec]

Recently pulled a VTVM from the recycle ,


A few safety issues to deal with ,
first a three core mains cable , with a decent ground to chassis .
Then Find something to replace the dirt oxide rectifier and mallory caps dating back to 1953 .
A BNC in place of the banana plugs seems appropriate .

The circuit doesnt look too far off being usable for audio , but most likely all hell will break loose when the feedback through the meter is cut .
I can run it on 110v for now via an autoformer , later if it turns out any good I can make a new PSU in a seperate enclosure .

 

[Tubetec]

Had a nice evening tinkering with the AV-2 ,
I found ,with a bit of adjustment I could fit a single gang metal clad back box in the space around the mains transformer , so providing sheilding to the nearby moving coil element , attenuator switch and coupling cap .
I removed the power switch from the front panel , as I didnt want mains voltages routed through signal areas .
Instead I'll place a toggle switch on the back panel .

I was able to carefully drill out the phenolic bannana plug spacer to accept a BNC socket at the input terminal , leaving the ground binding post /4mm plug as is .

Ive removed the metal oxide rectifier and original HT caps ,
seems worthwhile adding extra decoupling on the HT line between stages seeing as voltage drop wont be an issue .

Its definately going to need an output , I was thinking of switching out the rectifier/meter for a preset resistor to keep gain under control .Fortunately the redundant mains on/off switch is in just the right spot to do the job .

Tomorrow I'll do a cold check on some of the resistances , see how Heathkits specially selected components have held up after 70 years .
Im very glad I saved this humble device from the crusher , who ever put it together close on 70 years ago
made a good job of it .

 

[Tubetec]

I measured the attenuator resistances ,
693k
218k
69.4k
21.86k
6.95k
2.153k
701ohms
220ohms ,
So more or less everything has drifted up a small way ,

Selecting from the preferred values 22(xxx) , 68(xxx) , I doubt Id do much better , maybe if pot luck is on my side I might get a better match on a few values ,

 

[Tubetec]

Forgot to check the HT winding earlier , turns out its O/C ,

Theres an easily available Hammond transformer thats just right for the job ,
110/240 primary , 120V at 45ma and 6.3V 300ma secondaries
Don audio has them on back order for 32 euros .
https://www.don-audio.com/Hammond-Power-Transformer-262B6

 

[Tubetec]

False alarm, the HT winding is good after all .

I'll try half wave rectifier first but I think I'd prefer a bridge configuration .

 

[fallout]

I bought a small collection of VTVM's from a garage sale a few years back. Wound up getting them all up and running and back in service. I kept the Hewlett Packard 410B, it's actually a really nice VTVM and surprisingly very accurate. I sold the Eico's and RCA. I was surprised that people were still interested in these.

 

[Tubetec]

The HP 410b is one of the finest examples ,no doubt ,
Bruel & Kjaer have another tube based AC RMS meter thats got incredible bandwidth .

I got around to fixing most of the problems with the AV-2 ,
and there were many , both anode loads well beyond +20% tollerance
All the paper/foil caps were leaky and both tubes in this circuit rely on grid leak bias , so voltages were all over the place as well as leading to all kinds of unpredictable behaviour of the meter ,
The meter movement itself needed adjustment as it sometimes got hung up and didnt return to zero ,
minor adjustment to the set screws at the center of the movement fixed the issue .

I substituted a .02 coupling cap for the .002 initially , that was enough to excite motorboating in the PSU rail if the meter was triggered by touching the input terminal ,after the correct value was placed in the circuit no further traces of instabillity .

I added 2 extra sections of RC filtering in the HT path ,maintaining a similar voltage drop to the original arrangement but giving the 6AU6 its own decoupling stage .

Its well documented moving coil meters impart distortion on the driving circuit ,
perhaps in a reactive way ,a bit like an output transformer/speaker , it might provide some interesting color with an Instrument source Di'd .

I didnt mess with the attenuator switch and precission resistances even though some are beyond spec ,
I think I can live with the device calibrated for a 0db indication on the required range as needed .
Failing that I might try a recreating it with a modern switch and resistors , but where do you get a 216.2 ohm or a 6838 ohm .

 

[LoudSpeakerPro]

Creating precision resistors is easy ... get a hold of a parallel resistance calculator. A resistor slightly over the value you need, paralleled with one, two or three higher value resistors will get you the value you need, quickly and inexpensively.

 

[Tubetec]

Welcome to the forum LSP,

The resistors required here are actually close to standard value components (within around +/-2%)
So I was thinking of just selecting single resistors from batches of components with 5% tol.

I guess I could just take a mix n match approach and get a bit closer to the ideal values .

A calculator that expresses one number as a % of another is also a useful tool to check resistor tollerances .

 

[Tubetec]

I ended up replacing the single diode for a bridge in the psu .
Output voltage is almost exactly the same as half wave , correctly predicted by PSUDII
The spike at switch on goes around 15% higher , but it all settles back to where it should after a minute .
I increased the number of RC filter sections to make a total of 5 , the 5th just handling the 6AU6 anode and screen supply .

This has drastically decreased the noise of the unit , previously the meter gave an indication of residual noise on at least half the lower voltage ranges , now noise is only barely visible on the highest sensitivity position with the input unterminated .

I'm wondering about the best way to tap an output signal , there is some DC existing after the final coupling cap in the vicinity of the diode bridge/meter due to it being coupled back to 6AU6 cathode , so I cant just hook in there directly ,
Am I best taking the output directly off the 6AT6 anode , with its own coupling cap and bleed resistance ?
Is switching the meter out of circuit as simple as inserting the appropriate feedback resistor in its place ?

Another idea I had was to install a switch in such a way the attenuator is bypassed and instead a variable load resistance is presented to the input as the control is moved .

Seems like it should make a good enough general purpose bench preamp.

 

[Tubetec]

Ive a few interesting observations from fault finding this box of tricks ,
All coupling caps needed to be replaced as any DC leakage was enough to upset the grid bias and cause the meter to jump about .
I ended up swapping out the input coupling cap from a 600vdc metalised foil/paper to a polycarbonate tubular of approximately the same dimensions with a 2kv dc rating .For the other internal coupling caps I went with 200 or 250 v rated parts which were quite a bit smaller that the original parts .

There is a discharge path for the PSU caps although it takes around 10 minutes for voltages to drop away to safe levels after powering down , once the power is cut and the heaters stop conducting ,I noticed the voltages at the anode of the 6AU6 tube rises up over the space of around ten seconds to almost full HT voltage before slowly decaying again .

I added an NTC thermistor with a value of 50kohms @25c into the RC network feeding the 6AU6 screen/anode , not only does it give a gentle ramping up but after switch off it also limits abillity of the earlier caps to replenish the charge .
Im only drawing a very small current so the value of the NTC only changes from around 70kohms at ambient to around 30k while opperating ,
The series resistance of the NTC looses me an extra 20v at the HT supply node of the 6AU6 but the voltage at that point is still within spec for the unit overall . I monitored the anode and screen voltages over time and there right where they should be and show surprisingly good stabillity .

It would seem to me a correctly proportioned NTC device relative to current draw and anode load of the tube involved helps tame any tendency for the voltage to spike when turning on/off , while in use any change in resistance of the NTC due to variations in anode/screen current are so small as to be insignifgant , the anode /screen load resistors are always dropping the lions share of the voltage after the tube has started to conduct .

More investigations needed , not had a whole lot of bench time lately .

reposted schematic and voltage chart ,
Screen grid runs at a higher voltage than the anode on the 6AU6 .

 

[ruffrecords]

You could try adding a 1Meg across the HT supply. This should discharge it in about 20 secs at switch off.

Cheers

Ian

 

[Tubetec]

Well the path via screen grid to cathode to ground is only 1.5 meg or so already , but with a 5 section RC filter with 80uf caps a much longer time constant .

In any case the HT voltages are contained within the chassis its not so much of a worry .

Anyway the important lesson I learned here is large HT caps can recharge caps further down the line after mains power has been disconnected and the heater stops emitting electrons .The ntc thermistor appears to limit the voltage extremes your precious low level preamp tubes are exposed too with more or less no effect on an audio time scale ,
If I inject a snare beat into my meter, thats not going to have any impact on the value of the series thermistor ?

thats my guess anyway .... TBC