"Matching" of twin triodes.

[joebog1]

I have been progressing with my Goldenears build, and have come across a question, which is probably already answered by my question. Twin triodes, of which there are dozens, HOW well are they matched "in the bottle" ?.
So I am doing final wiring, and have built into the chassis, all the test points needed to setup this amplifier.
I don't know how many people are actually reading my posts, but I have this question regarding the cathode follower driver stages of the output. It is in regard to the 12BH7
( I have no idea what secret code the English use for this bottle ), but here is my question.
Setting up this amp requires plugging in ONLY the 12BH7, and setting the negative bias for the output stage to -38 to -40 volts on the cathode.
When I laid out the chassis I drilled a hole to add a 2mm test probe socket so I could read, without flipping the amp upside down, the voltage at the said cathode. At the time of drilling holes, I didn't have any thoughts on actually taking into account, the difference between each half of this valve, and chose to only add one socket for a test lead. In the readers experience,( you ) how closely matched are these twin triodes? .
I am thinking of wiring the right channel to pin 3, and the left channel to pin 8. I can set the bias for each channel and then swap the 12BH7's over right to left ( and vice versa ) and check the voltage again. That will show me any major difference between the
( two ) valves. Now, am I being a bigger audiophool than I already am ?, or will that benefit my aim to truly balance both halves of the 12BH7's, by taking into account the difference, and "tuning" midway so I have the same voltage each side ?.
( one valve is left channel and one right )
Yes I should have added two probe sockets ( cost about 8 cents) but I am loath to drill another hole with half the components and much of the wiring completed.

Just questioning my logic, probably more than anything.

ALL comments will be read and thought on.

Thanks in advance lads and lasses,

Joe

[Craig Sawyers]

I'm currently renovating some of my Racal radio gear, which is full of vaves - many of which are twin triodes of one flavour or another.

Ones which tested good on my AVO III were matched to maybe 10% on anode current and gm. New ones were very close indeed. I don't have the numbers handy, but they were a few percent different.

So are twin triodes matched? Well on the basis of measuring 6 in the last week - yes.

Good luck!

Craig

[joebog1]

Thanks Craig

Joe

[GrimJosef]

I don't think there's a direct equivalent of the 12BH7 in the ECC8x series. So the secret code the English use for the 12BH7 actually isn't a secret. It's '12BH7'. I think the Welsh, Scots and Irish use that code too .

The fundamental electronic properties of the valve (e.g., for a triode, the family of Ia vs Vak curves for a range of different Vgks) are determined by the electrode geometry and the emission characteristics of the cathode, aren't they ? Assuming the valve is built by fitting various sub-assemblies into the micas, as shown from about 3 minutes in here https://www.youtube.com/watch?v=dErsA6wFmlM, the geometry of the two entirely separate triodes inside a 12BH7 will depend on the exact dimensions of the individual sub-assemblies and on how accurately the mica holes were drilled.

I imagine that whoever was putting the valve together simply picked, say, control grid sub-assemblies from a rack of them. I'd be surprised if anyone had measured them all in advance and matched them into pairs with very close measurements. If they hadn't, and the sub-assemblies were just picked at random, then that aspect of the geometry wouldn't be any better matched within a particular valve than it would be between that valve and the next one off the production line. Perhaps valves made very close together in time would be better matched to one another than ones made months or years apart, during which time the grid-winding machine might have been re-set or even replaced, though.

Cathode emission couldn't have been matched precisely in advance as the final cathode chemistry is only established once the valve is assembled and pumped out. But again the two cathode sub-assemblies could well have been picked from the same rack of them, and those might have been dipped in the same bath of precursor chemicals. They would go on to experience the same vacuum and final chemical processing. So even if no deliberate steps were taken to match the cathodes I wouldn't be surprised to find that their properties ended up being close.

Cheers,

GJ

[Bazz4CQJ]

I've used ECC81's as standard valves for my Avo VCM 163. I cannot say that I've look at this in any great detail, but my impression is that with new valves, Gm is usually within 5%. That said, I've also seen new valves "burn in" quite a lot over the first 24 hours of use. With used valves, I've seen quite large variations between the two sides.

B

[turretslug]

I can see that new twin-triodes ought to show as pretty closely matched, but I wonder how closely this would hold true for well-used valves where the two sections had had somewhat different demands made on them over their life, say where one had been an input amp running at very low anode current and the other a cathode follower or driver stage running at substantially higher cathode current? Some applications might have had one triode pretty much cut-off or unused with cathode interface developing over time on that section. A bit academic really, but might rear its head when using "pulls" or uncertain history valves.

[joebog1]

Thanks all,
No pulls here. I have all brand new boxed valves for my amp, EXCEPT 12BH7's.
In Aus they were used for vertical oscillator/vertical_scanner for B/W TV's. Mostly 17" models. After that the TV,s had specific twin triodes known here as " twin triodes with dissimilar sections". Without being cheeky here, WHY are they called 12BH7 in England too ??

I try help others on the forum with requests for valves, but the gobbledygook the brits used for valve numbers, is over my head. I have some thousands of valves that were for TV service, that I will never use. But apart from postage, most NIB, free to whoever wants em. NO WARRANTY !!! offered as is.

I did some more work today, almost finished one channel input/phase_splitter_ driver stage on one channel. LOTS of connections with 18 valves!!. EEKK!! EEEKKK!!

Thanks to all that answered.

Joe

[Diabolical Artificer]

It depends what you mean by matched; to within 1%, 5%, 10% emission and both have similar gm figures as each other? Most new 12BH7A's I have tested are pretty damn close emission wise but there are outliars and to get a a match gm wise as well as I is trickier.

Andy.

[jjl]

Originally Posted by joebog1

Without being cheeky here, WHY are they called 12BH7 in England too ??

They're called 12BH7 because they were designed by Brimar who, at the time, were a division of Standard Telephones and Cables (STC) whose parent was the American ITT corporation.
Brimar's name is an acronym that comes from BRItish Manufacture American Range - they mostly concentrated on making copies of US developed valves with US numbering. Brimar did also create valves of their own design like the 12BH7 with US style numbering, many of Brimar's designs were not actually made by any US manufacturer.
Some info on the 12BH7 is here.

John

[GrimJosef]

Originally Posted by joebog1

... Without being cheeky here, WHY are they called 12BH7 in England too ? ...

I think simply because they weren't developed here in an independent way.

... but the gobbledygook the brits used for valve numbers, is over my head ...

If you mean designations like the ECC83 one then this surprises me. The first letter tells you about the heater (e.g. E for 6.3V, P for 300mA, U for 100mA etc), the next letter(s) tell you about what the valve type is (e.g. C for triode, F for small-signal pentode, Z for full-wave rectifier etc), the first number tells you about the base (4 for B8A, 8 for B9A etc) and it's only the next figure which is almost information-free - it just identifies the particular 6.3V B9A double triode you're holding. It's a system. It's informative. What's not to like ?

The 12BH7 numbering 'system' tells me almost nothing. As an example here is a sequence of valves with very close designations

6AN4 - small-signal triode, 7-pin base
6AN5 - beam power amplifier, 7-pin base
6AN6 - quadruple diode, 7-pin base
6AN7 - triode-hexode, 9-pin base
6AN8 - triode-pentode, 9-pin base

Their numbers are almost identical but in reality they're as different as chalk and cheese and putty and puffball and pimple cream ! Whatever it is, it ain't a system.

Cheers,

GJ

[David Simpson]

Joe, if you have access to an AVO VCM Maintainance Manual for the likes of the MK3 or MK4 or a CT160, or can download certain pages of several versions which are circulating the internet in pdf format? - - - Go to the Calibration Section & study the info on DC "Standardising" a CV455 or 491(ECC81 or 82) Valve. The end product is an Ia/Vg "Curve" on a piece of graph paper. AVO recommend "strapping" the two internal triodes together, but the proceedure can obviously be done seperately - producing plotted "curves". If both curves are very close over a wide range of working Ia's & Vg's then they are truely "Matched". Obtaining, with a valve tester, close Ia's for a given working Vg is just a decent indication of the possibility of full curve matching.
Gear needed - - A decent stabilised adjustable DC HT PSU, a decent stabilised adjustable DC LV PSU - for Vg's, and a spot-on steady source of 6.3V DC for heaters. Plus a MM or DMM for constantly reading Ia in Ma. A seperate MM or DMM for monitoring DC voltages is recommended if the PSU's voltmeter's arn't spot-on. Always tabulate your results starting with the highest Vg(hence lowest Ia), as those ECC(or Eq.) range of wee D/Triodes are prone to quickly go "Phut" with a flash, if Vg is applied at too lower a voltage.

Regards, David

[G.Castle]

Hello Joe.

My limited experience with dual valves is that they are normally a pretty close match, and living in the same home stay that way quite well too, its always a good idea to trim them for best match IMO, why put up with second rate when you can have first!

As for the type numbering question, there were other manufacturer type numbering systems in Europe too just to add confusion, the Phillips/Mullard system was mostly adopted because it did mean something.

One thing about your stock of valves for TV use you don't say, is that being an Aussie most setmakers over there used isolated power supplies with few exceptions, so I'd assume lots of them have 6.3 volt heaters, and you could be sitting a lot of Audiophool beer money! Worthwhile your investigation, line and frame output valves are getting popular for their "individual sound" for the guitar gurus for starters.

(For individual sound, read nonlinearity distortion).

BTW. Always enjoy reading your exploits, and good luck with your amp.

Greg.

[David Simpson]

Joe, here is a graph of a Telefunken ECC82's two Triodes. All right for simple pre-amp usage, but definitely not "matched".
However, the graph for two E282F Pentodes(see recent thread) reveals decent very close matching.
I've also scanned a page out of my VCM MK3's Manual copy. Such simple rigs can be extended & modified(I.e. make your own DC Valve Tester) to test & standardise, or match, almost any triode, tetrode or pentode.

Regards, David

PS. To all you keen VR/T.Eq./valve testing Cobbers & Sheilas out in Australia - Can't some of you standardise a few valves, if you have all the gear, and supply them to Joe & others ? DS

[David Simpson]

Sorry, forgot to add a picture of a T/F ECC88's decent matched curves. DS

[joebog1]

A new valve tester is next cab off the rank. I have an AVO I think mark I or II. It's got a sticky meter, and I had to repair one of the wirewound pots that had "broken" the join in the resistance wire. It's made up of three or four different gauges that look like they were originally welded together. From memory it was the "Set Zero" pot, which is two pots piggy backed together. I could never make sense of exactly how it worked and I suspect there is something else wrong. In any case I don't use it and have collected all the bits to make a new one. Something similar to the Sussex, but using valve PSU's.

Thanks David for the excellent engineering papers you have provided. From your graphs the valves you tested were "very closely " matched. If I get the same matching with my 12BH7's I will be very happy indeed. When I make up my tester I will be doing exactly what you have demonstrated. It's extremely logical and well thought out. I guess it's the same as my valve manuals, which as an aside have the American system with a small supplement added at the back to convert European to American system. It's very small print and a major PITA to find the equivalent American number, then go back to the manual proper to find the data.

To Greg, yes almost all are 6.3 ( or 12 volt ) heaters, and you are correct re: isolated power supplies.
Australia almost always had big power transformers with decent isolation. I am 68 years old, been in electronics all my life and I have never seen a valve radio or TV that wasn't isolated by a power transformer. Mind you, I only worked in the radio/TV trade a short while before going to mainframe computers so I haven't that much experience. I will do up a list of what I have, but I really don't want cherry pickers just cleaning out what's useful and leaving the dross.

Thanks again to all,

Joe

[Richardgr]

Hi Joe, I built myself a uTracer, which was delivered as a complete kit, excluding case, for €200.

I use that to curve trace tubes. I bought 300 CV4024, a special quality ECC81, and have been tracing them then advertising them with the trace graphs as part of the advert, matching manufacturers and styles of tubes.

(Unfortunately it has only been a labour of love so far).

[David Simpson]

Tak, Richard, for the veckert pictures of your uTester rig. This modern method of curve tracing is obviously the way forward in the 21st century. Compared to my old-fashioned analogue equipment and hand-drawing skills. All you valve testing enthusiastic vannen take note.

Regards, David