Tube Matching

[Rob11966]

Hi Folks,

I need some advice re: tube matching. I have just finished an amp based on the Dynaco 70. So, 2 x EL34 output tubes for each channel. I am currently using JJ tubes but I was thinking of upgrading to some NOS Mullards.

This is a silly and basic question I guess but what happens if the tubes (I am just talking about one channel now) are not matched? Is it a major problem or just a theoretical issue? I am totally ignorant concerning tube matching. The Dynaco circuit has bias pots - does this make a difference?

Thanks in advance,

Rob

[SY]

It depends on how badly they're mismatched. If it's severe, when you set the idle current via the bias control, one tube will have much more current going through it than the other. This means that one tube will be dissipating more than its share of power and could fail more quickly. That's why some amps have separate bias controls for each output tube (not just one for each channel).

In any event, mismatching will also increase distortion. If I had a choice between matched new production and unmatched NOS Mullard (or whatever), I'd use the former every time. Most of the NOS stuff is wildly overhyped and overpriced- the only NOS I use is stuff I bought new years ago or tubes no longer being made.

[Cycline3]

It is my humble opinion that you should NOT spend your money on the NOS mullards. The simple fact is - there are some great tubes in current production - and they keep getting better. Get a matched set of NEW mullards (tubedepot.com has them as well as others im sure) if you want the name. When you buy NOS - it's too pricey for uncertain results.

[richwalters]

Originally posted by Cycline3
It is my humble opinion that you should NOT spend your money on the NOS mullards. The simple fact is - there are some great tubes in current production - and they keep getting better.

Copy that.

New issue Tung-sols 6550 are way more robust than NOS KT88 and checking out a recen delivery showed lower thd than KT88's. The tung-sol glass alone can take far more punishment around the bench and in comparison, the KT88's are as thin as egg-shells.

My experience with Svetlana KT88's is don't punish them. An excessive power strain can change the internals. I have my doubts that the claimed performance isn't really worth the credit and sales hype. After equivalent 2 years operation at 500V (autobias) two Svet KT88's failed due to screen grid problems.

If the new issue tung-sols are so good then the old method of selecting 1 out of 3 for matching vanishes- But I would still keep my bias presets.


richj

[Geek]

Hi,

Independent bias controls - problem solved

[kevinkr]

The general consensus around here in ST-70 and my modified version of same is that the current EL34EH (New Sensor/Electroharmonix) is both better sounding and more consistent than the current crop of JJ EL34. YMMV

Some current JJ production has had quality control issues. (See posts by Jim McShane, myself and others on this subject.)

I recommend all power tubes be burned in and matched by the vendor if possible prior to purchase. This weeds out the ones prone to very early failure.

[Rob11966]

Hi Kevin,

Thanks for the info - when I searched the forum, it seemed to me that the JJ's got the best feedback (no pun intended) and that's why I went for them but I cannot really comment myself as they are the only tubes that I have ever used! Actually, I realised that my choice of tube would always be a matter for debate and opinion - I was more concerened with tube matching (whatever the brand) to determine if I could just by individual good quality NOS and use them without worrying too much about matching. It seems from what I have read so far that i may be able to do this but I risk shortening the life of one of the tubes and perhaps they may not sound quite as good. If the tube life is down by 10% i can live with that but if it is 50% less, then I guess that is a problem.

I am yet to get a good feeling as to how much of a practical, real life problem mismatched tubes are. Having said that, I appreciate what people have said about the NOS tubes. From my point of view, i have spent a lot of time trying to replicate the original construction (cloth wire etc) and I have this feeling that the amp 'deserves' some original tubes - this may just be romantic rubish though.

Cheers,

Rob

[Cycline3]

Originally posted by richwalters
New issue Tung-sols 6550 are way more robust than NOS KT88 and checking out a recen delivery showed lower thd than KT88's. The tung-sol glass alone can take far more punishment around the bench and in comparison, the KT88's are as thin as egg-shells.

The pic says it all!

[kevinkr]

The reproduction Tungsol 6550 is a very nice tube, and probably one of the best currently available 6550 - given what it costs and current production economies I would be surprised if it was nearly as long lived as the originals in service. I've had both new and heavily used genuine Tungsols and was surprised at just how rugged they were.

I haven't really found 6550 to sound better in an ST-70 than a good EL34..

Rob you must have pretty deep pockets to contemplate NOS Mullard EL34 - I certainly wouldn't spend $500 or more for a quartet of these for a garden variety ST-70.. Vintage Mullards are excellent performers, but IMO total overkill here.

The important thing is to get well matched pairs to minimize distortion in the output stage. It is impossible to know the impact on tube life of a mismatch unless it is severe enough that tube life is compromised. The mismatch will result in measurably higher distortion.

JJ are fine, just an observation from some old clients of mine regarding the EL34EH..

[Rob11966]

Hi Kevin,

I know the Dynaco-70 is a very common amp. My one was built for the ground up by me, so from my point of view it is quite unique and represents about 5 months of planning and work. That makes it a little easier for me to justify spending money on tubes. Having said that, I don't like the look of the price of them either and they don't seem to be getting too much great press in this thread.

Cheers,

Rob