Testing and Matching tubes 101 - Tube measurements shell game and rant

[Guest " "]

I had a discussion about tube measurements and was reminded how crazy the tube measurement shell game can get. Take some 6550 tubes which on my tester are expected to score a transconductance value of GM 11000. You see a lot of NOS tubes being sold as NOS if the tubes score within 90% of that number...lets say...10500 is grades a NOS tube or "test like new" tube. And if the seller has 4 of them that scores 90%, you now have a matched quad of NOS tubes and you can buy them for a primo price of 250 bucks a pair for tubes that you can buy new today any day of the week for 60 dollars a matched pair. SOme tube sellers only use an emissions tube tester to match tubes. No GM values, no bias current values, not even a regulated power supply.

I have to disagree with the methods described by many ebay sellers to match tubes. You can not match tubes using a tube tester along " (AVO 160, Hickok 539b & TV DU/7) ". You do need a device to host the tube that's being tested for matching. That device could be an amp or a tube tester. You then need a way of measuring bias current. You can buy a bias probe and meter on ebay, or you can make a bias probe cable and use a multi meter or you can modify your amp to behave like a bias probe by adding resistors which alow the measurements to be made. There are a few devices that can be used to match tubes but the cost $500, $900, and one even costs $3000. A good calibrated Transconductance tester (200 bucks on ebay) and a bias probe (90 bucks on ebay) and or your multi meter and bias cable (30 bucks for the cable) is all that you need.

So lets say you measure your tubes....and you make this simple....lets say you use a tube tester to host the tubes and in adition you use a bias probe. So you run 4 tubes thru the process and you find you have IP (bias current readings from probe) of 24.1, 23.9, 22.9, 23.1 and you find your tube tester says the transconductance values are GM 15000, 14900, 14000, 13900. what does this mean. well it means you have two matched sets. The pair that comprise of IP of 24.1, 23.9 with GM's of 15000, 14900 and the pair that comprise IP 22.9, 23.1 and GM's of 14000, 13900 are the two matched sets that must be used together . In the example I gave, we are testing 6550s' which have an expected transconductance value when new on my tester of 11000 which would make all tubes in this example score 126% or better.

So if your buying tubes that scores a "NOS" "test like new" score of 95%....you are buying a used tube that is scoring significantly below what a new tube will score.

So how useful is this information. Well, for the average Joe with 1 power amp and a small collection of 4 to 8 tubes, not useful at all.

If you have a half dozen amps that use 6l6 tubes and each amp has 8 tubes in it, then it's a worth while exercise to pull all 48 tubes out. Test them for GM and IP, label them, and them sort them by match pairs and quads and put them back into the amps. Often you will need to replace tubes that fail becuase they score too low or replace tubes that score too high and can't be matched with anything.

If you have someone provide you tubes, make sure there are labels on the tubes with their GM and IP values so that you know which tubes are a matched set. If you buy tubes and the labels just say 90% or some abstract number like 45/55 or clusters of letters like AFE, find out what does codes mean.

end of rant

[tube fanatic]

Makes a nice case for single ended, cathode biased amps! Put in the tube, measure the circuit parameters to see if they're within spec, and forget about it for the next 10,000+ hours!!!

[mike stehr]

Makes a nice case for single ended, cathode biased amps! Put in the tube, measure the circuit parameters to see if they're within spec, and forget about it for the next 10,000+ hours!!!

I can see driver tubes lasting 10,000+ hours...but common audio output tubes?

[tube fanatic]

Makes a nice case for single ended, cathode biased amps! Put in the tube, measure the circuit parameters to see if they're within spec, and forget about it for the next 10,000+ hours!!!

I can see driver tubes lasting 10,000+ hours...but common audio output tubes?

It all depends on how the output stage is set up. With carefully controlled filament voltage (i.e. does not exceed its rating under any line voltage condition), selection of its operating point, and proper ventilation, it's easy to achieve (I'm speaking about US made tubes as I have no familiarity with those of current manufacture). I've worked on numerous communications receivers from the 40s and 50s which used a 6AQ5, 6V6, or 6K6 in the audio output stage, which were operated 24/7, and had no failure of those tubes after 1-2 years! Similarly, in the past 15 years I have had no output tube failures in any of the single ended amps that I've designed and built for local audiophiles and family members who average 2 or more hours of use/day. I'm curious as to how modern tubes stack up against that kind of track record. So, if anyone has experience with that, please post!

[Guest " "]

Similarly, in the past 15 years I have had no output tube failures in any of the single ended amps that I've designed and built for local audiophiles and family members who average 2 or more hours of use/day. I'm curious as to how modern tubes stack up against that kind of track record. So, if anyone has experience with that, please post!

I like the SE sound and it is certainly a simpler design...the only gotcha is the power envelop....hard to get high WPC's using a SE approach unless you do what Cary does

on their V12's and parallel's the SE's and drives them with el-84's.

about your track record question....I have in my possession 3 Manley MB-350's which came out of a recording studio. They were in use every business day for the last 19 years. Each amp has 8 KT-90's. 350WPC capable but probally never really driven above 50WPC. Well one of the amps failed after 19 years and 1 power tube flamed out and took out two resistors and two capacitors.

Tube life has many factors....how much you beat them towards their max WPC.....heat....high bias......excessive plate voltage....lack of soft start....etc

[tube fanatic]

Similarly, in the past 15 years I have had no output tube failures in any of the single ended amps that I've designed and built for local audiophiles and family members who average 2 or more hours of use/day. I'm curious as to how modern tubes stack up against that kind of track record. So, if anyone has experience with that, please post!

I like the SE sound and it is certainly a simpler design...the only gotcha is the power envelop....hard to get high WPC's using a SE approach unless you do what Cary does
on their V12's and parallel's the SE's and drives them with el-84's.

about your track record question....I have in my possession 3 Manley MB-350's which came out of a recording studio. They were in use every business day for the last 19 years. Each amp has 8 KT-90's. 350WPC capable but probally never really driven above 50WPC. Well one of the amps failed after 19 years and 1 power tube flamed out and took out two resistors and two capacitors.

Tube life has many factors....how much you beat them towards their max WPC.....heat....high bias......excessive plate voltage....lack of soft start....etc

You are absolutely right about the power limitations of SE designs and the factors which control tube life. I have always preferred that route for up to 10-12 watts out (easily obtainable with 2 tubes in parallel) after which I've had to go to p-p due to lack of availability of affordable output transformers which could handle the current. And then there's the issue of the size and weight of the single ended output transformers too. Some argue that the distortion of SE pentodes is way too high; but, that is easily managed/reduced with some negative feedback at the ouput stage. The resulting smoothness, to my ears, is unbeatable. Regarding the tube life, one of my favorite amps is a triode strapped Philips 6BG6GA which has a 30 watt plate dissipation (basically a 6L6GC which was packaged for horizontal output service in a TV). With 250V on the plate, and idling at only 40ma, this tube doesn't even know that it's doing anything. I wouldn't be surprised to see a life span of 25000 hours or more on this one.

[Guest " "]

Regarding the tube life, one of my favorite amps is a triode strapped
Philips 6BG6GA which has a 30 watt plate dissipation (basically a 6L6GC
which was packaged for horizontal output service in a TV). With 250V on
the plate, and idling at only 40ma, this tube doesn't even know that
it's doing anything. I wouldn't be surprised to see a life span of
25000 hours or more on this one.

yeah and they are dirt cheap too.