[jea48]

I just re-tubed my PSE preamp, and it, well, sounds tubby, rolled-off, tubey.  Not a lot better than before I replaced its 15 year old tubes.

It's been a few years since I listened to it critically with good tubes, but I don't remember it sounding like this. 

I've gotten used to a very nice solid state pre, and then when that died, listened straight from the DAC to the amps, so maybe this is what it always sounded like, but it doesn't seem right and I'm not that crazy about it.

So my question is--do 6922s (these are 7DJ8s) need time to warm up, burn in, or otherwise sound right?

My experience with NOS tubes such as early to middle 1960s 6922/7308, Amperex PQ white label, Siemens all grey plate CCa/7308, and Telefunken 6922/7308 tubes is they need about 50 to 75 hours of burn in time to achieve their full sonic potential.
As always YMMV. My preamp is a Sonic Frontiers Line One.

As for new production tubes such as EH 6922, that I use in my Audio Research VT50 amp, burn in time is around 25 to 30 plus hours. Jmho a 6dj8 would never last as long, hours of life, as the 6922 tube in the ARC VT50 amp.

//

I read though the entire thread and I noticed no one mentioned the AC power line voltage feeding the preamp. IF, the manufacture's recommended AC line voltage for the preamp is, say, 115Vac and the users house mains voltage is 123Vac what would the preamp's power transformer heater secondary winding voltage measure?

If the DC power supply is not well regulated what would the DC plate voltage measure on the 6dj8/7dj8 tube? How hard did the designer of the preamp drive the recommended 6922 tube?

//

Does anyone have transconductance curves for the 6922/6DJ8 vs. the 7DJ8 as a function of heater voltage?  It seems with that information it would be pretty easy to figure out what I'm giving up in performance by running the heater voltage below its designed level.

I ran a test using a used EH 6922 tube. For the test a Hickok 6000a tube tester was used.
First I checked the tube for its' Mutual Conductance at the 6.3V Filament setting for the tube.

Test results,
section 1) 5300 MICROMHOS  (where 5500 is as new)
section 2) 5200     "

Next I set the Filament voltage selector switch to 5V
Test results,
Section 1) 4100 MICROMHOS
Section 2) 4500      "

As the tube aged, hours of use, I would image the spread would be greater.

For what it's worth there is a tube life test that can be preformed on a used tube where the filament voltage is decreased in the same manner as I did in my test above. Only difference is the tube bias dial setting is higher for the Tube Good / Bad test.
Jim

 

[Steve]

Ahhh.
The data sheet say LESS then 5% !  So it is better not to use 6v but a little higher.
But,  to be honest, if you use 6V the tube would not die in a very short time i guess.

There is no problem supplying 6.0 volts on a 6.3 volt filament in virtually all cases. However, there are extenuating
circumstances involving heavy plate current and increased frequencies into the high MHz, and depending upon the
tube, in which transit time becomes a problem.
 
1) Consider that the plate current draw means electrons leaving the cathode, which cools the cathode.
(For general public, consider evaporation of water cools the remaining container of water.) The lower
the plate current, the less the cathode is cooled by electrons leaving. I run my tubes in the 11A less
than typical plate current.

Secondly, only have two inputs from customers; they ran the 11A 24/7 for 3 1/2 years before replacing
the tubes, translating to approximately 28,000 hours of tube life.

Cheers
Steve

[Steve]

I read though the entire thread and I noticed no one mentioned the AC power line voltage feeding the preamp. IF, the manufacture's recommended AC line voltage for the preamp is, say, 115Vac and the users house mains voltage is 123Vac what would the preamp's power transformer heater secondary winding voltage measure?

Good points Jim. I would venture a guess that one of two reasons apply.

1) The filament voltages are regulated. I do in my preamps, but amplifiers, expecially the output tube, this is a consideration that needs to be addressed.

2) Some, many, do not understand the problem.

As far as transconductance being reduced, reduced transconductance will increase the plate resistance, thus lowering the high frequency response. Whether the reduction in hfr will affect one's perception will depend upon several factors.

Cheers
Steve

[jea48]

My experience with NOS tubes such as early to middle 1960s 6922/7308, Amperex PQ white label, Siemens all grey plate CCa/7308, and Telefunken 6922/7308 tubes is they need about 50 to 75 hours of burn in time to achieve their full sonic potential.
As always YMMV. My preamp is a Sonic Frontiers Line One.

As for new production tubes such as EH 6922, that I use in my Audio Research VT50 amp, burn in time is around 25 to 30 plus hours. Jmho a 6dj8 would never last as long, hours of life, as the 6922 tube in the ARC VT50 amp.

//

I read though the entire thread and I noticed no one mentioned the AC power line voltage feeding the preamp. IF, the manufacture's recommended AC line voltage for the preamp is, say, 115Vac and the users house mains voltage is 123Vac what would the preamp's power transformer heater secondary winding voltage measure?

If the DC power supply is not well regulated what would the DC plate voltage measure on the 6dj8/7dj8 tube? How hard did the designer of the preamp drive the recommended 6922 tube?

//

I ran a test using a used EH 6922 tube. For the test a Hickok 6000a tube tester was used.
First I checked the tube for its' Mutual Conductance at the 6.3V Filament setting for the tube.

Test results,
section 1) 5300 MICROMHOS  (where 5500 is as new)
section 2) 5200     "

Next I set the Filament voltage selector switch to 5V
Test results,
Section 1) 4100 MICROMHOS
Section 2) 4500      "

As the tube aged, hours of use, I would image the spread would be greater.

For what it's worth there is a tube life test that can be preformed on a used tube where the filament voltage is decreased in the same manner as I did in my test above. Only difference is the tube bias dial setting is higher for the Tube Good / Bad test.
Jim

EDIT:

 Correction:

For what it's worth there is a tube life test that can be preformed on a used tube where the filament voltage is decreased in the same manner as I did in my test above. Only difference is the tube bias dial setting is higher for the Tube Good / Bad test.

Only difference is the tube bias dial setting is higher for the Tube Good / Bad test.
[/I]

That should read,
Only difference is the tube SHUNT dial setting is higher for the Tube Good / Bad test.

Sorry about that.

[E55l2]

What do you (or those costumers) mean with tube life?  To know the tube life you have to measure the tube ( S, Ri, mu etcetera)

There is no problem supplying 6.0 volts on a 6.3 volt filament in virtually all cases. However, there are extenuating
circumstances involving heavy plate current and increased frequencies into the high MHz, and depending upon the
tube, in which transit time becomes a problem.
 
1) Consider that the plate current draw means electrons leaving the cathode, which cools the cathode.
(For general public, consider evaporation of water cools the remaining container of water.) The lower
the plate current, the less the cathode is cooled by electrons leaving. I run my tubes in the 11A less
than typical plate current.

Secondly, only have two inputs from customers; they ran the 11A 24/7 for 3 1/2 years before replacing
the tubes, translating to approximately 28,000 hours of tube life.

Cheers
Steve

[Steve]

What do you (or those costumers) mean with tube life?  To know the tube life you have to measure the tube ( S, Ri, mu etcetera)

Simply perception, which in the end is all that counts. Other customers probably do not power the 11A 24/7, so it is impossible to determine
the hours. 

I should probably power my personal 11A on 24/7, so I can measure the parameters for more
data. I will mark/favorite this string and reply again in the coming years with my results.    This string will be long gone.

From the RCA Radiotron Designers Handbook, 26 engineers, 1960:

Under normal conditions a valve should be operated with its filament or heater at the recommended voltage; in the case of an oxide coated valve it is possible to have fluctuations of the order of 10% up or down without seriously affecting the life or characteristics of the valve........

If the filament or cathode is operated continuously with a higher voltage than that recommended, some of the coating material is evaporated and permanently lost, thus reducing the life of the valve. Moreover, some of this vapour tends to deposit on the grid and give rise to what is known as grid emission when the grid itself emits electrons and draws current commonly known as negative grid current.......

If the filament or heater is operated for long periods at reduced voltages, the effect is a reduction in emission, but no damage is generally done to the valve unless the cathode currents are sufficient to exhaust the "space charge." Low cathode temperature is, therefore, permissible provided that the anode current is reduced in the proper proportion.

As noted above, under normal conditions, +/- 10% will not deplete the space charge (typical plate current is generally listed in the specifications sheet), and with less than typical plate current, 5% is just fine. 17% or more is another matter that must be weighed carefully. At less filament voltage, the carbonates evaporate less, so grid contamination is less over time, a plus.

Directly heated tubes with common cathode/filaments are a different story. Thoriated tungsten filaments with plus 3% increase voltage will increase emission by some 20%, while reducing cathode life by some 50%. Conversely, lowering the filament voltage by 3% will decrease emission by some 20%, but extend cathode life by some 50%. (Eimac Care and Feeding of Power Grid Tubes)

Cheers
Steve

[ACHiPo]

My experience with NOS tubes such as early to middle 1960s 6922/7308, Amperex PQ white label, Siemens all grey plate CCa/7308, and Telefunken 6922/7308 tubes is they need about 50 to 75 hours of burn in time to achieve their full sonic potential.
As always YMMV. My preamp is a Sonic Frontiers Line One.

As for new production tubes such as EH 6922, that I use in my Audio Research VT50 amp, burn in time is around 25 to 30 plus hours. Jmho a 6dj8 would never last as long, hours of life, as the 6922 tube in the ARC VT50 amp.

//

I read though the entire thread and I noticed no one mentioned the AC power line voltage feeding the preamp. IF, the manufacture's recommended AC line voltage for the preamp is, say, 115Vac and the users house mains voltage is 123Vac what would the preamp's power transformer heater secondary winding voltage measure?

If the DC power supply is not well regulated what would the DC plate voltage measure on the 6dj8/7dj8 tube? How hard did the designer of the preamp drive the recommended 6922 tube?

//

I ran a test using a used EH 6922 tube. For the test a Hickok 6000a tube tester was used.
First I checked the tube for its' Mutual Conductance at the 6.3V Filament setting for the tube.

Test results,
section 1) 5300 MICROMHOS  (where 5500 is as new)
section 2) 5200     "

Next I set the Filament voltage selector switch to 5V
Test results,
Section 1) 4100 MICROMHOS
Section 2) 4500      "

As the tube aged, hours of use, I would image the spread would be greater.

For what it's worth there is a tube life test that can be preformed on a used tube where the filament voltage is decreased in the same manner as I did in my test above. Only difference is the tube bias dial setting is higher for the Tube Good / Bad test.
Jim

Jim,
Great info!  Thanks!

AC

[ACHiPo]

Simply perception, which in the end is all that counts. Other customers probably do not power the 11A 24/7, so it is impossible to determine
the hours. 

I should probably power my personal 11A on 24/7, so I can measure the parameters for more
data. I will mark/favorite this string and reply again in the coming years with my results.    This string will be long gone.

From the RCA Radiotron Designers Handbook, 26 engineers, 1960:

As noted above, under normal conditions, +/- 10% will not deplete the space charge (typical plate current is generally listed in the specifications sheet), and with less than typical plate current, 5% is just fine. 17% or more is another matter that must be weighed carefully. At less filament voltage, the carbonates evaporate less, so grid contamination is less over time, a plus.

Directly heated tubes with common cathode/filaments are a different story. Thoriated tungsten filaments with plus 3% increase voltage will increase emission by some 20%, while reducing cathode life by some 50%. Conversely, lowering the filament voltage by 3% will decrease emission by some 20%, but extend cathode life by some 50%. (Eimac Care and Feeding of Power Grid Tubes)

Cheers
Steve

Steve,
Thanks for digging up the reference.  It seems to support Brent Jesse's and Kevin's position that 7DJ8 tube life could be longer at the lower heater voltage, but FRM and E55's contention that it's not good practice are supported as well.  Kinda makes me want to sit around the fire and sing Cumbaya. 



Thanks all!

[E55l2]

In Europe, where i live, the equivalent for an 6922 is the E88CC.

The E88CC is a long life tube, 10000 hours, datasheets (Telefunken, Siemens, Philips, Mullard) say 6,3V +/- 5%.

The Rca 6922 datasheet say 6,3V +/- 10% but not that it is a long life tube, so a normal life of 500-1000hours (after that performance is less good)

Ofcourse it is possible that the tube stays alive 20000 hours and 40000 hours maybe too but not with good performance and that is not what we want.  At least people who care about good performance don't like that.

[FullRangeMan]

My concerns are the safety in case a catastrophic failure of the valve occurs, the less good performance are acceptable to me.

[E55l2]

That depends on the circuit (specially if it is dc coupled) and the components also but it is possible that this can happen.

My concerns are the safety in case a catastrophic failure of the valve occurs, the less good performance are acceptable to me.

[Steve]

In Europe, where i live, the equivalent for an 6922 is the E88CC.

The E88CC is a long life tube, 10000 hours, datasheets (Telefunken, Siemens, Philips, Mullard) say 6,3V +/- 5%.

The Rca 6922 datasheet say 6,3V +/- 10% but not that it is a long life tube, so a normal life of 500-1000hours (after that performance is less good)

Ofcourse it is possible that the tube stays alive 20000 hours and 40000 hours maybe too but not with good performance and that is not what we want.  At least people who care about good performance don't like that.

The only reason I am posting now is to prevent the public from being misled with false information. Please reread my last posts, as I have some 58 years actual experience (45 years since college), including college lab and my own personal lab, and have explained some basic aspects of cathode operating principles.

Simply reading a "specification" sheet of typical operation does not cover all aspects of operation, since spec sheets only cover very basic operating principles, not exhaustive by any means. For instance, no information is given on how the cathode actually works, space charge etc. If I may, I suggest reading various texts such as RCA Radiotron Designers Handbook, Eimac Care and Feeding of Power Grid tubes, older Radio Amateurs Handbook, etc for how a tube actually works.

Next, I am not going take sides, to explore operating a 7.6 volt filament at 6.3 or 6.0 volts, as condition could vary tremendously, such as operating the tube at 2 ma, instead of 15 ma typical operation.

I am not sure where you come up with the 500-1000 hours, except to say that covers power tubes as well, and is under manufacturer's warranty. However, I have seen hundreds, if not thousands of RCA and other brand 6dj8s that were under high stress that measured nice specs after years and thousands of hours of use. Manufacturer's specs will be conservative.

My customer was reviewer Rich Weiner who wrote for Bound for Sound. Of course he is very careful about the sound of his reference system.

I also own my own products and continue to "review"/listening test them. However, over the years I have not keep a time clock on tubes hours.

You can read some reviews of one of my poor designs here.

http://www.sasaudiolabs.com/pream11a.htm

Cheers
Steve

ps. Please pardon me if I have crossed the line of civility.

[jea48]

Good points Jim. I would venture a guess that one of two reasons apply.

1) The filament voltages are regulated. I do in my preamps, but amplifiers, expecially the output tube, this is a consideration that needs to be addressed.

2) Some, many, do not understand the problem.

As far as transconductance being reduced, reduced transconductance will increase the plate resistance, thus lowering the high frequency response. Whether the reduction in hfr will affect one's perception will depend upon several factors.

Cheers
Steve

Steve said:
As far as transconductance being reduced, reduced transconductance will increase the plate resistance, thus lowering the high frequency response. Whether the reduction in hfr will affect one's perception will depend upon several factors.

Hi Steve,

I have an ARC VT50 power amp.

 For their ruggedness I use EH 6922 tubes for the input and driver tubes. One matched pair low noise per channel.
6550 power tubes are SED winged C 6550 power tubes.

The input and driver tubes for each channel have to be biased. If each section of the 6922 tube is not closely balanced/matched it is next to impossible to set the plate bias voltages within ARC specs, especially the driver tube.

Here is the bias adjust procedure for the VT50
http://www.audioresearch.com/ContentsFiles/VT50_BiasAdjust.pdf

Here is the wiring schematic for the VT50
http://www.audioresearch.com/ContentsFiles/VT50_SchemPL.pdf

Example, the ARC VT50 amp. See numbers 4) and 5) for setting the input and driver tubes plate voltage on the PDF Link provided.

It tells you to set the voltage on pins 1 and pin 6, on the input tube to +160Vdc and the driver tube to +170Vdc.( pins 1 & 6 are the plates of the 6922 dual triode tube).You can't set the two plate voltages of each tube individually. RV1 trim pot is for the input tube and RV3 is for the driver tube.

Next read note #7)

And there is the rub. Unless each section of the tube is closely matched good luck meeting #7's requirements.


I am not going to name the Vendor I bought the last closely matched low noise quad of EH 6922 tubes from. I will say he is well known and respected. I have bought from him before.  His close matching of the last set of 6922 tubes I bought from him might work as driver tubes in other audio circuits just fine, but not in an ARC VT50 amp.

Example:
Here are the tube vendor's test readings of some of the his tubes markings on each tube box.

Section 1)14700, section 2)14000

No way this tube could be biased in the VT50 as a driver tube let alone an input tube. And yes I tried.

Another one.
 section 1)14800, section 2)14300
For the driver it still would not meet the ARC specs.

another one.
section 1) 14000, section 2)14600
same problem as above.

another one.
section 1)14200, section 2) 14200.
 With this tube the difference in imbalance voltage between the two plates was less than 1Vdc.

Now I assume the vendor picked tubes with closely balance/matched plate currents and obviously checked for shorts and grid leakage. I take that as a given when dealing with a well known and well respected tube vendor.


Not knowing how it would turn out I tested several other used, that tested like near new, EH6922 tubes in a Hickok 6000a Mutual Conductance tube tester. (Note: These tubes were also bought as closely matched low noise tubes from well respected tube vendors.) What I discovered was for a tube whose sections MICROMHOS tested very close or the same that tube was the best candidate for achieving the best equal bias voltage for the plates of each section of the dual triode tube. This is a must for the ARC VT series tube amps. (Note, not so for the VT100 mrk 3 amp, different circuitry.)

Could you please explain why? Differences in Plate resistance? Ohm's law?
Again I have to assume the venders checked the tubes originally to make sure the plate current of the two sections of each tube were balance/close to one another.
So if working from that assumption, closely matched mutual conductance / transconductance of each section of the tube made the difference?

Jim

My apologies to the OP for getting off topic. 

 

[ACHiPo]

The only reason I am posting now is to prevent the public from being misled with false information. Please reread my last posts, as I have some 58 years actual experience (45 years since college), including college lab and my own personal lab, and have explained some basic aspects of cathode operating principles.

Cheers
Steve

ps. Please pardon me if I have crossed the line of civility.

Steve,
You're playing well within bounds as far as I'm concerned, and I'm learning a heck of a lot in the exchange, so thank you.

AC

[ACHiPo]

Jim

My apologies to the OP for getting off topic.

Jim,
I don't have any issue with you tagging onto this thread, but you might want to start another thread just so folks with a similar question in the future might be able to more easily find the information.

AC

[JakeJ]

Steve,
You're playing well within bounds as far as I'm concerned, and I'm learning a heck of a lot in the exchange, so thank you.

AC

Yup, no worries here.

Jim,
I don't have any issue with you tagging onto this thread, but you might want to start another thread just so folks with a similar question in the future might be able to more easily find the information.

AC

Yes but it is well within the context of the thread.  However, it is also an excellent topic for a stand alone thread.

[Steve]

Could you please explain why? Differences in Plate resistance? Ohm's law?

So if working from that assumption, closely matched mutual conductance / transconductance of each section of the tube made the difference?

If I understand your questions, why are not dual triodes matched more closely and what is ohms law.

I think this equation will help. First, some definitions.

Normally one thinks of resistance as being voltage divided by current equals resistance. This works for both AC (RMS) and DC voltages and resistors. The equations are:

E is voltage
I is current
R is resistance

E = I X R.    X = times
I = E / R      / = divided by
R = E / I

With tubes things are a little different. Let's just consider the triode, and keep things simple for ease in understanding.

Gm is the change in plate current with a change in grid to cathode voltage,,,, with a constant plate to cathode voltage. If we keep the plate to cathode voltage constant, and alter the grid to cathode voltage, the plate current will change.

Mu (amplification factor) is the change in plate to cathode voltage with a change in grid to cathode voltage,,,, with constant plate current. (Let's keep it simple.) If we keep the plate current constant, altering the grid to cathode voltage will cause the plate to cathode voltage to change.

Rp is the change in plate to cathode voltage with a change in plate current,,,, with a constant grid to cathode voltage. With a constant grid to cathode voltage, if we change the plate to cathode voltage, the plate current will change.

(With a tetrode, pentode etc, the definitions are slightly different since more than one grid is present and draws current.)

Mu (amplification factor) is theoretical gain. Gm is transconductance. Rp is plate resistance of a tube section, triode, tetrode, pentode etc.

Mu = Gm X Rp.    X = times
 
Gm = Mu/Rp.   / = divide by

Rp = Mu/Rp.

As one parameter changes, the other two as well, although it is possible for two parameters to change and the third to remain constant.

As far as sections not matching, let's check out the variables. Again, just for a triode.

1. The plate structure between sections may be sheped slightly different.
2. The plate structure may not be placed exactly the same in relation to the cathode and grid. This variance may be both horizontal and verticle.

3. The machine winding the grid wire for each section has tolerances, thus the grid wire for each triode section can and usually is a little different.
Slightly different spacing between each turn of wire.
4. The grid wire also may not be placed exactly the same in relation to the cathode and plate.

5. The cathode material may not be homogeneous across the nickel support.
6. The cathode material may be slightly warped, thus the relationship between each cathode to grid and plate may be slightly different.
7. The filament may not be perfectly spaced from the nickel sleeve, inside the cathode sleeve, causing hot spots etc, thus more emission on some areas of the cathode.
8. The filament itself may have hot spots, thus heating certain portions of the cathode more than other portions.

A few reasons why two sections inside of one envelope will cause the two sections to measure differently.

I hope I understood and properly answered your question(s) Jim.

Cheers
Steve

[E55l2]

RCA have a special range of tubes, the "red base". They are the special long life (10000h) versions of the normal range they make.

Also RCA uses +/- 5% for Vf and allow 10% IF less then 2% of it's operation time.

All the old normal tubes from my Tektronix or Hp equipment i tested have not so good specifications anymore (<<60%).
So i won't count on a very long life for a normal tube. Maybe a bit better then the 500-1000 i say before but after 3000h it is really a lot less then it was new.

[Steve]

Also RCA uses +/- 5% for Vf and allow 10% IF less then 2% of it's operation time.

All the old normal tubes from my Tektronix or Hp equipment i tested have not so good specifications anymore (<<60%).
So i won't count on a very long life for a normal tube. Maybe a bit better then the 500-1000 i say before but after 3000h it is really a lot less then it was new.

Your first sentence above is absolutely ridiculous except +/- 3% for Thoriated Tunsten filament types, and for pulse types that need proper filament voltage to prevent space charge depletion, and possibly horizontal power output types in TVs. See my previous posts.

Please specify, post links, where you actually get your bits of information as none of my RCA manuals, whether receiving manual or transmitting manual, RCA Radiotron Designers Handbook (26 engineers, 1400 pages on tubes, 1960), Eimac, Radio Amateur Handbook, Radio Handbook (750 pages) etc state such.
By the way, Telefunken was known for getting 100,000 hours on their 12ax7 "family" under typical conditions. Now if someone/company is running the tubes at maximum ratings (some do), life will be shortened.

I also had multiple Tektronix 535A scopes and tubes (50 years old, still have one scope left) that check/measure within normal range, using actual circuitry/typical manual conditions to measure specs, not a "tube tester". So I have trouble believing your information.

Cheers
Steve

[jea48]

If I understand your questions, why are not dual triodes matched more closely and what is ohms law.

I think this equation will help. First, some definitions.

Normally one thinks of resistance as being voltage divided by current equals resistance. This works for both AC (RMS) and DC voltages and resistors. The equations are:

E is voltage
I is current
R is resistance

E = I X R.    X = times
I = E / R      / = divided by
R = E / I

With tubes things are a little different. Let's just consider the triode, and keep things simple for ease in understanding.

Gm is the change in plate current with a change in grid to cathode voltage,,,, with a constant plate to cathode voltage. If we keep the plate to cathode voltage constant, and alter the grid to cathode voltage, the plate current will change.

Mu (amplification factor) is the change in plate to cathode voltage with a change in grid to cathode voltage,,,, with constant plate current. (Let's keep it simple.) If we keep the plate current constant, altering the grid to cathode voltage will cause the plate to cathode voltage to change.

Rp is the change in plate to cathode voltage with a change in plate current,,,, with a constant grid to cathode voltage. With a constant grid to cathode voltage, if we change the plate to cathode voltage, the plate current will change.

(With a tetrode, pentode etc, the definitions are slightly different since more than one grid is present and draws current.)

Mu (amplification factor) is theoretical gain. Gm is transconductance. Rp is plate resistance of a tube section, triode, tetrode, pentode etc.

Mu = Gm X Rp.    X = times
 
Gm = Mu/Rp.   / = divide by

Rp = Mu/Rp.

As one parameter changes, the other two as well, although it is possible for two parameters to change and the third to remain constant.

As far as sections not matching, let's check out the variables. Again, just for a triode.

1. The plate structure between sections may be sheped slightly different.
2. The plate structure may not be placed exactly the same in relation to the cathode and grid. This variance may be both horizontal and verticle.

3. The machine winding the grid wire for each section has tolerances, thus the grid wire for each triode section can and usually is a little different.
Slightly different spacing between each turn of wire.
4. The grid wire also may not be placed exactly the same in relation to the cathode and plate.

5. The cathode material may not be homogeneous across the nickel support.
6. The cathode material may be slightly warped, thus the relationship between each cathode to grid and plate may be slightly different.
7. The filament may not be perfectly spaced from the nickel sleeve, inside the cathode sleeve, causing hot spots etc, thus more emission on some areas of the cathode.
8. The filament itself may have hot spots, thus heating certain portions of the cathode more than other portions.

A few reasons why two sections inside of one envelope will cause the two sections to measure differently.

I hope I understood and properly answered your question(s) Jim.

Cheers
Steve

Steve said:
If I understand your questions, why are not dual triodes matched more closely and what is ohms law.

Nope that's not what I was attempting to ask. I guess I just did a poor job of asking what I was trying to find out.

Let me try again.
I know if the idle plate/cathode current of each section of a dual triode 6922 tube is not the same or close to the same it is not possible to bias the driver tube in the ARC Vt50 amp per ARC specs.

I think I understand the reason why when looking at the VT50 circuit design of the driver tube. That's why I asked is Ohms Law was at play. E=IxR

VT50 schematic wiring diagram.
http://www.audioresearch.com/ContentsFiles/VT50_SchemPL.pdf

If you look at the VT50 schematic wiring diagram for the driver tube you will see +330Vdc feeds each plate voltage drop 37.5K resistor of the dual triode tube.

Each plate set point plate idle voltage is +170Vdc ideally, best world.

A trim pot, RV3, circuit is used to vary (raise or lower) the negative cathode bias voltage of each section of the triode tube through a 681 ohm cathode resistor connected to each cathode of the tube which raises or lowers the +DC plate voltage measured at each plate. As you can see from the design of the circuit individual control of the +DC plate voltage is not possible. RV3 raises them both up or both down.

And there is the rub. Unless each section of the tube's plate/cathode current, load, is balanced, (closely matched), you can adjust RV3 that will set the plate voltage of one section of the tube  at +170Vdc but the other plate of the other section of the tube could measure, +190Vdc or +150Vdc, for an extreme example. The closer the match, balance of the plate/cathode current the smaller the spread of the two plate voltages will be. (Note: even in the extreme case above if the section plate voltage was adjusted by RV3 to +170Vdc and then the other ended up +150Vdc it could a problem and take out a 6550 tube or blow a screen resistor because of the design of the VT50. Not to mention the performance of the channel and sound of the channel.)

So, where does close matching of GM come into play? If each section of the tube GM are several hundred MICROMHOS apart will this cause/affect the plate bias voltage adjustment of the driver tube in the VT50 as described above?

Or was it the used tubes I tested that that were closely matched just better selected for balanced/matched for plate/cathode current for each section and was not the close MICROMHOS readings I tested on the Hickok 6000A tube tester at all that yielded the close differences of the plate voltages of each section of the driver tube? If that is the case then the vendor that sold me the last quad of closely matched low noise tubes did not select/find a closely balanced/matched tube/s with tube sections of each tube with equal plate/cathode current measurements.
Your thoughts?

Jim