crazy bill,
1. EF86 and RC coupling, single ended
A triode strapped EF86 that has a CCS plate load should have a gain of about 40.
(slightly less than 40 . . . often most limited by the output stage's g1 resistor to ground, and generally less limited by the very high impedance of a good CCS plate load; depends on Rg1 and CCS impedance, they are in parallel).
This gives lots of gain (if you need that much gain), good frequency response, low distortion, and a good square wave response.
I have never used an EF86, they seem a little expensive for me, but I might spring the $$ for them sometime.
I often use 12AY7, or sometimes ECC82 for my input/driver tube depending on the gain I need; and I RC couple it to the output tube (6L6, KT66, KT77, or 7591S in Ultra Linear or Triode wired mode).
I am mostly into push pull amplifiers now, but I still do single ended amplifiers.
I am contemplating building a 2 stage balanced amplifier, I might purchase two JJ EF806 for a mono-block; I can easily drive over to Eurotubes.com and pick up a pair (they vigorously re-test them from the manufacturer).
2. 6N1P or 6922 and Interstage transformer
A 6N1P has a plate resistance, rp of about 7700 Ohms.
A 6922 has a plate resistance, rp, of about 2600 Ohms.
You need enough primary inductance in the Interstage transformer, so that 7700 Ohms or 2600 Ohms does not reduce the low frequency bandwidth (low frequency fall off).
Example:
2600 Ohm plate resistance, rp
20.7 Henry has 2600 Ohm inductive reactance, XL
-3 dB at 20Hz
-1dB at 40Hz
Want to be -1dB at 20 Hz, then use 41.4 Henry primary inductance.
With 7700 Ohms plate resistance, then 7700 / 2600 = 2.96, almost 3 times the resistance.
With an rp of 7700, then In round numbers:
20.7 Henry x 2.96 = 61.3 Henry
With 61.3 Henry, and 7700 Ohms rp . . .
-3 dB at 20Hz
-1 dB at 40Hz
It takes a very good Interstage transformer to have good frequency response, and good square wave response (a good one might be expensive; only a few of the Interstage transformers I used had good square wave response).
Many Interstage transformers are very sensitive to magnetic fields, you may have lots of hum, unless you deal with that:
No magnetic steel chassis
Good spacing of the Interstage to the power transformer, and to the filter choke
Proper angular Orientation of the Interstage, versus the power transformer, and the filter choke
Happy designing, happy building, and happy listening!
1. EF86 and RC coupling, single ended
A triode strapped EF86 that has a CCS plate load should have a gain of about 40.
(slightly less than 40 . . . often most limited by the output stage's g1 resistor to ground, and generally less limited by the very high impedance of a good CCS plate load; depends on Rg1 and CCS impedance, they are in parallel).
This gives lots of gain (if you need that much gain), good frequency response, low distortion, and a good square wave response.
I have never used an EF86, they seem a little expensive for me, but I might spring the $$ for them sometime.
I often use 12AY7, or sometimes ECC82 for my input/driver tube depending on the gain I need; and I RC couple it to the output tube (6L6, KT66, KT77, or 7591S in Ultra Linear or Triode wired mode).
I am mostly into push pull amplifiers now, but I still do single ended amplifiers.
I am contemplating building a 2 stage balanced amplifier, I might purchase two JJ EF806 for a mono-block; I can easily drive over to Eurotubes.com and pick up a pair (they vigorously re-test them from the manufacturer).
2. 6N1P or 6922 and Interstage transformer
A 6N1P has a plate resistance, rp of about 7700 Ohms.
A 6922 has a plate resistance, rp, of about 2600 Ohms.
You need enough primary inductance in the Interstage transformer, so that 7700 Ohms or 2600 Ohms does not reduce the low frequency bandwidth (low frequency fall off).
Example:
2600 Ohm plate resistance, rp
20.7 Henry has 2600 Ohm inductive reactance, XL
-3 dB at 20Hz
-1dB at 40Hz
Want to be -1dB at 20 Hz, then use 41.4 Henry primary inductance.
With 7700 Ohms plate resistance, then 7700 / 2600 = 2.96, almost 3 times the resistance.
With an rp of 7700, then In round numbers:
20.7 Henry x 2.96 = 61.3 Henry
With 61.3 Henry, and 7700 Ohms rp . . .
-3 dB at 20Hz
-1 dB at 40Hz
It takes a very good Interstage transformer to have good frequency response, and good square wave response (a good one might be expensive; only a few of the Interstage transformers I used had good square wave response).
Many Interstage transformers are very sensitive to magnetic fields, you may have lots of hum, unless you deal with that:
No magnetic steel chassis
Good spacing of the Interstage to the power transformer, and to the filter choke
Proper angular Orientation of the Interstage, versus the power transformer, and the filter choke
Happy designing, happy building, and happy listening!
Last edited:
If you go with the IT coupled alternative, the rp of the 6922 looks like about 3k for each of the triodes.
So in parallel that would be ~1,5K. To avoid distortion, whatever load you put on the IT's secondary
should be at least 3X that. Normally easy to do for an IT, some resistive loading flattens the freq response.
I had a Decware clone on the bench here a while back. Lots of show but not much go.
Clipping at less than a Watt.
While I was interning the guy across the hall had a Perkin-Elmer IR Spectrum Analyzer.
He also had a Neutron generator as part of the experimental apparatus.
Sometimes we had to leave the building!
So in parallel that would be ~1,5K. To avoid distortion, whatever load you put on the IT's secondary
should be at least 3X that. Normally easy to do for an IT, some resistive loading flattens the freq response.
I had a Decware clone on the bench here a while back. Lots of show but not much go.
Clipping at less than a Watt.
While I was interning the guy across the hall had a Perkin-Elmer IR Spectrum Analyzer.
He also had a Neutron generator as part of the experimental apparatus.
Sometimes we had to leave the building!
Actually, 6N1P has an Rp of 4400R...
https://web.archive.org/web/20070707054433/http://www.svetlana-tubes.com/pdf/6n1p.pdf
Actually, I own quite a few Svetlana 6N1P tubes.
And I have the Svetlana 6N1P data sheet too.
The Svetlana data sheet is wrong. The u (Mu) is correct.
But they Swapped the numbers of Gm and plate resistance, rp. (swapped the numbers, but not the units uMhos and rp).
Svetlana had that error, and never corrected it; then they went out of business.
I have mentioned this before on various threads on Tubes / Valves.
More than once, I built single stages using the Svetlana 6N1P.
The data sheet u = 33 At least the Svetlana data sheet got that correct.
I used a plate load resistor of 3 x 7700 Ohms (RL = 23k Ohms), the stage has a gain of 25.
Gain = (23k/(23k +7.7k)) x 33 = 24.7
The Svetlana 6N1P plate resistance, rp is 7700 Ohms (not 4400 Ohms as listed in the Svetlana data sheet).
And u = Gm x rp. 4400 uMhos x 7700 = 34
So the Svetlana 6N1P transconductance, Gm is 4400 uMhos (not 7700 uMhos as listed in the Svetlana data sheet).
I rest my case.
Just my measurement results, and my calculations.
Your probes and oscilloscope measurements may vary somewhat, but not so much as the Svetlana data sheet errors.
And I have the Svetlana 6N1P data sheet too.
The Svetlana data sheet is wrong. The u (Mu) is correct.
But they Swapped the numbers of Gm and plate resistance, rp. (swapped the numbers, but not the units uMhos and rp).
Svetlana had that error, and never corrected it; then they went out of business.
I have mentioned this before on various threads on Tubes / Valves.
More than once, I built single stages using the Svetlana 6N1P.
The data sheet u = 33 At least the Svetlana data sheet got that correct.
I used a plate load resistor of 3 x 7700 Ohms (RL = 23k Ohms), the stage has a gain of 25.
Gain = (23k/(23k +7.7k)) x 33 = 24.7
The Svetlana 6N1P plate resistance, rp is 7700 Ohms (not 4400 Ohms as listed in the Svetlana data sheet).
And u = Gm x rp. 4400 uMhos x 7700 = 34
So the Svetlana 6N1P transconductance, Gm is 4400 uMhos (not 7700 uMhos as listed in the Svetlana data sheet).
I rest my case.
Just my measurement results, and my calculations.
Your probes and oscilloscope measurements may vary somewhat, but not so much as the Svetlana data sheet errors.
Mabye Putin dictated Svetlana to publish fake news, for whatever reason 🤔?
Best regards!
We need to remember that datasheet values are average values. The Russian 6N1P datasheet I use states u = 35 +/- 7. That's a big swing. Your specific tube may be anywhere in that range.
That, and you could have deviations high in one triode, low in the other triode- within the same bottle. Can be frustrating for some applications that count on roughly equal gain between sections for things like phase splitters without a CCS tail, or using one half per channel for gain and ending up with different volume levels.
Another benefit to the cathodyne versus a long tailed pair is that things like this don't cause extra work.
Gm 4400 uMhos
rp 7700 Ohms
Or . . .
Gm 7700 uMhos
rp 4400 Ohms
That is 1.75 to one (if that is the spread of those tubes, that is +/- 37.5%
Any tube that has that much range, I throw it out.
The Svetlana lists the maximum plate dissipation as 2.2 watts.
200V and 11 mA is 2.2 watts.
Look at Svetlana's own plate curves at this link, and graphically solve the Gm and rp for yourself:
https://datasheetspdf.com/pdf/1039296/Svetlana/6N1P/1
Yes, you can get higher Gm and lower rp, but only at voltage/current combinations that are higher than 2.2 watts plate dissipation. That means use it as a pulse tube, not an audio tube if you want those Gm and rp numbers.
I certainly tried more than one Svetlana 6N1P.
The Svetlana data sheet Gm and rp numbers are way off; but the plate curves are correct (it is hard to fool a curve tracer)
I used to have access to a Sony / Tektronix 370B curve tracer.
Your Mileage May Vary.
If you use 7700 uMhos and 4400 Ohms as your design guide, just be prepared to change the plate load resistor, and the self bias resistor, and be sure to Pulse the tube into the high dissipation area.
Just my opinion.
rp 7700 Ohms
Or . . .
Gm 7700 uMhos
rp 4400 Ohms
That is 1.75 to one (if that is the spread of those tubes, that is +/- 37.5%
Any tube that has that much range, I throw it out.
The Svetlana lists the maximum plate dissipation as 2.2 watts.
200V and 11 mA is 2.2 watts.
Look at Svetlana's own plate curves at this link, and graphically solve the Gm and rp for yourself:
https://datasheetspdf.com/pdf/1039296/Svetlana/6N1P/1
Yes, you can get higher Gm and lower rp, but only at voltage/current combinations that are higher than 2.2 watts plate dissipation. That means use it as a pulse tube, not an audio tube if you want those Gm and rp numbers.
I certainly tried more than one Svetlana 6N1P.
The Svetlana data sheet Gm and rp numbers are way off; but the plate curves are correct (it is hard to fool a curve tracer)
I used to have access to a Sony / Tektronix 370B curve tracer.
Your Mileage May Vary.
If you use 7700 uMhos and 4400 Ohms as your design guide, just be prepared to change the plate load resistor, and the self bias resistor, and be sure to Pulse the tube into the high dissipation area.
Just my opinion.
Last edited:
