Greetings everyone,
How does one choose the proper triode for a phase inverter? Here are the specs on a triode (6M11) I am considering:
Triode Class A Amplifier
Plate Voltage 125 V
Grid No. 1 from 120 Ω
Amplification Factor 58
Plate Resistance (approx) 10K Ω
Transconductance 8K µ
Plate Current 8 mA
Grid No. 1 Voltage for Ib of 50 µa -4.5 V
How does one choose the proper triode for a phase inverter? Here are the specs on a triode (6M11) I am considering:
Triode Class A Amplifier
Plate Voltage 125 V
Grid No. 1 from 120 Ω
Amplification Factor 58
Plate Resistance (approx) 10K Ω
Transconductance 8K µ
Plate Current 8 mA
Grid No. 1 Voltage for Ib of 50 µa -4.5 V
Well, since there are about a million phase inverter circuits, each one with different requirements, it might help to be a bit more specific. Diff amp? See saw? Cathodyne? Cross-couple? Paraphase?
Which circuit you pick might depend on the context: how much voltage swing do you need, what HT voltage do you have, hi-fi or guitar, how concerned are you about exact balance, what are the output valves and their configuration?
Alternatively, do the best you can with what you have and put up with whatever you get.
Alternatively, do the best you can with what you have and put up with whatever you get.
Well, application is Hi-Fi and feeding P-P AB1. I'll need about 75V P-P. The B+ for the "drivers" is around 200v.
Also consider what mode your are in (probably AB1, but some people run AB2), and what the drive current to the output tubes is going to need (easy to drive tubes or difficult to drive?)
The higher, the better. If you can manage 400V or so, a 6SN7 as a constant current diff amp (long tail pair) will easily swing 75V p-p and very low distortion. I don't know what the linearity is for a 6M11.
I set up tubes in a test jig and measure distortion using a sound card and an interface box (Pete Millett's is excellent for this purpose). See, for example, Morgan Jones's tests on 6SN7s and 417s in "Valve Amplifiers" 3rd edition.
Not so much. They'll give you a gross idea of distortion versus load, but in the end, you have to set up an experiment and measure. One advantage is that you can move the operating point and see how that affects distortion and the distribution of harmonics.
The 6M11 triode appears to be similar (higher gm, but same kind of curves) to a 12AT7. Should work well in an differential stage splitter (2nd harmonic cancels). The linear ramping gm curves will sum to a nearly constant gm.
Looks like maybe 4 to 9 mA should be linear (gm 100V curve curving badly above 9 mA and Mu 250V curve curving badly below 4 mA). So maybe 6 mA idle for each tube at 175 V actual plate voltage, using about 1 Watt plate diss. With 37.5 K load resistors (400V B+), a +/- 2 mA change would give +/- 75 volts change for a total of 150 V swing. A 1 mA change would give a total 75 V swing. A starting point anyway. Best way of course is to rig up a test stage with a varialbe B+ supply, variable load R, variable CCS tail and a sound card FFT to observe the results.
edit:
The 6M11 datasheet spec of 10K Ohm Rp at 8 mA appears to be incorrect. The GE manual gives 7250 Ohms at 8 mA. Mu = gm*Rp so that checks. The 37.5 K Ohm load resistors would then give around 5 X the Rp, an often used rule of thumb.
Looks like maybe 4 to 9 mA should be linear (gm 100V curve curving badly above 9 mA and Mu 250V curve curving badly below 4 mA). So maybe 6 mA idle for each tube at 175 V actual plate voltage, using about 1 Watt plate diss. With 37.5 K load resistors (400V B+), a +/- 2 mA change would give +/- 75 volts change for a total of 150 V swing. A 1 mA change would give a total 75 V swing. A starting point anyway. Best way of course is to rig up a test stage with a varialbe B+ supply, variable load R, variable CCS tail and a sound card FFT to observe the results.
edit:
The 6M11 datasheet spec of 10K Ohm Rp at 8 mA appears to be incorrect. The GE manual gives 7250 Ohms at 8 mA. Mu = gm*Rp so that checks. The 37.5 K Ohm load resistors would then give around 5 X the Rp, an often used rule of thumb.
Last edited:
Below are two, basically identical constructions of LTP phase splitter I have used with good results. The first one having 300 V supply voltage can give some 70 to 80 Vpp drive with total THD below 0,5 %.
The tube I have used for this is 6N1P, but I assume ECC81 will also work fine.
The second circuit uses soviet 6N7S or similar USA type 6N7. It has 400 V supply voltage and can provide up to 100 Vpp drive with total THD less than 0,5 % when "fine tuned" to optimum.
Both circuits have variable resistor to balance the output voltages or to "match" the output voltage for differing gain of the output tubes.
The tube I have used for this is 6N1P, but I assume ECC81 will also work fine.
An externally hosted image should be here but it was not working when we last tested it.
The second circuit uses soviet 6N7S or similar USA type 6N7. It has 400 V supply voltage and can provide up to 100 Vpp drive with total THD less than 0,5 % when "fine tuned" to optimum.
Both circuits have variable resistor to balance the output voltages or to "match" the output voltage for differing gain of the output tubes.
An externally hosted image should be here but it was not working when we last tested it.
Making the plate resistors equal and using a CCS as the long tail GREATLY improves the performance and eliminates the need for adjustment.
Here's how I designed that same sort of phase splitter. Here, you have the DC coupling between a SE preamp, and the LTP splitter itself. DC coupling gives the CCS the necessary operating voltage, so done. No need to force AC balance with unequal resistors or pesky variable resistances that can't be adjusted without an o'scope anyway, and which just might fail in operation.
Attachments
Ray, Miles's schematic is a good example. You don't even have to get that fancy with the CCS since the impedances at the cathode are pretty low.
I posted a couple months back a simple demonstration about the power of running a phase splitter with a CCS tail- I used grossly mismatched tubes (I think it was a 12AX7 and 6DJ8), yet the balance was perfect. It's the right way to do a long tail phase splitter.
I posted a couple months back a simple demonstration about the power of running a phase splitter with a CCS tail- I used grossly mismatched tubes (I think it was a 12AX7 and 6DJ8), yet the balance was perfect. It's the right way to do a long tail phase splitter.
Here's another similar method, with a nice low output impedance. You'll definitely want to get that B+ higher than 200V.