A long time ago I ran across some discussions of using choke loads for low B+ applications so I looked at simulations and of course the higher the inductance the better the low frequency performance. In the process however I discovered (OK I am sure rediscovered something obvious) that carefully chosen cathode resistor bypass can result in response quite close to that obtained with a larger choke. Going to diode bias (and I presume battery bias) provides some improvement but the right cap value (not too large not too small) can be even better. I also noted that group delay in the audible range remained below 12mS in all cases.
Might be something to keep in mind if you find yourself with a marginal choke. I presume that something similar should work for IT too.
Might be something to keep in mind if you find yourself with a marginal choke. I presume that something similar should work for IT too.
For a triode, the plate impedance, rp, is dependent on the impedance that the cathode sees.
A 300 Ohm cathode resistor without a bypass cap causes the plate impedance to be constant across the complete audio band.
A 300 Ohm cathode resistor with a 10uF bypass cap will cause the plate impedance, rp, to be constant from about 200Hz to above the high frequency limit of the audio band.
But at about 100Hz and lower, the plate impedance, rp will start to rise.
At 50Hz, the plate impedance will still be rising.
By the time the frequency is about 25Hz, the plate impedance will have stabilized at its new maximum value.
If you want the plate impedance, rp, to drive a choke or interstage transformer to low frequencies, use a larger capacitance bypass cap, perhaps 100uF, or more, depending on the triode's u.
With a triode u = 17, a 300 Ohm un-bypassed cathode self bias resistor, the plate impedance, rp will increase by(u + 1) x 300 = 5,400 Ohms.
A 6SN7 has an rp = 7,700 Ohms; but if the un-bypassed cathode resistor is 300 Ohms, then
the total rp will be 7,700 Ohms + 5,400 Ohms = 13,100 Ohms.
13,100 Ohms might be to high to drive some choke's or interstage transformer's primary inductance at low frequencies.
Your mileage may vary.
Just my opinions.
A 300 Ohm cathode resistor without a bypass cap causes the plate impedance to be constant across the complete audio band.
A 300 Ohm cathode resistor with a 10uF bypass cap will cause the plate impedance, rp, to be constant from about 200Hz to above the high frequency limit of the audio band.
But at about 100Hz and lower, the plate impedance, rp will start to rise.
At 50Hz, the plate impedance will still be rising.
By the time the frequency is about 25Hz, the plate impedance will have stabilized at its new maximum value.
If you want the plate impedance, rp, to drive a choke or interstage transformer to low frequencies, use a larger capacitance bypass cap, perhaps 100uF, or more, depending on the triode's u.
With a triode u = 17, a 300 Ohm un-bypassed cathode self bias resistor, the plate impedance, rp will increase by(u + 1) x 300 = 5,400 Ohms.
A 6SN7 has an rp = 7,700 Ohms; but if the un-bypassed cathode resistor is 300 Ohms, then
the total rp will be 7,700 Ohms + 5,400 Ohms = 13,100 Ohms.
13,100 Ohms might be to high to drive some choke's or interstage transformer's primary inductance at low frequencies.
Your mileage may vary.
Just my opinions.
Last edited:
I've seen this discussed in Van der Veen's book. Basically you're tuning the circuit to get a butterworth response; the passband gets a little wider but the cutoff gets sharper than with no cap. However, it is not clear to me why this would be any 'better' than just using an arbitrailty large cap to get an even wider response (with softer cutoff)🤷♂️
Thanks for your response. I am saving up for your power supply book. 🙂
My thinking is kind of a two birds with one stone situation. If the cutoff frequency is pushed just below the audible range the second order rolloff could provide a touch of infrasonic filtering for "free". The tricky part with choke loads is unexpected reactance in following stages. Of course if you plan on using local(ish) FB it could be very interesting.
A LR filter tuned to a steeper rolloff below F3 may have a resonance just above F3, which could be bad enough to cause boomy bass response in the amplifier.
A friend of mine had a SE 300B amp at one point. He bypassed the 300B Rk with a small value cap, and noticed big bass response. He later changed the bypass cap to a larger value (reducing the F3 down below the audio range) and found the bass response was flatter. I expect that was due to peaking in the filter formed by Lpri of the OPT, rp of the 300B, the value of Rk, and the value of Ck.
I'm pretty sure I've seen this in LTspice simulations. Should be easy enough to replicate...
A friend of mine had a SE 300B amp at one point. He bypassed the 300B Rk with a small value cap, and noticed big bass response. He later changed the bypass cap to a larger value (reducing the F3 down below the audio range) and found the bass response was flatter. I expect that was due to peaking in the filter formed by Lpri of the OPT, rp of the 300B, the value of Rk, and the value of Ck.
I'm pretty sure I've seen this in LTspice simulations. Should be easy enough to replicate...