Tracing the transmitter valve curves
I posted several times about my 4-65a SE Amplifier and also traced in the past the curves using my analogue curve tracer to get a view of the loadlines of this fantastic DHT in triode mode.
Now that I have the uTracer I traced again the curves including grid current and A2 anode curves which are very handy for this type of transmitter valve.
My tests were conducted with the following heating and bias setup:
- DC heater using Rod Coleman regulator @ 6V and 3.5A
- Cathode connection to the negative filament terminal of the regulator and valve.
There is no reference test point for this valve in triode mode and datasheet doesn’t provide the sort of bias point details I was looking for. I found the following to be quite good to represent a very healthy 4-65a/JAN-8165 valve:
- Vf=6V, If=3.5A
- Va=250V
- Vs=250V
- Vgk=-10V
- Ia=70mA
- Is=12mA
- Gm=3.2mA/V
- Ra=3KΩ
However, in triode operation Ra=1.6KΩ, Gm=3.4mA/V and Mu=5.4-6.
The curves are very linear as expected:
My A2 grid curves were limited to 100V and only generated +5V to +20V steps. I could have gone further but given the accuracy of the uTracer for A2 curves, I rather stayed on the safe side as we will see later why:
The accuracy is impacted on curves above +10V. Nevertheless the uTracer does a fantastic job to allow us to generate a very accurate model and adjust the shape of the curves in A2.
Here is my best attempt to match the grid current model in A2:
The model is very decent in particular if we consider that the area of operation during A2. Anode voltage will go down from 150V and the reflection of the grid impedance can help us better simulate the distortion of the grid driver under these conditions.
So, how will perform? I’m using the LL9202/100mA wired in 6K5:8Ω mode. The anode load is Zaa=Ra+N^2*(Rs+RL)=600Ω+812.5*(0.4Ω+8Ω)=7,425Ω. My amplifier operates at the following bias which I found optimal given power supply I have at hand:
The 4-65a can deliver in class A1 up to 4.7W of clean triode SE sound @ THD=3.57% when driven by a 120Vpp signal. If we push this triode to class A2, then:
H3 distortion component increases to about 0.25% and overall THD goes up to 4.6% to deliver 8.3W into the 8Ω load. The driver should be able to provide at least 160Vpp.
You can push the 4-65a to get 10W in this configuration:
Grid will swing positively up to +30V and driver should provide 180Vpp at least to deliver 10W @ 5%
Not only it looks good in paper, this triode sounds marvelous. Very detailed and with a powerful bass when coupled with the right output transformer. I’m keen to do same tests with the 814 and compare differences.