OK, so following the recent discusion on FET-based tube emulators, I did some more work on my own ideas. Friday afternoon at work, and all that.
First I took the dual gate MOSFET cascode one and compared it against a Mullard ECC83. After some playing with the component values, they gave similar traces, but it was extremely temperature-sensitive. Someone had borrowed the heat gun, but there was no need for it: even breathing on the MOSFET caused about 10V shift in drain voltage. That made me think that the variation between different MOSFETs would be even worse.
So, I tried the LND150 in a circuit somewhere between the Trioderizer and the scalable diode. It worked surprisingly well. The clipped waveforms were as good, if not better than the dual gate circuit, and temperature stability seemed much better. Also, the component count is lower.
The major drawback is that it leaks DC out of its "grid". In conjunction with the usual 1M grid-leak resistor found in a tube circuit (well, 910k, as I ran out of 1Ms) this creates the positive voltage needed for the LND150 to have the right bias point. But tube circuits often use a volume control, or a different value to 1M, as the grid leak.
So, I'm working on that one. For now, this circuit would only be usable in a design that took the DC into account: it might not work if you just plugged it into a tube socket.
The scope shots show the LND150 circuit working alongside the Mullard ECC83, also set up with a 100k plate and 1.5k cathode resistor. Both are running off the same 250V supply. Top trace is the input signal, and the bottom two are the outputs from the ECC83 and LND150. I will leave you all to figure out which is which.
To paraphrase the Monty Python sketch, "Bring out your FETs!"
First I took the dual gate MOSFET cascode one and compared it against a Mullard ECC83. After some playing with the component values, they gave similar traces, but it was extremely temperature-sensitive. Someone had borrowed the heat gun, but there was no need for it: even breathing on the MOSFET caused about 10V shift in drain voltage. That made me think that the variation between different MOSFETs would be even worse.
So, I tried the LND150 in a circuit somewhere between the Trioderizer and the scalable diode. It worked surprisingly well. The clipped waveforms were as good, if not better than the dual gate circuit, and temperature stability seemed much better. Also, the component count is lower.
The major drawback is that it leaks DC out of its "grid". In conjunction with the usual 1M grid-leak resistor found in a tube circuit (well, 910k, as I ran out of 1Ms) this creates the positive voltage needed for the LND150 to have the right bias point. But tube circuits often use a volume control, or a different value to 1M, as the grid leak.
So, I'm working on that one. For now, this circuit would only be usable in a design that took the DC into account: it might not work if you just plugged it into a tube socket.
The scope shots show the LND150 circuit working alongside the Mullard ECC83, also set up with a 100k plate and 1.5k cathode resistor. Both are running off the same 250V supply. Top trace is the input signal, and the bottom two are the outputs from the ECC83 and LND150. I will leave you all to figure out which is which.
To paraphrase the Monty Python sketch, "Bring out your FETs!"