BJosephs,
it is a push-pull topology exactly as it should be. The resistor provides protection by reducing the very high, nonlinear gate-source impedance and the amplitude.
Susan,
under practical load, the frequency response of the input transformer will almost certainly drop below 20 kHz.
it is a push-pull topology exactly as it should be. The resistor provides protection by reducing the very high, nonlinear gate-source impedance and the amplitude.
Susan,
under practical load, the frequency response of the input transformer will almost certainly drop below 20 kHz.
One could put in a 47K resistor on each MOSFET, but that won't be enough to protect the part for the BOOM!!!
I am curious as to where do you get the "... the very high, nonlinear gate-source impedance and amplitude." from?
... and???
Current Schematic
Susan
I believe you are using the IXTH20N50D as the output mosfet in your Zeus amplifier. I have eight of these MOSFETs but the fear that obtaining two match pairs is going to be an issue, given previous experience with matching vertical mosfets. The prospect of buying more to match, at $56.68AUD each (Digikey), puts that as a remote option. May I ask what parameter is most important matching and what tolerance is acceptable, eg. is it Vgs or transconductance? Like to know your thoughts and any suggestions.
Regards
Anthony
Susan
I believe you are using the IXTH20N50D as the output mosfet in your Zeus amplifier. I have eight of these MOSFETs but the fear that obtaining two match pairs is going to be an issue, given previous experience with matching vertical mosfets. The prospect of buying more to match, at $56.68AUD each (Digikey), puts that as a remote option. May I ask what parameter is most important matching and what tolerance is acceptable, eg. is it Vgs or transconductance? Like to know your thoughts and any suggestions.
Regards
Anthony
Dear Susan,
I have been referring to the same elegant circuit.
The insulated gate can hardly be characterized as anything other than a considerably doped, high impedance, nonlinear electrical interface that is in need of stabilization by resistive elements. The very thin dielectric layer is indeed vulnerable, but such hostile voltages can barely at all arise. For more common MOSFETs, a low value gate-source resistor is a natural part of the biasing arrangement.
Bandwidth is dependent on biasing and loading. Measurements data made under no load conditions are of little relevance.
Confusing information? I am not urging you to make changes to your working amplifier.
I have been referring to the same elegant circuit.
The insulated gate can hardly be characterized as anything other than a considerably doped, high impedance, nonlinear electrical interface that is in need of stabilization by resistive elements. The very thin dielectric layer is indeed vulnerable, but such hostile voltages can barely at all arise. For more common MOSFETs, a low value gate-source resistor is a natural part of the biasing arrangement.
Bandwidth is dependent on biasing and loading. Measurements data made under no load conditions are of little relevance.
Confusing information? I am not urging you to make changes to your working amplifier.
Hi Anthony,
I sort them by the current flow through a 1 ohm resistor, to get the best match to minimize the DC in the transformer. They need to be all on the same heat-sink as the values will change with temperature.
A few tens of mA difference isn't going to hurt with an EI transformer, might be a problem with a toroid.
You could gap EI transformers with a thin c. 0.1mm shim of copper without killing the bass-response too much as a "lets see how this works" measure.
I have hundreds of the parts, as they were going to be made obsolete in 2011 so I bought a lifetime supply.
For my daily use pair of PP amps and the SE line driver I got close enough matches from a tube of 30 but the variation between individual devices is large; I got a min/max spread of 512mA to 876mA.
I am NOT a seller however I have enough to do some matched sets if people are interested since it is now very unlikely that I am ever going to be able to sell my amps commercially.
Doing a quick search I see them from this place for under $10 each, which is a lot more palatable:
IXTH20N50D | IXYS Corporation IXTH20N50D Stock available from Infinity-Electronic.hk | IXTH20N50D with Best Price at Infinity-Electronic.hk
Best,
Susan.
This is what I refer to a 2:1 mode, when needing more power for driving bass speakers. All transformer windings are the same number of turns.
I recommend 4:1 for general use, which is substantially lower distortion. The transformer secondary output windings are half the number of turns as the input primary windings. It is technically 2:1 ratio, however I am referring to the amplitude between the PP primary windings which are +ve and -ve and the voltage across the output.
Thanks. This is drawn in DxDesigner and the curves and hop-overs are pass-thru components so the original is actually a live schematic and can be net-listed.
Susan,
I meant to say that 200 MHz for a sine wave is not a valid practical value. Bandwidth falls off rapidly with energy level. Signal transformers typically suffer from insufficient bandwidth.
The IXTH20N50D can withstand high gate-source voltages. Zener diodes do not protect against transients.
Regarding your preamplifier: a DC blocking capacitor produces lower distortion than a transformer. Have you considered an autoformer volume control instead of the transformer volume control?
Agreed.
TonyTecson,
the reverse transfer capacitance is much more harmful than the gate-source capacitance. The reverse transfer capacitance is always present, increasing with the drain-gate voltage / drain-source current, adding to the gate-source capacitance, and impacting transconductance. The low distortion of voltage followers stems from small amplitudes, low saturation and wide bandwidth in spite of a very large gate-source capacitance.
I personally don`t use voltage followers.
I meant to say that 200 MHz for a sine wave is not a valid practical value. Bandwidth falls off rapidly with energy level. Signal transformers typically suffer from insufficient bandwidth.
The IXTH20N50D can withstand high gate-source voltages. Zener diodes do not protect against transients.
Regarding your preamplifier: a DC blocking capacitor produces lower distortion than a transformer. Have you considered an autoformer volume control instead of the transformer volume control?
Agreed.
TonyTecson,
the reverse transfer capacitance is much more harmful than the gate-source capacitance. The reverse transfer capacitance is always present, increasing with the drain-gate voltage / drain-source current, adding to the gate-source capacitance, and impacting transconductance. The low distortion of voltage followers stems from small amplitudes, low saturation and wide bandwidth in spite of a very large gate-source capacitance.
I personally don`t use voltage followers.
Just in case you have missed the point in amongst the c. 1700 posted comments since 2004, as that is a lot to wade through, this thread is about using voltage followers.