jfet totem
Dear Euvl,
Nice that you brought up this jfet totem circuit.
I had built a version of it, for work not audio..
I would confirm that it works as stated. The DC balance of the working point is maintained from room temperature all the way down to 77K, where it is doing it's job also right now.
After a careful initial selection of singles, I have seen no more shift than a couple hundred mV
And I run it at low tension because of power consumption- one should consider cryogenic system capacity.
No special care in thermal coupling apart from close smd mount.
I applied a buffer afterwards. The high gain is useful to be converted to high speed after the series-shunt feedback as DRCG had suggested. (a current feedback amplifier again..) At gain 10 I have 150 MHz -3dB response now. But I have versions with 600MHz bandwith as well.
It has a very very nice dynamic response.We are at 2.4nsec rise/fall time now, slew rate above 4000V/usec, no overshoot / ringing at all at small signal level, some overshoot at 6Vpp output level.
Though I would not call it a low noise circuit. The low inherent device 'en' sees the additional contribution of the source resistance, and to this also the current source contribution is 'power summed'.
Ok, with these fets we are still below of 1.6nV/sqrtHz and with a very low 1/f corner.
Ciao, George
Dear Euvl,
Nice that you brought up this jfet totem circuit.
I had built a version of it, for work not audio..
I would confirm that it works as stated. The DC balance of the working point is maintained from room temperature all the way down to 77K, where it is doing it's job also right now.
After a careful initial selection of singles, I have seen no more shift than a couple hundred mV
And I run it at low tension because of power consumption- one should consider cryogenic system capacity. No special care in thermal coupling apart from close smd mount.
I applied a buffer afterwards. The high gain is useful to be converted to high speed after the series-shunt feedback as DRCG had suggested. (a current feedback amplifier again..) At gain 10 I have 150 MHz -3dB response now. But I have versions with 600MHz bandwith as well.
It has a very very nice dynamic response.We are at 2.4nsec rise/fall time now, slew rate above 4000V/usec, no overshoot / ringing at all at small signal level, some overshoot at 6Vpp output level.
Though I would not call it a low noise circuit. The low inherent device 'en' sees the additional contribution of the source resistance, and to this also the current source contribution is 'power summed'.
Ok, with these fets we are still below of 1.6nV/sqrtHz and with a very low 1/f corner.
Ciao, George
They are "tres cher" but Mouser is now carrying the Interfet IF3601 and IF3602.
Datasheet is somewhat skimpy -- no neat graphs -- and I can see why as the dissipation is only 300mW. 750mS @ Idss. Ciss and Crss are hundreds of pF.
The gm is a bit discombobulating.
Tomorrow will start a thread with some of the characteristics.
Datasheet is somewhat skimpy -- no neat graphs -- and I can see why as the dissipation is only 300mW. 750mS @ Idss. Ciss and Crss are hundreds of pF.
The gm is a bit discombobulating.
Tomorrow will start a thread with some of the characteristics.
Last edited:
One thing about that circuit -- the input impedance is essentially R1, or 100 Ohms. If you eliminate R1, or make it a high value, the maximum input impedance is approximately R2||R3. (Approximately).
Great amplifier for MC, but for measurement purposes it loads the DUT.
