On my Q17 I have 680uV DC offset with C7 1uF Mundorf Evo Silver-Gold and C2 Rubycon Black-Gate NX.
Regards,
Tibi
Almost any part is difficult to get now. Some parts are listed with the price 4 times higher than January 2020.
If you wanna match positive and negative rails on the opamp, than use resistors instead Zenners. Each zener is feed by a good quality CCS at ~4mA, but the difference in voltage is given by N and P mosfets used. IRF610 and IRF9610 are advertised as complementary, but they are not, beside transconductance and many other parameters, Vgs is far from been the same. Replace D2 and D3 with resistors in range of 3750ohm to get simetric operational power supply. Use some multi-turn 5K pot to trim.
Regards,
Tibi
If you wanna match positive and negative rails on the opamp, than use resistors instead Zenners. Each zener is feed by a good quality CCS at ~4mA, but the difference in voltage is given by N and P mosfets used. IRF610 and IRF9610 are advertised as complementary, but they are not, beside transconductance and many other parameters, Vgs is far from been the same. Replace D2 and D3 with resistors in range of 3750ohm to get simetric operational power supply. Use some multi-turn 5K pot to trim.
Regards,
Tibi
Still on this offset story, another idea.
Can we sort the IRF960 / IRF9610 with the available parameters for the best matched?
Otherwise, we weld a measured P-Chanel and then, we try several N-Chanel in order to try to be as close to the balance. OR we weld only one that we measured and we deduce the best by measuring the others (without welding them).
Stef.
Can we sort the IRF960 / IRF9610 with the available parameters for the best matched?
Otherwise, we weld a measured P-Chanel and then, we try several N-Chanel in order to try to be as close to the balance. OR we weld only one that we measured and we deduce the best by measuring the others (without welding them).
Stef.
Attachments
There should be some typical distortion value for this circuit. A deviation for the worse is a signal that something is not working as it should. Distortion measurement and the use of an oscilloscope are necessary to adjust the amplifier.
Matching mofet is not the best solution. Simplest solution, If you like to have opamp power supply symmetry, is to use resistor instead zener.
Any difference between opamp in+ and in- is reflected at the output in DC offset.
Anyhow, 40mV is low, but can be even lower.
Regards,
Tibi
THD as a whole does not mean much, as long you do not know the harmonic distribution of this.
In other words, in low dosage and properly "prepared" HD is good and I need HD.
Regards,
Tibi
The tube amplifier or class A are a good example.
For the offset.
I don't think I will make the change. It is this breaking the head for not much and it will make a big truth on the PCB.
Out of curiosity, can we do it on only one of the power rails (the V +) and leave the zener on the negative rail?
If this is a reliable and secure method to have an offset setup, one can imagine replacing a single zener and instead adding the resistor and the trimmer on another PCB version.
I had fun putting the old man OPA627 on the board socket. It's not good at all.
Stef.
For the offset.
I don't think I will make the change. It is this breaking the head for not much and it will make a big truth on the PCB.
Out of curiosity, can we do it on only one of the power rails (the V +) and leave the zener on the negative rail?
If this is a reliable and secure method to have an offset setup, one can imagine replacing a single zener and instead adding the resistor and the trimmer on another PCB version.
I had fun putting the old man OPA627 on the board socket. It's not good at all.
Stef.
The one I use is https://ro.mouser.com/ProductDetail...=sGAEpiMZZMutXGli8Ay4kJNmOIz0qTNK8meaMEmKCYk=
Regards,
Tibi
Regards,
Tibi
Hello Tibi,
I am now building a new Q17 again, this time I have reduced R3 from 150 to 100 ohms. This increases the current in the cascode. I have loaded the Space models of the ZVN and ZVP types into the simulation and I got the impression that they play cleaner with a little more current. So just from the simulation evaluation, in reality I have to build it up first.
Tim
I am now building a new Q17 again, this time I have reduced R3 from 150 to 100 ohms. This increases the current in the cascode. I have loaded the Space models of the ZVN and ZVP types into the simulation and I got the impression that they play cleaner with a little more current. So just from the simulation evaluation, in reality I have to build it up first.
Tim