I think of Rg through Rf as one load.
Back to topic: Here's an idea to eliminate the TL071 and replace with an extremely simple BJT/JFET combination transistor -- the CPH5905 -- a $0.57 part with a 1dB noise figure (the TL071 is 18nV/RtHz @1kHz). No feedback, so no negative impedance wonkiness. CPH5905 is in production, unlike the BF862 which is EOL.
You should get around 20dB gain -- and it's good to 1MHz. V2 and R3 can be adjusted to get ~2.5Vds.
While the CPH5905 is surface mount, I place them on a 6-pin adapter.
Attachments
Looking at available poly-switch the primary objection is the fact that they are smd which is something I am trying to avoid.
This one is an Elektor publish.
Two scales.
100μV-20V
600μV-60V
Bandwidth up to about 2Mhz.
Two scales.
100μV-20V
600μV-60V
Bandwidth up to about 2Mhz.
Attachments
Last edited:
It came with a 90 degree scale (0-50 uA). The meter can go to a 100 degree sweep but it requires the removal of the zero adjuster which does not seem like a good idea so a compromise of 95 degrees was used.
To determine the linearity of the meter a current was applied (at 5 volts) that produced 100% deflection and the the voltage was reduced in 1 volt increments to verify the linearity. The results:
5 Volts, 100%, 51.4uA
4 Volts, 80%, 40.8uA
3 Volts, 60%, 30.5uA
2 Volts, 40%, 20.3uA
1 Volt, 21%, 10.2uA
The linearity is very good through most of the range and it degrades in the lower 30% of movement which is to be expected. All in all this looks suitable for the project.
