I actually simply modified the previous design you provided earlier to remove the transformer.
So does your earlier design also create an arbitrary phase shift to the voltage?
In my design the quadrature generating R-C network is
in series with the tested impedance Z so that the
current I is the same: Vrefc is in phase with I (and Vrefs in quadrature), no arbitrary phase shift. This is simple but may introduce quadrature error when Z is very high and the current very small, so it is better to feed Z and the reference/quadrature generator in parallel, with the same voltage (something as you did but taking the reference input from V (after R2) and not from the source). But then if you want to generate a Vrefc in phase with V you need an active device.
yeah, an active differentiator. But the problem with a generic op amp is that one cannot use multiple test frequencies due to oscillations at higher frequencies. I intent to use a max test frequency of 100khz.
What are some choices for good op-amps? Would TL0XX make the cut.
thankz.
yeah, an active differentiator. But the problem with a generic op amp is that one cannot use multiple test frequencies due to oscillations at higher frequencies. I intent to use a max test frequency of 100khz.
What are some choices for good op-amps? Would TL0XX make the cut.
thankz.
I have been using LT6220/LT6221 in a similar application at 1-2MHz, these high speed precision opamps will perform much better than TL0XX. (NB. To be stable they will need a not too big feedback resistance (e.g. R9 rather 1k than 10k, also add a small 5-10pF cap in parallel with R9, see "feedback components" section in the datasheet).
Other suggestions:
comparator: LT1394
switch/multiplexer: above 1MHz I was using LMH6570 which is much faster than ADG1634 (but more expensive per channel) and also features internal input buffers solving problems pointed out by unitedatoms. (cf. an alternative cheaper per channel: AD8182, also buffered)
All these components work well with a +/-5V supply.