I build up 2 channels on the PCB, call them Channel 1 and Channel 4
Channel 1 is build according to the list on the thread, see above.
Channel 4 i modified to increase stability etc. allthough difference seems minimum.
Upper frequency using a resistive load seems around 1MHz.
When using inductive loads it will increase the overshoot at MOSFET turn off time possible
damaging it, so snubbers should be added to protect the MOSFET.
I believe in the original design (quadratron) there was a snubber config (R7 and C8) in
combination with the D4.
So all in all a nice design and certainly an asset on the bench.
But this design will not overcome the spikes inherent to using inductive loads as no snubber
exist, so care must be taken to not damage the used MOSFET and/or the scope as
spikes of several hundreds of volts are easily obtained.
Also the isolated frontend will be compromised when using a grounded FG at the input and
using a grounded scope on the backend / dut during the measurement fase.
Video here:
https://www.youtube.com/watch?v=F7HCUil5Rcc&feature=youtu.be My mods on Channel 4 done are:
# C3 was changed from 100nF ceramic as decoupling to 47uF tantalum as buffer.
The 100nF decoupling was moved directly on the MCP1403 chip pins 3 and 6 (+ / -).
# D1 schottky 1n5819 was changed to faster schottky bat 46 (faster MOSFET turn off).
# R3 was changed from 1 Ohm to 0 Ohm (faster MOSFET turn off)
# R4 was changed from 10 ohm to 4.7 Ohm (faster MOSFET turn on).
# D4 was changed from 1N5804 to MUR460 (lesser overshoot during MOSFET turn off).
# additional; better decoupling (100nF ceramic) of the 7805's input/output might be needed.
Itsu