If I understand right, the 414x SMU has the LO terminals of its four channels interconnected, i.e. all channels have the same reference. I would like to measure a device with four terminals (A,B,C,D) with this SMU. I plan to use SMU channel 1 to apply a voltage between A and B (Vab) and measure the flowing current (Iab), and SMU channel 2 to measure the voltage drop between C and D (Vcd). To do this, I would connect A to HI_ch1, B to LO_ch1, C to HI_ch2 and D to LO_ch2. Since the LO points are interconnected, I would be externally connecting my B and D chip terminals, which may create a loop inside my chip. Even if I use four-terminal measurement in channel 2, I would need to connect, close to the DUT, sense-LO_ch2 to LO_ch2, which automatically would connect points B (LO_ch1) and D (LO_ch2) again.
Is that correct, or I am missing something? In case it is correct, couldn't then I perform the measurement with channel 2 without interconnecting terminals B and D? Since channel 2 will only be measuring (supplying 0A as a current source), could I connect only sense-LO and sense-HI in channel 2 to avoid the interconnection?
Yes, all four LO terminals are isolated from ground, but not each other.
If you use channel 1 to apply a voltage between A (HI) and B (LO), then all other channels must reference B. Note you could also reference A (LO) if needed by connecting it to the LO terminal and then applying a negative voltage -(-Vab).
If you then connect channel 2 to C (HI) and D (LO), then both LO terminals (B and D) must reference the same voltage.
Since your device is only 4 terminal, you could just connect Channel 1 HI to A, Channel 1 LO to B, Channel 2 HI to C and Channel 3 HI to D. You'll need to do some simple math to figure out the voltage, but then there's no risk of shorting leads together (unless of course you set the output voltage to 0).
There's no way to avoid the interconnection because internally all the LO leads are tied together. Remote sense is just going to make the HI terminal voltage change; there's nothing you can do for the low terminal.
Let me know if I didn't answer any of your questions or if you come up with any more.
Thanks for the answer John.
Your solution of using channel 3 and only the HI terminals sounds feasible. Just an additional question:
In that case, with only the HI terminals of channels 2 and 3 sensing voltage, would it still be possible to do remote sense via interconnecting the HI and HI-sense terminals close to the DUT? Or given that LO/LO-sense are not connected, remote sense (to get rid of cabling effects) wouldn't work?
As long as the part internally references whichever pin you connect to LO, I recommend hooking all negative remote sense leads to said pin and each positive remote sense pin to the respective channel. The differential voltage from each channel to the common low pin will then be correct.
If you leave one of the remote sense leads disconnected, there's an internal pull resistor from each sense lead to the respective HI/LO pin, but that's neither recommended or electrically what you want to do.