Page 1 of 1

Output connections

Posted: August 30th, 2010, 11:39 am
by pevans
I have a triax to alligator cable connected to the OUTPUT HI triax connector on the rear panel.
What clips should I connect to the DUT?

Re: Output connections

Posted: August 30th, 2010, 11:52 am
by Dale C
The Model 237-ALG-2 triax to alligator clip cable has three alligator clips on the clip end.
1. RED is the output HI or center conductor.
2. BLACK is the guard or the inner shield.
3. Green is the earth ground or outer shield.
With the 236, 237 or 238 the DUT should be connected to the RED and GREEN clips.
This is with the LO to GROUND Link on the rear panel in place.
LO is also on the SENSE LO triax connector. It is the inner shield. But you would need an additional triax to alligator clip cable.
Or LO is available as a banana jack. But this would take another banana to alligator clip type cable.
The Model 237-ALG-2 is the most convenient method.

Re: Output connections

Posted: September 23rd, 2010, 2:48 am
by nooblerz
Is it ok to short the BLACK and GREEN clips to the DUT chassis? My setup requires reverse bias current measurement. Thanks

Re: Output connections

Posted: September 23rd, 2010, 10:02 am
by Vince W
Assuming you are talking about the OUPUT HI triax,
Don't short the BLACK lead to GND. This is a driven guard which is designed to negate leakage currents in cabling or fixturing. It is best to leave BLACK unterminated, but run the triax out as close to the DUT as possible. If the DUT chassis is the LO side of the measurement, it is reasonable to connect the GREEN lead to DUT chassis.

For SENSE LO triax, the inner shield is the OUTPUT LO, so it would be OK to connect its BLACK lead to DUT chassis. Without knowing about how the LO side of the measurement within the DUT is related to the DUT chassis, I'm not sure of the best approach.

If the DUT LO side is isolated from the DUT chassis, it is better to connect OUTPUT HI to the DUT HI, and OUTPUT LO (in the SENSE LO cable) to DUT LO, and both GREEN clips to DUT CHASSIS (GND) as an electrostatic shield.

Re: Output connections

Posted: March 16th, 2017, 11:02 am
by spazos
Hello, I see this is an old thread but I have a similar issue. As part of a Noise characterization system for unpackaged semiconductors, I was looking to include a switch in my system that allows to perform I-V measurements through an SMU or Noise measurements through an amplifier stage and a signal acquisition system. The idea is to avoid connecting and disconnecting the sample and the equipment to measure these two characteristics.

The problem lies on the fact that the noise characterization stage is fully Coax (BNC connectors), and the probe station and SMU are TRIAX (2636B). Among the available adapters, I have TRIAX to BNC with SHIELD tied to GUARD. My noise measurement system has LO tied to GND (hence LO in the SMU is connected in the backpanel to GND, i.e. nonfloating). Therefore, this renders that the SMU HI GUARD is shorted to GND and, hence, to the SMU LO.

Can this damage the SMU's FORCE/GUARD outputs? Are there safety margins to be able to use this connection?

I've seen that floating the GUARD would be an option but a) I don't have that adapter currently available and b) it is not recomended for currents below 1nA.

Thanks in advance for the help.

Re: Output connections

Posted: March 16th, 2017, 11:24 am
by Vince W
Where are the BNC/TRIAX adapters and the switch in your system? Is the switch a TRIAX switch (recommended)?

I would recommend maintaining a TRIAX connection from the SMU, through switches and to the probes when using your SMU, and only transition to coax when connecting the noise measurement equipment to the probes. So the adapters would go between the noise measurement equipment and the TRIAX switch. This way, you will maintain low current sensitivity when switched to the SMU.

Keithley sells 708B and 707B switch frames with 7072 and 1714A switch modules - both of these switch modules types are TRIAX.

If you short the TRIAX outer shield (LO or Chassis, depending on the model of SMU) to the TRIAX Inner Shield (Driven Guard) of the SMU, you will be shorting the Guard amplifier, but will likely not damage it, as it is current-limited. However, you will never get a good measurement with the SMU while having them shorted. This is not the way you would want to leave your system.

In general, it is better to use BNC/TRIAX adapters which drop the TRIAX inner shield (guard), so that the BNC outer shield is the same as the TRIAX outer shield, and BNC center lead is the same as the TRIAX center lead. Keithley sells 7078-TRX-GND (TRIAX M to BNC F) and 237-BNC-TRX (TRIAX F to BNC M) adapters, both of which drop the guard connection.

Re: Output connections

Posted: March 16th, 2017, 12:35 pm
by spazos
Let me first thank you for the quick response.

Attached you'll find a raw block diagram of the measurement setup.

Regarding your first question, the switch is installed in the amplifier stage metal box (all conections go through this stage). The switch is NOT triax.
I was hoping not to add a switch frame to the system as it would impact negatively on the noise floor of the noise characterization stage (hoping to reduce the connected equipment to a minimum). Right now, without the SMU connected, the noise floor is within our requirements, but I'm afraid that adding cables and a switch frame would degrade this figure.

Thanks for the info on the current limitation of the guarding amplifier. I understand that there would be no risk to damage or degrade the SMU, which is my main concern. I expect the isolation to be a problem without the guard operating, but would need to test this in order to understand if it would still allow to characterize the DUT at least in the required current range.

I've found earilier today the adapters you are referencing, but sadly we don't have them right now. We only have POMONA 5299 (used in the fixture as shown in the attachment) and 7078-TRX-BNC (drops TRIAX SHIELD and connects GUARD to BNC SHIELD, if I'm correct). We will look forward to acquire those adapters you suggested.

Thanks again

Re: Output connections

Posted: March 20th, 2017, 8:35 am
by Vince W
I did not mean to say that you should continue to use the adapters you currently have. While shorting the guard to outer shield may not damage the guard amplifier, it is a potentially dangerous practice, and one that Keithley does not support. As you noted earlier, because you could encounter a situation where the outer shield of the cable is exposed to guard/output voltages, it can be dangerous to do this!

If there is any way to implement your internal switch box while keeping TRIAX connections everywhere but the noise measurement box. In other words, arranging the switch circuitry so that it switches the axial lead, but keeps TRIAX connections between the probes and SMU, you will likely have good low-noise performance AND good low-current sensitivity.

Keithley also has solderable TRIAX connectors for your noise measurement switch box: 237-TRX-TBC and 7078-TRX-TBC.