SMU Guard connections

[nickstorton]

Hi All,

I've moved this question here as this is a more appropriate board.

I'm looking to perform some low current measurements on my DUT and therefore am looking at including the Guard wires on my DUT evaluation PCB. I have two related questions on this issue.

1: The pinout of my PXIe-4141 says that there are two guard pins and I have verifed this with a DMM, they are connected together and drive at the same level as the Channel HI. However the recommended cable DB25F-DB25F Low leakage cable only has one of these guard pins connected through. Is this correct or is there an error with my cable? I have checked all 4 channels and they are all the same.

2: Does anyone know of any good information / best practices for the guard tracking on my PCB?

Thanks all,

Nick

[Vsenior]

Hi Nickstorton, 

Hope you're keeping well? 

I can't tell from the help file if these are suppose to be wired through but looking at some internal ressources it looks like these cables should be wired through. I'm going to see if I can find one of them to test for you and I'll let you know what happens. 

Thanks

Vsenior

[Vsenior]

Hi Nickstorton, 

One other thing, which of the guard pins is the one you had reading as wiring through? It would be good to have a better idea of what I'm looking for. 

Thanks

Vsenior

[Vsenior]

Hi Nickstorton, 

I've been doing some research on this cable and the use with the PXIe 4141. My internal circuit map, backed up with a record of an escalation to a Precision DC product specialist, tells me that not all the wires are directly wired through the cable. 

Referencing this help file for the pinouts: http://zone.ni.com/reference/en-XX/help/370736Q-01/ni_dc_power_supplies_help/4140_4141_connections/

Guard pins 1, 4, 7, 10 are connected through the cable, while guard pins 15, 18, 21 and 24 are not connected through the cable. Pins 1 and 15, 4 and 18, 7 and 21, 10 and 24 are however connected inside the connector. 

Does this tie in with what you find?

Thanks

Vsenior

[nickstorton]

Hi Vsenior,

Yes I'm very well thanks and how about yourself?

Thanks for looking into this for me. Your findings on the cabling tie in exactly with my findings, All connections are made on the FP connector of the SMU (except pin 13 N/C) but as you say  guard pins 15, 18, 21 and 24 are not connected through the cable. What I'm unsure about is this, according to this help file http://zone.ni.com/reference/en-XX/help/370736Q-01/ni_dc_power_supplies_help/gaurding/ the Force HI should be the center conductor of the triax cable with the guard next and then the Force LO is the outer conductor. What I don't know is whether the Sense HI should also be guarded in the same way? If yes is it connected to Guard pins 1, 4, 7 or 10 at both ends of the connector or should it be connected to guard pins 15, 18, 21 or 24? or should it not be guarded at all? Sorry three questions in one sentence there. I've done a bit of research on this but cannot find a definitive answer on what should and should not be guarded. This also extends to my question on the guarding on the PCB, does guarding stop at the end of the cable? should it just loop around the Force HI trace? or should it guard both the Force and Sense HI traces? Oops done it again 3 in 1. Is there any possibility of getting detailed information on the construction of the cable as this may give me a clue?

Thanks for looking into this for me and I hope I might see you again soon, maybe next time I'm at CSLUG,

BR,

Nick

[Vsenior · Accepted solution]

Hi Nick, 

Yes thanks, very good, moved to AE about a year ago from ISR. 

So having had a think and a look at the connection diagram (which I'm sorry to say I can't send you in it's entirety) and I think I can help with the confusion. 

The cable we're talking about actually isn't made up of coaxial cables, it's made up of twisted pairs with drain wires and an outer conductive sheath. The drain wires are inert and don't get wired to anything, so that leaves the twisted pairs and their sheaths. There are 8 twisted pairs and sheaths in the bundle that gets wired to this connector, meaning 24 potential connections, with pin 13 not wired to anything making up the count to 25.

Now bear with me I'm pretty sure this makes sense: twisted pair number one is wired to pins 2 and 14, which are the two HI pins for channel 0 in and out. The shield for that twisted pair is then wired to pin 1 (a guard pin). So that guard pin guards the two HI pins at once. 

Twisted pair two is connected thus: one half of the twisted pair is connected to pin 3 (Sense LO), the other half is not connected and the shield is wired to pin 16. Now as I understand it, the LO wires don't need so much guarding, so the extra wire I suppose is more like a drain wire here. 

The pattern for the wiring continues like that. Half the twisted pairs are the HIs for Force and Sense for a channel guarded by a shield wired to a guard, and the other half are wired with both LOs and an extra drain. 

I don't know why we don't use COAX when our help file explains it using COAX, but I would suspect is has something to do with cost (one of the AEs who is an electrical eng graduate tells me COAX is far more expensive than twisted pair wires, but he also says it's far harder to strip down for soldering so I suppose that would make manufacturing lots of these cables very expensive and difficult). 

Now as for the PCB stuff, I'll be honest it's not something we have much NI documentation on, but from what google tells me, guard lines can be incorporated into a PCB and from what we've seen in the cable, it looks like we want to guard the HIs for both Force and Sense. I'm a physics grad myself so PCB design isn't something I've got too much experience with. 

When is the next CSLUG meeting (by the way love the group mascot, sea slugs are beautiful!)? I'm presenting at the CLD summit in early September at the Newbury office if you're going to that?

Thanks

Viv

[nickstorton]

Hi Viv,

Hope you are enjoying life in AE.

Thanks for the info it all now makes sense. Twisted pair is indeed much cheaper than Triax, and easier to work with, it's also less prone to breakage when the cable flexes. The fact that the Force HI and Sense HI are guarded together with the same guard helps with the PCB guard trace layout as well. Force LO and Sense LO don't need to be guarded so that bit makes sense as well.

The next CSLUG is 17th Sept, Unfortunately I won't make that one as I'm on holiday for the first half of Sept. Maybe the next one after that.

Thanks again Viv for your help with this,

BR,

Nick

[evancox10]

Thanks for the useful info here, there's not a lot on this cable online, so this was a big help for me. I have a separate but related question about this cable, and figured I'd ask it here in order to consolidate information.

We are building a custom PCB that would interface to a PXIe-4144 via the DB25F-DB25F low-leakage cable. Part of the application involves measuring very low currents, down to the limits of the hardware (<10 nA). Has anyone found a suitable PCB-mount 25-pin DSUB connector that will maintain the low-leakage/shielded characteristics of the cable necessary for these types of measurements? One of the parts I'm considering for this is Norcomp's 183-025-113R561. Does anyone, from NI or otherwise, have a suggestion for a better one? If this information can be shared, what does NI use internally on the board? (I could take our card out to check, but it's currently in use.)

Second question, the markings on the front-panel indicate that the cable shield is connected to chassis/earth ground, NOT the common LO terminal. If I have this correct, then I believe we should continue the shield to the PCB, but do NOT connect it to circuit ground there, as that would form a ground loop. Correct?