Cabling for Fluke 732A/B, etc.

[Hawaii596]

I'm getting ready to buy some interface cable for an array of Fluke 732A's, a 732B, and a Datron 4910.  I am connecting them into a switching matrix unit to 'bump' them (do HI to HI comparison).  I've been looking at cabling and found what looks like it might be a good price cabling that should do a good job.  It is a Quad Shielded dual coax RG-6.  Basically two RG-6 coaxial cables attached together (i.e.: dual coax).  The center conductor is solid bare copper, AWG18.  Some that I have found is copper clad steel for the center conductor.  I'm going to continue to shop around for solid copper.  It has four layers of shielding and is considered audio cable (although I'll be using it for 10 VDC).  I also did find some by Dayton Audio of this type.  The price seems fairly in expensive (about $75 or so for a 250 ft roll).

Just wanted to get some feedback from folks who have done this.  Does this seem like a reasonable type cable, or are there some other recommendations?

[Bryan]

Twisted Shielded Pair was what I was taught on, solid core about 16-18 AWG. 
Fluke still offers some decent pomona tipped, I forget the part, starts with 5440 or 05440.  I have some others with Fluke name on them, numbers like PCA-0110 about 12", PCA-0116 about 18" and some others PCA-xxxx.  These appear to have multistrand conductor and have spade type terminals that appear to be a copper alloy and are shielded.  I have been very happy with them.

[michthai]

Test Leads Probes and Clips
5440A-7002 Low Thermal Copper EMF Plug-In Cables     
5440A-7003 Low Thermal Copper EMF Plug-In Cables

[Hawaii596]

I have some of them and am famiiliar, but unfortunately that will not work for this application.  I use those for high accuracy measurements currently.  I am setting up a standards lab with a bank of about ten 10 VDC references.  Preliminary work includes automatically running comparison measurements 24/7 and running them all through a switching matrix card.  So I need no connectors on the switch matrix card end of the cables (involves probably a couple of hundred feet of cabling).  But I need good, precision cabling (similar to what is in the 5440 cables).  The cable I mentioned in my original posting is dual coax, pretty much just like found in the 5440 cables (two RG series coaxes with terminals on the ends (banana or spade)).

I appreciate the input, and hope this wasn't a negative reply.  What I want to do though, is use cabling similar to that found in the 5440 series leads and create an installed harness to automatically run the series of measurements.  I will be doing hundreds per day automatically using a Labview program that I'm currently writing, and automatically storing them into an Excel file.  Then I can set up some ongoing statistical plots of which cells drift more or less than others, eventually do a "hot cal" using a shipped standard from Fluke, and establishing absolute uncertainties for each cell and for all of them in combination (hopefully sub ppm with this mountain of data and RSS combination of the entire array).

[scottbp]

Twisted Shielded Pair was what I was taught on, solid core about 16-18 AWG. 
Fluke still offers some decent pomona tipped, I forget the part, starts with 5440 or 05440.  I have some others with Fluke name on them, numbers like PCA-0110 about 12", PCA-0116 about 18" and some others PCA-xxxx.  These appear to have multistrand conductor and have spade type terminals that appear to be a copper alloy and are shielded.  I have been very happy with them.

I cannot find any reference to the aforementioned PCA series cables on either the Fluke or the Pomona website.

I've been using the Pomona model 1756 patch cords, which have gold plated copper spade lugs, but their fatal flaw is that the spade lugs are so thick you can barely stack two with the binding posts of a 3458A or 5700A screwed all the way out to the stop.

The 5440B cables are good, but their fatal flaw is that the "bananas" are too long for most binding posts and thus fit very loosely, and they have a tendency to bounce out due to the spring loaded banana guard, so I don't use them unless I have to.

The Fluke 8508A-LEAD set is excellent for low voltage and resistance measurements, but their fatal flaw is that the leads extending past the plastic block on the ends is too short, and they are fragile on top of that, meaning making any repairs once the spade terminals break off makes the leads that much shorter. To add insult to injury, Fluke charges close to $1,000 for the 8508A-LEAD kit. For this reason I call them "Gucci speaker cables".

What I'm looking for is an ideal substitute for the Fluke 8508A-LEAD set that doesn't cost so darn much!

[Hawaii596]

I also posted my query on Linkedin in the Metrologically Speaking group, and I got a response from a person (he may be on here too).  The individual worked at Fluke's Primary Lab for quite a few years.  He said they used to use Belden type 8762 cabling.  It is twisted pair AWG20, tinned copper stranded type.  It has shield, and is not too expensive.  I just bought a spool a couple of hours ago. 

As for the Fluke 5440 series, I know exactly what you mean about the cable bouncing out.  I've lost it, and will need to look it up again, but there is a part number for replacement copper banana springy piece (the actual copper banana that contacts the female banana.  I have a bunch of aging sets of 5440 series that were all getting really bad (way too expensive to replace the cables).  So I bought a bag of the replacement banana contacts, I believe for around $3 per contact or so (not sure).  It really rejuvenated the cables. 

Back to topic - - I did put in a requisition for the 8762.  I think that should be good for my installed harness.  When I finally get back to making real measurements, I will likely use my old standby Fluke 5440 cables.  Another thought I may explore is making some home made cables with this spool of Belden 8762.  If I can find some copper (or maybe gold plated) double banana Pomona's or something like that, I may experiment with them, and maybe attach a guard terminal to the shield.  I'll have to wait and see how I like them.

We own a couple of 8508A's, but I am not familiar with the lead se that goes with them.

Thanks for the inputs.

[metrologygeek]

I've got an old (obsolete) Keithley 1483 low thermal EMF connection kit which includes gold plated lugs of various sizes and configurations. I also use the Belden 8762 20 ga wire, and it works very well. I have a number of four-conductor assemblies, also. All of the cables used in primary temperature on the physical side where I work are constructed in-house, and I believe that over on the other side of the house in primary electrical they use exactly the same configuration of cables. Luckily (for me) I have enough stock left in the Keithley kit to last a few more years 'cause Keithley discontinued it quite a while ago. I would not recommend using bananas for your high accuracy low-thermal DC measurement setup.

[Hawaii596]

No, I'm not using bananas for the automated 'bump' measurements.  You brought up a good point though.  I remember a "LOOOOONNNNGGG" time ago, some lab where I worked used to have a little Keithley kit with various copper stuff (including copper spade lugs).  Any thoughts on where to get something like that.

Back to original topic... I will be using switch matrix card (so screwed in at that end).  But at the Fluke 732A end, I was going to wrap around the terminals.  I'm a little skiddish to do it that way.  I don't think there is a proper spade lug to use for that.  The only way I would consider using bananas is if I could find something all copper.  Any thoughts about best connection method from the 8762 cable to 732A terminals?

[Bryan]

Perhaps gold plated spade lugs & silver solder?
I'm looking at it here http://www.alliedelec.com/catalog/pf.aspx?FN=487.pdf and sort of get the impression it may be a solid core?  I notice some of the others specify stranded while the 8762 does not.
I'd be curious to know what you get.

Scott I don't know where/when the Fluke PCA labeled cables originated but they do have Fluke & the part # on some heatshrink on them.  They are much like the Belden 8762 mentioned in this thread with lugs attached, stranded conductor.  I'm guessing they date back to the 80s.

[beadwork]

The lead kit for the 8508A works really great (8508A-LEAD).  I would highly recommend you purchase one.

The kit comes with:
4 2 way shielded PTFE cable, with low thermal EMF gold plated 4mm spades.
1 2 way High Current cable, with 4mm spades.
2 Locking adaptor (1 Red & 1 Black) 4mm terminal post to satety socket, gold plated.
1 General purpose CAT II DMM lead and probe kit.
2 4 way Shorting link pcb.
1 Hexagon key.

[BamaKid]

One of our metrologists provided this response:

"For the type of work being addressed, I am a firm believer in solid copper clamped under a binding post. I avoid banana jacks/plugs for precision DC work.

I make my test cables from vintage Guildline type SCW low thermal interconnecting cable. It's solid copper, 22gage, twisted pair with a braided copper shield. Our spool, which has a 1978 date on it, originally held 300 feet of cable. If it outlasts me, it won't be by much. I have not found a replacement.

Our Dataproof scanner, which I use for automated tests of our 732Bs, was purchased with similar solid copper leads."

[Hawaii596]

It's kind of an exciting time for my DC references.  Just got power to the copper mesh screen room (maybe overkill, but I got it for free - and it looks cool).  Just got my LabVIEW FULL real software, almost done writing the program, got the new GPIB card, just ordered the cable, put all those Chinese batteries into the DC references and getting ready to start at least a year of drift plot numbers. 

One of the final burrs in my britches (so to speak) is my Datron 4910.  I've schematically reverse engineered the UPS section, but you can't get the batteries any more.  It has four cells and an auxiliary section.  I won't go into all the sordid details, but it looks basically like I will need to put some of them outboard.  I have five cells up and running.  This will add four more.

I also have my Fluke 792A plugged in (need to send it out in a while).  I also have an old Fluke 540B that I was thinking of using to start doing ILC's with.  Use the 792A as my standard and the 540B as a comparison unit to justify my numbers.  The 540B has a dead mercury battery, and I will need to build a new NICAD battery pack by hand.  I also have a Datron 4920 (I think) that has some problems.  I may repair that as well.  The 792A is obviously the standard.  But which, if either of the 540B and 4920 should I resurrect?  Or should I resurrect them both for comparison standards?

[Bryan]

Down at the bottome of the page is some wire

http://www.guildline.com/Datasheet/Guildline6664CDatasheet.pdf

[scottbp]

Nice... The Guildline 6664C is the exact same thing as a Dataproof 164B OPT 2 scanner, in fact I bet Dataproof made it for Guildline.

We have a Dataproof 320A, which has 32 sets of copper binding posts on the back. Our standard cells (a Datron 4910 "4-banger") are connected via unshielded, un-tinned, un-twisted Belden 6795 22AWG telephone wire. We take our voltage readings with Dataproof Voltref, and track the drift on an Excel spreadsheet.

I dug around in our warehouse and found our old Fluke 7105A DC measurement system and lo and behold, it had the Fluke PCA type cables in the bottom drawer.

The thing is, the internal conductors are tinned copper, and they have gold plated spade lugs which are crimped and soldered on. I was taught early on that dissimilar metals (copper + tin/lead) create a thermocouple junction which adds error to DC measurements.  

If the tinned conductors and soldered spades only add negligeable error, and I thought it would work better than bare copper telephone wire, I would go ahead and whip up some more cables, because I already have a couple rolls of Belden shielded twisted pair tinned stranded cable similar to the ones the PCA cables are made of, one is 8760 which is 18 AWG, and the other is 8737 which is 22 AWG.  

[Hawaii596]

Got the DC references up and running.  They are still settling in (only been taking data for about a week; and cells have been on line for about a month or so).  Using the Belden cable as suggested.  But I got a shift in one of the cells last night of about 12 uV.  I am changing up cabling configuration etc., and fine tuning (still characterizing oven behavior, etc.).

I wanted to get some thoughts as to optimum GUARD and GROUND configuration.  I have twisted pair shielded cable for the voltage.  So that is the HI and LO.  I am using the third un-insulated lead as GUARD connection at the standard end.  Then at the other end of the cables, I have all of the GUARD's tied together, and connected to the GUARD terminal of the HP 3458A, which I am using to do BUMP readings.  Originally I had all of the GUARD shorted to GROUND at each of the DC references.  But I disconnected it last night, so the GROUND terminal on the front of each reference was floating.  It was after I did that, that I got a -12 uV shift in the relative output of one of the references (but only one of them).  Maybe this tells me I have a poor quality ground on that one that is offset by that much.  I have nine references.  I am using #1 as the reference unit.  Then I am doing "BUMP" readings of 2 through 9 against 1 (i.e.: 1vs2, 1vs3, 1vs4, etc...1vs9).  My 1vs4 reading shifted.  They are all powered through a single power strip.  I am running readings through a LabVIEW program I wrote.

So I'm trying to figure out if my Cell 4 just did a shift, and potentially has a problem, or if something I just changed in my cabling may have caused the shift (i.e.:  disconnected GND from GUARD on cells).

I don't want to think one of my cells just shifted, which means I have to fix it - which also means waiting until it shifts again).  I tried replicating my old cable configuration and manually running the checks, but I got the same shifted readings.

One other possibility I need to entertain is that perhaps it is the switch card I am using.  I think I'll check that next.

Any thoughts from experienced people appreciated.