DIY Low EMF cable and connectors

[quarks]

The other trick is to make or buy an adapter board that brings out those child-safe recessed banana sockets to dangerous binding posts... Male plugs on one side, binding post on the other.  There are photos on EEVblog...

about adapters have a look

https://www.eevblog.com/forum/testgear/lking-for-gold-plated-tellurium-copper-banana-plugs/msg246371/#msg246371

https://www.eevblog.com/forum/testgear/lking-for-gold-plated-tellurium-copper-banana-plugs/msg327291/#msg327291

[cellularmitosis]

Maybe some of you guys find this trick useful:

I threw some normal crimp spade lugs in citric acid and after 2 weeks the relevant tin layer was etched away, yielding lowest cost pure copper (according to an exemplary manufacturer >99,9% electrolytic copper) spade lugs suitable for nanovolt-stuff.
Of course hydrochloric acid and oxygen peroxide would etch way faster, but i didnt have that stuff lying around.

Echo88, I've replicated your results using citric acid.  I just cut up a few pieces of some cheap crimp connectors as a test.

If I recall, this was 60 grams of water, and 12 grams of citric acid powder (i.e. 20% by weight).

I got lazy about removing them, so they were in there from 8/16 through 9/6, but they were mostly stripped after about 7 to 10 days.  You do need to agitate them to ensure all surfaces are exposed to the acid, daily agitation would be good.

(the large welding cable connector is just for color -- it came as bare copper from the factory and is used as a color reference only).

[cellularmitosis]

Next into the bath: some 1/4 inch spade connectors (Morris #11520) from amazon.com.  These are the right size to accommodate e.g. a pomona 3770.

Unfortunately, the crimp loop was a bit larger diameter (16-22AWG, or 1.25mm) than I'd hoped for (I'm hoping to be able to crimp to a single CAT5 strand directly).  I can probably use a hammer, punch, and anvil to do the job.

See you in a few weeks!

Hrmph, it appears these are already "no longer available" on amazon.

Here's a google shopping search.  They are available for $13 / 100 pack: https://www.google.com/shopping/product/14631136190881433596?rlz=1C5CHFA_enUS708US708&biw=998&bih=1454&q=Morris+11520+spade&oq=Morris+11520+spade&sa=X&ved=0ahUKEwiMn83lppTWAhVkl1QKHTuZCNEQ8wIIvwEwAA

[cellularmitosis]

It turns out these connectors follow a naming convention, e.g. SNB1.25-6, OT2-4, UT1.5-3, etc.

SNB, OT, UT describe the kind of connector (fork vs ring, both SNB and UT appear to be forks, OT appears to be a ring).

The first number describes the diameter of the crimp opening (e.g. 1.25 mm).

The second number describes the fork / ring opening clearance (-6 is 6.2 mm, about 1/4 inch).

You can find these by the thousand pack for under $10 on alibaba and toaboa, but at those prices they are likely to be brass rather than copper.

I find it hilarious that unplated copper connectors sell at a crazy premium compared to tin-plated-copper connectors.  In theory, they should be cheaper (one less manufacturing step).

The size we want (to accommodate a pomona 3770) is the 1/4 inch stud (-6), and for crimping to CAT5 / telephone wire, I think you'd ideally want a 1mm crimp opening, so you're looking at either SNB1-6 or UT1-6.  Domestically, the closest I've been able to find is SNB1.25-6 (this is the size of the Morris #11520 connectors).

I have seen UT1-6 available on taobao, but copper vs. brass is going to be luck of the draw there, and minimum quantity will likely be a bag of 1000.

[manganin]

copper vs. brass is going to be luck of the draw there

But there are also honest sellers:

Product Name: Non Insulated Fork Terminal
Model: SNB1.25-6S
Material: Metal
Color: Silver Tone

http://www.ebay.com/itm/311026743780

[SZA263]

Try also this:
http://www.tme.eu/gb/details/bm01132/non-insulated-terminals/bm-group/bm-01132/

[cellularmitosis]

Looks pretty good

https://world.taobao.com/item/546621717699.htm

0.01C, these arrived today, and they in fact are solid copper.

[cellularmitosis]

With these (and a few other items), I discovered a little hitch in the taobao game. 

Some sellers will ship the items directly to you.  When this happens, you just pay once, and you know the total cost up-front.

Other sellers ship the item to some sort of exporting warehouse.  So the amount you are paying upfront isn't the total cost -- a few days later you get hit with additional shipping charges, and they seem to be fairly expensive.  Unfortunately, I haven't been able to figure out a way to detect which situation will happen ahead of time.

In this case, the binding posts which look like a copper pomona 3770 clone were shipped directly to me, and the cost was known up-front.  But with these latest bare-copper binding posts (which don't have banana jacks), I got hit with an additional $13 shipping charge, which made the other posts a bit of a better deal.

The copper pomona 3770 clones ended up $36.77 for 8 posts (about $3.83 each plus $6.13 shipping), making them effectively $4.60 each.  https://world.taobao.com/item/40823926055.htm

These most recent bare-copper posts were $33.11 for 10 posts (about $2.85 each plus $4.63 shipping), but with an additional $13.13 warehouse shipping charge, making them effectively $4.62 each.  https://world.taobao.com/item/546621717699.htm

I don't want to paint a rosy picture of taobao -- so far I've paid for two items which haven't arrived in over a month (a few cheap banana plugs, but also a pair of LT1088 chips which were $18 each, ouch!) -- likely I just got scammed.

[0.01C]

With these (and a few other items), I discovered a little hitch in the taobao game. 

Some sellers will ship the items directly to you.  When this happens, you just pay once, and you know the total cost up-front.

Other sellers ship the item to some sort of exporting warehouse.  So the amount you are paying upfront isn't the total cost -- a few days later you get hit with additional shipping charges, and they seem to be fairly expensive.  Unfortunately, I haven't been able to figure out a way to detect which situation will happen ahead of time.

In this case, the binding posts which look like a copper pomona 3770 clone were shipped directly to me, and the cost was known up-front.  But with these latest bare-copper binding posts (which don't have banana jacks), I got hit with an additional $13 shipping charge, which made the other posts a bit of a better deal.

The copper pomona 3770 clones ended up $36.77 for 8 posts (about $3.83 each plus $6.13 shipping), making them effectively $4.60 each.  https://world.taobao.com/item/40823926055.htm

These most recent bare-copper posts were $33.11 for 10 posts (about $2.85 each plus $4.63 shipping), but with an additional $13.13 warehouse shipping charge, making them effectively $4.62 each.  https://world.taobao.com/item/546621717699.htm

I don't want to paint a rosy picture of taobao -- so far I've paid for two items which haven't arrived in over a month (a few cheap banana plugs, but also a pair of LT1088 chips which were $18 each, ouch!) -- likely I just got scammed.

for your reference

http://www.howtotao.com/buy-from-taobao/

https://world.taobao.com/item/45907925781.htm

http://www.taobaoguides.com/2016/07/06/the-best-way-to-ship-taobao-packages-from-china-to-usa-and-eu-countries

[cellularmitosis]

But remember Cellular has his null meter coming, which will lead to a much better Keithley 155 / Fluke 845a / or ?? + 7-decade KVD plus some 732's also. He's got the infection, which means he'll be very broke (but very accurate on DCV measures) in no time.

So it turns out MisterDiodes can predict the future...

Edit: derp, wrong picture

[branadic]

A question to those that are familiar with the topic. Is it really necessary that the outer insulation is PTFE too? I found cable for RTDs, with PTFE insulation on the single wires, 4 wires twisted and shielded, but with PFA for the outer insulation. So the question is, does it really make any difference if the outer insulation is non PTFE?

-branadic-

[The Soulman]

A question to those that are familiar with the topic. Is it really necessary that the outer insulation is PTFE too? I found cable for RTDs, with PTFE insulation on the single wires, 4 wires twisted and shielded, but with PFA for the outer insulation. So the question is, does it really make any difference if the outer insulation is non PTFE?

-branadic-

You do the math,it's on the outside of a braided shield. 

Here is something that could belong here, a quote from myself from the ltz1000 thread but didn't get a reply there so re-post it here:

Zener voltage reading by SMA connectors? SMA is not good for low thermals, as connector package is usually gold plated brass. You want only copper/copper connections with minimal amount of junctions (best is single twisted wire direct from zener to DMM binding posts) to get accurate microvolt-level measurements.

Ok not the best, but realistically what thermal-gradient do you expect there to be on that connector?
And how does the base metal of the connector matter as all connections are made on the same gold surface? When mating the two (assume crimped )connectors it would be copper-gold-gold-copper?

My gut tells me to be more concerned in areas where larger thermal-gradients do exist such as the connection to the kovar legs of the heated reference and to the oxidized binding posts of some warm ancient dmm. 

Not that I'm a expert, just asking to learn something. 

This low thermal connections are very important for accurate readings, but is starting to smell like some audio related behavior. 

[Edwin G. Pettis]

"This low thermal connections are very important for accurate readings, but is starting to smell like some audio related behavior."


The answer is that thermal EMFs, usually in microvolts, becomes important when attempting to accurately measure PPM, at lower accuracy, say ~25PPM or higher, the thermals becomes less important as they are a proportionately smaller than the measurement's signal value in general.  In practice it is a good idea to minimize thermal EMFs at all times as they represent an uncertainty factor in the measurement which can vary.  For example, a resistor with ±.0025% tolerance, a couple microvolts of thermals aren't very important, at ±.001%, now they've become significant as they could represent 2PPM, 3 PPM or more of error,  a null meter is the best instrument for detecting this.

PPM measuremenvts are almost always at microvolt or lower potentials so thermal EMFs become very important, they are not part of the audiophoole mythology. 

[MisterDiodes]

But remember Cellular has his null meter coming, which will lead to a much better Keithley 155 / Fluke 845a / or ?? + 7-decade KVD plus some 732's also. He's got the infection, which means he'll be very broke (but very accurate on DCV measures) in no time.

So it turns out MisterDiodes can predict the future...

Edit: derp, wrong picture

You thought I was kidding...   I hope your '845 works well, these nullmeters will open up a new low-ppm world of discovery for you.  Be patient with it at the low end though...What you think is the needle floating around at 1uV and 3uV range is probably a thermal somewhere in your setup....There will be plenty of that.

[BU508A]

Hi branadic,

[...]
does it really make any difference if the outer insulation is non PTFE?

In my opinion: no. There are two or three main reasons, why using PTFE:
- high insulation resistance
- heat resistance
- low dissipation factor

https://www.druflon.com/ptfeprop.html

For the outer insulation, PFA is very fine.

Good example: inner insulation: PTFE, outer insulaton: FEP (cable is RG 400 U)
http://www.ebay.com/itm/162634086255

Andreas

[BU508A]

I just stumbled over this page:
https://goo.gl/kfkKWN (link goes to google books and sorry the site is in german. But the table shown is interesting, chapter 0.2.3.9).

Does this table suggest, that a smart combination of copper, silver and gold will give you some kind of compensation effects?
All metals were paired to copper.

from the table:
thermal emf copper - gold:  +0,1µV
thermal emf copper - silver: -0,2µV

Thanks,

Andreas

[Echo88]

There is no smart combination of different metals to compensate for thermal emf, instead:

-just use the same batch of copper wire for your whole construction,
-crimping instead of soldering,
-where soldering cant be omitted use symmetrical components to cancel the resulting thermal emf between copper-solder
-keep the whole construction isothermal/couple all your measurement stuff thermal tightly.

The Keithley 147/148 and all other low thermal emf devices can be used as a good example how to do it.

Another way of thinking: To efficiently use different metals to cancel thermal emf youd have to knew the exact temperature distribution across your construction in extremely high resolution, which isnt possible i think. At least i dont know if its ever been done in this way and no thermal emf equipment manual ever suggested doing it your way.

[branadic]

Hi,

I wonder which diffusion barrier is used when gold is plated over copper. Nickel seems to be a bad idea, but does it really influence to measurement? It's only a few nm of thickness.
Is it just galvanic gold without diffusion barrier in between? What about copper-palladium-gold? Any ideas?

-branadic-

[MisterDiodes]

Of all the connections, clean copper to clean copper is overall best - but does have it's practical limits because it has to be kept clean. A gold plate, thick enough to not wear quickly is a good choice that does add a slight amount of noise, but very small.  "Low Thermal" brand has that construction - gold directly onto copper, and thick enough.  Gold over nickel can be used also without too much ill effect and you can get away with a thinner gold flash.  Never had a huge problem with nickle sub-plate in our setups, and we've never had a problem with direct gold over copper (when the gold is thicker, say 50u").  A very thin (cheap) gold flash directly on copper is not the best.

As for insulation, also consider EFTE wire - a cousin to Teflon but much stronger and higher dielectric strength, very slightly less volume resistivity (10^17 vs 10^18).  We use this stuff a lot and also the Conetic braid and flex conduit for precision measures that need good mag shielding and tough enough to hold up to daily use:

http://custommagneticshielding.magneticshield.com/viewitems/magnetic-shielding-for-wiring-applications/inter-8-weave-cable

Remember that noise can be coming in to your DUT on that sometimes dirty power mains ground, so pay attention to how you use / or don't use that - and use a clean quiet dedicated lab "ground" reference as required.   Keep your test leads twisted and short - and remember that the moment you connect those leads to your circuit then your instrument and all the ground loops in between now become part of the whole system measured noise.

[martinr33]

Other than that tricky noise point, plating layers don't matter too much. What does matter is that the final layer is low thermal relative to the base layer. The thermoelectric voltages are largely balanced across the layer: most of a copper-nickel junction voltage is canceled out by, say, a nickel-gold junction on the other side, leaving the copper-gold voltage as a residual.  Of course, that assumes that both sides of the layer are at the same temperature, which is OK if the layers are thin.

Where we have a bigger problem is across the terminals.  The jacks themselves can get quite warm, and any temperature differential here will expose thermoelectric voltages at the jacks.

I think that my meter jacks are gold plated, so I am facing gold no matter what..

[MisterDiodes]

I think that my meter jacks are gold plated, so I am facing gold no matter what..

None of that matters much as long as you've got the same combination of material balanced throughout your circuit.  Your binding posts are mounted on a good thermal conductor panel (not plastic) - that helps them stay at the same temp.  You're not using banana plugs, you're using a direct clamped wire or crimped spade / ring at the front of the panel to match the (usually) crimped ring connection on that binding post behind the panel (posts usually not soldered behind the panel for precision equipment).  Keeping those nuts snug helps keeps those joints in better balanced thermal contact with the panel surface.  That doesn't work as well with banana plug systems - especially the shrouded types.

Keep everything balanced and as close to the same temp as possible and you should be getting some repeatable measures.

Don't forget - if worse comes to worse you get rid of the binding posts and go for a direct connection...but that's a story for another day.

[Neomys Sapiens]

On the matter of the teflon 'cables' many of you seem to use:

most of the material seen here are technically not cables, but hookup wiring material.
I enclose the supplement which contains the overview of the sub-standards for the most common type of such wiring material, which is MIL-W-22759. A manufacturer's designation is for example Raychem Spec-55.
As you can see from the document, this comes in copper and copper alloy as base material and in several different platings (silver, nickel, tin).
And the standard as well as the commercial implementation does cover twisted pairs, shielded twisted pairs and shielded quads. The shielded versions have a outer isolation layer over the shield, which does not yet make them a cable in the formal sense.
There was also a series by Raychem called Spec-99, which used extra-fine stranding. Some of the nicest, most supple isolated wire that i have encountered. Unfortunately, only made to order.

Maybe this helps when when obtaining surplus PTFE/ETFE material.

[cellularmitosis]

Welp, the citric acid progress seemed to stall out, and after about a week I got impatient and picked up some muriatic acid at the hardware store.

A word of caution: this stuff is scary strong.  Only use this outdoors, wear gloves, don't breathe the fumes, and have a gallon of water and a box of baking soda within reach at all times.  Use a pipette or turkey baster to transport the acid -- don't try to pour the bottle!

Initially, the reaction was quick and vigorous -- the outer, shiny coating bubbled and was stripped in about 15 minutes.  This left a very dull, darker gray inner layer, which took about 40 hours to completely dissolve and reveal the copper.

I may need to lap these smooth, as the acid leaves a matte finish with high surface area (which will make slightly inferior thermal contact and will invite oxidation faster).

Off the top of my head I'd say the end result is about $0.34 per connector (~$14 for the connectors, ~$15 for the acid, ~$5 for baking soda).

Neutralizing the acid took a while.  By spooning in a teaspoon of baking soda, and then waiting for the foaming to die down, then repeating, it took me about 10 minutes (and probably 1/3 of the box of baking soda) to neutralize about 1/2 cup of acid.

[lukier]

I've ordered a bunch of these copper binding posts (https://world.taobao.com/item/40823926055.htm), but it seems I did something wrong on Taobao - should learn more languages.
The order was paid correctly (VISA), including the shipping (2 x 40 yuan for that) and sent by the seller but the parcel is still in Shenzen for the last two weeks (LX480192824CN).

Did anyone encounter the same problem?

[CalMachine]

Procured some parts to make a cable for older Keithley NanoVoltmeters (180,181,182).  There are other models that use the same type of military style connector, as well

Connector - AIT6E16-11PC-B30 (Gold Plated, Copper Alloy, Crimp connection)
Cable - Belden 5500FE (~1.5m)
Spades - Bare copper crimp spades

I could have gotten a bare copper version of the connector...  But, that isn't a normally produced item and would have been made-to-order with a MOQ, which I would have been no where near close to.   

Short values are compared with the K1488 Low Thermal shorting plug.   Only ~130 nV deviation from the plug leaves me very satisfied with this cable