Low thermal EMF solder

What are you soldering to what? If it's a part lead to a PC board, or a wire splice, it doesn't matter. As long as the connection blob is isothermal, any thermoelectrics depend on the incoming/outgoing alloys of the leads, and not the solder.

According to this article: the magic recipe is 85% lead and 15% tin. That's approximately the mixture of solder used to repair copper/brass automotive radiators (before RoHS). You might ask at a radiator shop or an antique car restoration buff.

I found one source: There are probably more.

If you have an electric solder pot, propane furnace, or home foundry, you could make your own solder out of pure tin bars and some old lead fishing weights. Hint: Don't breathe the fumes.

That's low EMF against copper.

The effectiveness depends on the physical characteristics of whatever he's soldering. If there's no temperature difference across the copper to solder boundary, there should be no thermocouple effect.

Googling, I find that the recommended solder mix is 70% cadmium and

When one of the engineering techs tested high positive for cadmium poisoning, the company had everyone involved get a blood test for heavy metal poisoning as well as cadmium. My cadmium and lead levels were rather high but not excessive. At the time, I thought I had caught a cold or flu that would not go away. It turned out that it had something to do with my handling and testing NiCd batteries and possibly the cadmium loaded repair solder. Not much could be done except to take zinc supplements and avoid further exposure. It took about 6 months to shake the cold. In the meantime, I discovered that crimping the bolometer leads with a brass ferrule produced less temperature related issues than with the cadmium solder. Anyway, I strongly recommend NOT using the Cd70/Sn30 solder mix.

Also depends on the allowed error level, because nothing is truly isothermal, especially if there are any active components.

Joe Gwinn

This link will work equally well:

Note: MT-Newswatcher wraps all long URLs, ignores the < > brackets, and requires users to hit the command key in order to click on its own wrapped URL's. Please complain to the author (Simon Fraser) as this has been a chronic problem for quite a while.

The active components have internal aluminum or gold wirebonds that have huge thermoelectrics against silicon and against the leadframe material.

If you solder an IC lead - maybe alloy 42, maybe a copper alloy if you're lucky

- to copper traces on a PCB, the inherent thermoelectrics of that will blow away any effects of a little solder. The gradient across the solder will be microkelvins, picovolts.

What sort of errors are you budgeting? What's the signal represent?

The best thing to do to keep thermal noise and offsets down is to put a cover over the sensitive stuff and a fat copper plane below it, to keep it isothermal, and keep local power dissipation gradients low. The effects of solder will be way, way below other thermal offsets.

Google links are huge for some reason. There's a shorter "real" link under there somewhere, or use TinyURL.

--

I'm repairing a couple of old Holt 323A Audio Voltage standards. The thermocouple circuitry requires a low EMF solder (specified in the manual, and contained cadmium). The insstrument was sold with a hank of the right solder inside, but has long since been used or lost in someone's toolbox.

I've tried using several formulae of solder, but nothing I've tried will allow the themocouple InAmp to behave. I used to repair & calibrate this same model back in the late 60s and 70s, and remember that the proper solder had to be used or the thing drifted all over the place. Thus, I'm back to looking for the "right stuff".

Thanks for those who've suggested alternatives. I'll keep trying untill something works,

Cheers, Dave M

If you can reduce the 'fit' of the conductor to the hole it is being soldered into, the cross sectional thickness of the solder is reduced, and thereby its factoring in this equation. No?

IOW, is there any way to insure that the 'gap' between the conductor and walls of the hole it is being soldered into is minimal?

These guys in India list it (70:30 Cd:Sn), bottom of the page.

Would be interested to know what/whether they quote for a small amount. I know others have had success in getting reasonable quantities of special alloy wires from India.

Sn and Cd are not particularly expensive metals.

Best regards, Spehro Pefhany

Might have trouble actually importing it (into the US), however.

On Sat, 19 Jan 2013 14:59:00 -0800, Jeff Liebermann wrote: (...)

Here's a possible source for Leeds and Northrup "thermal free solder" PN 107-1-0-1. According on one reference, it's 70% Cd, 30% Sn.

Some interesting reading on thermal EMF: Looking at Table 4.10, it seems that pure silver and pure gold have the same thermoelectric potential as cadmium/tin solder. Too bad silver melts at about 900C which would require welding instead of soldering.

Hi Dave, did you ask on volt-nuts yet?

Wow, does that organization still exist ?

Maybe I am thinking of another.

Jamie

I cannot believe you claim to be in (or a part of) the electronics industry, but are unaware of what the most common rags are.

I cannot believe you claim to be in (or part of) the electronics industry, AlwaysWrong.

I don't think Simon still maintains MT-Newswatcher, as the last update to the webpage was in 2006.

I've been looking for a replacement, but haven't really found one whose user interface I like.

Joe Gwinn

I'm not the original poster, so I don't know his application.

But I do know that low-EMF solder is widely used for low-level measurements. For instance, thallium copper used to be used in place of pure copper, if memory serves. Doesn't seem to be used anymore, from google searching.

Don't know why they are so large. Probably contains my life history.

I tried to use the link in my original post, and it didn't work either.

However, the link in the copy I automatically email to myself did have a working link.

Joe Gwinn

I do far more in it than you do.

And you'll never catch up, RetardRW.