15

I need to run coaxial cables into a cryostat, and because copper conducts too much heat, I decided to use semi-rigid stainless steel cables. As far as I can tell, both the shield and the inner conductor are made of stainless steel. I want to connect this cable to electronics in the cryogenic region, but the connection must be non-magnetic. My standard electronics solder does not wet the stainless steel at all.

Are there special fluxes or surface treatments so that I can solder to stainless steel?

Update: Thank you all for your comments and for the answer! I tried different soldering approaches and wrote my own answer below. I'll wait for a few more days, and then accept the answer with the highest vote.

As for RF or DC: There are several cables, some with DC, some with RF, some with both. The DC needs to be extremely stable (10 µV), but only needs to be precise on a 10 mV level. This stability requirement makes thermovoltages very relevant, which is why we want everything that sees the temperature gradient to be made of the same material. The RF has to be as clean as possible (I don't have numbers, here). I think the special problems with sending RF and DC voltages into a cryostat could best be separated into another question.

Martin J.H.
  • 651
  • 1
  • 5
  • 13
  • 1
    [Soldering Stainless Steels](http://www.bssa.org.uk/topics.php?article=150) - you need aggressive cleaners and fluxes – RedGrittyBrick Jun 26 '13 at 12:18
  • 2
    Not all stainless steels are non-magnetic and those that are non-magnetic to magnets don't always exhibit this "neutrality" when operated in AC mag fields upwards of 50kHz. – Andy aka Jun 26 '13 at 12:36
  • 4
    I would crimp it instead of soldering. But then again, I have never used SS for this so I could be completely wrong. –  Jun 26 '13 at 12:52
  • I thought about crimping, but crimping to the shield is tricky. – Martin J.H. Jun 26 '13 at 12:57
  • I think you are talking about Spot welding. http://en.wikipedia.org/wiki/Spot_welding Then this question should be migrated to mechanical engineering. – Standard Sandun Jun 26 '13 at 16:11
  • Do you need to think about the thermocouple you're creating? – Scott Seidman Jun 26 '13 at 22:45
  • @ScottSeidman - No. The DC bias the thermocouple produces will not affect the RF content. The system almost certainly has DC blockers in it too, anyways. – Connor Wolf Jun 26 '13 at 22:55
  • What surface prep have you done? You likely need to lightly sand the surface you are planning on soldering to. – Connor Wolf Jun 26 '13 at 22:56
  • @ConnorWolf -- how do we know this is an RF job, and not something like a sensor? – Scott Seidman Jun 26 '13 at 23:09
  • @ScottSeidman - Because if you're trying to connect to a sensor in a dewar, you would use cryowire. Stainless coax is a pain in the ass to work with, and I really doubt anyone would use it if they don't have to. – Connor Wolf Jun 26 '13 at 23:11
  • @ScottSeidman I do worry about thermovoltages, but I think the inner conductor is stainless steel as well. Basically everything that sees a temperature gradient is either non-conducting (G10) or stainless steel (cables, beam tube). – Martin J.H. Jun 27 '13 at 07:55
  • @ScottSeidman What is cryowire? – Martin J.H. Jun 27 '13 at 08:21
  • Don't know. I think it might be braided. – Scott Seidman Jun 27 '13 at 10:25
  • @MartinJ.H. Wires which have been through multiple cycles of cryogenic (ultra low temperature) treatment. Some audiophiles believe this improves the quality of sound from an audio system. Some audiophiles probably also believe in tidal effects of the moon on phase shift in an amplifier. – Anindo Ghosh Jun 27 '13 at 10:29
  • @AnindoGhosh - I think you meant to ask me, not Scott. It seems there are multiple things called "cryowire". There is the audiophile hocus-pockus you mentioned, and there is [wire for cryogenic uses](http://www.lakeshore.com/products/Cryogenic-Accessories/Wire/Pages/Overview.aspx), which is colloquially called "cryowire", at least in the lab I work at. Basically, it's wire that tries to maintain high electrical conductivity, while having low thermal conductivity. – Connor Wolf Jun 28 '13 at 02:02
  • 2
    Basically, when you're working with cryogenic systems, you can get significant and problematic thermal leakage along any mechanical connections between your cold-stage and any vacuum feedthroughs, which are typically at ambient temperature. Considering that you rarely have more then 1/2-1W of thermal capacity on the cold-head most cryopumps (note - I'm ignoring intermediate stages here), even the thermal conduction along a thin copper wire between the cold-head and ambient can significantly degrade your thermal performance. – Connor Wolf Jun 28 '13 at 02:05
  • @AnindoGhosh - Derp, I think **he** meant to ask me, not scott. And I can't type. – Connor Wolf Jun 28 '13 at 02:11

4 Answers4

10

Unfortunately, I don't have the tools that are needed for spot-welding or for proper crimping (Anindo's answer), and I fear that the dielectric of the coaxial cable won't survive the temperatures that are needed for brazing (Brian's answer). The solution that worked for me was using a specialized flux, as Scott and RedGrittyBrick suggested.

For future reference, these were my attempts at soldering to the surface of the stainless steel shield of the coax cable:

  1. PbSn solder with a flux core (standard electronics solder): The solder did not wet the surface at all, and consequently did not stick. Sanding the surface with very fine sandpaper did not help.
  2. Applying a more agressive F-SW 21 flux (3.1.1, brand "Lavar") that I found in our workshop A tiny bit of solder stuck, but not a lot. Surface was cleaned beforehand with acetone, then sanded with very fine sand paper.
  3. Wrapping an (uncoated) copper wire around the stainless steel cable, then covering it in standard electronics solder: This created a surprisingly good mechanical/electrical joint. In the picture it is shown on the left. You can see that the solder does not wet the stainless steel surface, so the connection is probably made by the copper wire sitting on the stainless steel surface. One experimenter claims to have used such a joint without problems for years, but I doubt it is very stable over time, as oxygen can sneak in and corrode the contact.
  4. Dissolve zinc in H-Cl to create ZnCl flux. I used 10-20ml of 37.5% hydrochloric acid and dropped a few tiny pieces of Zinc in it (less than 1 g), and waited 30 minutes until it stopped bubbling (well, actually there were still a few bubbles left). This worked as a flux, but I needed to put a lot of it onto the stainless steel surface until the solder started to stick nicely. This flux led to a lot of discoloration during the soldering (see middle section of the cable on the picture). Maybe it would have worked better had I cleaned the surface or sanded it down before. Note to self: When your local chemist recommends using a fume hood, then use a fume hood.
  5. Using special stainless-steel flux: I managed to get a small sample of BrazeTec Soldaflux Z, a strong, but corosive 3.1.1a flux. This worked great, even without any prior surface treatment! Only a tiny amount of flux is needed. There is a nice, smooth transition between solder and the stainless steel surface (see right side in the picture). The downside: The flux expires after 6 months and can only be ordered in 1 kg amounts (ca. 70 €).

5 cm stainless steel coax cable with different solder attempts

It is sometimes claimed that the acid in the aggressive fluxes, together with moisture, carbon dioxide, and time, will lead to corrosion of the lead of standard electronics solder. Maybe using lead-free solder, such as SN96.5 AG3 CU0.5, is a better idea.

Martin J.H.
  • 651
  • 1
  • 5
  • 13
  • You could use a flux remover to prevent or slow down the corrosion. I'm not sure if a standard off the shelf prep will work with this stuff – Scott Seidman Jun 28 '13 at 12:41
  • I believe the flux is water soluble, so it should wash off. But supposedly the chloride of the acid is dissolved in molten solder and stays in it when it solidifies. – Martin J.H. Jun 28 '13 at 14:18
3

Silver brazing, with borax flux, works well on stainless steel.

(It is sometimes called silver soldering but don't confuse it with those electrical solders that have a small percentage of silver. It is much closer to brazing in both strength and temperature)

The downside is the temperature you need - 650-700C, a dull to medium red heat. Clearly we're not talking about turning up the soldering station a notch, but this sort of temperature can be achieved with an inexpensive propane/butane torch.

However there may well be other reasons why this solution is not practical in your application!

3

Here are some approaches that were used for wiring up stainless steel - In both cases, the applications were high current industrial equipment, so applicability to a sensor may be debatable:

  1. Soft silver "collar" ring on SS316 shield, with copper spade connector crimped over the silver ring under high pressure: Apparently the silver forms a better gas-free colloidal bond at both junctions, than direct copper on steel. Similar, smaller silver collar ring on each inner conductor, again surrounded with the crimp portion of a brass spade terminal and high pressure crimp. The inner steel "wires" were about 8 mm diameter, and the shield was just over 22 mm diameter. Conventional crimping was done on the copper spade terminals.
  2. Spot welding of mechanically flattened (hammered) copper-plated steel cable wound in a spiral onto steel conduction bars (not sure whether the conduction bars were SS316 or SS304, but it was austenite steel for cryogenic use).

Both instances were done in-factory (different factories), and were experimental efforts. Both systems have been operational for a year, and several years, respectively without any repairs needed. In both cases, my involvement would rate as "curiosity", so I don't have any additional first-hand knowledge of the decision process.

Anindo Ghosh
  • 50,188
  • 8
  • 103
  • 200
2

I was in a similar situation once, and called the experts, who were extremely helpful. The fine folks at Kester got me in touch with the right technical person, who was nice enough to send me a small sample of the appropriate flux for my situation, no charge.

Scott Seidman
  • 29,274
  • 4
  • 44
  • 109