Ti Round shank current tests

[Swede]

WSM and I have been networking a bit on these, and while I was working with them the other day, I decided to do some real tests on them to see just how much current they can carry, and how much heat they'll produce as a result.

 

The experiment started with a number of round Ti shanks, from 6mm to 1/2" (or 12.2mm) most not yet welded onto an electrode. Some were empty, some were 100% filled with tin, the latest was the one I made that used mashed stranded copper wire + solder to create the fill.

 

The power supply used was an EMS 10-60, 10v 60A constant current supply. These types of supplies can handle a dead short without damage, which is good, because that is essentially what I was doing to these shanks.

 

http://www.5bears.com/perc/rounds/shankt03.jpg

 

The shanks were mounted in a vise, and a pair of #2 Cu cables supplied the juice. A type K thermocouple was attached to the middle of the shank using PTFE plumbing tape.

 

http://www.5bears.com/perc/rounds/shankt04.jpg

 

http://www.5bears.com/perc/rounds/shankt05.jpg

 

All of the temperatures are in degrees Celsius. I applied increments of current to the shanks, and allowed the temp to more or less stabilize. Some of the results were surprising. I list the voltage as well as the current, because it is an indicator of how conductive the shank is... a LOWER voltage at a given current between shanks means that its resistance is lower, and will heat less in use.

 

I'll point something out now that I found to be important... hugely important, in the end. All of the steel hardware alloy used is 316, and 316 is a terrible conductor of electricity. I found that how the Cu cable is applied to the shank is very important for heat. I'll explain this in a bit.

Edited November 2, 2012 by Swede

[Swede]

First up was a 1/2" shank 100% filled with molten tin. The cable mount on the end looked like this:

 

http://www.5bears.com/perc/rounds/shankt06.jpg

 

The numbers here are VOLTS, AMPS, and TEMPERATURE

 

0.18 -- 10 -- NA

0.35 -- 20 -- NA

0.52 -- 30 -- 48

0.70 -- 40 -- 73

0.91 -- 50 -- 100

 

This seemed really hot for this large shank. I noticed that the heat was very concentrated at the top, with a TON of the heat radiating up the Cu delivery cable. That was when I noticed how the cable was tightened... it had to flow through a large 316 SS nut. I swapped the mounting around to look like this:

 

http://www.5bears.com/perc/rounds/shankt07.jpg

 

Got the following results:

 

0.11 -- 10 -- NA

0.26 -- 20 -- NA

0.38 -- 30 -- NA

0.50 -- 40 -- 45

0.61 -- 50 -- 66

0.69 -- 60 -- 78

 

HUGE difference in voltage and heat. Conclusion: Be sure the Cu lug on the top is in direct contact with the high-conducting fill.

 

The next test was a small 6mm OD Cu and solder-filled shank.

 

0.53 -- 10 -- NA

0.72 -- 20 -- 66

0.82 -- 25 -- NA

0.91 -- 30 -- 109

1.07 -- 40 -- 120+ and rising

 

Now, on to the non-filled round Ti shanks... First up was a 1/2"

 

0.43 -- 10

0.69 -- 20

0.83 -- 30 -- 80

0.96 -- 40 -- 120+ and rising

1.08 -- 50 -- 180+ water sizzled on contact.

 

and finally, the smaller 6mm OD. I knew this wasn't going to carry too much current in its unfilled state:

 

0.16 -- 10

0.33 -- 20 -- 72

0.54 -- 30 -- 130+ and rising

[Swede]

http://www.5bears.com/perc/rounds/shankt08.jpg

 

Conclusion - obvously, the filled shanks performed MUCH better, whether the fill is mashed Cu or tin or a mix. The path from the power cable to the fill is critical. I had been using stainless hardware for corrosion resistance, but in the future, I am going to use solid brass threaded machine bolts or all-thread. These will be green after a run, but that's OK, the power lugs are green as well, and require cleaning afterwards. There's an added benefit of the brass stud soldering actively to the fill, which is something that is impossible with stainless steel... it will not soft solder, and relies on mechanical contact.

 

Pictured above is one of the Kynar compression fittings which was the goal for this from the outset. I need to re-check what sort of temperature Kynar can handle - I want to say it'll be fine at about 120C, but it's definitely going to soften, and I'd like to keep my stuff up there below 100C. I have (in the past) used heat sinks, small fans, and a combination of the two, and they do help a lot in keeping the temps down.

 

The #2 Cu welding cable didn't even get warm in the middle of the cable run. Any heat felt was from the junction between cable and shank.

[pdfbq]

I'm not sure if the figures (watts) needed to heat up your shank will be about the same in a real situation.

At 5.2V 50 amps might give different figures (but I am not sure here).

I understand you can not put 5.2V over this setup directly (that wil make the glow white probaby)

 

But what I read is that a non filled round Ti shank of 1/2" can take 0.83 -- 30 -- 80 or almost 25 watt without getting hotter than 80C^o

 

Means when your cell runs at 50 Amps at 5.2V this shank can handle a little less than 10% loss.

If one knows the resistance of the shank one can calculate the loss. You know the inner and outer diameter of the shank?

[WSM]

What I'm taking from this is that the figures show titanium to be conductive but still resistive to current flow (not unlike nichrome [nickel chromium alloy]).

 

If the stainless steel hardware is an issue, I suppose it's possible to run a long copper core (beyond the top of the tubular titanium lead) and clamp the power lead directly to it (rather than to the stainless hardware). The solder fill is still neccessary to get better physical contact with the interior of the titanium tube for good electrical conduction. All of this to run through modified PVDF compression fittings to prevent or control salt creep. Proper venting of cell vapors and/or salt spray will help prevent corrosion of the electrical connections and other susceptible components outside the cell.

 

I like having access to the data that Swede is collecting. It'll prove useful to those wanting to design their ideal cell setup.

 

Thanks Swede.

 

WSM

[Swede]

You're welcome. It's better to know now rather than later, watching a PVDF fitting loosen or melt... or char. US Plastics lists Kynar fittings as good to 275 f. or 135 C. I think < 100 C is a good goal. If a drop of water boils on your shank, then it's too hot.

 

pdfbq, the wattage will certainly be different in a real setup, but the resistance of the shank won't change (except with temperature, a bit), so the heat generated by the passage of current through the metal should hold true. In real life, the electrolyte will definitely act as a bit of a heat sink, drawing heat from the shank into the liquor. How much depends upon how deep the shank goes into the electrolyte. But the fact remains that probably the hottest part of the shank is going to be that which is enclosed in the Kynar fitting. The plastic is going to insulate, creating a hot spot at that location.

 

Heat sinks for rounds in a standard diameter (like 1/2", or 12mm, or 8mm) should be available, and it might be worth looking in to.

 

I took all my current filled shanks and replaced the hardware with brass. A 316 SS stud, which was cast in place, unscrewed with ease, leaving behind a perfect 1/4" x 20 threaded hole. Into that went a fluxed brass threaded rod, and the shank was heated up to the point of good flow and the brass threaded rod soldered right in. The good part about that is the stud is now permanently affixed in place.

 

After all the experimentation, I am torn between two options:

 

1) Cu wire installed somewhat tightly, and solder or lead + tin is added to 100% fill - topped with a stud; or...

2) An excess of Cu wire is installed, and a tamper or punch is used to really pound it down. Then, only the top portion is soldered in with a brass stud.

 

Option 2 creates millions of contact points inside the tube under pressure. Option 1 sounds great, but you have a monolithic insert, and an oxide or chemical layer could build up between the tube's inner wall, and the casting. I'm leaning towards option 2 myself. I like the notion of the Cu jammed in there really tightly. Copper (or bronze) wool would be ideal. And you'd save $$ on solder or similar, which has gone crazy price-wise in the last couple of years.

Edited November 6, 2012 by Swede

[Swede]

One other thing I forgot with regard to option 1 - Ti is like Al in that it expands aggressively with heat. If the tube out-expands the monolithic core, you could lose a LOT of the contact between core and tube, and of course the Ti oxidixes so easily, it could form an oxide skin that would prevent good continuity between core and tube inner wall.

[Swede]

One last comment on this over-studied topic.

 

Forget option 2 above. I tried it., It was grossly labor intensive and took a long time, and didn't work as well as a solid fill. When I passed current, the area of solid fill was MUCH cooler than the tamped Cu. It was very obvious.

 

In fact, these anodes are expensive enough to merit a solid fill. I'm not even going to mess with Cu wire. Here's the secret - use a BRASS stud, and outside the anode tube, flux and tin the stud in the area where it will be inside the tube. This is critical, or the stud will not solder itself to your melt. Fill the tube all the way to the top with your melt (lead, tin, plumber solder, any mix of these), and with a torch handy, and with heavy gloves, get ready.

 

Put a stainless nut on the stud to act as a stop. Slowly press the stud into the melt, realize it'll actually want to float, and press down until the nut hits the top of the tube. Excess melt WILL overflow; we want that, we want the tube to be 100% full. After it's cooled down, unscrew the stainless nut, clean up excess flux, clean the threads, and replace with a brass nut. It'll be ready to go.

 

You can use 2" long brass screws (saw the head off) or find some brass all-thread. Anything from 0.1875" to 6mm - 1/4" is fine, with 1/4" fitting inside most of the tubes we;d use.

Edited November 7, 2012 by Swede

[Swede]

Some final data points.

 

Shank A was the tedious copper-shred filled topped with solder and a brass stud. This is the one that felt much hotter in the shred portion, about 75% of the tube, while the solder portion was much cooler.

 

8mm OD, approximately 0.315". The thermocouple was over the solder portion, and has artificially low temps. The Cu shred end was blazing hot, well above 100C at 30 amps.

 

The numbers here are VOLTS, AMPS, and TEMPERATURE

 

0.59 -- 10 -- NA

0.69 -- 20 -- 37

0.80 -- 30 -- 58

0.88 -- 40 -- Experiment halted

 

Anode B, 8mm, solid fill, well-engineered stud section with pre-tinned brass stud.

The solid tin job was made as best as I could, including a small brass ferrule which centered the stud and was soldered in along with the stud. Note the voltage difference between anode shanks A and B.

 

0.11 -- 10 -- NA

0.24 -- 20 -- NA

0.31 -- 30 -- 64

0.41 -- 40 -- 83

0.46 -- 50 -- 103

 

With enough of such a shank submersed, you're going to get a heat sink action that should hopefully draw much of the heat away. I'd guess this 8mm OD shank should be good to 50 amps.

 

There IS another solution to excess heat... use a pure PTFE compression fitting instead of Kynar. PTFE can handle 3X the heat of Kynar, and retains strength to 400+. Unfortunately, such a fitting in this size is going to cost about $35 to $50, vs. $5.

[WSM]

Some final data points.

Shank A was the tedious copper-shred filled topped with solder and a brass stud. This is the one that felt much hotter in the shred portion, about 75% of the tube, while the solder portion was much cooler.

8mm OD, approximately 0.315". The thermocouple was over the solder portion, and has artificially low temps. The Cu shred end was blazing hot, well above 100C at 30 amps.

The numbers here are VOLTS, AMPS, and TEMPERATURE

0.59 -- 10 -- NA

0.69 -- 20 -- 37

0.80 -- 30 -- 58

0.88 -- 40 -- Experiment halted

Anode B, 8mm, solid fill, well-engineered stud section with pre-tinned brass stud.

The solid tin job was made as best as I could, including a small brass ferrule which centered the stud and was soldered in along with the stud. Note the voltage difference between anode shanks A and B.

0.11 -- 10 -- NA

0.24 -- 20 -- NA

0.31 -- 30 -- 64

0.41 -- 40 -- 83

0.46 -- 50 -- 103

With enough of such a shank submersed, you're going to get a heat sink action that should hopefully draw much of the heat away. I'd guess this 8mm OD shank should be good to 50 amps.

There IS another solution to excess heat... use a pure PTFE compression fitting instead of Kynar. PTFE can handle 3X the heat of Kynar, and retains strength to 400+. Unfortunately, such a fitting in this size is going to cost about $35 to $50, vs. $5.

 

Since the compacted copper strands were a wash, how about a single, solid bare copper wire, say #12 or #10 AWG and fill the open spaces around it with solder? I can see it being more conductive because of the copper and having better metallic contact with the titanium tube ID because of the solder fill (best of both worlds?). This notion would probably work better with tinned copper wire.

 

Another question; I wonder how much current the filled 1/2" OD titanium tube can handle?!! It would be nice to know if someone were wanting to build a high current beast of a cell, using tubular leads through compression fittings.

 

Thanks for sharing your test results so far, Swede.

 

WSM

Edited November 10, 2012 by WSM

[WSM]

With enough of such a shank submersed, you're going to get a heat sink action that should hopefully draw much of the heat away. I'd guess this 8mm OD shank should be good to 50 amps.

There IS another solution to excess heat... use a pure PTFE compression fitting instead of Kynar. PTFE can handle 3X the heat of Kynar, and retains strength to 400+. Unfortunately, such a fitting in this size is going to cost about $35 to $50, vs. $5.

 

Another option not mentioned is using multiple shanks for larger electrodes. This could work if you have lots of a smaller size titanium tubing and where mounting plus using multiple fittings in your setup is not an issue for you.

 

There are many ways this can work for us.

 

WSM

[FrankRizzo]

Since the compacted copper strands were a wash, how about a single, solid bare copper wire, say #12 or #10 AWG and fill the open spaces around it with solder? I can see it being more conductive because of the copper and having better metalic contact with the titanium tube ID because of the solder fill (best of both worlds?). This notion would probably work better with tinned copper wire.

 

Another question; I wonder how much current the filled 1/2" OD titanium tube can handle?!! It would be nice to know if someone were wanting to build a high current beast of a cell, using tubular leads through compression fittings.

 

Thanks for sharing your test results so far, Swede.

 

WSM

 

That sounds like a doable, and less epensive idea. Tin even some stranded wire that was a close match to the ID of the tube, fill the tube partially with molten tin (keeping the tube hot enough to maintain the molten state), then push the tinned wire into the tube to displace the molten tin to fill the empty space surrounding the copper wire fill.

Edited November 9, 2012 by FrankRizzo

[WSM]

Hi Frank,

 

I like the idea. You might have to use a rosebud (large torch head for oxy-acetylene outfits) to heat it all that much or maybe one of those thin metal flame spreaders that occassionally come with propane torch kits as an accessory. Be sure to flux the copper strands well and use plenty of heat to get the solder to fill the strands completely, and I see your idea working effectively.

 

I think making the copper strands an easy slip fit or even a sloppy fit will make for less solder wasted (by a possible overflow) and a better fill where no gaps or bubbles in the wire and solder will happen, i.e., more metal contact with the titanium tube ID. I suppose my thinking was less of replacing the tin fill and more of making it more conductive (thus lowering the heat from resistance).

 

WSM

Edited November 10, 2012 by WSM

[Swede]

I've only got time for a quickie post, then off to work.

 

There is a cool benefit to using a round shank. The following pic is self-explanatory:

 

http://www.5bears.com/perc/rotate.jpg

 

I tried this by simply taping some index cards to a pair of pencils, and rotating them in place. The best setup for variable spacing like this is to have them set up in the close state a bit farther apart than you'd normally have them. Then, when you rotate, the electrodes actually line up perfectly at a "mid point" of the rotation, while further rotation once again has them not perfectly parallel.

 

If anyone is curious, try it with an artificial setup. It really works.

 

The huge benefit of this is for folk with a fixed voltage supply. Obviously, closer spacing = more current. If current is too high, simply rotate them apart until you are happy. And obviously, electrodes to do this have to be welded as shown, with the shank as close to one edge as possible.

[graumann]

One other thing I forgot with regard to option 1 - Ti is like Al in that it expands aggressively with heat. If the tube out-expands the monolithic core, you could lose a LOT of the contact between core and tube, and of course the Ti oxidixes so easily, it could form an oxide skin that would prevent good continuity between core and tube inner wall.

 

A solution to this would be something we once did at uni, have the shank filled with a conductive liquid and the leads extending into the shank, I believe for one reason or another some sort of ferrofluid been use successfully.

[WSM]

A solution to this would be something we once did at uni, have the shank filled with a conductive liquid and the leads extending into the shank, I believe for one reason or another some sort of ferrofluid been use successfully.

 

What's the conductivity of it compared to copper?

 

If I were to consider a conductive liquid, I confess the first one that came to mind is mercury (Hg). I don't think I'd do that (adding Hg vapor to the toxic brew in my cells just doesn't feel right )!

 

WSM

[graumann]

actually our first though was gallium (hugely conductive), but for obvious reasons it was discounted. I can't find the notes with the exact fluid, but it was selected for very low thermal expansion and was a surprisingly good conductor. I'll let you know asap if I find the specs.

[Swede]

That's an interesting thought, graumann. So long as the fill is close to the conductance of lead or tin, and not too expensive, it might work nicely.

 

After all this smoke and mirrors, I like Frank's idea best as a compromise between labor, cost, and performance. I dissolved some Rosin flux in methanol, and used a paint brush to coat sections of stranded Cu wire. When you carefully strip welding cable, what comes out is a number of "sticks" of Cu wire, like 7 to 11 inside a section of #2 Cu welding cable. Each of these sticks of wire has integrity, so if your Ti tube is smallish, you can test to find out how many of these sticks can be inserted with ease.

 

Next, the shank is 1/2 filled with tin, and the selected number of Cu sticks is inserted all at once, with a propane torch keeping the melt hot. If you get overflow, great, otherwise, fill to the top, keep the heat up to absolutely ensure the Cu is tinned and part of the insert. The rosin will catch fire, and there will be a bit of a flare about 8" tall at the top of the tube. But it goes out in about 10 seconds.

 

Finally, install your pre-tinned brass or Cu stud into the melt, using a stainless washer and nut for positioning. The stainless wil NOT stick to the solder, and can be easily uninstalled. A gentle wire brush, and you're done.

 

An acid-based flux wouldn't flare, but it makes me nervous to have acid flux deep inside the shank.

Edited November 13, 2012 by Swede