Wire bonding, instead of spot-welding (like Tesla)

spinningmagnets

Moderator
Staff member
Joined
Dec 21, 2007
Messages
12,952
Location
Ft Riley, NE Kansas
I wondered how those "fuse wires" that connect the individual Panasonic cells to the bus bars are connected. I supposed that they could have been spot-welded in some custom way, but apparently the wire is pushed against the surface to be bonded, then the wire is vibrated at 60 kilo hertz (60,000 cycles per second?) for 100 milliseconds (1/10th of a second?). It is a "friction weld" process.

Its a "clean" process that requires no cleanup after, and also a "cold" process that cannot damage the cells' internal chemicals. Any heat generated is "just enough" and it is very localized on the skin of the outer surface. Is there any way for use to do this in a garage for ebikes?

https://chargedevs.com/features/a-closer-look-at-wire-bonding/

tesla-cells-wire-bonding.jpg


Greenpower using aluminum buses and aluminum fuse wire, two fuse-wires per cell.

https://www.linkedin.com/pulse/battery-pack-same-tesla-welding-process-jess-chang
 
spinningmagnets said:
I wondered how those "fuse wires" that connect the individual Panasonic cells to the bus bars are connected. I supposed that they could have been spot-welded in some custom way, but apparently the wire is pushed against the surface to be bonded, then the wire is vibrated at 60 kilo hertz (60,000 cycles per second?) for 100 milliseconds. It is a "friction weld" process.

Its a "clean" process that requires no cleanup after, and also a "cold" process that cannot damage the cell chemicals. Any heat generated is "just enough" and it is very localized on the skin of the outer surface. Is there any way for use to do this in a garage for ebikes?

https://chargedevs.com/features/a-closer-look-at-wire-bonding/

tesla-cells-wire-bonding.jpg

I actually think it would not be that difficult to design a spot friction welder. You could pull apart an old diesel peizoelectric fuel injector and use that as the oscillating mechanism. Then it comes to engineering material science to determine the shape of the mandrel to press onto the wire and fixing it to the vibrator. And a microcontroller of course.. And lots of trial and error
 
I've never heard of friction welding, but this sort of thing definitely could be spot welded with a precision spot welder. I could probably weld wires that small with a delicate touch and some time using my sunstone engineering welder which goes all the way down to 1 watt second of power. I weld 0.15mm nickel with 110 watt seconds.

What would be interesting to know though is what the resistance is like in those tiny wires are? And what material is it? The model S for example can draw up to about 500 HP from the battery, which is around 333 KW's. So that is up to around a 4C pulse on the batteries. So those little tiny wires would have to be able to withstand bursts of around 12-13A, and perhaps continuous currents of around 1.5C or ~5A (roughly the current for super charging).

I'd guess the little wires would have to be lower resistance than nickel, and maybe some kind of alloy with some copper in it? I guess the key is to pick a low resistance alloy but one that will burn out in extreme conditions like a short circuit situation which causes the metal to act like a fuse. Maybe its like zinc or nickel plated copper?
 
The bond wire is aluminum.

This process could not be matched by resistance welding and maintain the cell can temperatures achieved by ultrasonic welds.
 
I read the "Beyond the simple connections, Tesla describes a design method that also uses aluminum wire connections as fusible links that allow the expected current to pass, while breaking the connection in an overcurrent condition like a short circuit."

I assumed this was not talking about those cell in the picture but was describing some other patent or process. I have a hard time believing that those tiny aluminum wires could handle the kind of current we are talking about without huge voltage drop. Many companies file multiple patents but may not actually use every single patent (e.g. they may have filed patents for an aluminum fuse and a nickel plated copper process).

Maybe one of the members here that has actually taken apart a Tesla pack can chime in and test/inspect the fusable link material. And also report on the diameter of the material so we can calculate the resistance.
 
For an ebike battery pack, I'm sure the wire would have to be thicker. Because ebike batteries are so much smaller, perhaps there is more current per cell (compared to a Tesla)?

A couple of years ago, there were threads on re-purposing transformers from trash microwave ovens, custom re-winding them, and adding a few other cheap components to make a garage spot-welder. This last year the prices of "entry level" spot welders from China have come down.

Now, I am intrigued at the possibility of someone finding an ultrasonic device that is readily available, and repurposing it in some way to make it a wire bonder. This idea may not be feasible, but I thought I'd just throw it out there to see if someone with experience with ultrasonic consumer products might know of a viable candidate
 
amberwolf said:
:?:

https://www.google.com/search?q=spot+friction+welder&tbm=isch&tbo=u&source=univ&sa=X&ved=0ahUKEwiauKjJ6a3MAhVG2GMKHXz9DCwQsAQIIg&biw=955&bih=927

Friction stir welding is not really want you want for a battery, but Elon does use it at both SpaceX and Tesla for fuel tanks, frame parts... :))
 
spinningmagnets said:
For an ebike battery pack, I'm sure the wire would have to be thicker. Because ebike batteries are so much smaller, perhaps there is more current per cell (compared to a Tesla)?
Not necessarily, remember those wires are really emergency fuses to protect the pack if a cell goes closed circuit, and are sized to just exceed the max designed discharge load for the cell.
I believe there are Youtube vids of some folk shorting a module to test these wire fuses, and other posts on here where the size and current limit of the Tesla wires were estimated at ~15 amps.
Portable Ultrasonic wire welders are readily available for industrial use, and i gather are not uncommon in the electronics assembly industries.
 
As far as fuse-wire diameter is concerned, the characteristic that hasn't been identified yet is the time element. How fast do you want the wire to melt/separate during a short and max over-current condition? A fatter wire will take longer to blow, but it will also have a lower voltage drop during its intended daily use.

By this I mean that...a fatter wire that is aluminum or copper will conduct reasonably well (for an ebike pack) but still blow long before there is cell damage to the P-string. In a single-cell internal short, it is as if that cell is replaced with a solid wire. That cell is already toast, so we are trying to protect the rest of the cells in the P-string from cascading into a downward spiral of heat failure. The bad cell will continue to spew heat and smoke, but it may be possible to save the rest of the pack, or...at the very least avoid a large fire, created by 60 cells eventually shorting from heat.

The diameter of fuse-wire and the material used (Al or Cu), affect how we might find a way to use the method of untrasonic bonding for an ebike pack.
 
I was looking into this about 8 or so years ago... the ultrasonic heads are available on the secondary market. Back then I was looking at $8K to about $14K to set up an ultrasonic wire welding station.

This will get you started: https://www.sonobondultrasonics.com/welders-bonders/metal-welders

Look on eBay for Branson horns, transducers and such. I just looked again, and prices have come down significantly. There are transducer/drivers for $125 to $250... makes it interesting to design a machine around them.
 
That's brilliant, bigmoose! This is exactly why I started this thread. I had hoped that someone knew of some ultrasonic device that had an element which could be re-purposed into a wire-tip vibrator.

Once we identify the most common device that has this part buried in its guts, we can keep an eye out for used units on ebay. For instance, I was able to source a 1500W microwave oven transformer for only $10 after a month of patiently waiting and searching in my spare time.

Luke mentioned that the Tesla wires are aluminum, and I can only guess that it is because Al has a lower melting temp than a copper wire. I have seen "thin and narrow" steel and nickel foil used as an individual cell fuse, but that was only because it was easy to spot-weld.
 
t Luna Cycle has announced that all their fusion batteries are being wire bonded in the USA. http://www.electricbike.com/luna-cycle-announces-wire-bonded-ebike-battery-packs/
 
mr.electric said:
t Luna Cycle has announced that all their fusion batteries are being wire bonded in the USA. http://www.electricbike.com/luna-cycle-announces-wire-bonded-ebike-battery-packs/

Slowly catching up with their competition, I have a few reservations about the fusible link in a potting compound but good on them for catching up.
 
Friction stir welding wouldn't work with a battery. You need to have the welding bit plunge into the battery to weld and it wouldn't work.... Now ultrasonic welding is a different story. That would definitely work I think...
 
Don't count your chickens before they hatch...with the so called "Ultrasonic Welding", you do move something very rapidly against the other surface to create HEAT and weld the two pieces together. Usually you do produce a lot less heat than with something like a spot welder but you do have to move/vibrate the wire and/or the battery (60K hertz I read above) and you can't move one part against the other without transmitting some vibration to the battery. You are going to vibrate the battery some and the long term effect of that vibration is unknown at this point...it probably won't affect the battery.

I applaud Luna for trying it and hopefully my old conservative nature is unwarranted.
 
Rube said:
mr.electric said:
t Luna Cycle has announced that all their fusion batteries are being wire bonded in the USA. http://www.electricbike.com/luna-cycle-announces-wire-bonded-ebike-battery-packs/

Slowly catching up with their competition, I have a few reservations about the fusible link in a potting compound but good on them for catching up.


Just curious what competition they are "slowly catching up" with?

It's radically ahead of any ebike pack I've personally seen to date. Survives extreme vibe and salt fog and G-shock and has integrated safety functions I've never seen outside a Tesla pack.
 
with the so called "Ultrasonic Welding", you do move something very rapidly against the other surface to create HEAT and weld the two pieces together. Usually you do produce a lot less heat than with something like a spot welder but you do have to move/vibrate the wire and/or the battery (60K hertz I read above) and you can't move one part against the other without transmitting some vibration to the battery

60K hertz is 60 kilohertz, or...60,000 cycles per second. Its not so much the fact that there is vibration occurring, it is the distance that the two components are moving. The faster you vibrate the two parts, the less distance they need to move in order to achieve enough friction in order to create enough heat to weld the two parts together.

The amount of milli-seconds of vibration needed is very adjustable so that anyone who spends the money to buy the equipment that does this, can be assured that there is a very solid electrical connection, and the amount of heat generated is very localized and focused so that absolutely no heat reaches the inside of the 18650 cell.

As to vibration penetrating the cell, the mass of the connecting wire that is vibrated is very small, especially in comparison to the cell it is being attached to. Therefore, the cell experiences no measurable vibration. The mass of the cell has no time to react to the small wire vibrating a tiny distance about 60,000 cycles per second.

If any vibration ever did penetrate the cell, I am unsure if that would cause any damage, but...just to be on the safe side (because I don't know everything) Luna has thoroughly tested the prototype packs made this way. Heat, cold, extreme moisture, drop-testing, vibratory cycling, etc...It's my understanding that they followed the wire-bonding protocols used by Tesla (who has been very open about their patents), and it has worked out fairly good for them so far.
 
liveforphysics said:
Rube said:
mr.electric said:
t Luna Cycle has announced that all their fusion batteries are being wire bonded in the USA. http://www.electricbike.com/luna-cycle-announces-wire-bonded-ebike-battery-packs/

Slowly catching up with their competition, I have a few reservations about the fusible link in a potting compound but good on them for catching up.


Just curious what competition they are "slowly catching up" with?

It's radically ahead of any ebike pack I've personally seen to date. Survives extreme vibe and salt fog and G-shock and has integrated safety functions I've never seen outside a Tesla pack.

I assume it refers to Em3ev, as they have had cell level fused links for some time, but looks to be spot welded tabs rather than wire bonding. Apples, Oranges etc.
 
Sorry for reviving this thread after years, but I'm studying those battery packs and I'm interested in wire bonding connections. I understood how the connection is made but my question now is: How do you think those connections can be removed? What if I would disassemble a module and I would like to re-use the single cells?
 
the wire bond can be removed with a dremel with a blue or red grinding wheel in seconds.
 
Back
Top