Tabless design cylindrical cells tests

I have access to this awfully expensive Micro-TIG system with the hand operated touch retract TIG torch from Amada Weld Tech. At the time, it was one of only two such systems on the market. The other is from Sunstone Welders. Today, such a system has been surpassed by Lasers in virtually every respect. The biggest disadvantage of the Micro-TIG technology is that it is practically unable to weld aluminium.

Micro Arc Welding Torch - MacGregor | Amada Miyachi

It has a lot of power so it can virtually weld up to 1mm of Cu, but you are practically limited by can thickness of the 18650/21700 format, so you can reliably weld a Cu strips with the thickness of 0.5mm. No additional chemistry is needed, the system automatically brings a protective argon atmosphere thru the torch to the weld spot.
 
Last edited:
TIG welding usually requires a shielding gas. Does the Micro-TIG system not use it?
 
Yes, a bottle of argon is connected to the device and the device automatically dispenses the argon. The way it works is that at the end of the torch there is an approx 8mm metal tube with an approx 2mm thick TIG electrode with a sharp tip end in the middle. When the torch is pressed against the material, the system automatically applied argon thru the tube into the weld spot, the weld takes place with automatic retraction of the electrode. It all takes several seconds and after that you can remove the torch and immediately make another weld.

One of the limitations of this particular system is that the electrode inside the tube is not exactly centered, but can move sideways, which leads to not being able to place the weld exactly on the desired spot.
 
Pajda said:
In your setup will be the biggest problem with the fixture, where at 40A the transient resistance here cause additional heat generation on both terminals which will rise measured temperature of the cell body. For such currents it's better to weld or carefully solder some kind of strip or wire (I am using welded CU strips, but some thicker Ni will be also fine)
Click to expand...
IMHO what type of holder and the gauge of the connected can have a big effect on the final temp, i.e., the total mass of the contacts. Standard 0.2mm copper strip isn‘t going to pull a lot of heat away on its own but it could sink more than some compression holders, especially if large gauge wire is soldered to the strip. Sounds like you’ve got some great heat sinking from the strips.

The type of welds and how good they are will change things too IMO.

With the Ampace web site page for the JP40 saying 45A max (no temp limiting) I’m thinking that about 75°C at 45A is what Ampace was getting for their testing. Unless their cutoff is higher but the full datasheet lists 75°C as the limit. Confusingly though the A0 edition of the full datasheet seems to have 40A as the true continuous rating…but perhaps this is an older datasheet?

I’ve been getting about 78°C max (calibrated type-k thermocouple, cell wrap off) at 45A down to 2.5V from 24.0°C, +/-2°C still air starting temp for all of the differently wrapped JP40’s.
 
Last edited:
Pajda said:
It has a lot of power so it can virtually weld up to 1mm of Cu, but you are practically limited by can thickness of the 18650/21700 format, so you can reliably weld a Cu strips with the thickness of 0.5mm. No additional chemistry is needed, the system automatically brings a protective argon atmosphere thru the torch to the weld spot.
Click to expand...
Ahh…so not the spot-welding some might use…yea, that can weld on some serious copper to sink more heat. Wow, very jealous!
 
Last edited:
Btw. using Cu strips (in my setup standard it is 11 mm wide with a thickness of 0.5 mm so ca 5.5 mm^2 cross-section) creates another problem that the voltage sensing in Kelvin (4-Wire) connection does not directly touch the cell body, but the welded Cu strip where the actual current flow causes small but visible voltage drop. I tried to solve this issue by cutting a hole in the Cu strip so the voltage sense needle of a standard cell fixture can touch directly cell body, but the above mentioned inaccuracy of placing the welds with Micro-TIG makes this almost impossible.

So in my setup with massive Cu strips (let's say an ideal current path) I can achieve up to 10°C lower cell body temperatures at high currents around 40A. On the contrary, there is a slightly larger voltage drop in my discharge charts and thus a lower discharge energy in table results.

@CamLight has a also very good point about the influence of the actual temperature sensor used for the measurement. Professional instruments use miniature T-type thermocouples (they at least promising slightly better resolution in the typical battery testing range vs. K-type). Where the actual sensor is a metal ball with a diameter of 0.2mm and therefore combining very low heat capacity with excellent thermal conductivity. This allows the sensor to react very quickly to temperature changes. Therefore, the housing is usually more important than the principle of the sensor. For example if you use a sensor in a plastic housing like TO-92, it is quite likely that this will not follow the temperature changes caused by high discharge rates.
 
This is what normally what I do
 

Attachments

  • IMG_8120.jpeg
    IMG_8120.jpeg
    2.7 MB · Views: 40
  • IMG_8117.jpeg
    IMG_8117.jpeg
    2.6 MB · Views: 39
  • IMG_6855.jpeg
    IMG_6855.jpeg
    1.9 MB · Views: 35
  • IMG_6858.jpeg
    IMG_6858.jpeg
    846.4 KB · Views: 38
Pajda said:
@CamLight has a also very good point about the influence of the actual temperature sensor used for the measurement. Professional instruments use miniature T-type thermocouples (they at least promising slightly better resolution in the typical battery testing range vs. K-type). Where the actual sensor is a metal ball with a diameter of 0.2mm and therefore combining very low heat capacity with excellent thermal conductivity. This allows the sensor to react very quickly to temperature changes. Therefore, the housing is usually more important than the principle of the sensor. For example if you use a sensor in a plastic housing like TO-92, it is quite likely that this will not follow the temperature changes caused by high discharge rates.
Click to expand...

I agree. The tiny bead-welded and “butt joint” welded thermocouple tip types are necessary for good thermal coupling and short response time. Even the “flat” thermistors, with a relatively fast (for a thermistor) response time are pretty bad compared to a decent thermocouple. And you are so right…the larger case types, especially the TO-92 cased sensors, are essentially useless. They have only a tiny contact patch against round cells and verrrrrry slow response times.

As always, @Pajda can always be counted on to have great equipment. 🙂 I haven’t found any practical difference in the performance, accuracies, and response times of type-t thermocouples (versus type-k) but @Pajda is right about the specs. Both types are used in many different industrial, commercial, and research applications.

Type-t are designed for cryogenic (verrrry cold) applications and have a typical accuracy of about +/-1°C. Type-k have a wider and higher temperature range but, typically, a +/-2°C accuracy.

I use +/-1°C calibrated Fluke type-k thermocouples so I don’t worry about the accuracy (mine were about 0.6°C off at 0°C and 100°C) but it’s something anyone should consider (along with thermal coupling/resistance, response time, durability, etc.) when choosing how to measure cell temperature.
 
Last edited:
CamLight said:
I agree. The tiny bead-welded and “butt joint” welded thermocouple tip types are necessary for good thermal coupling and short response time. Even the “flat” thermistors, with a relatively fast (for a thermistor) response time are pretty bad compared to a decent thermocouple. And you are so right…the larger case types, especially the TO-92 cased sensors, are essentially useless. They have only a tiny contact patch against round cells and verrrrrry slow response times.

As always, @Pajda can always be counted on to have great equipment. 🙂 I haven’t found any practical difference in the performance, accuracies, and response times of type-t thermocouples (versus type-k) but @Pajda is right about the specs. Both types are used in many different industrial, commercial, and academic research applications.

Type-t are designed for cryogenic (verrrry cold) applications and have a typical accuracy of about +/-1°C. Type-k have a wider and higher temperature range but, typically, a +/-2°C accuracy.

I use +/-1°C calibrated Fluke type-k thermocouples so I don’t worry about the accuracy (mine were about 0.6°C off at 0°C and 100°C) but it’s something anyone should consider (along with thermal coupling/resistance, response time, durability, etc.) when choosing how to measure cell temperature.
Click to expand...
Can you give me a link from the sensors you recommend?
 
ZEUS-FL said:
This is what normally what I do
Click to expand...
Along with @Pajda’s description of his low thermal resistance setup this is a great opportunity for all to see how three different setups can all get great comparative data but slightly different results. The details matter. 🙂

And even if any setup has an offset (due to voltage drop) or slower temp reading response time, as long as the testing methodology is sound and…hugely important…incredibly consistent then the results of that tester can be used to compare the performance for those cells. We just have to be careful when comparing results for different testers. Again, the details matter.

My cell connections use a compression-type jig with a 12mm long slightly rounded tip 6mm diameter copper rod pressed into a depression in a 1mm thick copper plate and soldered (with the main wire leads soldered to the plate) for the positive and a 1mm thick copper plate for the negative (cell wraps always removed). I don’t remember what I measured for the resistance of the setup but it was micro-ohms and low enough that I was plenty happy with the setup.
 
Last edited:
ZEUS-FL said:
Can you give me a link from the sensors you recommend?
Click to expand...
The calibrated thermocouples? Sorry, I don’t remember where I bought them, it was years ago (I stocked up when I actually had some spare cash 😂). One of the big lab equipment companies though.
 
Last edited:
Check this other test to a single EVE with the new setup.

- YouTube

Enjoy the videos and music you love, upload original content, and share it all with friends, family, and the world on YouTube.
www.youtube.com www.youtube.com
 
Pajda said:
It has a lot of power so it can virtually weld up to 1mm of Cu, but you are practically limited by can thickness of the 18650/21700 format, so you can reliably weld a Cu strips with the thickness of 0.5mm. No additional chemistry is needed, the system automatically brings a protective argon atmosphere thru the torch to the weld spot.
Click to expand...
What does "no additional chemistry is needed" mean?
 
Pajda said:
I have access to this awfully expensive Micro-TIG system with the hand operated touch retract TIG torch from Amada Weld Tech. At the time, it was one of only two such systems on the market. The other is from Sunstone Welders. Today, such a system has been surpassed by Lasers in virtually every respect. The biggest disadvantage of the Micro-TIG technology is that it is practically unable to weld aluminium.

Micro Arc Welding Torch - MacGregor | Amada Miyachi

It has a lot of power so it can virtually weld up to 1mm of Cu, but you are practically limited by can thickness of the 18650/21700 format, so you can reliably weld a Cu strips with the thickness of 0.5mm. No additional chemistry is needed, the system automatically brings a protective argon atmosphere thru the torch to the weld spot.
Click to expand...
Pajada,
With a Glitter 811H (my welder) I can weld copper 0.4 + nickel plated steel in sandwitch.

Check on this video when I am doing that on top of a EVE 40PL while I was making a pack. minute 39:41 seconds.

 
ZEUS-FL said:
Pajada,
With a Glitter 811H (my welder) I can weld copper 0.4 + nickel plated steel in sandwitch.

Check on this video when I am doing that on top of a EVE 40PL while I was making a pack. minute 39:41 seconds.

Click to expand...
Did you find that using flux significantly reduced the power needed? And the flux didn't help with spot welding the copper directly?
 
ZEUS-FL said:
Pajada,
With a Glitter 811H (my welder) I can weld copper 0.4 + nickel plated steel in sandwitch.

Check on this video when I am doing that on top of a EVE 40PL while I was making a pack. minute 39:41 seconds.

Click to expand...
Forgive me if you said it somewhere in that long video. Why aren't you cutting your own copper patterns so you can use a single piece of copper for those 4 cells instead of 4 separate strips that you have to layer?
 
Jamesavery22 said:
Forgive me if you said it somewhere in that long video. Why aren't you cutting your own copper patterns so you can use a single piece of copper for those 4 cells instead of 4 separate strips that you have to layer?
Click to expand...
That’s a good question! I went with separate strips instead of a single piece of copper mainly to keep the pack as light as possible. A solid copper sheet would add extra weight that wasn’t necessary for this build. This pack is intended to be used in an RC jet.
 
You must log in or register to reply here.

Similar threads

P
    • Like
Cycle life tests of mixed-technology cells
Replies
8
Views
1,779
Pajda
P
P
    • Like
Cycle life tests of High Energy density cylindrical cells
4 5 6 7 8
Replies
187
Views
50,988
BlueSwordM
BlueSwordM
P
    • Like
Cycle life tests of High Power density cylindrical cells
2 3 4 5
Replies
119
Views
28,166
yantra yatra
Y
P
    • Like
Calendar ageing tests of mixed cylindrical cells
Replies
13
Views
3,571
Pajda
P
GreenTecAuto
Valence U27-12XP Lithium Ion Battery $370 per Module
Replies
1
Views
1,271
GreenTecAuto
GreenTecAuto
Back
Top