A123 26650 100 Amp discharge test

[wb9k]

My colleague is near finalizing his solderless 26650 pack building blocks kit, and it's looking really nice. This a a 3-D printed kit. After 16 previous iterations, the design is very compact and mechanically robust.Still optimizing materials and print density, but watch for pics here soon.

A few days ago in another thread, I was challenged as to whether the A123 26650 can deliver 100 Amps of steady current (exactly 40C for these cells). I thought I had seen the cells do that, but was foggy on details. So, as part of a test to see what the printed kit could withstand, we ran a 100 Amp discharge test on a 4S1P pack of M1B 26650 cells. Load device was a B&K 8520 programmable DC load. I logged the results with the Fluke 289 DMM:



As you can see, this little pack delivered 100 Amps for about 40 seconds with no trouble. Loaded voltage is still well over 50% of unloaded voltage, so it would appear that the cells can handle even more than this. The test was halted because the PLA plastic parts that the pack terminals run through began to melt. This is where the voltage goes high again, around 45 seconds. The battery is then disconnected, which is why voltage falls to 0 at the end of the curve. Rick is going to try printing the parts that melted in nylon to give them better thermal stability. The final kit will handle plenty of current, but the weak link in this test was the cell mounting and connecting apparatus, not the cells themselves. The cells got rather hot after this test--almost 60C, their max operating temperature. Active cooling would be required to do this kind of a thing on a regular basis. A 100 Amp discharge will drain a 2.5Ah cells from full to 0% SOC in 1.5 minutes.

Here are a couple images of the test unit:





I've done 45C discharge tests on the M1A 32113 (a 4.5 Ah cell at 200 Amps) and found that to be around the limit for that cell. M1B chemistry performs a bit better. It may be possible to get even greater than 50C out of them. Stay tuned....

 

[arkmundi]

Nice work. Thanks. So what's the plan with the kit? Going to sell or go the open source route for distributed manufacture?

On another note, when's the new A123 14ah prismatic pouch cell going to be in full production and available for purchase? And will that be made available through your new distributor StorTronics?

 

[wb9k]

Nice work. Thanks. So what's the plan with the kit? Going to sell or go the open source route for distributed manufacture?

On another note, when's the new A123 14ah prismatic pouch cell going to be in full production and available for purchase? And will that be made available through your new distributor StorTronics?

I'm trying to get my friend to become a member of the board. He'll have to fill you in on his long term plans.

Regarding the 14Ah cell, I have no idea when you might be able to get it. I found out about StorTronics here on this board....I knew a new sales channel was in the works, but had no idea what it would look like. Curious to hear reports about folks dealing with StorTronics. They appear to be well-established, so that's a good sign. They're local to us here. Maybe I can drop in on them sometime.

 

[gensem]

Can you run lower discharge tests like 10c and 20c to check the voltage drop?

 

[wb9k]

Can you run lower discharge tests like 10c and 20c to check the voltage drop?

I imagine we could do that. We'll probably be running more tests in the next week or so when the next set of parts is printed. Could be as soon as tomorrow. I imagine we'll be characterizing the whole package thermally as well. Thermal measurements today were taken quickly with an IR camera.

 

[Doctorbass]

Thanks for the test.

The A123 specs are 30c continuous and 60C burst 10 sec.

That's for a reason of max thermal limit reached.

for sur ethe A123 cell can do 100 continuous! but the cell temp get so high than it damage it and it loose capacity and can not do the rated cycles.

I have a report where they tested the cells and determined the 30c with the max temp the cell reach in a normal environment of use ( 23 degree c normal humidity, not enclosed.. etc...

The continuous C rate is not determined if the cell will or not explode... it,s depending on the heat it will produce and will thatit also can dissipate. and this factor depend on teh internal resistance as well


At 100A a A123 26650 will dissipate 0.01 * (100A*100A) = 100watts of heat! and in 40 sec that's 4kJ of heat! per cell.

That's a great proportion of energy loss in heat vs electrical energy out of the cell :lol:

Doc

 

[wb9k]

Thanks for the test.

The A123 specs are 30c continuous and 60C burst 10 sec.

That's for a reason of max thermal limit reached.

for sur ethe A123 cell can do 100 continuous! but the cell temp get so high than it damage it and it loose capacity and can not do the rated cycles.

I have a report where they tested the cells and determined the 30c with the max temp the cell reach in a normal environment of use ( 23 degree c normal humidity, not enclosed.. etc...

The continuous C rate is not determined if the cell will or not explode... it,s depending on the heat it will produce and will thatit also can dissipate. and this factor depend on teh internal resistance as well


At 100A a A123 26650 will dissipate 0.01 * (100A*100A) = 100watts of heat! and in 40 sec that's 4kJ of heat! per cell.

That's a great proportion of energy loss in heat vs electrical energy out of the cell :lol:

Doc

I agree you would need some serious active cooling of the cells to hope to do anything like this on a regular basis if the cells were expected to last very long, and even at that they will heat up quite a bit. I wanted to point out however, that pulling 100 Amps from these cells does not cause voltage to sag to unusable levels, which is what others predicted would happen.

 

[brownrdb2]

I am currently printing out new end plates to run the test with ABS and then hopefully ABS with Nylon inserts. the ABS test will be easy, but I need to get my new extruder online to go above 240C which is required for Nylon.
The next test for the 26650 batteries will be done in stages to test low to high C rates. The test that was posted was just to prove the cells could do it

 

[gogo]

Great! Welcome to the E-S forums, brownrdb2. These 26650 batteries and your enclosures might make a good short-burst pairing.

 

[ramboman]

I used Solderless Power Tubes years ago... unfortunately there is no alternative today...
I have to build two 10s2p packs for my Mosquito...
Either soldered together or assembled with your 3D kit or even between two wooden plates...
What would you suggest ?

 

[gensem]

Looking forward for the building blocks.

 

[ramboman]

Several questions about your project...
=> do you solder something to the a123 cells or not ?
=> if not... do you improve contact with silver paste or not ?
=> if you use bolts (soldered or not) are they copper, brass or other ?
=> what is the size of those bolts ?
What I found on ebay is similar to what I use for Headway... but Headway has bolts... a123 has no bolt... so useless for a123 :
http://www.befr.ebay.be/itm/2-PAIRE-PLASTIC-HOLDER-FOR-A123-26650-BATTERY-LI-ION-LI-POLYMER-LIPO4-/160862705605?pt=LH_DefaultDomain_71&hash=item25742a17c5
=> is your project similar, but solving connecting issue ?
=> would you propose the cell connectors too like Headway does :
http://shop.rc-electronic.com/e-vendo.php?shop=k_emcotec_e&SessionId=&a=search&SearchStr=headway
Thanks for your feedback

PS: angled mounts may be usefull like on my B-25...

 

[brownrdb2]

The connector system that you have attached is interesting, but will not work with cells that do not have a post welded to the cell ends (If I understand the way this one works). The current design that I am working on uses compression and copper foil (with contact goop to avoid oxidation) to form to the contact area of the cell. The contacts were sized to insure that the negative terminal couldn't short out to the case if cell tolerances were an issue. The hardware used to for the terminals are 6-32 X 3/8" bolts that are retained in the plastic ends. The copper foil wraps around a special cavity in the end cap and lays flat to the back of the plastic. The copper foil can be soldered , but for all of my testing I just used mechanical attachment to some copper bars. There is very little voltage drop. Pulling 240 amps out of the 4P pack was only limited by the cells (Since A123 26650 can put out some serious current, I figured if it could handle them there would be no worries). I have been looking into tin plated foil tapes, but finding one thick enough (my foil is currently about 4 mils thick) has kept me from switching over.

The cells are not soldered or welded. Compression contacts that are used are robust enough to allow high current draws with little voltage drop.

No silver paste has been used in any of the testing, but it is an interesting idea (I've got a friend on this board who has been playing with nano-tubes that may be an option in the future).

The screws that I am using are zinc plated steel (Since they don't carry any current). I have looked at brass foil as a replacement for the copper foil, but do not want the increased resistance. The washers that come into contact with the copper foil are brass.

The thought behind this pack was to make it so no soldering or welding was required. It had to be easy to assemble, the cells had to be replaceable, high current capability, and there needed to be a small clearance between the cell bodies to avoid a possible connection between cells and help with cooling.

So far the testing has shown the battery holder has met the design requirements. I am currently performing a cost analysis and trying to reduce the bill of materials for the kit.

The cost reduced version passed the drop test yesterday, so I am starting to prepare the final documentation. Just working on the final cell configurations now. I am looking at configurations of 2X2, 2X4, 2X5, and 4X4. If anyone has any suggestions for needed configurations, I would welcome the input.

Hope this answered your questions

 

[ramboman]

Excellent
Please add 2*6...
=> because 6s2p a123 is equivalent to 5s LiPo
=> because most large brushless run under 10s LiPo i.e. 12s2p a123

 

[gensem]

brownrdb2 how are you going to route the jst balancing wires?

Do you plan on selling cells aswell?

 

[ramboman]

I don't have a 3D printer
This is the basis for a 4s1p a123 solderless pack...

 

[dnmun]

if your electrical connections all involve the pressure that is created by forcing the contacts up against a plastic surface then expect the connection to fail as the plastic softens under the heat generated by the resistance in the contact where the power is delivered.

you are using a weird copper connector on the outside which is mostly just a waste of metal and does not provide a secure connection because as the plastic softens then the connection will lose the compression of that single nut so it will heat up more and more rapidly. you would be better off if you used just simple 10AWG if you wanna go to 100A. but use ring terminals on the connector and clamp the ring terminals between 2 nuts, and not use the plastic as backing to tighten up the connection. that is a fail imo.

 

[ramboman]

May be... may be not...
My first packs were eight 6s1p Solderless Power Tube... based exactly on the same principe... they never failed
There is no plastic... only wood and copper...
But I am open to a more professional approach.

 

[wb9k]

The compression is provided by a foam pad, not the plastic itself. Regardless, ABS is totally stable at even 60*C, the max temp for the cells. Thus, if you don't exceed the limits of the cell, you don't exceed the limits of the plastic either. Nylon is even more heat-stable, which is why that material is also being considered for really extreme applications. The first test shown here was printed in PLA, which proved to be too easy to melt and has not been used since.

 

[liveforphysics]

The compression is provided by a foam pad, not the plastic itself. Regardless, ABS is totally stable at even 60*C, the max temp for the cells. Thus, if you don't exceed the limits of the cell, you don't exceed the limits of the plastic either. Nylon is even more heat-stable, which is why that material is also being considered for really extreme applications. The first test shown here was printed in PLA, which proved to be too easy to melt and has not been used since.

Google plastic creep. Even at room temp materials like ABS, Nylon, and PLA all have clamp loads approach zero clamp force given enough time. The relaxation effects are amplified by the thermal expansion differences causing relative motion and changes in pressure in the assembly.

Also, just as something to remember when selecting plastics for battery use, nylon absorbs moisture, rapidly drops many orders of magnitude in it's electrical resistance, and begins corroding apart. Poly-carbonate and a handful of other materials also do this to a lesser extent.

When you look at the datasheet electrical specs for plastics, they are useless for knowing what the material will be like once it's saturated with water vapor, which starts happening the moment it's in the atmosphere.

 

[brownrdb2]

<a href="http://s284.photobucket.com/user/brownrdb2/media/Cell%20Cap%20Back_zpst6yqpxzu.png.html" target="_blank"><img src="http://i284.photobucket.com/albums/ll13/brownrdb2/Cell%20Cap%20Back_zpst6yqpxzu.png" border="0" alt=" photo Cell Cap Back_zpst6yqpxzu.png"/></a>

<a href="http://s284.photobucket.com/user/brownrdb2/media/Cell%20Cap%20Front_zps3k51twcx.png.html" target="_blank"><img src="http://i284.photobucket.com/albums/ll13/brownrdb2/Cell%20Cap%20Front_zps3k51twcx.png" border="0" alt=" photo Cell Cap Front_zps3k51twcx.png"/></a>

I am new to posting on this forum, so hopefully the pictures I have uploaded show up. If not, I will try a different method.

I greatly appreciate everyone's input and suggestions. The issue of creep and overall strength of the plastic have always been an area of concern. I spent a good deal of time discussing aspects of this design with a few seasoned mechanical engineers and got some rather good advice. Plastics changing shape over time and temperature is just an unfortunate fact of life when it comes to this material. The only thing that you can do is create features to reduce this risk. The base plate of the Cap is 6mm thick with a hexagon infill to help strengthen it and reduce the overall weight. The Cell support rings are 10mm high and act as a rib support system to help maintain the shape better under load and temperature. The total height of the structure is 16mm. The maximum distance from the center of the screw used for compression (not including the outside diameter of the washer) and the center of the battery terminal is 20.5mm. Since the pads compress in about 1.5mm the plastic would have to creep or bend a significant amount interfere with the connection to the cell. I do plan on putting this through some accelerated life testing (cycling from the temperature extremes the cell can handle) to see what kind of movement the plastic will experience. I can see that it would be better to trap the bus bar or foil between two hex nuts and avoid using the plastic as the base for the connection. I think that on the next build I will give that a go.

I will be looking at adding in a sense harness (currently I'm using some molex connectors) but I need to pick up the JST crimper to be compatible with most of the cell logging and BMS components out there.

Thank you once again for your input and questions. Revision 20 is now under way

 

[brownrdb2]

 

[brownrdb2]

 

[brownrdb2]

Sorry, now the pictures came thru

 

[dnmun]

reimagine already. too complicated. get the entire parts count to 4 pieces. why do you need bolts to begin with? why do you have to use a separate piece to connect those bolts together using fasteners? it is called a backing nut FYI.

instead your serial connector could be a single wire or stamped copper strap from the anode of the one can over to the case of the other can. mount it on the face of the inside aperture in the end piece. for the terminations, bring them out as one piece from the inside face of the end plate through the side. use copper wire and solder a ring terminal or just solder to the wire for the power connection.

you don't seem to care about balancing. so sense wire to the middle connector inside is not needed. but the sense wire could be soldered to the connector before assembly and guided out through a hole in the face.

since the conductors are inside then just wrap it end for end with strapping clamp or even have a big C-clip to hold the end pieces on the ends of the cans under pressure.

reimagine simpler. how to form the ends of the connector as they are stamped so the connector has proper orientation of the contact face when it is in contact with the can. BOL