1000kW DC powersupply - Help

[wb9k]

Hi wb9k, ..are any of those packs rated at 20kW/kg ?

No, not on paper anyway. The 20 Ah pouch cell is rated for 20C continuous, which by my calculations is 2.64 kW/kg. However, it's possible to get much more from them (about in the ballpark of the stated target here) with the tradeoff being reduced calendar life. How bad the reduction in calendar life is depends on lots of details. I would think thermal control would be the hardest aspect. You want the cells to be relatively warm at the start of the run, otherwise cell impedance will be higher than optimal. At the same time, you don't want them to overheat from the abusive run conditions. From the sound of things, you'd want to pre-chill just about every other part of the system while keeping the batteries as warm as possible without allowing them to overheat (60C max) at any time. Not easy, but that's the nature of such an extreme design I suppose.

http://issuu.com/chargedevs/docs/issue_13/62?e=3746341/8631268 This link takes you to a recent issue of Charged magazine with an article on the Buckeye Bullet. They discuss their rationale for using A123 cells for the project. May or may not be the absolute best cell for this job, but I would think they would be on the short list.

 

[liveforphysics]

No, not on paper anyway. The 20 Ah pouch cell is rated for 20C continuous, which by my calculations is 2.64 kW/kg. However, it's possible to get much more from them (about in the ballpark of the stated target here) with the tradeoff being reduced calendar life. How bad the reduction in calendar life is depends on lots of details.

Once your voltage drop is below 50% of unloaded cell voltage, drawing additional current only decreases the power output of the cell.

The 50% voltage sag point occurs somewhere pretty close to 20C from my own testing on those 20Ah pouch cells. Keep in mind, while extracting peak power from the cells, an equal amount of energy is going into heating your cells internally as is going towards propelling your vehicle down the track.

I've found in practice, designing around not exceeding ~30%-ish of pack voltage to be sag tends to be around the threshold where things still work well for making big power. Drawing the cells down below ~30% sag tends to be generating a LOT more cell thermal issues in exchange for very little additional system performance.

If you want to break records on the drag strip though, as much as I don't like working around the atrocious QC related issues hobby-grade cells have, it seems few alternatives can match it. Even cells that boast 200C discharge or whatever often make a worse drag pack because they have such poor specific energy that RC LiPo still puts more power into a smaller lighter box even with only 100C discharge or whatever.

I haven't tested this exact cell, but it seems it may offer somewhere around a usable >8kW/kg in a well designed pack for a drag racing application.
http://hobbyking.com/hobbyking/store/__45143__Turnigy_nano_tech_A_SPEC_G2_5000mah_4S_65_130C_Lipo_Pack.html

 

[halcyon_m]

What happened to the SAFT 200-400C cells we were discussing or other similar pulse-power cells?

LiveForPhysics brings up a good point about the 50% rule (peak power from a cell comes at or above 50% of the open-circuit terminal voltage). Also, thanks for the 70% (30% sag) rule-of-thumb.

Maybe we can compile a list of candidate cells? Or.. maybe this is getting too much down in the noise of a just-for-fun exercise and call it 'plausible'.

 

[major]

I thought Davide had a good method of looking at power from cells here: http://elithion.com/wp_short_discharge_time.php He did not consider the hobby grade Lipo worthy to include on his list. But it is interesting to note where the A123 pouch is.

 

[liveforphysics]

What happened to the SAFT 200-400C cells we were discussing or other similar pulse-power cells?

This .pdf will show you.

http://www.dtic.mil/cgi-bin/GetTRDoc?AD=ADA494956


Essentially, they are 200C discharge cells with 1/4 the energy density, so as far as mass/volume of battery vs power output, RC LiPo over ~50C discharge slaughters them.

Remember, if we took a great new cell like the Sony VTC5 with 2.5Ah of capacity and say for a given application it's able to handle 35A discharge (14C). If the cell were rated for 1Ah rather than 2.5Ah, then the same 35A discharge becomes an impressive sounding 35C discharge... This is what we see with the SAFT cells and many others. I think the KERS A123 18650's were something like 0.8Ah or 1.1Ah per cell or something. They encourage as poor of energy density as possible due to rules regarding maximum pack energy storage. This makes it pretty easy to claim a 50C or 100C cell, but it doesn't mean it performs any better in a given application than a 25-30C discharge cell with 2-2.5x the capacity.

 

[liveforphysics]

He did not consider the hobby grade Lipo worthy to include on his list.

If he was discriminating on cell safety and manufacturing QC, he should have booted at least half the cells off the list.

If he wanted to include the cells offering the greatest performance for a drag application, as far as anything I've tested, the >100C discharge hobby cells are the highest power density cells I've seen by a pretty healthy margin.

 

[halcyon_m]

[A]s far as anything I've tested, the >100C discharge hobby cells are the highest power density cells I've seen by a pretty healthy margin.

Do you have any data points you could throw out at us? I see the test from the SAFT cells got 14kW/kg at 20C initial temperature, but considering a 3S, 2600mAh LiPo, rated at 130C max, (229g), the power density is 16.4kW/kg NOT considering any internal resistance. If you figure they rate using the 70% rated voltage rule, then that figure drops to 11.4kW/kg, which is notably below the 14kW/kg of the SAFT cells. I'd love to get some real data to put this to rest though.

 

[Hillhater]

[A]s far as anything I've tested, the >100C discharge hobby cells are the highest power density cells I've seen by a pretty healthy margin.

Do you have any data points you could throw out at us? I see the test from the SAFT cells got 14kW/kg at 20C initial temperature, but considering a 3S, 2600mAh LiPo, rated at 130C max, (229g), the power density is 16.4kW/kg NOT considering any internal resistance. If you figure they rate using the 70% rated voltage rule, then that figure drops to 11.4kW/kg, which is notably below the 14kW/kg of the SAFT cells. I'd love to get some real data to put this to rest though.

Agreed, but I suspect finding any reliable .high discharge ( 100+C ), test data from HK Nanotech lipo, will be a futile exercise !

 

[Punx0r]

It'd be a bit odd if hobbyking RC lico, being some of the cheapest available also turned out to be some of the best performing. Aren't there any superior RC lico pouches from more specialist manufacturers? I always figured Hobbyking stuff was near the bottom end of the RC market.

 

[halcyon_m]

It'd be a bit odd if hobbyking RC lico, being some of the cheapest available also turned out to be some of the best performing. Aren't there any superior RC lico pouches from more specialist manufacturers? I always figured Hobbyking stuff was near the bottom end of the RC market.

A battery fab facility is a hard thing to start up, so that's likely why there are few boutique batteries out there. Although, if you can tell from the growth of the RC hobby, they are demanding better and more powerful batteries, particularly in the competition sector.

When conducting a test, though, it's important to consider that the lead wires on each cell (10awg, most likely) are 1mOhm/foot, so for every 6" of 2-wire lead, that's another 1mOhm. At the currents we are talking about, that's a significant addition to the impedance. Any data we see should have voltage measured at the terminals of the cell to ensure fair cell comparisons that aren't tainted by the lead wire length or contact resistance of the connectors.

 

[wb9k]

Once your voltage drop is below 50% of unloaded cell voltage, drawing additional current only decreases the power output of the cell.

The 50% voltage sag point occurs somewhere pretty close to 20C from my own testing on those 20Ah pouch cells. Keep in mind, while extracting peak power from the cells, an equal amount of energy is going into heating your cells internally as is going towards propelling your vehicle down the track.
I've found in practice, designing around not exceeding ~30%-ish of pack voltage to be sag tends to be around the threshold where things still work well for making big power. Drawing the cells down below ~30% sag tends to be generating a LOT more cell thermal issues in exchange for very little additional system performance.

You raise a good point here. I have tested single A123 M1 32113 cylindricals to 50C (200 Amps) before hitting 50% sag, as have others here who have posted charts of the results. I would suspect the reason for such poor performance from a pouch with exactly the same chemistry and base materials is that the level of compression was inadequate for the level of current draw. At these power levels, it becomes very important to have heavy, even compression. Many users who demand ultra-low impedance and stratospherically high C-rate capabilities (like Formula 1) still opt for cylindrical cells because they are frankly a bit more capable in that realm--but not nearly so much as the numbers you're throwing out here would suggest.

I know you're not a fan of A123, but the fact remains that our cells are particularly good at dumping (and sinking) enormous amounts of current and that they remain very popular among users who need those capabilities combined with decent calendar life.

 

[Hillhater]

Polyquest is a reputable name in RC lipo, and has been used in EV drag racing .
Here is a test chart of some 90C cells ..curtesy of John Metric of "Assault n Battery" drag Miata fame.
http://www.diyelectriccar.com/forums/showthread.php/assault-n-battery-build-thread-74539.html
Note how the voltage is quick to drop below 3v even at only 77C and warm .
but I think John has some better cells now.

 

[halcyon_m]

This is very interesting to see that even at 340A, it falls off quickly, but then rebounds to 4V, presumably where ~80% capacity can be extracted. When they say 'burst 90C', they really mean it. At 300A, I calculate with rough numbers for weight from a 3.5Ah pack on HK, that this cell only gives 5.2kW/kg, a pretty far cry from the numbers we were throwing around earlier.

 

[Hillhater]

remember that curve is a ..Discharge, Rebound, and Recharge curve.
but how do you get 5.2 kW/kg ?
The cell seems to "average" about 3 volts at 340A =1020 W,... for a few secs
and typically those cells weigh 95-100gms
so that would be 10-11.0 kW/kg.

I cannot find the graphs, ( lot of data on his facebook page) but JM is now running some newer chemistry cells that hold 3.5 cell voltage @ 300A and weigh 95gms.
so again ~~ 11kW/kg.
.... and that is not pushing them to their max output, so maybe 15+ kW/kg (burst ?) is possible on those.
He does mention 3.0 volts @ 450A

 

[Punx0r]

Do we know that the polyquest cells are different to HK cells? There seems to be some debate about whether Haiyin cells are a superior product, or the same in a different wrapper.

It'd be interesting to have high discharge specs for all the different A123 cells

 

[Hillhater]

Haiyin are a large manufacturer, producing cells under many labels ( possibly some of the HK products ?)...
But they have certainly produced some poor performing pouches too...if you remember EVMetro's drag car build a year or two back, using Haiyin 6Ahr pouches of supposedly 67C capability, .....those simply dropped half their voltage under a 30C load !

Any progress with this design/build ?

 

[Punx0r]

This isn't getting built