Looking for help with an unusual battery application

busted_bike

100 mW
Joined
Jun 24, 2010
Messages
42
I have an application in mind for a battery pack as short-term energy absorption. It has little to do with EVs, but I figured I’d ask the question here since the folks on this forum have, collectively, as much battery knowledge as I’ve seen just about anywhere.

Anyway, I’m working on a circuit which will occasionally feed current back onto its power supply bus. Left unmanaged, this condition would result in huge voltage spikes on the power bus which would destroy all sorts of things. One management strategy I'm considering is to detect this condition and engage a load circuit to burn off the excess power resistively. But on thinking through it, a battery pack seems like a more elegant solution.

When the load circuit deposits energy on the power bus, the cells will naturally absorb it and the bus voltage will rise. When the circuit begins consuming energy again, the battery will discharge to supply the load and the bus voltage will drop until the main power supply re-engages to hold up the voltage. The aspects of this solution that I’m unsure about are as follows: the battery would be connected to a live power bus at a relatively low working voltage (lets say 3V/cell), nominally <10%SOC when not absorbing or supplying energy (i.e. most of the time).

Though I’m not necessarily looking for recommendations for a particular cell or chemistry at this point, long working lifespan is an important feature which the newer Li-Ion chemistries seem to lend themselves well to. Am I setting myself up for spectacular failure? This seems different enough from most battery applications I’m aware of (where the battery is charge online and discharged offline) that I’m afraid I’m missing something. Are there any similar applications that I’m not aware of? I’d feel a whole lot better if others were/are successfully doing something like this.

A couple of other notes, which might help to clarify:
• It is acceptable to employ a full-time circuit to monitor and balance at low current.
• Per-cell LVC / OVP is also acceptable, though neither of these conditions should occur if the circuit is working properly.
• It is acceptable to tune the power supply circuit so as not to trip its OVP over the working voltage range of the battery.
 
Depending upon the load/power requirements, I'd look into ultracapacitors instead of batteries. No need for cell management and cant overcharge, etc. Unlimited number of charge/discharge cycles. Maxwell is the place to start. Pricey...
 
I'm guessing he is talking about storing a useful amount of energy. Ultracaps really suck for energy storage, but would excel at all the other requirements.

Give us an idea of the current peaks in the spikes you're trying to absorb, and the scale of the energy.
 
Thanks - unfortunately, I seem to have left out some relevant information: the duration during which the load can act as an energy source isn't well bounded. In the extreme case, it could be hours. Worst case current spikes are around 100A, operating Voltage is flexible... but lets say 24V-27V operating range as frame of reference.

Though I haven't crunched the numbers (I will, now that you've mentioned ultracaps), I strongly suspect the $/watt-hour for ultracapacitors wouldn't make sense.
 
Absorbing 100A for hours is going to take a really big and expensive battery. Plus the circuitry to regulate the charge voltage to the pack. And if the pack fills while the load is still pumping juice, you are going to have a real problem.

A 200 Ah cell runs around $250. You will need a minimum of 8 of them. Multiply that by your load generating time if you need more than 2 hours. Also, how long between load and unload cycles?

Perhaps dumping the load power as heat or mechanical energy is not such a bad idea. What you need to do is find a constructive use for that heat/power...
 
texaspyro said:
Perhaps dumping the load power as heat or mechanical energy is not such a bad idea. What you need to do is find a constructive use for that heat/power...

Yep, that was my first intuition. Once I wrap a control loop and some diagnostic circuitry around a heating element and then add some fans (which also need a monitor circuit) to manage the heat, it's about the same cost per watt as a battery. I've explored a couple of options for doing something useful with the heat... but they aren't exactly free either. If you have any ideas other than a big dumb power resistor as a heating element, I'm willing to listen. (Something that runs hot enough to shed heat radiatively?) The low voltage DC bus doesn't help in that respect - if I were dealing with AC line power, a really inexpensive power dump is essentially the equivalent of a hair dryer element.

The ~2kW figure was peak, not continuous power. In terms of energy capacity, 0.5kWH would be plenty... which would imply that I need to use cells that are happy charging at at least 5C. And the vast majority of the time (>>90%), the load circuit will be a net energy consumer and the battery will be dormant. So back to my original question... what are the reliability concerns with running a mostly empty battery pack off a live DC bus and expecting it to absorb (and then return) energy on demand? Are other similar applications that I'm not aware of?
 
How about powering an electrolytic hydrogen generator? One can have endless fun with stochiometric mixes of hyrdrogen and oxygen. :twisted:
 
This would take your 5C charge with a chuckle, and transition to discharging with minimal change in voltage, as the Ri is under 1mOhm/cell.

If you wanted a usable 0.5kwhr capacity, get 8 of these packs:

http://hobbycity.com/hobbyking/store/uh_viewItem.asp?idProduct=11957&Product_Name=Turnigy_nano-tech_5000mah_4S_45~90C_Lipo_Pack
 
A123 26650 M1 3.3V, 2.3Ah cells are rated for 4.5C continuous charge 10A for 2.3Ah cell and can accept more for bursts, 30C continuous discharge, 52C burst. They are robust, safe and have excellent cycle life. RC LiPo is high performance, but 3V is towards the bottom of it's operating range and it could never be considered as a safe chemistry. Around 60-70 M1 cells should be sufficient. Drop me a line if you want a pack made up to your requirements, they are 6USD a cell spot welded with copper sheet.
 
texaspyro said:
How about powering an electrolytic hydrogen generator? One can have endless fun with stochiometric mixes of hyrdrogen and oxygen. :twisted:

Heh. I'm no chemist... what sort of input bandwidth does a system like that have? Can you ramp hydrogen production up as an energy sink and then back down again in, lets say, less than 100us? Because that's about how much time I have before I start to let the magic smoke out of stuff.

Also, LFP & cell_man - I appreciate the suggestions on specific cells... but I'm (pleasantly) surprised that nobody has expressed any concerns about running the pack on a live power rail at very low SOC. That doesn't somehow qualify as abuse?
 
There are already people in this thread who know a lot more than I do, but it seems to me if the charge time is not well bounded, you're going to need something that can sink power indefinitely anyway. If your energy storage mechanism has any practical limit on how much it can store without damage, you probably need to plan for what happens when it fills up. So your lowest cost solution is probably just the big dumb power resistor and fans.
 
Back
Top