inductance problems w long wires? Make them parallel connect

[Hummina Shadeeba]

inductors add up in series but are less than any one of the inductance when wired in parallel just like resistors.
So..
Is there a reason I shouldn't make my long connections parallel to reduce inductance and cut down on my need for more capacitors?

All I can think is that doing so maybe would slightly imbalance the cells due to the inconsistent resistance of the inconsistent wire length

 

[jmz]

Even better, if you place the return wire with the opposite current (or a group of three wires in a three phase system that sum to zero) right next to each other, the loop area of the magnetic fields will be constrained and the inductance will decrease.

 

[Punx0r]

It's said a lot that twisting a pair of wires reduces their inductance, but the last time I researched this the answer seemed to be that it's not really true: twisting the wires ensures they are close together, which does reduce inductance, but only like a figure-8 cable. The downside to twisting is the wire length required to reach from point A to B increases a little, which increases inductance.

You do see a lot of twisted pair cabling, but AFAIK it's solely for reducing susceptibility to external noise.

 

[Deafcat]

run all your power in Teck conduit, problem solved :wink:

 

[Hummina Shadeeba]

it sounds like that teck stuff is shielded but none-the-less there's still going to be inductance in my wires due to them being long.

I'm hoping someone can confirm that this battery layout will reduce inductance since the long distance connections are parallel connections instead of series.


damn i dont think the photo will show. imagine a series chain of 18650s but with two cells in parallel between each series connection and those parallel connections would have the long wires instead of the series connections. going by the definition of inductance in parallel vs series it should be way better..but maybe there's other problems that will arise.

View attachment 1

this pic is missing the second connection between the two paralleled cells. so two long wires instead of the one

 

[bobfandango]

inductors add up in series but are less than any one of the inductance when wired in parallel just like resistors.
So..
Is there a reason I shouldn't make my long connections parallel to reduce inductance and cut down on my need for more capacitors?

All I can think is that doing so maybe would slightly imbalance the cells due to the inconsistent resistance of the inconsistent wire length

Yeah, that certainly will work. Taken to an extreme, you could run something like a dozen small wires. You've just (re)invented a litz cable! A litz cable is a cable with a number of small gauge solid conductors, each individually insulated and all of them put in an outer sheath. Solid conductors have lower inductance than stranded, so hence the solid. Many in parallel reduce overall inductance, and also allows the cable to be more flexible than a single equivalent solid conductor. Audiophiles are nuts over them...

 

[Hummina Shadeeba]

thank you!
what you are describing is maybe a better method of reducing the inductance than what I am trying to describe and I might do it instead.

Let me better explain my goals:
I'm hoping to build a battery pack of (48) 18650 cells (12s) under my skateboard and I want to put half the batteries at one side and half at the other..and I want to have the least inductance in the battery wires so as not needing any more capacitors.

The pack will be comprised of 4 sets of 12 cells with each set spaced to the outer edges of the skateboard as far apart as possible. This will require long battery wires. I can either do each set of 4 as a pure series of 12 cells and then connect the 4 sets via the litz-like wire as you describe, which sounds possibly the best way, or I could have each of the packs of 12 cells connected as in my picture with 2 cells in parallel connected to 2 more sets in parallel with series connections...and end and start each of the sets with a parallel connection bridging to its similar polarity cell in another of the 4 packs.

what do you think about what I think?

http://www.electronics-tutorials.ws/inductor/parallel-inductors.html

 

[Punx0r]

It didn't sound right that Litz (or multi-stranded insulated) wire would reduce inductance and a quick Google search suggests this isn't so.

Hummina, I don't think it matters how you arrange your cells - the inductance of your wires will be determined by total cross-sectional area and separation distance between positive and negative wires.

 

[Hummina Shadeeba]

my google searching brings me to fundamental electronics laws about inductance stating that parallel inductors produce the reciprocal of the sum of the reciprocal of each inductor while series connections add up. hugely different. I'd like to get things confirmed one way or the other.

Also in formulas for inductance i dont see amps and I guess the formula just defines the wire and to figure what voltage spike an inductor produces the current is added later?

also is it true that batteries in parallel dont pass current..or at least very little?



my pic above is missing the needed second connection between the paralleled cells with the long wires.

And no point wrapping the two paralleled wires closely together because they're the same polarity?



Can I assume half the current goes in each of the long parallel connection?


ill put up the right pic soon

 

[bobfandango]

It didn't sound right that Litz (or multi-stranded insulated) wire would reduce inductance and a quick Google search suggests this isn't so.

Hummina, I don't think it matters how you arrange your cells - the inductance of your wires will be determined by total cross-sectional area and separation distance between positive and negative wires.

Yeah, of course that's right now that I think about it more... So, yes, inductors in parallel will have lower series inductance overall, that is only true if those inductors are independent! That is, they each are not coupling into one or more of the others. However, when you run two wires next to each other with the current running in the same direction, of course(!), there is little difference between that and a single conductor. Magnetic field created by the current is proportional to the current.... Doubling the conductors, and halving the current through each still results in roughly the same magnetic field since current is in same direction....

 

[jmz]

The purpose of Litz wire isn't to reduce inductance, it's to increase the cross-sectional area available for conduction in the presence of the skin effect at extremely high frequencies (50 kHz+).

Anyone who thinks they need Litz wire for speaker cables is a moron.

For your application, if you really are getting sufficient droop across your battery wires to require capacitive decoupling, you have problems with resistance, not inductance. If you are drawing high frequency current at hundreds of kHz then something is going very wrong. Place the required decoupling capacitors (if any) at the input to the converter, and forget about what the batteries are doing.

 

[Hummina Shadeeba]

Thank you both for your help. I'll think about it.

How about my wiring, please confirm this is ok.

 

[Punx0r]

Heh. There's an awful lot of B.S. spouted in Audiophool circles

IIRC inductance of the battery wires can be a problem for RC controllers and there are warnings against extending the wires for that reason, but the normal practice seems to be to add a little bank of electrolytic capacitors just before the controller.

It's interesting to talking about the theory behind wire inductance, but I think jmz is right: in this real-world application you don't need to worry about it. If it's an RC controller, add some input caps and be done with it

 

[Hummina Shadeeba]

so no free lunch with my parallel connections inside a series circuit as in my pic, and the paralleled insulated wires such as litz wire isn't going to do it either.


so how about if my (4) packs are comprised of only series connections and then they are paralleled going to the controller...that would still just be a bunch of series inductances added up as well right?!!

Is there no way for me to decrease inductance with my (4) distantly packs of 12s other than simply keeping the opposite current wires paired together?

where's the free lunch? If i cant find it the wire lengths are so long that I'll very likely need to add capacitors going into the speed controller. Or I could just go without adding them and...if my capacitors aren't getting hot assume the best??

 

[Punx0r]

The problem wouldn't be the caps over-heating, it would be voltage spikes on the DC rails in the controller blowing up your FETs.

 

[999zip999]

Voltage spikes my controller and c.a. is set at 80amps my c.a. shows 135 amps my controller really spikes 135 amps with100v fets ? Or is just some reading bounce and don't worry ?

 

[Hummina Shadeeba]

The problem wouldn't be the caps over-heating, it would be voltage spikes on the DC rails in the controller blowing up your FETs.

if the fets are getting warm or hot would it be an indication of the possibility of them blowing up and I could worry about adding capacitors then? I'd have thought the caps would be getting warm or hot at least before the electricity gets to the fets...but maybe it's a cool voltage issue or very quick and it wont show up as heat in either? I heard that caps will get warm if they're overly-taxed and will break sooner if they are getting warm or hot. Arent caps the fets security from spikes and if the caps are cool I dont have to worry about the fets?

 

[Punx0r]

No. The transient voltage spikes that kill FETs are short and no warning signs (such as heat) will be present. If you are using an RC controller I would follow the manufacturer's advice on adding input caps, otherwise, you're over-thinking it.

 

[Hummina Shadeeba]

So there's no thinking possible and I should just take the manufacturers' suggestions? It'd be nice to figure out what inductance the pack produces and reduce it if possible. I've got these cells sitting here with nothing to do till my motors arrive anyway. I have an LC meter but imagine I can't use it on batteries without it blowing up

since I have two chains maybe orienting them beside each other so as to reduce the inductance or be creating more....like to figure it out if possible. Any way to figure inductance of batteries?

 

[Grizzl-E]

You have an lc meter? Than hook it up to the two lengths of wire required to bridge the gap between the cells and get a nanohenry reading. I'm betting your meter will have trouble getting an accurate reading on such small inductance.

Are you planning on running very close to the voltage rating on your fet's? You are talking about such small lengths of strait wire here that the inductance on the dc side should be irrelevant.

 

[Hummina Shadeeba]

I can put my LC meter to my battery leads?! I didn't think so. That's awesome.

People in the model airplane world say if u extend your battery wires more than 4 inches u should at..220 caps I think it was. But bikes don't care. I don't care much at this point at least I'm not worried, but I'll be definately seeing what I can get with the LC meter moving e batteries and their wires around. Supposedly ur slowly killing ur esc not following the 4 inch rule.

I can get an inductance reading with a wire that has no current, and it's very precise, maybe inaccurate but precise, .. how will adding a current effect the inductance? I never see inductance formulas that incorporate amps. I'm guessing it's just a linear relationship.

I'm wondering how pairing different parallel pack's wires will effect things. Pairing current going in opposite directions. Ideally I guess the wire would go straight out and back keeping beside itself with a 180 turn at the end. I like this tool

 

[Grizzl-E]

No no no that's not what I said. Just test the wires. No batteries connected. Sorry.

 

[Hummina Shadeeba]

Ooo. Good to read!!. Disappointing. Luckily I'm on here so much. That could've been bad. So no current possible with the LC meter and I'll just estimate/simulate. I guess a battery can just be considered a wire and I'll just follow the direction of current