Nov 12, 2008, 05:01 AM
homo ludens modellisticus
Careful!

# too long battery wires can kill ESC: precautions, solutions & workarounds

I have also copied (parts of) useful posts in this thread into this opening post.

... therefore, lengthen the motorwires if possible, not the battery wires.
However, if you cannot but extend the battery wires ...

Contents
• Problem
Voltage spikes higher than battery voltage
• Example four costly controllers consecutively cremated
• Solution I: simply lengthen the motor wires instead
• Solution II: add extra capacitors, rules of thumb
• Capacitor type
• Capacitor polarity/orientation!
• How & where (not) to add extra capacitors
• DIY capacitor pack pictures
• Capacitor & pack suppliers
• Expert/manufacturer opinions & their rules of thumb
They all say the same ...
• Cause, explanation, water hammer/knock analogy & video, theory
References
• Measurements & scope traces

Problem
Conclusion from the links below, all controller manufacturers say the same:
Too long battery wires will kill your ESC over time!
Voltage spikes higher than battery voltage will destroy components and/or the standard input capacitors (large cylindrical thingies in thin colourfull shrink wrap) will be destroyed over time because they get warmer/hot. Using thicker wire will not help, it's a wire inductance problem, not a resistance problem. See Lenz's law. Inductance is what makes sparkplugs spark, 15-25kilovolt.

This goes for all makes, they all use the same principle (except SLS controllers, they use sinusoidal commutation instead of trapezoid and tested 150m extension without extra capacitors). However, lengthening the motor wires may lead to radio interference. Give the three of them a twist to prevent this.

Example
Four costly controllers consecutively cremated
ESC Failure and Overheating - RCG

Solution I
Keep battery wires short, lengthen the ESC-motor wires.
That's hardly critical because there's already a lot of wire in the coils in the motor itself. If the motor-ESC wire eventually gets too long, it will not harm motor and/or controller. May cause interference though, give the motor-ESC wires a twist. Always a good idea to do that anyway.
Also, motorside wire-inductance is a good thing, it takes care of high frequency Pulse Width Modulation harmonics motorside, by smoothing the current.
Assuming you're not using a core-/iron-less motor (= low inductance, needs high PWM frequency).
Motorwire induction (coils) is good, battery wire induction is bad.

The freewheel/flyback diodes in the power FETs also help in smoothing motorcurrent.
Even though the FETs switch off, the diodes keep the current flowing through the motor and motor↔ESC wires.

Motorside: voltage chopping and smoothing of current

Some ESC manufacturers advise against lengthening the wires motorside. But that's only because motor wires have a tough insulating varnish/resin coating that has to be removed before soldering. Use the aspirin trick mentioned in opening post:
(Re)winding and building motors (sticky)

Noise from the motorwires is hardly an issue, here induction is good, in combination with PWM frequency it smoothes the current.

But what if you can not extend the motor wires?

Solution II & rules of thumb, calculation spreadsheet
Controllers already have capacitors on-board to compensate for 'standard' length battery wire inductance. But that may not be enough capacitance when adding extra battery-wire.
If you have to lengthen the battery wires, for whatever reason, add extra electrolytic capacitors in parallel with ESC, never in series with ESC. As a rule of thumb, for (but nut on!) every 4inch/10cm extra length/distance between battery and ESC, add 220uF extra capacitance near the controller (electrolytic condensators, voltage the same as the capacitors already installed, low ESR type) (Ludwich Retzbach, German e-flight author&editor, the 'R' in LRK).
When using two battery packs in series the extra length of wire between the packs counts as well.
Better to use several smaller caps (in parallel) instead of one biggie. Smaller caps can shed more heat and total inductance will be lower (inductance per cap is lower and those inductances are paralled to boot ). See attached pictures below.
Also keep the positive and negative wires as close to each other as possible by taping them together (twisting the wires will give extra length!). When the wires are close to each other then the series inductance will be reduced, because the current is going in opposite directions in each wire (and therefore producing opposite polarity magnetic fields which cancel out). For example, 13AWG wires separated from each other by 1 inch have about 4 times higher inductance than if they are bound together. (Bruce Abbott in www.rcgroups.com/forums/showthread.php?p=11594609#post11594609)

Or even remove insulation of one of the wires??? And/or thinner insulation? The less distance between the two copper cores, the lower the inductance.

Extra capacity based on current and wire length, calculation spreadsheet

Influence of ESC max.current rating on choice of capacitance

Capacitor type
The main spec you need is low impedance and low ESR (Equivalent Series Resistance). I think the only thing you will find at radio shack will be general purpose caps, not low ESR. The ESR value of a cap is not printed on it, you will have to look up the manufacturers spec sheet. The Rubicon ZL series mentioned in the Schulze instructions is a good one and is available from newark/farnell. The Panasonic FM series is another good low impedance cap and is available from digikey
www.farnell.com/datasheets/2161.pdf
industrial.panasonic.com/www-data/pdf/ABA0000/ABA0000CE108.pdf

You would want the voltage rating on the caps to be significantly higher than the battery voltage. Same voltage rating as the manufacturer installed caps. Higher voltage rating is no problem. (thanks jeffs555).
If you try it with longer wires and no extra low esr caps it may work for a while, but the longer wires put an extra load on the original cap. The extra load shortens the life of the original cap and it will eventually fail, probably catastrophically. (thanks jeffs555, from www.rcgroups.com/forums/showthread.php?p=11996840#post11996840.

Capacitor polarity/orientation!
Electrolytic capacitors, like batteries, have a (+) and (-) lead! Solder them in the wrong way and they will got hot, pop open or even explode. Nasty fumes and stains from the liquid. Don't ask how I know

How & where (not) to add extra capacitors
The capacitors compensate for the effect that wire length (induction) has on the ESC. Therefore all capacitors as close as possible to or on the ESC board, direct across plus and minus terminals.
Do not distribute them along the battery wires! (noise_suppression_101)
Capacitors near/on battery are useless, the battery itself already is a huge capacitor.
By distributing them along the wires between battery and ESC they will be less efficient, or even useless.
This would only add extra weight, take up extra space and it would only introduce extra points of failure. ...

www.matthias-schulze-elektronik.de/guide/gfutc-de.pdf (English)
YGE controllers, extra capacitors and their location, installation manual, nice pics, click to enlarge:
www.yge.de → products → accessories

DIY pictures

Capacitor & pack suppliers
CapacitorsComplete packs

Expert opinions & their rules of thumb
They all say the same ...

Cause, explanation, water hammer/knock analogy & video, theory
First a watery analogy: water running in a pipe and through a tap. Now turn off the tap quickly. You may hear a loud knock/shock sound (voltage spike) in the pipe (wire). The water (current) wants to continu flowing, but it can't, for a moment the water pressure (voltage) is much higher than the static water pressure (voltage). It's the same for a current that's switched off, because of the inductance it wants to keep on flowing, voltage gets higher. This is also what causes sparks (brush fire) in a brushed motor.
wikipedia...Water_hammer

Water hammer video
I like the instantly-stopping-a-moving-freight-train-at-the-front analogy in this video.
Also notice the location of the anti-surge device.

The controller is like a watertap that's switched on and off very fast (8,16, 32kHz PWM voltage chopping) to get the desired current. Turning off the current, in combination with the battery wire inductance, causes voltage spikes because the current wants to continue on its course (ref. inertia of the moving watercolumn). These voltage spikes are higher!!! than the battery voltage. The input capacitors (cylindrical aluminium 'barrels') take care of these spikes (they reduce the wire inductance). The longer the wires, the higher the voltage spikes induced in the wires, the harder on the input capacitors. They will get warmer, heat up and explode and the rest of the controller will feel the full brunt of the voltage spikes. This is caused by the wire inductance, not by wire resistance. Therefore, using thicker wire will not help much, it's not a bad idea either, but extra capacitors are the solution, thus reducing/compensating the wire inductance. Or longer motor wires instead of long battery wires.
More on voltage spike and switched inductive load
PWM chopping - Wikipedia
Lenz's law - Wikipedia
Flyback diode - Wikipedia

1 - Full throttle, no PWM--2 - Part throttle, PWM chopping

Scope traces from www.aerodesign.de/peter/2001/LRK350/index_eng.html
Click to enlarge.

References
1. hyperphysics.phy-astr.gsu.edu...indeng.html#c2
Shows that energy stored is proportional to inductance, L, and proportional to current squared, I²
2. www.ee.scu.edu/eefac/healy/indwire.html
Shows that the inductance of a straight wire is proportional to the length of the wire times the natural log of the length of the wire ...which is close to being linearly proportional.
(References thanks Panther3001, post #685)

Measurements & scope traces

Vriendelijke groeten Ron
brushless motor building tips & tricks
diy brushless motor discussion group

cap caps capacitor capacitors heat heating overheating hot hotter pop popping burst bursting melt melting molten explode explosion explosive

### Images

Last edited by Ron van Sommeren; Dec 11, 2018 at 04:57 PM.
 Nov 12, 2008, 06:58 AM Crash Master Odd that in the industrial drive world we use input reactors to assist knocking the spikes from the incoming line. I've yet to see this on any ESC, but I have seen it on servos and BEC's. Is there any manufacturer who adds or suggests a torroid on the input lines?
 Nov 12, 2008, 08:58 AM homo ludens modellisticus Thread OP That would only increase the problem because induction in the lines causes the spikes. Industrial controllers are AC voltage fed, our controllers convert DCinto some kind of AC by chopping. The chopping is the cause of the long battery wire problem. What are 'input reactors'? Vriendelijke groeten Ron Last edited by Ron van Sommeren; Aug 11, 2015 at 12:32 PM.
 Nov 12, 2008, 10:23 AM Reduce the drama... Here's a PDF from a manufacturer http://www.ab.com/drives/techpapers/...%20Drives1.pdf
 Nov 12, 2008, 10:25 AM Reduce the drama... And a presentation here http://www.mtecorp.com/tutline/reactors.pdf Says they are essential for proper operation of Variable frequency drives
 Nov 12, 2008, 10:27 AM Reduce the drama... And finally one more article http://www.idscontrols.com/publicati...ansformers.htm
 Nov 12, 2008, 07:10 PM Registered User Industrial drives are used on 3 phase usually.The rest are usually AC of some kind. Ours are DC. That may be the change point. Chopping DC into the 3 phases. Industry is trying to stop excessive inrush currents. Use a inductor. DC.... capacitors...... slow down a voltage rise on the cheap drive parts used.
Nov 13, 2008, 03:25 AM
Crash Master
Quote:
 Originally Posted by cyclops2 DC.... capacitors...... slow down a voltage rise on the cheap drive parts used.
Which a series inductor would do, along with the capacitor across the input... An 'L' filter.

In a VFD, most large drives use a DC choke / reactor in series with the bus capacitors to help filter the bus voltage ripple and reduce the reflected voltage notching on the incoming line due to current pulses.
 Nov 14, 2008, 06:35 AM homo ludens modellisticus Thread OP The RCGroups links were dead, fixed it. Prettig weekend Ron
 Nov 14, 2008, 09:32 AM homo ludens modellisticus Thread OP Excellent Roger, I added a link to your post to my first message, for copy/pasting the text in the future. Prettig weekend Ron
Nov 14, 2008, 04:26 PM
Registered User
Quote:
 Originally Posted by Ron van Sommeren This goes for all makes, they all use the same principle (except www.sinusleistungssteller.de , they use sinusoidal commutation instead of trapezoid, they tested 70 meters without capacitors).]
Ron: The MaxCim line of motors [unfortunately no longer in production ] did not have any known limitation on battery-side wire length. Whether the fact that they are sensored makes that difference, I don't know. But for this reason, I'll never part with any of mine. Tex.
 Nov 15, 2008, 07:08 AM Crash Master Well put Roger. I know we fight the 'distributed inductance' of long line and load leads with AC-sourced VFD's routinely. The typical 1st level solution is to concentrate a large inductance close to the drive, thus limiting the current change... but again, this is for AC. In the DC section, an inductor is used in conjunction with the capacitors to limit ripple currents. But, of course, AC and DC circuits have different problems and solutions
Nov 15, 2008, 07:16 AM
homo ludens modellisticus
Quote:
 Originally Posted by Sabrejock Ron: The MaxCim line of motors [unfortunately no longer in production ] did not have any known limitation on battery-side wire length. Whether the fact that they are sensored makes that difference, I don't know. But for this reason, I'll never part with any of mine. Tex.
The switching (coils off) causes the problem, this is not influenced by sensored or sensorless.

At WOT, life is much easier on the ESC and the input capacitors, than at mid-range. There's not so much switching (PWM chopping) going on in the ESC.
https://www.rcgroups.com/forums/show...postcount=2661
http://www.aerodesign.de/peter/2001/...DY-BL_eng.html
http://www.aerodesign.de/peter/2001/...er_so_eng.html

Prettig weekend Ron

ESC output at WOT