Who adds capacitors to their power train

[Carapau]

Most ESC manufacturers recommend you keeping the battery to ESC cable length under a certain amount- approx 20cm ie from where the cable exits the battery's heat shrink to the ESC. If you go over this then they recommend adding extra capacitors. So, most large MRs I have seen do not have any extra capacitors but the power leads are normally well over 20cm by the time they exit the battery, go to the power distribution board and then into the ESC. So, who adds capacitors, who doesn't and what are your thoughts?

 

[Benjamin Kenobi]

I don't, never thought of it.

 

[SoCal Blur]

From what I've read is that not adding capacitors in that situation can result in premature failure of ESCs over time. I think its a good precaution, certainly.

 

[maxwelltub]

I heard this too, but all the reading I have done is in reference to single engine prop planes. Not sure but I think having multiple motors that are firing their poles at different intervals might help alleviate this issue. But I am far from an expert on this. On this build site Blayne adds extra capacitors and has a neat way of doing it. http://gimb.al/

 

[fpvroo]

Most ESC manufacturers recommend you keeping the battery to ESC cable length under a certain amount- approx 20cm ie from where the cable exits the battery's heat shrink to the ESC. If you go over this then they recommend adding extra capacitors. So, most large MRs I have seen do not have any extra capacitors but the power leads are normally well over 20cm by the time they exit the battery, go to the power distribution board and then into the ESC. So, who adds capacitors, who doesn't and what are your thoughts?

I don't think it's too much of an issue, because multicopter esc are usually oversized (10A hovering, 40-60A ESC), so the capacitors are already quite big. Also the ESC's are in parallel, so they smooth each other out in a way. Also multicopter usually use low kv motor which is easier on the ripple current. And the internal resistance of todays components is much lower than it used to be (lipos, esc's, motors).
Also large lipos have got lower resistance as well..

That way there is a lot of headroom...

 

[Carapau]

Current draw is not necassarily the issue. Rapid changes in current can also cause the issue and this is where MRs can exacerbate the problem as far as I can tell.

Ben, I guess you havent added them because on the S800 there isnt the space and one would hope (perhaps a little bit optimistically) that DJI would have designed this into their ESCs.

 

[Benjamin Kenobi]

Quite right mate. Haven't thought about it because I don't have to only having a little S800. The leads are less than 21cm.

Im surprised we haven't seen this issue discussed before if its recommended by manufacturers.

 

[fpvroo]

Current draw is not necassarily the issue. Rapid changes in current can also cause the issue and this is where MRs can exacerbate the problem as far as I can tell.

Ben, I guess you havent added them because on the S800 there isnt the space and one would hope (perhaps a little bit optimistically) that DJI would have designed this into their ESCs.

Yeah you are right it's the current ripple that you get, that can be an issue. But it's more likely an issue when the esc is working near it's limit. Because then during a peak current moment the voltage drops too low (if the internal resistance of power train is too high) and the esc control circuit and Bec (if there is one) running on 5v or 3,3v shuts down (esc failure).

 

[Electro 2]

Always, but on the radio side of the FC's BEC. With 40 years of electrical design behind me, I am unaware of a reason the input side of ESC would have to have a short length cable. No circuit model supports this. If the wire is of adaquate size, (gauge), the physical source impedance of the battery itself dominates the input side of the equation. This would be true for AC as well as DC. I put the ESCs on the booms underneath the motors, zero issues to date. Over 500 flights on two aircraft over 18 months.

 

[Bartman]

the way I understand this is that the electrons have momentum like water flowing through a pipe. the ESC's are basically extremely high frequency switches that open and close the flow of electrons to the motors gazillions of times per second and each time the ESC's close the switches the electrons bang up against the switch like water knocking in a pipe. the element within the ESC that feels the momentum of the electrons is the capacitor with it feeling the spike and flow of electrons in motion a gazillion (or more, yikes!) a second. it is said the capacitors have a finite ability to withstand these electrical spikes and that they will eventually fail. the term thrown around for the effect is "inductance" and a lot of people buy into it.

my personal guess is that a one wire, one switch system would feel this effect the worst but that the multiple paths from center that we have in our circuits act as big capacitors of sorts adding to the circuits' ability to absorb fluctuations. i've tried esc's out at the motors and haven't seen any related failures but i don't have high hours or extreme use with which to judge.

i'd certainly welcome a more educated explanation as to why this is or isn't something to be concerned about.

thanks,
bart

 

[Electro 2]

the way I understand this is that the electrons have momentum like water flowing through a pipe.....
...the term thrown around for the effect is "inductance" and a lot of people buy into it.....

This is correct. I don't think it's the best analogy, but it's basically correct. That said the only inductance present here of any significance is the motor coils themselves. The interconnecting wires, on both input and output sides, are of waaaaay too low an inductance to matter in this circuit model.

my personal guess is that a one wire, one switch system would feel this effect the worst but that the multiple paths from center that we have in our circuits act as big capacitors of sorts adding to the circuits' ability to absorb fluctuations. i've tried esc's out at the motors and haven't seen any related failures but i don't have high hours or extreme use with which to judge.

Reasonable length wires, on either side, simply aren't an issue. The DC losses from an excessively long run or too small a size wire, can be an issue, but can be dealt with. (Bigger wire!)

 

[gtranquilla]

That might help to explain why Quadrocopter USA is selling a $700 PDB/ESC package for the Cinestar 8 ARF kits! A close look reveals some small caps mounted on the PDB! But first of all, can you provide a source for your original statement re: 20cm restriction so i can investigate further.

 

[Carapau]

So if reasonable wires, on either side isn't an issue, why is that some of the top ESC makers all recommend keeping the battery to ESC cable under a certain length? Companies such as Kontronik (the best of the lot imo), Castle, Jeti etc? These guys use designers who really know their stuff and would not put some restriction like this for a reason surely?

 

[Bartman]

So if reasonable wires, on either side isn't an issue, why is that some of the top ESC makers all recommend keeping the battery to ESC cable under a certain length? Companies such as Kontronik (the best of the lot imo), Castle, Jeti etc? These guys use designers who really know their stuff and would not put some restriction like this for a reason surely?

the applications they deal with are going to be almost entirely single motor, single power circuit set-ups. i'm not an expert but i'd guess the single circuit model will be the most restrictive with the strongest effect and so require the most concern. don't quote me but I'd guess it has something to do with it. most 30A ESC's have single capacitors where 40's have two. it's a small weight penalty to go to the 40's but the extra cap may be worth the effort.

if anyone's interested i've got a few hundred black anodized aluminum motor mounts (compatible with Tiger MT 2814, 4008, MN4012, KDE 4012, and all similar sizes) with an arm built into them for an ESC to be mounted. i've also got a couple of thousand boom clamps for 7/8" tubing (or 19mm, almost identical size) molded from black nylon plastic. i can list them in the classifieds area if anyone wants to try mounting ESC's by the motors. the motor mounts were designed to be used for single motors or coaxial arrangements. these are leftovers from my abandoned effort to manufacture my XY8 frames and to run an online store (Aerdro.com)

bart

 

[Bartman]

pic if anyone's interested

 

[Electro 2]

That might help to explain why Quadrocopter USA is selling a $700 PDB/ESC package for the Cinestar 8 ARF kits! A close look reveals some small caps mounted on the PDB!

I put buckets of caps on the PDB, but not for the benefit of the ESCs. It helps to minimize/arrest, i.e. provide a low shunt impedance, any high-frequency AC trash that might find it's way into the electronics. As for the ESC manufacturers that require caps, I have no idea. Never used any other than whatever was soldered on the ESC board itself, and no issues of any sort.

 

[gtranquilla]

There is very little difference between the Hobby level ESC and the Industrial level VFD = Variable Frequency Drive. The only differences are: 1) Size and 2) the VFD has a built-in full wave rectifier on the front end to convert 3 phase AC input to the required DC voltage/current that is required by it's internal "ESC"! In the industrial world these VFD switch off and on so quickly that the create a lot of emi, electromagnetic interference. In other words, high frequency electricity tends to propagate in all directions and even continues into the air after reaching the end of a wire. This is the same basic concept used to produce a radio frequency carrier wave, i.e., a high frequency signal generator!

In summary it quite conceivable that emi from ESCs can travel back thru the battery power supply wires to the PDB. By adding the appropriate sized capacitors in the appropriate configuration it is possible to specifically sink only these high frequency low energy interference signals to a common point..... Unfortunately there is no true "earth" ground on an aircraft..... but a so-called "common" ground point still might have some value in helping to prevent emi from affecting the adjacent escs.

The optical isolation ESCs only prevent emi from travelling back through the servo signal wiring.

 

[coreyperez]

So, my question is, how do we know what type, size, etc capacitors to use in our build? I'm doing a 680PRO and a 960 or T15 in the future. I'd like to get all the stuff while in the states to take back to Korea to finish my builds.

Corey

 

[jes1111]

I guess the only way to truly understand the role of capacitors in the power feed is to remove them all and go flying - see what happens. I'm not going to try it with my rig - any volunteers?

I treat the capacitors like bearings - parts that are likely to wear out or destroy themselves - they get inspected before and after every flight (tell-tale bulging in the top of the cylinder). Also I replaced all the factory-fitted capacitors with top-grade Panasonics right out of the box. These capacitors are dirt cheap, even the very best ones, so why worry? - have a stock of them and replace them on a regular basis. If you're running longer than usual power wires you can add additional caps at the ESC end or replace them with higher voltage rating (which are physically larger). There's a slight advantage in using multiple capacitors (rather than a single larger one) - they'll dissipate their heat more effectively.

Panasonic FR series ("Low ESR" type) are the ones to go for - capacitance value around 330-470uF (the least important factor), voltage rating (the most important factor) 35V or more, I went with 50V - the higher the better (but this is what determines the physical size of the capacitor - higher voltage rating ones are larger). http://uk.mouser.com/ProductDetail/...=sGAEpiMZZMtZ1n0r9vR22f2plZjWOZXTu0LqHMfIqXE=

 

[Bartman]

fwiw, there is a mention in the new Mikrokopter BL 3.0 speed controllers that references a feature to make mounting them at the ends of the booms possible.

I2C-Bus with

integrated optocoupler allows you to put the BL-Ctrl to the end of the rigger

i don't know why or how an opto coupler would make this any different but it's interesting that they specifically mention this detail in the wiki entry. it actually says "opocoupler" and I corrected it to "opto", maybe it is an "opocoupler"? it's above my pay grade