PDA

View Full Version : How to size capacitor for Speed 600


jackhy
12-10-2005, 09:31 PM
My Graupner Speed 600 motor says it is suppressed, but to eliminate radio interference one should install a capacitor between the poles to eliminate possible radio interference. This sounds contradictory to me and no mention of capacitor size is indicated or in the Hobby Lobby catalog. What size capacitor should I use on this motor also using a 30A ESC? Also on a prior plane using the same type motor, a 250V capacitor apparently failed and I blew out a 40A ESC. Electric planes ain't very simple or better than nitro I feel! My first post, Jackhy

phat23
12-10-2005, 11:51 PM
If I remember right, the Graupner 600 has internal caps. It wouldn't hurt to add an external one across the poles. Use 47-100 nanofarad ceramic disk type.

jackhy
12-11-2005, 03:57 AM
to phat23: Many thanks for the capacitor info for my speed 600! Great quick response to my first post! Jackhy

slipstick
12-11-2005, 10:46 AM
Just for future reference, the size of motor has no effect on the capacitor needed. The same 10-100nF ceramic disc capacitors will work on any brushed motor.

Steve

crash_alot
12-11-2005, 12:05 PM
If I remember right, the Graupner 600 has internal caps. It wouldn't hurt to add an external one across the poles. Use 47-100 nanofarad ceramic disk type. My Graupner 600s do not have internal capacitors. The values of supression capacitors are .01 to .047 uF (microFarad) not nF. NanoFarad capacitors are way too small for this purpose.

jackhy
12-11-2005, 01:42 PM
TO- Slipstick and Crash alot, Steve: Thanks to you also for capacitor advice and your 10-100nf seems to contradict with reply from crash-alot who says NF caps are too small. So a 47-100 (.047) MF must be a larger cap. than a similar 47-100 nanofarad type? Also the Graupner box states that suppressers are built in, but external caps are recommended. With past failures, I'd rather not fly the 600 without an external cap. or a recommend size range!

crash_alot
12-11-2005, 02:43 PM
TO- Slipstick and Crash alot, Steve: Thanks to you also for capacitor advice and your 10-100nf seems to contradict with reply from crash-alot who says NF caps are too small. So a 47-100 (.047) MF must be a larger cap. than a similar 47-100 nanofarad type? Also the Graupner box states that suppressers are built in, but external caps are recommended. With past failures, I'd rather not fly the 600 without an external cap. or a recommend size range!
1uF (microFarad) = 10 to the -6 Farads
1nF (nanoFarad) = 10 to the -9 Farads
1pf (picoFarad) = 10 to the -12 Farads

My Graupner 600s came with .047uF capacitors.
Capacitor markings of 104 = .01uF and 474 = .047uF

phat23
12-11-2005, 04:04 PM
Maybe this will clear things up.

From RCG E-Zone FAQ -

Q. How many capacitors do I need on the motor and what values?
A. EFlight and Ezone discussions
It is a good idea to fit at least 2 capacitors to most brushed motors to cut down on the radio interference which the motor may generate. The same capacitors can be used for all brushed motors, the value does not change with the size or power of the motor. The capacitors should be soldered from each motor terminal to the motor case. For extra security against interference you can also fit a third capacitor between the 2 motor terminals.
Note: many Graupner Speed xxx motors have the first 2 of these already fitted internally. They are not easy to see but you can tell if your motor has them because each terminal will have a small wire poked through it. These just need to be soldered when you solder the motor connections.
In all cases the value for these capacitors is not very critical. It is most commonly between .001uF (microfarad) and .1uF. Another way of saying these same values is 1nF (nanofarad) to 100nF. Some people have suggested values as low as .000047uF or 47pF (picofarad). What is important that the capacitors are CERAMIC types. These types give the most effective suppression at our radio frequencies. They should have a voltage rating at least equal to your battery voltage. This is not normally a problem since most capacitors of the right type have ratings of 100V or more.
The most commonly used value seems to be .047uF / 47nF for all 3 capacitors.

jackhy
12-11-2005, 06:03 PM
Thanks phat23, I believe your reply is the most complete I've seen especially for Graupner motors and the voltage reference (not read before). I hope others read this thread to solve the issue. I don't even want to ask what the deuce a "farad" is! It might be like Noah asking the Lord what a Cubit is when making the Ark!!!!!!!!!!!!!!!

Will Hicks
12-11-2005, 07:54 PM
Actually a disc capacitor marked 104 is .1 mfd. A disk capacitor marked 474 is .47 mfd.

The first two digits are the number value; the third digit is how many zeros are to be added to the first two numbers.

104
100000.0

By moving the decimal place over six places to the left, you'll arrive at the value of the capacitor in mfd; .1

crash_alot
12-11-2005, 09:06 PM
Actually a disc capacitor marked 104 is .1 mfd. A disk capacitor marked 474 is .47 mfd.

The first two digits are the number value; the third digit is how many zeros are to be added to the first two numbers.

104
100000.0

By moving the decimal place over six places to the left, you'll arrive at the value of the capacitor in mfd; .1
OOPS! Will is correct, I was off by an order of magnitude - sorry. Must have been a senior moment. The advantage of .1 vs .001uF is the ammount of brush noise energy absorbed by the capacitor. The .1 or .47 absorb more brush noise at lower frequencies than the .001. Your call, I use .1 and .47s.

slipstick
12-11-2005, 10:19 PM
My Graupner 600s do not have internal capacitors. The values of supression capacitors are .01 to .047 uF (microFarad) not nF. NanoFarad capacitors are way too small for this purpose.
What do you mean ?
10nF = .01uf
47nF = .047uF
100nF = .1uF

Capacitors are the same value whether you write that value in nF or uF. It's just that many of us find it easier not messing about with decimal places.

BTW every Graupner Speed 600 I've seen has had 2 internal capacitors fitted. As I said in that RCG EFlight FAQ entry I wrote years ago, they are not easy to spot.

Steve

crash_alot
12-12-2005, 01:29 AM
What do you mean ?
10nF = .01uf
47nF = .047uF
100nF = .1uF

Capacitors are the same value whether you write that value in nF or uF. It's just that many of us find it easier not messing about with decimal places.

BTW every Graupner Speed 600 I've seen has had 2 internal capacitors fitted. As I said in that RCG EFlight FAQ entry I wrote years ago, they are not easy to spot.

Steve
As I mentioned earlier, I screwed up. I was off by an order of magnitude and meant to say .1uF and .47uF. These values give better dV/dT performance than the ones you recommend. If Graupner 600s have internal caps, why did the 2 I bought have caps and installation instructions in the box?

slipstick
12-12-2005, 10:32 AM
As I mentioned earlier, I screwed up. I was off by an order of magnitude and meant to say .1uF and .47uF. These values give better dV/dT performance than the ones you recommend. If Graupner 600s have internal caps, why did the 2 I bought have caps and installation instructions in the box?
I think you'll find I recommended up to 100nF which is exactly the same value as the .1uF you recommend. If people want to use 470nF/.47uF it won't do any real harm, the precise values really make very little difference to the RF supression, but it's often quite difficult to find real ceramic capacitors that large. The type (ceramic) is vastly more important than the value.

I'd be interested to know the Graupner model # of your Speed 600s. I must have had at least 20 over the years and all mine have been internally supressed. I've had a few off-brand (not Graupner) 600s that have no supression but if there's a genuine Graupner like that I'll have learnt something.

Steve

crash_alot
12-12-2005, 03:33 PM
What do you mean ?
10nF = .01uf
47nF = .047uF
100nF = .1uF
-------------------------Steve
Steve,

My recommendation is .1 to .47uF (100 to 470nF) not .01 to .1uF as your post indicates. Also, your FAQ mentions .001uF, really not adequate for this application.

Looks to me we are dealing with some cultural issues here. Here in the US capacitors of this type are specified by their manufacturer in microfarads not nanofarads. The folks who supply these devices specify them in uF not nF. Also, capacitors other than ceramic can be used.
See the following:http://dkc3.digikey.com/PDF/T053/1080-1081.pdf

Sorry, I toasted my Graupner 600s some time ago and no longer have them. I do have some of the supression capacitors that came with those motors. I have attached a picture of the Graupner cap (big round yellow metal film and not a ceramic) and polyester .1 and .47uF supression caps.

Matt Kirsch
12-12-2005, 03:55 PM
Is the size of the capacitors that critical? I've always understood that to not be the case, that the size really didn't matter. Only the presence of the capacitors is important, which is why the actual values of the capacitors is not really specified.

The capacitors don't actually conduct electricity. That's why you don't get a dead short by adding the capacitors from each terminal to the motor's case, and directly across the two terminals.

crash_alot
12-12-2005, 04:56 PM
Is the size of the capacitors that critical? I've always understood that to not be the case, that the size really didn't matter. Only the presence of the capacitors is important, which is why the actual values of the capacitors is not really specified.

The capacitors don't actually conduct electricity. That's why you don't get a dead short by adding the capacitors from each terminal to the motor's case, and directly across the two terminals. Matt,

A capacitor is basically 2 plates of electricaly conductive material seperated by an insulator. It has infinate resistance to DC currents. It is this property that lead to the term "blocking capacitor." They do however pass AC currents and drop AC voltages. Motor brush noise is complex AC voltage and current. Capacitors have a property known as capacitive reactance (Xc). Xc =1/(6.28 * Frequency * Capacitance). Xc is expressed in Ohms and directly substituted for R in Ohm's Law. For example: E = I * Xc. In this case E is in AC volts and I is in AC amps.

I would not say supression capacitor values are critical but, they should not be taken too lightly. Without doing the mind numbing math, a .1uF capacitor has lower Xc than a .001uF at any given frequency. As a result, a .1uF absorbs more brush noise than a .001uF. Generally speaking in this application, more is better. There are however, practical limits. Hence the range limit of .1 to .47uF.

Supression capacitors provide protection against 2 things. The first is interferrence from motor brush noise. The other is preventing the destruction of an ESC's motor drive MOSFETs. In older technology, like relays and bipolar transistors, motor brush spikes were of less concern and .001uF caps served well. Today's MOSFETs are high current, high speed, and more sensetive to brush noise spikes. The Xc of a .1uF slows down a fast spike better than a .001uF. I have designed and built my own brushed ESCs and have smoked my share of MOSFETs due to supression capacitor issues.

Hope this sheds some light on the subject.

slipstick
12-12-2005, 06:34 PM
Fortunately this sort of bypass application is not value critical. Over the years I've used a range of values from .001uf/1nF to .1uF/100nF and found that almost anything works. I settled on .047uF/47nF myself purely because there's no point stocking a large range. I always use ceramics not polyester or any other type simply because their performance at RF frequencies is better and I'm interested in supressing radio interference not interference at some other frequency.

If you really want to get into the techy stuff:
The total impedance of a real capacitor is considerably more complex than just Xc. Real world capacitors also have inductive reactance Xl and Equivalent Series Resistance ESR to be taken into account. With increasing frequency Xc reduces but Xl increases. The net effect is an impedance curve that is always notch or v-shaped with a minimum at a particular frequency (called the FSR - Series Resonant Frequency) and increasing impedance at higher and lower frequencies. Increasing the capacitance does usually lower the minimum impedance (which is equal to ESR) but also lowers the frequency at which it occurs. For RF supression we ideally want this FSR/low impedance somewhere around our radio frequency because that's the main noise we're trying to get rid of. But because of this notch effect a capacitor value that's too high may end up with a very low FSR and actually result in worse interference rejection at the real frequency of interest than would a lower value.

Steve

crash_alot
12-12-2005, 09:21 PM
Fortunately this sort of bypass application is not value critical. Over the years I've used a range of values from .001uf/1nF to .1uF/100nF and found that almost anything works. I settled on .047uF/47nF myself purely because there's no point stocking a large range. I always use ceramics not polyester or any other type simply because their performance at RF frequencies is better and I'm interested in supressing radio interference not interference at some other frequency. In my post I said I would not call these capacitor values critical. You address only the issue of RFI. I address both RFI and MOSFET protection. In reality, aside from the difference between .01 and .1, we are in violent agreement.

If you really want to get into the techy stuff:
The total impedance of a real capacitor is considerably more complex than just Xc. Real world capacitors also have inductive reactance Xl and Equivalent Series Resistance ESR to be taken into account. With increasing frequency Xc reduces but Xl increases. The net effect is an impedance curve that is always notch or v-shaped with a minimum at a particular frequency (called the FSR - Series Resonant Frequency) and increasing impedance at higher and lower frequencies. Increasing the capacitance does usually lower the minimum impedance (which is equal to ESR) but also lowers the frequency at which it occurs. For RF supression we ideally want this FSR/low impedance somewhere around our radio frequency because that's the main noise we're trying to get rid of. But because of this notch effect a capacitor value that's too high may end up with a very low FSR and actually result in worse interference rejection at the real frequency of interest than would a lower value.

Steve Extremely low inductance .1uF supression capacitors with good VHF performance are available. I can send you data sheets if you like. They are a little on the large side though. I stock a wide range of many different capacitor types. Not a real big problem there. The supression caps I use are a good compromise among RF supression, ESR, and price. All of the brushed motors I have (GWS, Graupner, Maubichi, and a couple others not coming to mind right now) either have or recommend .1uF or larger. In your impedence analysis, what do you do with motor inductance? Just a thought.

slipstick
12-13-2005, 12:26 PM
In my post I said I would not call these capacitor values critical. You address only the issue of RFI. I address both RFI and MOSFET protection. In reality, aside from the difference between .01 and .1, we are in violent agreement.
Yes as usual we're quite close. On the rare occasion when I've had problems with MOSFET punch-through I've always found Schottky diodes get rid of the spiking very effectively but in my experience most modern MOSFETs are sufficiently protected by their internal diodes. That leaves the caps to deal with just the RFI.

All of the brushed motors I have (GWS, Graupner, Maubichi, and a couple others not coming to mind right now) either have or recommend .1uF or larger. In your impedence analysis, what do you do with motor inductance? Just a thought.
Oh I've never gone as far as doing a complete mathematical circuit analysis, that's way too much like work.

BTW I'm not sure what everyone recommends but the internal caps fitted to at least some of the Graupner-rebadged Mabuchi motors are .047uF/47nF. That's another reason why I chose that as my standard value, I guessed Mabuchi knew more about it than I did ;). Where did you find the recommendations ? I had a look on the GWS site and was amused to find that the FAQ has just got the first paragraph stolen from the RCG FAQ entry that I wrote, but missing the part about values. It's also got the complete text of several other of my RCG FAQ entries....I guess that's a sort of fame :(.

Steve

crash_alot
12-13-2005, 03:48 PM
slipstick,
Try this: http://rocky.digikey.com/WebLib/BC%20Components/Web%20Data/2222%20338%202%20Interf%20Sup%20Film%20Cap.pdf (http://rocky.digikey.com/WebLib/BC%20Components/Web%20Data/2222%20338%202%20Interf%20Sup%20Film%20Cap.pdf)

There is another manufacturer, Murata I think, with similar specs but, I seem to have lost the link.

I have made PCBs that attach to the back of SP400 motors to allow the use SMD capacitors. I found some surface mount parts with good looking inductance and frequency response characteristics. BTW - I put a complete ESC on a PCB the size of a SP400. Had better eye sight back then.

slipstick
12-13-2005, 05:53 PM
Yes, nice caps and made in a range all the way down to 1nF ;).

But what I was asking was where you found the motor manufacturers recommendations. It just seemed odd that Mabuchi would recommend .1uF or greater and then fit .047uF.

I haven't bothered making my own ESCs for years now and of course all this debate will be pretty much irrelevant soon because brushless motors/controllers are getting so cheap that brushed motors in EFlight will be restricted to RTF toys.

Steve

batman
12-14-2005, 06:37 PM
"On the rare occasion when I've had problems with MOSFET punch-through I've always found Schottky diodes get rid of the spiking very effectively but in my experience most modern MOSFETs are sufficiently protected by their internal diodes. That leaves the caps to deal with just the RFI."

Does this mean that it no longer nesessary to fit shotky diodes to brushed motors?

slipstick
12-14-2005, 06:43 PM
If the ESC comes with a Schottky then obviously you should fit it (I haven't seen one like that for a few years but they may still exist). If it doesn't then the only time I might bother is if I have lots of wire between the ESC and the motor, particularly on twins.

But that's just me, I expect you'll get plenty of other opinions.

Steve

crash_alot
12-14-2005, 10:32 PM
"On the rare occasion when I've had problems with MOSFET punch-through I've always found Schottky diodes get rid of the spiking very effectively but in my experience most modern MOSFETs are sufficiently protected by their internal diodes. That leaves the caps to deal with just the RFI."

Does this mean that it no longer nesessary to fit shotky diodes to brushed motors?
Schottky diodes are not necessary. They are a good idea, but not necessary. Modern MOSFETs have a reverse p-n junction diode connected internally across the transistor's source and drain for protection. The gate also has ESD protection diodes. The ESCs I design and build have Schottky diodes across the motor terminals built in. I suspect ESC manufacturers do the same thing. An ounce of prevention is worth a pound of cure.

fabricator
12-15-2005, 12:10 AM
Holy Moses! You guys are scaring the crap outta me! the only thing I know about Schlotzky's is the deli went out of buisness.:eek:

crash_alot
12-15-2005, 12:40 PM
Holy Moses! You guys are scaring the crap outta me! the only thing I know about Schlotzky's is the deli went out of buisness.:eek: fab - all is well no need to be scared. Life is good. There are other Schlotzkys!

"It is alive!" - Bill Murray

fabricator
12-15-2005, 10:18 PM
fab - all is well no need to be scared. Life is good. There are other Schlotzkys!

"It is alive!" - Bill Murray

I feel better now.:p