What is the problem

[PatMc]

Posted by Erfolg on 10/09/2012 08:43:10:

How would I check for a short?

Measure the ohms across each coil, looking for an inbalance?

Or something else?

The resistance is so low that a normal ohmmeter wouldn't measure the difference between a dead short & normal.

If there was a short the motor wouldn't run.

[PatMc]

Posted by Erfolg on 10/09/2012 08:43:10:

How would I check for a short?

Measure the ohms across each coil, looking for an inbalance?

Or something else?

The resistance is so low that a normal ohmmeter wouldn't measure the difference between a dead short & normal.

If there was a short the motor wouldn't run.

If you want to check for an imbalance you could drive the motor with a fixed rpm electric drill at the same time measure the generated voltage across each coil in turn. You would need a meter capable of measuring very low AC for this.

[PatMc]

Posted by Erfolg on 10/09/2012 08:43:10:

How would I check for a short?

Measure the ohms across each coil, looking for an inbalance?

Or something else?

The resistance is so low that a normal ohmmeter wouldn't measure the difference between a dead short & normal.

If there was a short the motor wouldn't run.

If you want to check for an imbalance you could drive the motor with a fixed rpm electric drill at the same time measure the generated voltage across each coil in turn. You would need a meter capable of measuring very low AC for this. Most digital multi meters would do. The voltages will vary but not by very much.

This still won't pick up the situation of a strand or two of a multi strand winding being open circuit. But in this case the motor would run very hot below it's normal max current rating.

[WolstonFlyer]

I would check the resistance across the 6 different pair combinations of the three wires, they should all be very low resistance and all the same or very close. Also test for continuity between each wire and the metal cage (if you can find a spot that is bare metal), there should be no circuit. If there is then there is a short somewhere.

 

You could spin the motor up using a battery drill (very carefully) with your meter( set for "AC volts" ) on any two leads and you will see a small AC voltage...like .125v or perhaps lower. If you have nothing on a pair of leads, then you have an "Open" winding...

 

 

A different simple way is to short the wires together (with no battery) while turning the motor by hand. You should feel a bit of an increase in turning resistance when you short any two wires together and even more resistance with all three shorted. If you don't feel any difference then a winding is either already shorted or has an open circuit.

Edited By WolstonFlyer on 10/09/2012 10:40:23

[Toni Reynaud]

Hello all,

A nice explanation of when an ESC gets hot and why may be found at http://www.giantshark.co.uk/forum/viewtopic.php?f=15&t=689

Full throttle is cool, part throttle is hot.

Toni

[Erfolg]

I have tried to measure the resistance.

The only reading I could get was on the 2k scale of 0.001 on scales such as 2m get a result of 0.

Is it possible that the motor is a source of a problem, or do the results perhaps indicate that it is OK to use. Changing to a different type of motor is a bigger job,

[WolstonFlyer]

The k stands for kilo or 10^3 and M is Mega 10^6

So your 0.001 is 1ohm, I guess a more accurate meter is needed as PatMc suggested.

Do all of the combinations of pairs show the same 0.001 ?

I have just realised that I have the exact same Keda Thumrun 2834/18 motor sat in a box at home so can see what readings I get from it later tonight it you want me to try?

[Erfolg]

yes, all the same

Any help much appreciated.

[Chris Bott - Moderator]

Erfolg two thoughts spring to mind that I don't think have been mentioned before.

I don't like coincidences either so would be looking for something causing the ESC's to go.

So:-

1. Have the battery to ESC leads been extended at all? Sometimes long battery leads can be a problem. This is made worse of course while a wattmeter is in line.

2. Could there be an intermittent short in the motor? There are lots of insulated strands of copper crossing over each other, possibly rubbing against each other and rubbing through the insulation. I've had one motor which started acting strangely. I could eventually see very slight sparking when viewed running in a dark room.

[WolstonFlyer]

I will give it a go later although I am not sure if my meter will be accurate enough either. I would need to use a 4-wire (Kelvin) resistance measurement rig.

In the long run it might work out quicker to spend £15 and try a new motor especially as you have swapped the ESC a couple of times already.

 

Edited By WolstonFlyer on 10/09/2012 13:41:23

[Peter Beeney]

WolstonFlyer, - Thanks for your answer. It’s a bit long-winded, I must admit, but I try and start at the beginning, so to speak, and then make it up, ‘on the hoof’. I have try and convince myself that I know what I’m talking about, and most of the time that is not easy……
And yes indeed, we had a scope at work, but unfortunately this was one of the items of interest that I was not allow to keep when I retired; and that’s a long time ago, now…. Maybe I will purchase one one day, purely from the hobbyist point of view.
I’m sure it’s possible there are some good examples of oscilloscope traces of motor currents on the web somewhere, but I haven’t looked.

Maybe, as Erfolg sort of suggested, we should strike a thread that we can throw all this stuff around in without fear of conflicting with the Original Post. If we started by including a description about ‘How Electric Motors Work‘ we could possibly get two or three different ideas at least, par for the course, which we can then pick the bones out of. We could call it something flamboyant, such as ‘The Squirrel’s Nuggets’, (is that a pub?), perhaps, on the assumption that there might be some useful kernels of information squirrelled away in there, also a title which might tend to stick for future reference. Might be useful, perhaps some knowledge can be helpful, inasmuch you don’t have to know anything about the clutch to drive a car, for instance, but if you do, you might tend to generally treat it with respect. Also if it starts to fail you are instantly able to recognise this.

Pat, - Many apologies for my rather tedious tirade, but it’s as I always think when the next piece of junk mail turns up, do I really need to read this? As I said, we shall still have to amicably agree to disagree here, but I’m fairly happy with my theory. If you are as happy with your explanation then to make it more interesting, and short, sweet and to the point this time, I’m prepared to put my big money where my mouth is, and say that I’ve got some cash that's saying that the current in the armature of a brushed motor is AC when the motor is running; but that’s only because it’s easiest to explain. For the record, that’s observing the current at a given point, but that could be anywhere on the armature, and by AC I mean the current is first going in one direction and then the other for half of each complete revolution of the the armature; … which is of course what AC always is anyway, just DC going in the other direction from time to time.

And with the utmost and greatest respect, Pat, do you think it might be worthwhile having a small recap here? Could save yourself a chunk of money, who knows? Or I might be able to buy an oscilloscope, perhaps?

Totally without prejudice, please, I’m really only wishing to make the point that there maybe an alternative answer to this question. And in my simple way, I consider that if I start off by having the wrong ideas about the workings of these things, it certainly might be quite difficult to answer the next question down the line.

PB

[WolstonFlyer]

I agree about a new thread. The theory stuff is not helping the original post.

I am lucky to work somewhere that has a lot of very clever electronics & engineering people (a Uni) so I am just arranging access to a scope and trying to find a willing electronics expert to help (I work in administration IT so don't normally use any of the labs).

If I can run the experiment and get nice results I will post results on a new thread.

[Chris Bott - Moderator]

I would expect the waveform to look a bit like a modulated RF waveform. The outer "envelope" running at a frequency that varies with motor speed. With a much higher rate PWM signal running inside this envelope. The on off ratio of the high rate PWM controlling the effective voltage applied, with the much lower frequency being the actual switching between windings.

[WolstonFlyer]

I am half expecting something like this... perhaps. With the frequency increasing with throttle setting.

Anyway leaving it alone until (if) I can do the tests with a scope.

[Chris Bott - Moderator]

I agree WF, I see something like that, but with that waveform chopped up with a high frequency PWM inside your waveform.

Your 3ph waveform might be at say a few hundred Hz, with the PWM running at say 10KHz.

Edited By Chris Bott - Moderator on 10/09/2012 17:12:35

[Chris Bott - Moderator]

For those that really can't sleep at night I've found a long document about controling brushless DC motors.
Of interest is fig 11 showing some waveforms.

Brushless motor control made easy (easy if you're that way inclined!)

To be honest, we really don't need to know any of this and we're way off thread here.
If you would like me to Erfolg, I could try to separate out the two discussions here into two threads?

[Erfolg]

I am happy with things as they are Chris.

The ESC discussion is interesting in its own right.

[Peter Beeney]

Erfolg, - My very humble apologies, I’m afraid I rather hijacked your thread somewhat in a rather ungentlemanly way. I certainly didn’t really intend to do that!

It seems to me the 3 phase brushless motor is a bit complicated to sort out in the first instance, so as you suggested, would there be any milage in starting with the very basic stuff and working upwards? In a separate thread, of course. There are occasional questions about BEC’s, linear and switched-mode, ESC’s and motors, so I’m sort of thinking that not everyone is totally familiar with these things. But, by the same token if there is only little interest it may not be worth doing, I’m not sure it’s possible to fully describe all these functions in a few sentences; it takes at least a small but finite amount of time to write a post. Although I personally think it’s quite interesting it can also be perplexing, permanently so in my case it seems, from the recent evidence here we have a modicum of confusion just with AC and DC; and yet without these very fundamental but vital components no electric models would be possible. You will now be able to ponder at length the fact that when your elevator servo moves briefly to make a small correction to flight, the spinning motor will be subjected to DC and AC at the same time!

So, would this be a live wire, with a few shocks along the way, or simply become just a dead duck?

I have no idea.

PB

[Radge]

Oooh! 30amp Mystery Pentium ESC's seem to have a batch fault(allegedly), I've just used one (new unused) on a motor test, correctly set up and checked it would only allow the motor to start at half throttle then give a 1 second burst and die. Not a set up or programming fault it went in the bin. I googled the fault only to discover many other modellers have had a bad news experience with these ESC's. Pointless reporting the fault to 'Janet' in Hong Kong(why do they always use English female names?) as a replacement would cost less than the return postage.

[WolstonFlyer]

Ahh so could be a batch of faulty ESC's then!

Just for anybody following this for ESC info out of academic interest.

This PDF from Microchip Technology Inc has some great information about how our ESC's work. 

Edited: Oops, this is the same document that Chris has linked to a bit earlier.... should have read all of the above posts properly first, sorry.

Edited By WolstonFlyer on 11/09/2012 16:30:57

[Erfolg]

My Pentium ESC is not new. I think it was purchased from either GC or Robotbirds, some years back. It had been used a number of times, probably about 20.

I have now wired in a HobbyKing Blue 30A, I have used a number of these devices successfully to date. I have run the set up a few times now in testing. The amps drawn are quite low, as are the watts at about 145w @14A with 10v.

I keep charging the Lipo to 4.2v per cell or 12.6v. On load the voltage is sagging down to 10-11v instantly. There appears to be a slight puffiness of the plastic bag making up the cell. Not puffed as I have seen, yet not completely flat. Being about 2 years old, I guess it is nearing its lifes end.

[Swissflyer]

Posted by Toni Reynaud on 10/09/2012 11:28:17:
A nice explanation of when an ESC gets hot and why may be found at http://www.giantshark.co.uk/forum/viewtopic.php?f=15&t=689

Full throttle is cool, part throttle is hot.

Toni

Toni,

Thanks, the material quoted is the best description I ever saw of why ESC’s work harder (heat up more) at low throttle settings. Strange but true, thanks.

I try to oversize by 50% for exactly those reasons.

Erfolg, if your multi-meter is a modern DVM type, you may find it has a 500mV or similar measuring range. You are in luck if there is as you can check the low resistances in brushless motors very easily.

The trick is to put a 12V 20 Watt Halogen car light bulb in series with a 3S LiPo.
With this set up about 1.5-1.7 Amps will flow. Measure & write down this current.

Mostly BL motors have windings in the 10-100 milli ohm range and they are wired in a Y configuration. You can put your LiPo & light bulb in series to measure the voltage dropped across two windings with your DVM.

If there are no shorts, then all the voltages you measure on the three possible combinations will be the same.

I see your KEDA 2834/18 is specified at 77m Ohms (.077 Ohms)

**LINK**

Your LiPo and light bulb will supply a constant current (for practical purposes) & you will measure the current you wrote down (about 1.5A) x 77m Ohms x 2 = 230mV.

If all is well you will get this Voltage across all windings.

If you have a short somewhere you will find 230mV across one pair of wires and something lower on the other two pairs.

This simple technique has revealed some “horrors” to me, even with new motors!

Sorry this is long to read! It is much easier to do.

I hope your ESC problems will soon be fixed

KR Mark

[John Olsen 1]

PAt asked the following questions a while back now...sorry, I haven't looked in on the thread for a while, so here are some answers

How can the motor see AC when there is only pulsed DC ?

Why would DC saturate the inductor (motor coils ?) ?

I don't understand your ref to protection diodes. Where are these diodes ?

Where does the 22v come from when the supply is 11v ?

So, starting with the first...I don't have a circuit diagram of an actual ESC to look at, but the way this sort of thing is done is that each motor connection is a bit like one phase of a three phase power system. The motor is very like a standard star connected three phase motor, except with a permanent magnet rotor instead of a squirrel cage or a wound rotor like a big one would have. Now, what we do in the ESC is that for each phase (motor wire) we provide two switches. One can connect the phase to the positive side of the supply, and the other can connect the phase to the negative side of the supply. Obviously we must not turn on both switches at once although I have seen what happens when this occurs on a motor controller capable of 12 Amps from a 600 Volt DC supply. By switching each switch on alternately we can generate a square wave. Now, we are also doing the same on the other two phases, although at different times. Because all the waveforms are the same, apart from the timing, the star point in the middle of the windings will end up being at an average voltage which is half the supply. So if the battery was say 12 Volts, the star point ends up at six Volts. Now, as each winding is connected in turn to the positive and negative supplies, that winding will see either plus six Volts or minus six Volts applied to it. Caveat...there will probably be times when both switches are turned off, see below.

OK next, question...if the motor winding has a DC voltage applied to it, the current will rise to a value determined by the Dc voltage and the resistance of the winding. The resistance of the winding is very low, only milliOhms, so even a small DC Voltage will lead to a large current. Now, even a pulsed DC Voltage has a DC component, it can be regarded as being a DC voltage plus an ACwaveform. For instance, if we pulse our 12 Volt supply on for 1 millisecond and off for 11, the coil will effectively see 1 Volt DC plus a 12 Volt peak to peak square wave. 1 Volt DC across say 10 milliOhms will lead to a current of 100 Amps flowing, although probably not for long. The saturation in itself is not a bad thing, except it means that the efffective inductance has fallen very low so it will not take long for the current to get too big. Practically we don't have to worry about this since the ESC designer will have made sure that the circuit provides an AC waveform to the coils. This will not be a sine wave.

Protection diodes...OK, when we switch current to an inductance, it is fine when we turn the switch on, Current begins to rise exponentially. When we turn the switch off, the current tries to continue to flow. From the motors point of view, this is OK, current in the windings is what makes it go, but for the switch the attempt to interrupt the current to a coil can be fatal. The collapse of the magnetic field in the coil creates a voltage spike, the magnitude of which depends on how fast we try to stop the current flowing. So we have to protect our switching devices from over voltage spikes caused by switching the current off. You may have seen a diode in circuits where a relay is driven by a transistor. That diode protects the transistor against overvoltage (Back emf if you like) from the relay coil. In our example above with a winding connected to two switches, one to positive and the other to negative, we would provide two diodes, one across each switch. These would be connected so that they are reverse biased when the switch is off. They may not appear as a visible item in the circuit, should you dismantle an ESC, since the construction of a power MOSFET conveniently contains a diode between the source and drain with exactly the orientation needed.

Ok, the last question...each connection on the motor is connected in the ESC to a pair of switches in the ESC that can connect to either the positive or negative supply. So taking two of the connections, the maximum voltage one way is when one is connected to positive and the other is connected to negative. That would be 11 Volts. But at the opposite point in the cycle, the connections will be reversed. That will also give 11 Volts, but the other way around. So a scope connected across the coil would show an AC waveform of 22 Volts peak to peak. This will not be a sine wave, it will not even be square since in fact the switching devices will not be on for half the cycle. But that is how you can get an AC waveform that is a higher peak to peak Voltage than the DC supply it came from.

One reason that the switches will not be on for half the time is that I gather that the Esc senses the Voltage in the coil to know where the motor is. To do this it will need to sense at a time when the power is turned off. So the switches for any one phase can be in three states...one on, the other on, or both off.

Someone further up mentioned using PCM waveforms to approximate say a sine wave...This could be done, but I doubt if it is actually done in a practical ESC. Too much complication for little gain. Even big controllers for three phase motors usually use quite simple approximations, since the inductance of the windings smoothes out the actual current flow through them. The funny waveforms do make the motors sing a bit.

John

[Peter Beeney]

So, after all that, is the current in the armature DC or AC? Sounds a bit like a suitable subject for a new mini-poll! I have to say, I’m not quite sure how I interpret it in the above.

And also with the greatest respect, but I think I might also also be prepared to debate a little some of the other statements too, because I’m of the opinion that if the currents in all three branches of an AC star connected circuit are balanced, the voltage at the centre connection will be zero. This is because adding the voltages at any one instantaneous point will always be zero. Likewise if you measure the three input currents together this too always amounts to zero. Also you can say on paper that the wave form of a single positive half wave of one cycle might be x volts peak-to-peak with the negative half cycle; and you can draw it on paper or see it on a scope. But the peaks only exist one at the time, so could you measure it with say a voltmeter? And anyway if you take the average voltage or current over one cycle the result is again zero, the positive and negative half cycles cancel each other out.

It all begins to get quite interesting, or should I read that as confusing……

PB

[John Olsen 1]

The current in each winding will be AC. It will not be a sine wave.

When you measure the Voltage at the star point of the motor, where are you going to connect the other lead of the Voltmeter? If this was a three phase mains system, then yes, I would expect the star point to be at a reasonable approximation of the same Voltage as the neutral point of the supply, even if they were not connected. An ESC does not usually provide a neutral point, so there is nowhere for us to measure to.

The three input currents to the motor must always total zero, (Neglecting stray capacitance effects) since they are summed at the star point, and the sum of the currents at any node must always equal zero. The sum of the voltages around any loop in the circuit must also equal zero. (Kirchoffs laws.)

What a Voltmeter will measure will depend on the particular instrument. A moving coil instrument will measure the mean voltage, but will be calibrated to indicate the RMS Voltage of a sine wave with that mean value, and so will be incorrect if the waveform is not a sine wave. If you know what the waveform is you can calculate the error and corect for it. A digital instrument may be a true RMS reading, or it may be reading a mean voltage...if your instrument does not come with details of what it reads, it is best to assume that it may only be accurate for sine waves. It may also be limited in the frequency range it is useful for.

A scope connected across a pair of motor connections will show the Voltage applied there, if the star point is accessable we could also look at the Voltage across one winding. However this would apply quite a bit of stray capacitance to one point in the circuit and I am not sure that I would be game to do this.

John