12 kw rc motor

[Arlo1]

If these are still lost..... here they are.

 

[Arlo1]

Well blew another 24 fet. I was running burties timing controler... I was trying to get my other computer running to turn the settings down but its 1/2 crashed and I have yet to get paremeter design on this one. So tonight I will get parameter design set up on the new one then I will try to get some customer work done tomorow and get onto fixing one of the blown controlers.

I can tell you this THIS ENGINE SCREAMS! I found a spot at -25 deg and found I could rip way past any rpm I have ever herd from this and it just kept reving I let off because I was scared!
I will either try the map a/b trick with 5deg difference or my second magura hooked up to the tps input so I can hold the rpm and go up and down till I get a sweet spot. I figure one more try till I unwind my second motor and decide what to do with it!
Sorry no videos I had to many things to work with today and nobody around to film my popcorn machine!

 

[Arlo1]

Just re-read the timing controler manual and -25 means I was advanced 25 deg!

I also see it doesnt do anything for timing off the start the motor needs to spin to a certain speed before it takes control. I found lots of starting issues today. I am not so sure anymore. I think I will try one more time with low phase currents then try a whole different aproach! After all my tests my motor was cold and Its not even liquid cooled yet! I think this motor needs some advance but my settings we proly all wrong!

 

[Arlo1]

So blew another controller!
Unwound one of my collossus motors and it has 4 turns per tooth of 32awg x 186 peices in parallel. One whole phase was 11 feet off 186 pieces of 32 awg I tried winding two teath and the most 14 awg I can fit is 9 turns. And luke as usual you were right my hands are killing me!
My plan is for a 9 phase motor but a quick calc says it will be wound for 10000 rpm so..... Should I do it? Or get some 16 awg? The plan was like this wind for the same kv and run 3 controlers then It will be 3 times the resistance on each phase so 3 times better for each controler... But with 9 turns its likely only 2 times the resistance maybe less because of the thicker wire? Pictures will come later!

 

[Arlo1]

Yeh You are right I should do it and run the next controller at 20 s so its not at the limit of the fets! 9phase 100kv run at 84v max it is!

 

[bigmoose]

Arlo, here is a link to what I think may be this green epoxy we see on all these stators:
http://www.monti-inc.com/pdfs/hysol_coating_powders.pdf

Note that the green stuff has a use temp of only 130 oC. Yest it cut's through at 330 ish oC. I am assuming at that temperature it cuts like butter.

I am wondering if you guys using 200 oC wire and going to all this trouble should be using good slot paper, like a Nomex 410 calendared aramid paper on top of the green stuff for extra protection.

"Nomex Type 410 is a calendered insulation paper which offers high inherent dielectric strength, mechanical toughness, flexibility and resilience. Type 410 is the original form of Nomex paper, and is widely used in a majority of electrical equipment applications. Available in 11 thicknesses (2 to 30 mils, or .05 to .76 mm), Type 410 is used in almost every known electrical sheet insulation application. Nomex 410, will maintain 300 V/mil (12 kV/mm) dielectric strength for several hours ar 400°C (750°F)."

 

[Arlo1]

Arlo, here is a link to what I think may be this green epoxy we see on all these stators:
http://www.monti-inc.com/pdfs/hysol_coating_powders.pdf

Note that the green stuff has a use temp of only 130 oC. Yest it cut's through at 330 ish oC. I am assuming at that temperature it cuts like butter.

I am wondering if you guys using 200 oC wire and going to all this trouble should be using good slot paper, like a Nomex 410 calendared aramid paper on top of the green stuff for extra protection.

"Nomex Type 410 is a calendered insulation paper which offers high inherent dielectric strength, mechanical toughness, flexibility and resilience. Type 410 is the original form of Nomex paper, and is widely used in a majority of electrical equipment applications. Available in 11 thicknesses (2 to 30 mils, or .05 to .76 mm), Type 410 is used in almost every known electrical sheet insulation application. Nomex 410, will maintain 300 V/mil (12 kV/mm) dielectric strength for several hours ar 400°C (750°F)."

Well it is a liquid cooled stator. but the windings will be getting hot.
One day I will have a sevcon to run one of these motors! So that motor I will wind with the paper. But I already know a stupid infineon controller can't hurt collossus! I do think you have a good point the winding wires will be getting hot and they will cut though the coating like a hot knife!

 

[Arlo1]

I need a dual trace oscilloscope... I still dont belive how the hall sensor timeing could be right? I feel it was maybe right for rotation one direction and not the other. But.. I need to draw the stator vs magnet diagram again. What I will do is put 9 hall sensors in this motor so I can have a proper hall set for each controller. The halls need to be proper to begin with because burties bord can only start adjusting ~1100 rpm!

I am open to sugestions. But the goal is to keep the engine reving to 7000-8000 rpm max and have the most resistance possible. I would rather have resistance in the windings then use inducters to add some back in.

 

[Arlo1]

Thud mentioned to me replacing the magnets. And making it a 24 magnet 12 pole pair motor but I was thinking about reving it higher so maybe 32 magnets grooped together 2 at a time to get 8 pole pairs 16 poles total. Thats will lower the electrical RPM then we can run it harder and spin it faster with a good controller!
Edit: Biff I was just looking at the second picture and I am not sure I understand the hall sensor placement! If I was to chose it would be at the end of one phase then 6 later for the next then 6 teath later for the last one!

 

[Biff]

You could separate the hall sensors by three teeth. Between 18 and 1, 3 and 4, 6 and 7. You could only run that motor with two controllers, unless you used paralel strands as different "motors". That should work, but maybe not as well as running off separately wound teeth.

If you double the resistance, just to make it more controllable, I think you will be dissapointed. The motor will only be capable of half the power, since 2x the heat would be created for a given current. And it might not save your controller anyways. Inductors are really easy to make with some welders cable and any tube to wrap it around, for an air core. , you can make itany inductance you want, and the heat is produced where you can easily get rid of it.

-ryan

 

[Arlo1]

Hey ryan the stock hall sensors were between 1 and 2 7 and 8 13 and 14 So could that be the problem?
I did a volage drop test on stock collossus my x5304 and my 18 turns on 2 teath so I could compare numbers. I will get it loaded asap.

I am actualy thinking about rewinding it series parrellel 8 turns a tooth and they can be switched each tooth to 8 in series for start and then 2 set of 4 in parrel for hi speed! Then just run one controler.

Somthing was very very wrong in stock form because it could only make 1.5 hp at the wheel and my bmx mad 6.58 hp at the wheel with a x5304 both runnning a cheep infinieon controller that always uses hall sensors.

If I just keep it 3 phase is the sensor location the same as you just sugested?

 

[Arlo1]

Here is the video showing the voltage drop method of testing phase resistance. It compares stock collossus a stock x5304 and a collossus with 2 teath wound with awg 14. Now take note this shows only one phase of collossus rewound with 18 turns total and not hooked to any other phases in WYE or DELTA so it will be wich would make it 25% lower.
[youtube]5L6JcgRvJYE[/youtube]

 

[Arlo1]

Word of the day " INDUCTANCE" It is what absorbs and smooths out the PWM from the fets into usable power. The problem from what I gather is the further away the battery voltage is from the max rpm for that voltage, the more inductance you will need. Make sense?

 

[Arlo1]

You could separate the hall sensors by three teeth. Between 18 and 1, 3 and 4, 6 and 7. You could only run that motor with two controllers, unless you used paralel strands as different "motors". That should work, but maybe not as well as running off separately wound teeth.

If you double the resistance, just to make it more controllable, I think you will be dissapointed. The motor will only be capable of half the power, since 2x the heat would be created for a given current. And it might not save your controller anyways. Inductors are really easy to make with some welders cable and any tube to wrap it around, for an air core. , you can make itany inductance you want, and the heat is produced where you can easily get rid of it.

-ryan

This you meen for collossus in stock from? The sensors were between 1&2 7&8 and 13&14 so I am sure they were off.. That will make for a lot of problems right there. As for inductance.... Do I want Xnumber of turns on each phase wire? Do the phase wire inductors need to be away from each other? How big of Dia can the coils be or how small? How many turns per phase does it need?

 

[Biff]

The Stock colossus is 18teeth 20 poles, but the winding pattern is still the same and hall sensor location should be the same aswell.

So I believe the correct locations of the sensors for delta terminated winding would be between 1-2, 4-5, and 7-8, that would give you slightly advanced timing for counter-clockwise rotation, if you wanted clockwise rotation you would put the sensors in 2-3, 5-6 and 8-9 for best performance.

Teeth 13 and 14 are the same position relative to magnets as 4-5 so putting the sensor there is just as good, and maybe gives better options for wire-routing or something.

So in summary, if colossus is terminated delta, the halls are in the right location, but the stock motor should work better turning counter-clockwise because the hall timing is not exactly neutral. If you are running the motor clockwise the position sensing would be slightly behind, which could give poor performance. The delay could be calculated but since the actual installation of the sensors can affect the timing quite a bit, I would want to scope the sensors before saying that is the cause of the problem.

-ryan

 

[Biff]

As for inductance.... Do I want Xnumber of turns on each phase wire? Do the phase wire inductors need to be away from each other? How big of Dia can the coils be or how small? How many turns per phase does it need?

I like to use carpet roll tubing to wind air-core inductors. you want the inductors to have the same number of turns. I just use an on-line calculator to estimate the inductance to start.

http://www.daycounter.com/Calculators/Air-Core-Inductor-Calculator.phtml

Wind something which should have about 15uH of inductance and put one beteen the controller and each phase lead. If things run better, take a turn off each inductor and see how that works. Once you figure out what number of turns works, you can measure the inductance that you have (or calculate it) and then make an Iron core inductor with similar properties if you want to save space.

As far as distance between the inductors, I haven't experimented much with that. With iron core inductors it shouldn't matter since the flux is "contained" in the iron. with air core I haven't found any issues with having them pretty close together, The inductors I have been using are about the size of a 1l water bottle, and I have them all sitting together in something the size of a milk crate.

-ryan

 

[Arlo1]

Thanks biff i think the thing is we are spining this motor backwards from what it was designed for.
What would be the easiest on a controller? What about a newtral timing and adding the advance with burties bord?

 

[Biff]

Thanks biff i think the thing is we are spining this motor backwards from what it was designed for.
What would be the easiest on a controller? What about a newtral timing and adding the advance with burties bord?

The Problem is that neutral timing is really hard to get (upon first inspection) with that motor, What need to do is take the picture, and line up one magnet with the middle tooth of one of the phase coils, then find the best place to put a hall sensor so that it is exactly in the gap between 2 magnets. I don't think that there is a gap between teeth that lines up exactly between magnets for this combination, so you would either have to build a mounting system allows you to put a sensor which isn't between teeth, or grind away part of a tooth so that you can put the sensor where you want. With a WYE configured motor, you always just put the hall sensor exactly halfway between the centers of 2 phase coils, which for this design is always in the gap of a tooth. For the Chrystalite motors which are something like a 48/51 pole/tooth combination or whatever it is much harder to picture where to put the hall sensors (or start your winding pattern if the hall sensor locations are already there).

Burties timing thing should be able to take care of all the problems once the motor is spinning fast enough for timing thing to take over. Having a little timing advance is a good thing, so I would try moving the halls one tooth over and see if that helps, or just swap 2 phase leads and find the hall combination to get the motor spinning the other way and see if it helps performance, If you are a really good backward rider you might be able to tell if the performance is better, or just get your bike to pull your car or something (you will need to get a fixed gear or the bike).

Is it possible that you just have the wrong hall combination? I know sometimes you can get a wrong combination and the motor seems to work ok, just not great, but in fact the halls are not correct, fixing them gets you much better performance.

-ryan

 

[Arlo1]

I like to use carpet roll tubing to wind air-core inductors. you want the inductors to have the same number of turns. I just use an on-line calculator to estimate the inductance to start.

http://www.daycounter.com/Calculators/Air-Core-Inductor-Calculator.phtml

Wind something which should have about 15uH of inductance and put one beteen the controller and each phase lead. If things run better, take a turn off each inductor and see how that works. Once you figure out what number of turns works, you can measure the inductance that you have (or calculate it) and then make an Iron core inductor with similar properties if you want to save space.

As far as distance between the inductors, I haven't experimented much with that. With iron core inductors it shouldn't matter since the flux is "contained" in the iron. with air core I haven't found any issues with having them pretty close together, The inductors I have been using are about the size of a 1l water bottle, and I have them all sitting together in something the size of a milk crate. -ryan

When it says coil Dia is the the distance of diameter of one loop? And coil length is the total lengh of all the turns of the coil? So if its the total lengh of all the coils you want to keep the gauge of wire as small as you need to make the inducance as hi as possible so you can make the total inductor as small as possible.
Luke mentioned an air core has no losses in efficency what about an iron core? It would proly take some energy away kinda like a iron stator?

 

[bcwrench]

I need a dual trace oscilloscope... I still dont belive how the hall sensor timeing could be right? I feel it was maybe right for rotation one direction and not the other. But.. I need to draw the stator vs magnet diagram again. What I will do is put 9 hall sensors in this motor so I can have a proper hall set for each controller. The halls need to be proper to begin with because burties bord can only start adjusting ~1100 rpm!

Arlo,

Just happened to drop in to ES. Do you still need access to a scope? I have a Fluke 97 and an old CS1021 Kenwood analog if you still do need one. Ring me at Island D Machine if you do. I will be off Monday (11th)

 

[Arlo1]

Yup thanks.
Im in the market for a uber cheep multi channel oscilloscope (pc based is fine) I dont need nothing to fancy. But I want to scope 4 or more things at once. I posted in the oscilloscope thread as well. Once I get timing issue worked out I will build a set of inductors and scope before and after them to see how they build up and relase their energy. As well as how they off set the timing!

I am looking at how to build the most compact inductors possible right now as well. Looks like magnet wire and loops wound right next to each other are more efficent.

 

[Arlo1]

So as I was watching YouTube to learn more about induction I found this video and it talks about inductors next to each other and how if wound the same direction they would help increase induction and opposite directions they would reduce induction so then I realized that with collossus the AaA winding will lower the induction then a AAA winding (not that that's possible but...)
http://www.youtube.com/watch?v=GXcxswDcUbI&NR=1 [youtube]GXcxswDcUbI[/youtube]

 

[Biff]

When it says coil Dia is the the distance of diameter of one loop? And coil length is the total lengh of all the turns of the coil? So if its the total lengh of all the coils you want to keep the gauge of wire as small as you need to make the inducance as hi as possible so you can make the total inductor as small as possible.
Luke mentioned an air core has no losses in efficency what about an iron core? It would proly take some energy away kinda like a iron stator?

Coil Diameter is the diameter of the pipe you loop around

Total length is the length of the pipe which you wrapped, not the length of the wire.

Iron core has eddy current and hysteresis losses as frequency and amplitude increases, very hard to calculate.

-ryan

 

[Nuts&Volts]

Don't want to side track the discussion, but would two matched (by resistance and inductance) motors aligned end to end be easier on a single controller? The motors would have pole to pole alignment and one set of halls would be used for timing. Essentially make it one long motor.

-Kyle

 

[Arlo1]

Don't want to side track the discussion, but would two matched (by resistance and inductance) motors aligned end to end be easier on a single controller? The motors would have pole to pole alignment and one set of halls would be used for timing. Essentially make it one long motor.

-Kyle

Run in series maybe a bit but not enough. And then for 7500 rpm you would need 200 volts! The problem from what I have been lerning from others is the low inductance and that will go up in series as well but if its 100x to low as a single motor then two in series it will be 50x to low!