Who is an expert on brushless motors

Thanks Jeremy for the circuit.
Very interesting. I see that you sense each coil, while I sense the complete motorcircuit. i.e. 140 Ampere. I am considering the 250 uOhm 3 watt Vishay shuntresistor. ($14 for 10 + postage ) What is the advantage for sensing each coil?
I am using the Microcontroller PIC16F877A, but am trying to program the PIC18F4320. Lots of fun, trial and error.
Bert

 

Hope it was useful, Bert. My circuit doesn't sense each winding, perhaps the diagram is a bit confusing. There are three sets of power stages, the FETs, driver chips and DC DC converters, but only one set of commutation capacitors and shunt resistors. The shunts I'm using are fairly cheap and work well. The voltage across the shunt is fed back to a comparator in the controller chip and can respond quickly enough to shut an overcurrent pulse down mid-pulse in a uS or less. This is a big advantage over a microcontroller system, because it's fast enough to shorten pulses to limit current on a pulse-by-pulse basis, making the controller pretty near short circuit proof. Because a microcontroller has to rely on getting an over-current signal, processing it and shutting down it can't operate as fast if there is a sudden increase in current (even if driven by an interrupt it's going to take a couple of clock cycles to shut down).

Some of the Chinese controllers I've used have two current limit circuits to get around this problem. They use the microcontroller to control the mean current limit, via an A/D fed with the shunt voltage, then they use an extra comparator set to a threshold above the normal current limit that shuts the whole controller down quickly in the event of an over-current spike. It's not ideal, as the microcontroller then needs to reboot to start the motor again when the current spike has been suppressed, but it does get around the current spike problem.

Jeremy

 

Thanks for the warning. The Micro I am using is 20 Mhz and I hope that the speed of sensing an overload will be less then a microsecond. I will have to check the interrupt and the time it takes. If not, I probably will follow your suggestion.
Thanks, indeed great fun to fiddle around and make things working.
Bert

 

Jeremy, I have never in my life strugled so much as with my motors. Unless I have a good and reliable motor system, I cannot dare to take anything onto the sea.

Here is my problem.
a) I have only a scope with one beam. The second beam burned out and parts are no longer avialable.
b) The one motor (A) I have normal magnetic sensors in the motor at 120 degrees. What I don't know does it matter at what excact position the sensors are glued. i.a.w. The 3 coils are wound , each parallel via 4 ancres, making it in total 12. The sensor maybe at the beginning of the coil or at the end, 30 degrees further or earlier.
c) The second motor (B) I have latched sensors at 120 degrees, also I did not check whether at the left or the right hand side of the coil at 120 degrees.

Now it comes. Motor (A) runs beautiful in forward mode, from 0.5 rev per second to the full speed of 3400 revs/min. I even managed 1 rev per minute, but that is a laugh, as the motor every 5 seconds jurks, 30 degrees further. I scrapped that part again in my software. No extra heavy currents at all, at full speed. But in reverse mode , at approx 50 % PWM, the motor start taking heavy currents on. Also in reversed mode, if I stall the motor, I also have to reset the micro. That defeats the object while at sea.

Motor (B) runs up to 70% in forward mode fine, but then it also start taking on heavy currents. In reverse mode , it is a disaster. Straight away I start drawing heavy currents at 10% of PWM.

I shaffeld my software in the corner and used the MicroChip software AN857
. Also, exactly similar problems. My software even runs slightly better than that of Microchip.

I suspect that it has to do with the crossover point of the EMF (EMK) and because of the 13 magnets, whereby I may have the sensors on the wrong side of the coil. Although 120 degrees is still appicable.

Because I don't have a double beam scope, I cannot see phase differences.

Question: What happens if the motor start to run faster, what happens to the EMK crosover point? I assume it start shifting slower. What are the possibble changes in behavious, in forward mode versus reverse mode?

Jeremy if you could help me, I truly would appreciate that.
Bert

 

Hi Jeremy,

You don't have to answer my question anymore. I got mad, sat down and worked very long hours, until I painstakingly found what my problem was. It is a combination of issues. I use in my upper row MOSFETS DC-DC floating transformers. With 4N25/35 as interface. The 2 - 6 uSec switch on/off delay from the opto-coupler is one of the delays I had to compensate. Further as soon the pull up on the Microchip for reverse mode was used , it influenced the ADC. There were a few more small issues, but at the end I am pleased it is working to my satisfaction.
Sorry to have bothered you and thanks for all the help you have given in the past.
Bert

 

jeremy does not seem to be around at the moment I have sent him a pm about his electric boat and got no answer....??? Hope he is ok and not having more surgury ....

 

Thanks Pistnbroke, Sadly to hear that Jeremy needed surgery. I wish him a speedy recovery. He is such a Gentleman in all his replies and answers. Lets hope he is just recoveriing in the Swiss Alps.

 

I di not say he was having surgury I said I hoped he was not ....

I think he had some on his arm some months ago

 

Thanks, understood.

I still hope he is in the Swiss Alps, nothing nicer to be in one of those Swiss pubs high up in the snow white alps, sipping a "gluhwein" with family.

You Pistnbroke, also have a nice Christmas and a prosperous New year. We will be probably like you, sitting after church on the beach, burning away at 30 degrees.
Bert

 

a Church what is that ??? You must be on another planet..

 

Merry Christmas Pistnbroke, I just came back from the other planet. The people there were very lonely and it seems bitter. I like it here more. May you have a beautifull day, lots of sun and a nice dinner.
Bert

 

Bert, sorry, I've been away from here for some time, busy designing a house we're going to self-build and dealing with all the stuff that goes with getting planning, services etc to the plot.

Hopefully you've sorted your problem, sorry I wasn't around to offer advice.

 

Hi Jeremy, that is fantastic. I also made my own drawings of a double story house and build it. I did 100% of the electrical and plumbing, maybe 5% of the bricklaying and 50% of the woodwork. De builder helping me was a fellow who could not speak English, we communicated with arms and legs, but we did it. Lucky the local helpers were able to speak English, otherwise it would have been a disaster. I wish you therefore also lots of fun, put as many mains plugs in the kitchen and in your workroom as you can.

To come back on the motors. I ordered 2 extra motors in May 2011, they still haven't arrived from China, they do not wish to tell me when I will get them, or what the reason is for the delay.
The other one is working beautiful in forward motion, but in reverse, at 80% it picks up some heavy currents. The reason is most likely the 14 magnets versus the 12 ancres and the position of the WRONG sensors which were supplied by the supplier. They were not latched types. Thus I have quite some play between forward and reverse. However I can use the motor without problems. I changed the PWM from the top 3 MOSfets bridge, driven by your suggestion of isolated trafo's and opto couplers at the top, in having only the PWM at the bottom 3 MOSfets. It had a vast improvement on performance. First I used 12 Volt batteries, which worked also very well, but I had to replace them after an half year. That was not good enough to my likings.

The second motor with latched sensors, I figured out that one sensor is not a latched type, also as an experiment, I glued the sensors at 30 degrees x 3.
Just to see what the performance was. Now I am busy to glue 4 more sensors, all 2 sets of 3 at 120 degrees, using the 2 good latching sensors to see what happens, having the sensor mounted at the left and at the right hand side of the coils. Just to see the difference in performance.
i.e. between ancres 3/4 - 7/8 - 11/12 and a set of 3 between 2/3 - 6/7 - 10/11. Lucky there is quite some space between the ancres this is in view that the motors are 3 KW each.

I hope, to eliminate the heavy currents in reverse in that way.
Bert
P.S. Somebody asked when I build it. It was build 1996, first the small 36 m2 wooden house, thereafter I build the double story main house of 230 m2 and at the end connected the two with the deck and interleading wooden construction. I sold it, in 2005 when we moved to our new house in Littel Brak River.

 

Well at last I have the motors running beautiful in forward and backward motion . The fault was not in my software or hardware, but because the sensors are so small, the supplier had various types mixed in his bin. In the one motor I had 2 correct sensors of the latched type and one which was not latched, but the histeresus in the middle point.

The other motor had 3 different sensors. By replacing all faulty or wrongly supplied sensors, both motors are running now to perfection.

My only excuse is, I designed the printed circuit boards, made the printed circuit boards, wrote the software, who would have thought that the supplier would have shipped me a mixed bunch of sensors. The correct type was ordered. (The sensors are very small and the markings does not mean anything) I was looking at a fault from myself somewhere.

Moral of the story. One can only use latched type of sensors, i.e. when the magnetic field changes, the output is either high or low, but not going slowly from zero to plus or visa versa.

Have all of you a fantastic new year celebrations. Thanks for the support and advise given
Bert