How to make an over-current fuse/switch to avoid damage to a brushless motor

Sorry, i meant current peaks. Shouldn't the ESC input capacitors be smoothing the pwm pulses out? That is their purpose afaik.

 

Noways, those capacitors are for smoothing the plus line to the battery. Measuring the current is always at the minus side. If one would smooth the pwm pulses, to the coils, your efficiency would go through the roof. Bert

 

Just to come back on this topic. The capacitors are as a buffer between battery and controller, to smooth the current bursts, which affect the lifetime of a lead acid battery. Not a lithium battery, they can handle high current bursts. I have a 68.000 uF/63 Volt between my motor/Esc and battery. Ideally a super capacitor would even be more helpful. Bert

 

Here some photo's from the oscilloscope I took with the mobile phone camera.
Screen-A, pwm half open and motor running, screen same time base. easier to measure the mVoltage.
Screen-B, pwm half open and motor running, normal time base, but because of the different speeds between camera and oscilloscope, it is shifted.
Screen-1 , pwm hal open, note the motor running, 1.39 Ampere
Screen-2 , pwm full open, current 3.43 Ampere
Screen-3, motor stalled, 8.04 Ampere, but because i use as protection bulbs/globes in serial, no spikes, but just shows you the voltage in milli Volts over the shunt resistor.
Bert

 

EK, Are you sure, you do not have an over current protection in your ESC? I see most of them has it and can be set with software. Normally they use a magneto resistor and just measure internally the current without a shunt resistor. Those magneto resistors are reasonable fast. One of the IC pins are then used to verify whether it has changed. Bert

 

My ESC has over temp and over current protection. It is not software settable. It is much higher than 30 amps (150 amps or something like that). I do not want to see how well it works...

 

Hi EK, I can understand your concern. You don't want to buckle your shaft. My concern is the substantial higher peaks due to the pwm currents and is even much higher than the 3.4 or 8 Ampere. Photo 1 shows the peaks at 1.39 Ampere, Photo 2 shows at full speed pwm and the Photo 3 shows at stalling (with 200 watts bulbs/globes in serial limiting the current). I have to do much more verification at 14 Ampere etc. In the meantime I take Porta his project as the priority no 1, now my 2 motor controllers are working again. Bert

 

Hi Porta,
The reason why my home made Esc is so important. 1) I can take your motor type and get very accurate information from the ADC via the 2 milliOhm shunt resistors. Should they not high enough in resistance I use EK's test pcboard and connect that on my ESC. 2) The advantage is that now I can with the PIC18F2431 create a single input ADC and track every step what is happening to it. Also with your type of motor the current should not peak as badly as with a brushless PWM driven motor. The software was already done a few weeks ago and it now so nice to hook a normal motor on and see also what is happening in all the registers. The bad news is that i will not be able to get it all finished before I am off to the Elephant National park. Bert

 

Bert, have a good trip to ENP!

Here is a company (no longer in business AFAIK) which I was in contact with, many years ago, in SA: http://disabilityworld.org/03-04_01/access/wheelmaster.shtml
Maybe you knew of them.

"Also with your type of motor the current should not peak as badly as with a brushless PWM driven motor."

Maybe it would peak badly if it was a way larger motor being run at a tiny fraction of capacity in my application, seems to me... Example- if I ran 1 hp brushed at 5% range, and had a stall situation without a gear box, wouldn't it give a bad peak? EK's motor might not have that bad peak if it was more closely matched to the maximum load, without the huge overhead capacity- is what I am thinking... RC motors cannot run at full capacity for more than a few minutes even with excellent cooling, because of heat, bearings, magnets, etc.- so EK's adaptation was a clever way to circumvent.


PC

 

Thank you. Indeed I am looking forward to that trip. I am not looking forward to my birthday. At our age, we rather try to forget those days. However via the grapevine, I understood that my wife and kids bought me a very nice present. That makes it up.

Incredible fantastic solution. I am not too sure whether I also not will, one day, be sitting in a wheelchair and their solution is brilliant. I will remember and store the website.

Hold it your guys, are we changing the criteria again? Not that I mind, but it will only take longer. EK changed the voltage from 12 to 16.8 Volt and maybe higher(I thought only 4 x 3,30 = 13.2 Volt . Are you planning to use a pwm for a possible future motor controller instead of your resistance system at present? It is fine with me, but conclusions from the oscilloscope screen indicates a 10 Ampere pwm current may give as much as 25 Ampere peak. What I cannot do, regretful, is to measure the time for the initial peak, before it has subsided.
Whether the ADC is so fast that it register the initial peak, apart of the subsided peak current. My oscilloscope is quite old, probably 25 year old. I have a second one which is from the 1960's. For sure that one will definitely not help me.
Bert

 

No, I am not changing anything. A toggle and resistor arrangement is the simplest possible novice or user friendly solution, that I can see. Just as lever like doors are easier than twisting a door knob. Even a sliding linear throttle speed control can present issues for the physically challenged.

PC

 

That pleases me, as I was making progress. Tomorrow I may be able to put some pictures on the website, to show you what the currents on the screen looks like with different motors. The program seems to run flawlessly. Bert

 

Hi Porta, EK, Here the pictures of the oscilloscope screen.
I am a little surprised on what I see on the oscilloscope. The time base is 1 ms and the vertical is 5 mV per block. I expected a reasonable flat voltage over the shunt resistor. The 2 milliOhm shunt resistor gives by 2.04 Ampere for motor-A a reasonable flat reading. 2.04 Ampere gives a reasonable flat voltage of +/- 4 milliVolt.
Motor-B has a pulsed signal from 0 to 6 milliVolt at an average of 1,31 Ampere, but the peak at the top is 6 milliVolt. (theoretical 2,6 milliVolt)
Motor-C also has a pulsed result 0 to 5 milliVolt at a current 1.58 Ampere and only a pulse with a timebase of 5mS.
As soon I labour motor A, the voltage increases nice and reasonable flat, but not for motor B, very high pulsed. It means that the ADC which is faster than the pulses, will have a difficult task, as it either measure too low, or if the 2 are not coinciding with each other the result may not be right.
My conclusion is that I will have to create a better test platform. It was extremely difficult to labour the motor with my right hand and trying with my left hand to keep the camera up and with my nose trying to click and make a photo. For that reason photo B5 is very blurred and shaken. Also we may have to flatten with an electrolytic capacitor the pulse, without that it slows down the cut out of the system at less than 50 uSecond.
Porta, is there a way to see whether your motor gives a flat reading or also a pulsed reading.
Bert
Explanations to the photo's.
Please move with cursor over the photo and it will tell you what photo it is.
Photo
Motor-A1 2.04 Ampere +/- 4 mV Theoretical 4,08 mV
Motor-A2 zero line = 0 mV
Motor-A3 2,24 Ampere +/- 4,5 mV Theoretical 4,48 mV
Motor-A4 2,91 Ampere +/- 5 mV Theoretical 5,82 mV
Motor-A5 3,98 Ampere +/- 8 mV Theoretical 7,96 mV
Motor-A6 3,80 Ampere ? mV Theoretical 7,60 mV
Motor-B1 1,31 Ampere +/- 6 mV peak Theoretical 2,62 mV
Motor-B2 4,61 Ampere +/- 16 mV peak Theoretical 9,22 mV
Motor-B3 5,47 Ampere +/- ? mV peak Theoretical 10.94 mV
Motor-B4 7,85 Ampere +/- 26 mV peak Theoretical 15.7 mV
Motor-B5 10 Ampere +/- 35 mV peak Theoretical 20 mV

Porta, one can see, as soon I labor the motors sever, the frequency drops and the peak goes up, except in a reasonable flat curve like motor A. How do we calculate that in the the algorithm properly? We need to do quite some thinking.

 

Porta and EK, You both have to promise me something. You are working with the average reading from the ampere meter, while because of the speed the ADC is reading the peak currents. I am going now for a nice break in the Elephant park, when I am back I will make the units as you have set the currents. i.e. EK, I am happy with your 30 Ampere, I may make it 32 Ampere, but should the units be too sensitive that you both will post the IC back to me for me to make the changes. (weight IC is 2 gram) I have now found out what my mistake was, I have misinterpreted the left and right justifying of the ADC reading, thus when I am back I will make the units for you.
Bert

 

Hi Porta and EK, Here you can clearly see what I mean with previous thread. The timebase for Motor-C1 is changed from 1mSec to 5mSec on the oscilloscope, On Motor-C2 it is clearly seen, whereby the timebase is back to 1 mSec, that the average current is 1,5 Ampere, (3 milliVolt) while the peak is in excess of 2 times i.e. 7 milliVolt.
The ADC is very fast and will probably do the check 5 to 6 times and therefore it is better to add something to the reading than to subtract something to the reading which then would give a overflow bit set.
The motor is my third different motor and it proves that I need to get to know your motor behavior, whether it is like Motor-A, or Motor-B and C. The way you could do a simple test, is to get hold of a 5 to 10 watt 0.1 Ohm resistor and first measure the voltage and current through the resistor and thereafter place 4 x electrolytic capacitor of 4700uF each parallel over it. This would give you a smoothing of 4 x 4700 x 0.68 divided by 1000000 = 12 milliSec , which then will give you an reasonable accurate peak current. If you don't have give me a private email and I will make a plan for you. Bert