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

Morning Porta and EK
I found a couple of P MOS, slightly higher Drain to Source resistance 14 and 20 milliOhm, but EK you can have your 20 Mhz and Porta, you can have your 2 Voltage circuits.

I am getting overseas visitors and don't know how my time will go and whether I will be able to find the time to get into the Internet.

I have for each of you made a proposed flow diagram and circuit layout.
It is on your shoulders to give me the best possible currents expected.

Any question you may have, I will answer.
Bert.

 

Hi Porta,

Attached the Schematic Circuit diagram as I see it, that it will work. Also the flow diagram.Please note, you will be surprised and wonder why I go back to the stalling loop and have it twice. The reason is, that I do not believe, we should have all the time, the extra 1 second delay in the loop.
Also decide on the Voltage break. 18 - 21,5 Volt and 21,5 Volt upwards or 18 - 22 and 22 Volt upwards or other Voltage break.
With respect to the circuit diagram. The maximum Voltage allowed on the gate is 20 Volt and I work with 12 Volt, thus we need a little trick to get there. I trust that it will work, when time I will do a test.
Bert

 

Hi ElectricKyak,
Attached the proposed circuit and flow diagram at a speed of 20 Mhz crystral. Please propose the stalling current, I propose the same as your motor can handle: maximum current. Maybe plus 5 %. Also I will try and do some extra loop for light/heavy laboring. I assume 20 % more than your maximum current under load. i.e. 75% of the above maximum current.

Trust that you will think about it, what is the best. Bert

 

O.K. We will do. You need at least 5 milliOhm to make a difference for the ADC. Only one row means 2,5 milliOhm and that is not enough by 25 Ampere to make an impact. Fluctuations and influences from magnetic fields from high currents (I think) will then start playing a roll. I found some P Mos type and also 8 and 20 MHz crystals. Can't give you a DC to DC isolated converter as I need them for myself. I have major problems with my boat. The cheap aftermarket automotive relays I am using are giving me major headaches. I am going to replace them with a IRFP064N MOSfet and a DC converter. But I need 3 and have only 2. Sorry though luck for you. But it give you an idea, should you want to make a couple of units more, after we ironed out all problems and unforeseen hiccups, what is possible.

No idea what the total trip time will be. I will only be able to verify this, after have all the steps and program is completed. However at 20 MHz, although Microchip does not guarantee this, it could be as fast as 200 - 500 nano Seconds for the ADC. Thereafter it depends how bad I am do doing the programming, whether I do 10 or 100 clock cycles. However I am convinced it will be fast enough to protect your ESC and motor.
Bert

 

Good morning EK, Porta, I don't know whether I am still good enough to do assembler language, like I did 30 years ago, when it was my hobby.

But am I wrong in saying, why do we bother with all those different currents? Is it not so, that if you hobble along on the river at a certain speed and I record the current at certain intervals, and should you hit a branch or stall, that the deviation is so drastic, that is differs vastly from the continuous recorded current level and then action must be taken. This means, that at any given moment and at any time and any speed and with any boat, with a dropping battery voltage, you could with the same unit with the same settings, with 12 Volt or 24 or 36 Volt, detect that there is an overload or stalling or laboring.

What I don't know is in how hard and fast, do you both hit the accelerating or lower the speed pedal. Is this something to chew it over and see what I can still find from my 30 year old notes, about mathematical calculations in assembler.
Bert

Bert

 

Hi Bert,

In my case, I agree.

I was about to write saying please don't implement the dotted portion of the EK flow diagram. A single and simple trip point is fine.

Tracking multiple currents or potential motor labouring is not necessary in my system. For me, if the motor is running around 25 amps and lower then it is fine. If it exceeds that then something has gone wrong and I want it to trip as fast as possible. Even labouring is not an issue as I operate at such a low percentage of my motor's capabilities.

As for accelerating speed...I accelerate only at a moderate pace. I expect the motor L/R and ESC's RC time constants should be sufficient to avoid false triggering.

So, I'm very happy with the current simple design. Two rows of 4 parallel resistors = 5 mOhms. If my arithmetic is correct that can trip at 31 amps (ADC=8) or even 27 amps (ADC=7). That is roughly 20% or 10% over my maximum. I lean towards 31 amp trip point but 27 amps is fine too depending what works best for you.

Also for clarity (since we talked about many things in this thread), my system is nominally a "12 volt" system. I use either a 12v AGM battery or a 4S Lithium battery. I don't expect to use anything higher.

I hope this helps.

 

Perhaps I mis-understood you Bert.

Are you suggesting to not even have a set point as I just posted above? But rather maintain a record of the readings and if they suddenly change by a certain (drastic) amount then shutdown? I believe this would require maintaining a moving average with a time constant sufficient to handle startup and acceleration in order to work reliably. This is what I suggested earlier but you felt it was too difficult so I'm a little confused as to what you are now suggesting. I still prefer the moving average approach provided the trip point is around the +25% level but it is definitely more complicated than a single set point approach.

If you are suggesting a higher trip point (ie double the current reading) to simplify things then sometimes it will be better and sometimes worse than a single set point approach. I am less happy with it then.

 

Thanks Bert.

PC

 

Good morning EK,
Soooooooorrrrrry, I did not pick that one up. I don’t know whether I am good enough and able to come up with an algorithm for this small processor, which covers all possibilities. i.e. a rope wrangles slowly around the prop, which takes between anything and a few millisecond. The same for some weed, which gets onto the prop and then get off again. In other instances, it stays on the prop. Then there are situations like a rock or piece of wood, or a sandbank, which instantaneous stops the prop.
Storing previous readings is not a problem also working out the average. We have enough registers and also EEPROMS memory positions.
Thus I will only for the time do a prevention of the maximum allowed current to protect the esc and motor. Nothing else. It was just wishful thinking. Bert

 

Good morning EK. Porta
I made some changes in the software program and I no longer deduct the current in 8 bit form, but transfer the current to be deducted in 10 bit and deduct this from the 10 bit measured result, and then left justify to get it in the 8 bit registers. i.e. 31 Ampere minus 20 Ampere = 11 Ampere negative. If an overload within 4 uSec it will be 40 Ampere – 31 Ampere = 9 Ampere positive and the system must shut down. The advantage is that I now work with 4.8828 milliVolt per step and thus by 31 Ampere, I maybe, maybe, maybe could work with a row of 4 x 10 milliOhm parallel and save at 20 Ampere “sailing” current 1 watt.

I am worried about small electromagnetic fluctuations, but it is worth a try. I finished quickly the second controller and can thus do some various tests. I don’t want to expose the controller I made for the boat (refer item in photo in the background.) and then have nothing when I do something stupid. Also, the ICD3 development platform does not support the PIC12F675 and thus I will do all the testing with the large brushless motor and do some stalling and measure and debug with the PIC18F2431 and if it truly true that the peak current is less then the 31 Ampere by only 20 Ampere cruising current, which I think will be higher, then I use the information and load it in to the PIC12F675.

The reason also for lower measuring current Voltage, EK you have now a P MOSfet with 14 milliOhm and not a N MOSfet at 8 milliOhm and at 20 Ampere it means 5,6 watt loss in the MOSfet versus 3.2 Watt in the N MOS type. Maybe if you are nice to me, I will put 2 P MOSfets parallel and then the power dissipation is halved, only then 2.8 watt total + 1 watt in the resistor gives you a total consumption of 3.8 watt loss at 20 Ampere. And if you are really nice to me, I put 3 parallel. But l think space is limited and it will start playing a roll.

For you Porta, it does not matter so much, as the current is only 2 Ampere and the dissipation is then 2 x 2 x 14 milliOhm = 56 milliWatt and 2 x 2 x 40 milliOhm measuring current = 160 milliWatt.
Bert

P.S. My apology, I have a new mobile phone, Samsung Note 4, and I have no idea, why the picture is up side down, when clicked onto the picture for enlargement. Sorry for that.

 

Good morning EK. Porta
We are getting slowly to the final specifications.
Porta: 24 Volt, P MOSfet IRF4905 , 20 MilliOhm Source to Drain, internal 4MHz oscillator, 2 seconds start up time in total.
1st verification point Greater than 19 Volt or lower
Delay loop 1 sec
2nd verification point, check for greater than 3.7 Ampere stalling current > stop motor, set red flashing LED
Delay loop another 1 sec
3rd verification point, check for greater than 1,95 Ampere overload > set blue flashing LED
4th verification point, check again for Voltage Lower than 19 Volt.
Start looping via 2, 3 and 4th Points.
If lower than 19 Volt,
Jump to second series of verification points.
5th verification point, check for greater than 2.25 Ampere stalling current > stop motor, set red flashing LED
6th verification point, check for greater than 1,80 Ampere overload > set blue flashing LED
Start looping via 5th and 6th until it is time to go home.


Electric Kyak: 12 Volt, 2 parallel PMOSfets MOSfet IRFB4710 , 14 milliOhm each, parallel 7 milliOhm. 20 MHz, ADC needs 1.6 uSecond + required minimum acquisition time for 6 KOhm input , (120pF Cap x (1 + 7 + 6 KOhm ) +temperature compensation 0.05uSec/oCelcius for 25 – 50 degreesC = total 18 uSecond. Refer page 47 , PIC12F675 technical specification.
1st verification point: 35 Ampere
References: thread 11,19,20, 30
Trust that we can now complete the circuits.
Bert

 

Porta, EK,
I am struggling, I am getting different values, when I labor the 3 brushed motors, much wider gap than in thread 11,20,30 . running 1,7 Ampere, laboring 3.35 Ampere etc. Also stalling current is more towards the Voltage /Ohmic resistance.

Also I need to make a plan to get a better Oscilloscope view on the real current for the brushless Motor, when labored. I really think that 31 Amp and 35 Ampere are too low. I have a smaller 3,5 KW brushless motor and must see whether I can not hook an ESC onto it, working on the back EMF and not with sensors.

Tomorrow we are out with the visitors, I will carry on, on Thursday. Also we have the US open tennis championship and we are watching Kevin Anderson of South Africa giving Murray of Britain a hard time. Nice weather in New York.

Bert

 

Porta, Are you absolute sure of the current ratings in my proposal of thread 56, whether I should those use???.
I measured the following with just 3 different motors.
Motor A, 12 Volt, free running 1,95 Ampere, slightly laboring 4.05 Ampere and heavy laboring 6,30 Ampere.
Motor B, 24 Volt, free running 1,89 Ampere, slightly laboring 4.20 Ampere and heavy laboring 7,45 Ampere.
Motor C, 24 Volt, free running 1,78 Ampere, slightly laboring 2,85 Ampere and heavy laboring > 10 Ampere. Bert

 

Hi Efficient Kayak,
May I do give a proposal. I configure the Microprocessor for 31 Ampere and with 2 rows of 4 x 10 milliOhm shunt resistance. Should it trip to quickly by 31 Ampere, you could short circuit the one row of 10 milliOhm resistances and then it will trip at 62 Ampere, without changing the software.
Which is sufficient to protect your ESC and motor. Alternative you find a 0.5 Farad small capacitor and place that over the resistors, to smooth it out and delay it slightly for speed changes, whereby the PWM moves up before the motor has increased speed and have build up more inductive resistance.
Please let me know whet you think about the above proposal.
Bert

 

I will some more real use testing in the next couple of days to pin down these estimated values. Thanks, Bert.

PC