Current Controller for RC ESCs
- Thread starter rhitee05
- Start date
rhitee05
10 kW
Still a ways to go, but with the sensor proof-of-concept out of the way I'm pretty confident it will work, just a matter of building and testing it.
I got a package in the mail a couple days ago with some very nice presents courtesy of Thud:
Those are the Mk2 version of my sensor. I'm going to use one of those to build a 2nd prototype very soon, and if all goes well that will be the final sensor design. Next I have to do a little bit of breadboarding to test some elements of the circuit. I'm pretty confident it will work based on SPICE, but I like to check things in real life to be certain. Once I'm happy with that, I'm going to send for a handful of PCBs and build one or two full-up prototypes for testing.
I got a package in the mail a couple days ago with some very nice presents courtesy of Thud:
Those are the Mk2 version of my sensor. I'm going to use one of those to build a 2nd prototype very soon, and if all goes well that will be the final sensor design. Next I have to do a little bit of breadboarding to test some elements of the circuit. I'm pretty confident it will work based on SPICE, but I like to check things in real life to be certain. Once I'm happy with that, I'm going to send for a handful of PCBs and build one or two full-up prototypes for testing.
Danny Mayes
1 kW
Very exciting stuff! 
D
D
liveforphysics
100 TW
rhitee05 said:
Thud rocks. Super craftsmen extraordinaire, and it's impossible not to like him.
Since this device can be equally helpful for folks with infinion-style controllers, having 2 throttle output options would be very nice. Something for all the folks running ebike controllers to the extreme, and something for folks running RC controllers.
rhitee05
10 kW
liveforphysics said:
Yeah, I had thought along those lines. It should be easy enough to accomplish, just a matter of providing 0-5V output instead of the servo tester output.
There do seem to be a few people who've made a habit of blowing up Infineons.
liveforphysics
100 TW
rhitee05 said:There do seem to be a few people who've made a habit of blowing up Infineons.
Thud
1 MW
HEY!
I have only vaporised 9 xieChang's & 3 hv160's.....I can stop at any time so its not a problem!
....I am not addicted to plasma, I just like the way it smells
(substitue cocain for plasma & its an old R.Dangerfeild joke!)
rhitee05
10 kW
I wasn't going to name anyone, but if we need to create a support group I'm sure we can put one together. If anything, I think LFP can teach us that melting cheap Chinese controllers is just a gateway drug to melting more expensive hardware! Do you keep a graveyard around your house, or is it more like a trophy wall?
This reminds me of my electronics class back in college - the professor was a little bit strange. On the day when we talked about why you can parallel FETs but not BJTs, he set up a little demo. Blowing up a couple of BJTs proved his point. He still repeated the demo a few times just for fun!
If anything, I think LFP can teach us that melting cheap Chinese controllers is just a gateway drug to melting more expensive hardware! Do you keep a graveyard around your house, or is it more like a trophy wall? This reminds me of my electronics class back in college - the professor was a little bit strange. On the day when we talked about why you can parallel FETs but not BJTs, he set up a little demo. Blowing up a couple of BJTs proved his point. He still repeated the demo a few times just for fun!
Rc speed controllers react differently but some (hobbywing) can go up to 500hz control pulses
Those are the best to put inside a control loop like a governor for helicopter or speed control on a multi rotor helicopter.
The faster the input the faster they ramp up the output current
I think that most can accept up to 250hz but their firmware delay the output
Do search on rcgroup on multi rotor helicopters
Those are the best to put inside a control loop like a governor for helicopter or speed control on a multi rotor helicopter.
The faster the input the faster they ramp up the output current
I think that most can accept up to 250hz but their firmware delay the output
Do search on rcgroup on multi rotor helicopters
rhitee05
10 kW
Useful info, thanks. Hopefully during beta testing I'll be able to figure out what timing some of the more popular ESCs will accept, like the Castles that many people use. It won't be hard to adjust, though. Timing is set by an RC network so the frequency can be changed by swapping the value of a resistor.
liveforphysics
100 TW
rhitee05 said:I wasn't going to name anyone, but if we need to create a support group I'm sure we can put one together.
This reminds me of my electronics class back in college - the professor was a little bit strange. On the day when we talked about why you can parallel FETs but not BJTs, he set up a little demo. Blowing up a couple of BJTs proved his point. He still repeated the demo a few times just for fun!
BJT's parallel just fine, as is done in virtually every amp over ~50-100w. You just need to put a 0.1ohm or 0.05ohm 5-10watt resistor in series with each one to make them balance despite the inherent silicon differences.
rhitee05
10 kW
Of course. BJTs need a little help to share current, wheras MOSFETs generally do just fine on their own. The difference is that the on-resistance of FETs have a positive temperature coefficient, which tends to reduce current as they get hotter; in BJTs the Vce saturation voltage has a negative temp coefficient with the opposite tendency. Negative feedback versus positive feedback, basically.
rhitee05
10 kW
Just in case anyone was wondering, I haven't abandoned the project!
Been busy with other things, but in my limited time I've been finalizing the circuit design and laying out boards for the prototypes. I think I'm still going to do some tweaking before I send them off for manufacture, but the first layout iteration is complete. The current design is a 2-board stack, but I'm considering making it a 3-board stack. The advantage would be a more compact footprint and modular functionality.
Here's the feature set for the prototype:
- Custom hall-based 3-phase current sensor. The sensor should have fully-linear range of >300A, >200A, or >150A depending on the sensitivity of the sensors used. Sensor is fully isolated and insulated from the phase wires - the current design just has 3 0.25" dia holes which the phase wires pass through. I have some 8 ga wire which passes through easily.
- Interfaces to standard 5V e-bike throttle, hall- or pot-based. 5V supply is provided to the throttle. Currently both gain and offset adjustments are included.
- Option for either pass-thru throttle or closed-loop current-based throttle (selectable by jumper).
- Throttle filtering with separate up/down ramp times if desired (ramp-down time <= ramp-up time). Filtering can be disabled by removal of a capacitor.
- Option for throttle "boost" which provides temporary 10-30% (set by fixed resistor) increase in throttle gain. Boost is activated by a momentary pushbutton.
- Phase current limiting via 2 separate limits (works for either direct or closed-loop throttle). One limit has a fast, aggressive response, the second has a slower, integrating response which is less aggressive.
- Includes servo interface directly to ESC with adjustable min/max pulsewidths. Pulse rate is set via a fixed resistor (50 Hz nominal). Prototype is designed to drive ESCs only, but future versions will have a 0-5V output to drive a standard controller instead.
- Current design generates 5V supply directly from the battery. Future versions will probably make that optional, either internal 5V supply or taken from an external BEC, and will probably use a switching regulator rather than the current linear regulator.
The prototype does not have forward/reverse capability, low-speed lockout, or LED indicators but those may be added later. I had to cut off the feature set somewhere and the above list included most of the features which were easy to add. The prototype also does not include any additional safety features. I am open to suggestions about what people would like to see there. I think safety features are important, but might also annoy the user, so a balance is needed and feedback will guide the decision.
More to come - comments appreciated!
Been busy with other things, but in my limited time I've been finalizing the circuit design and laying out boards for the prototypes. I think I'm still going to do some tweaking before I send them off for manufacture, but the first layout iteration is complete. The current design is a 2-board stack, but I'm considering making it a 3-board stack. The advantage would be a more compact footprint and modular functionality.
Here's the feature set for the prototype:
- Custom hall-based 3-phase current sensor. The sensor should have fully-linear range of >300A, >200A, or >150A depending on the sensitivity of the sensors used. Sensor is fully isolated and insulated from the phase wires - the current design just has 3 0.25" dia holes which the phase wires pass through. I have some 8 ga wire which passes through easily.
- Interfaces to standard 5V e-bike throttle, hall- or pot-based. 5V supply is provided to the throttle. Currently both gain and offset adjustments are included.
- Option for either pass-thru throttle or closed-loop current-based throttle (selectable by jumper).
- Throttle filtering with separate up/down ramp times if desired (ramp-down time <= ramp-up time). Filtering can be disabled by removal of a capacitor.
- Option for throttle "boost" which provides temporary 10-30% (set by fixed resistor) increase in throttle gain. Boost is activated by a momentary pushbutton.
- Phase current limiting via 2 separate limits (works for either direct or closed-loop throttle). One limit has a fast, aggressive response, the second has a slower, integrating response which is less aggressive.
- Includes servo interface directly to ESC with adjustable min/max pulsewidths. Pulse rate is set via a fixed resistor (50 Hz nominal). Prototype is designed to drive ESCs only, but future versions will have a 0-5V output to drive a standard controller instead.
- Current design generates 5V supply directly from the battery. Future versions will probably make that optional, either internal 5V supply or taken from an external BEC, and will probably use a switching regulator rather than the current linear regulator.
The prototype does not have forward/reverse capability, low-speed lockout, or LED indicators but those may be added later. I had to cut off the feature set somewhere and the above list included most of the features which were easy to add. The prototype also does not include any additional safety features. I am open to suggestions about what people would like to see there. I think safety features are important, but might also annoy the user, so a balance is needed and feedback will guide the decision.
More to come - comments appreciated!
rhitee05
10 kW
Anybody still paying attention, or did everyone move over to the CA thread?
rhitee05
10 kW
Good to know!
Almost ready to get a prototype going. Changed to a 3-board stackup, as it was getting hard to fit everything into the footprint I wanted (5 cm x 5 cm). Last step before I ship it out the door is to put a BOM together, so I know what parts I need to order along with the boards. That will also give me an idea of what these suckers are going to cost.
Almost ready to get a prototype going. Changed to a 3-board stackup, as it was getting hard to fit everything into the footprint I wanted (5 cm x 5 cm). Last step before I ship it out the door is to put a BOM together, so I know what parts I need to order along with the boards. That will also give me an idea of what these suckers are going to cost.
liveforphysics
100 TW
rhitee05 said:
The CA can't do a thing to protect an RC controller. It can only limit battery current, which does nothing to save an RC controller (which never even see's battery current).
Thud
1 MW
it all sounds great rhitee05.
I been digging out of a huge hole at work myself lately. Looking foward to seeing the beta unit in testing.
Let me know if there is any thing i can do to help.
T
I been digging out of a huge hole at work myself lately. Looking foward to seeing the beta unit in testing.
Let me know if there is any thing i can do to help.
T
parabellum
1 MW
rhitee05 said:
I have 3 RC ESCs waiting for your controller. Can also do some crude beta testing on my conversion scooter if necessary. :?
gtadmin
10 kW
Why is that, Luke? I understand the that the motor doesn't see battery current, as PWM acts as a multiplier of the current in the phase leads, but isn't that (approximately) inversely proportional to the PWM duty cycle? To explain using round figures, Ibatt = 100A, dutyCyclepwm = 10%, phase lead current = 100/0.1 = 1000A? And you can approximate PWM duty cycle as (actual motor speed)/(No load motor speed). <-- or should that be theoretical maximum speed?liveforphysics said:
So, too much current in phase lead --> back off the throttle --> drops the battery current --> drops the phase lead current. Or is there something else going on that I have missed?
And rhitee05, I'm still following too 8)
Cheers,
GT
panurge
10 kW
Hi rhitee05
I'm still following your progresses with interest too!!!!!
A promising work 8)
Jules
I'm still following your progresses with interest too!!!!!
A promising work 8)
Jules
rodgah
10 kW
Im still interested too....
I would run it in combination with the CA-RC if possible....but thinking about it, it would not be very easy to implement. I would think that battery current limiting and throttle control done by the CA (if it works as expected) and this device to do a hard throttle cut when the phase current limit reaches its set threshold. I am a little lost to whether my applications can really get much of an advantage from either since Ive yet to smoke an ESC, but current based throttle is definitly an advantage of either option.
Im up for some testing anyway.....Im interested to see how the phase current differs from the battery current and what set of circumstances creates the problems with blowing an esc due to phase amps....
:wink:
I would run it in combination with the CA-RC if possible....but thinking about it, it would not be very easy to implement. I would think that battery current limiting and throttle control done by the CA (if it works as expected) and this device to do a hard throttle cut when the phase current limit reaches its set threshold. I am a little lost to whether my applications can really get much of an advantage from either since Ive yet to smoke an ESC, but current based throttle is definitly an advantage of either option.
Im up for some testing anyway.....Im interested to see how the phase current differs from the battery current and what set of circumstances creates the problems with blowing an esc due to phase amps....
:wink:
rhitee05
10 kW
Good to know there's still a lot of interest!
The key word there is approximately. I think that in some/most ESCs you can program the throttle range, so input throttle of 20% doesn't necessarily mean 20% duty cycle. More importantly, the ESC might be limiting battery current, in which case the duty cycle would be lower than the throttle input implies. To make that work at all, you'd really need a dumb ESC that translated the throttle directly to duty cycle, no questions asked.
My device will implement the current limit in two different ways, based on the fact that there are two main ways you can destroy a controller with too much current. A really large pulse of current will tend to melt whatever the weakest link in the chain is (frequently a FET lead) as if it was a fuse. For that, I have one limit with a fast response time and a hard trigger - over the limit and the throttle is pulled to zero more or less instantly (actual response time of the ESC will depend on the pulse frequency of the servo signal). The second way is to simply get it too hot, which is more dependent on the average level of current over time. For this I implemented a second limit which works on the integral of current and has a progressive action - that is, the further the integral goes over the limit the further the throttle will be pulled down. The idea here is to allow an equilibrium where the current could be maintained at the limit continuously without the on-off stuttering action of a hard limit. I'll see how this works in practice during testing and I've intentionally set up the circuit so that various parameters can be adjusted as needed.
gtadmin said:
The key word there is approximately. I think that in some/most ESCs you can program the throttle range, so input throttle of 20% doesn't necessarily mean 20% duty cycle. More importantly, the ESC might be limiting battery current, in which case the duty cycle would be lower than the throttle input implies. To make that work at all, you'd really need a dumb ESC that translated the throttle directly to duty cycle, no questions asked.
rodgah said:
My device will implement the current limit in two different ways, based on the fact that there are two main ways you can destroy a controller with too much current. A really large pulse of current will tend to melt whatever the weakest link in the chain is (frequently a FET lead) as if it was a fuse. For that, I have one limit with a fast response time and a hard trigger - over the limit and the throttle is pulled to zero more or less instantly (actual response time of the ESC will depend on the pulse frequency of the servo signal). The second way is to simply get it too hot, which is more dependent on the average level of current over time. For this I implemented a second limit which works on the integral of current and has a progressive action - that is, the further the integral goes over the limit the further the throttle will be pulled down. The idea here is to allow an equilibrium where the current could be maintained at the limit continuously without the on-off stuttering action of a hard limit. I'll see how this works in practice during testing and I've intentionally set up the circuit so that various parameters can be adjusted as needed.
dontsendbubbamail
10 kW
- Joined
- May 28, 2008
- Messages
- 718
gtadmin said:
There are pages and pages of discussion on this topic in the last couple of years. I implemented a PID battery current controll and can attest to the lukes statement that it will not save your RC esc. Using an Eagle Tree data logger, I could see the PID limiting the battery current just fine, but my esc still blew. My brain hurt too much trying to understand the technical reason causing the issue, so I just moved on to adding sensors to a one of Lynn's controllers. A sensored controller has been a good solution for me.
Bubba
gtadmin
10 kW
Doesn't this device do this though?
Thanks guys for your answers, I'm off to try and find those pages ....
Similar threads
