Thread Tools
May 07, 2017, 05:28 PM
Registered User
Thread OP
Discussion

ESC idle power


Greetings, I've souped up the blower motor in my Camry with a much more powerful SK 5045 450KV.

2.5x more powerful Camry blower motor & impeller (1 min 10 sec)


I can now get a strong breeze even in the back seats.

Unfortunately, the idle current draw from the HobbyKing BlueSeries 30A ESC is too much. It draws 58mA during idle and the total draw measured on the car battery is 78mA. I have a Yellow Top deep cycle battery with 140 minutes reserve capacity (58.3 Ah), so it will last 31 days before being drained. The other annoyance is the high frequency PWM commutation noise as you can see on the oscilloscope.

Does anyone know where all that power is going? Is it the BEC? If it is, could it be disabled and instead connect my own 5V regulator to power both my control board and the ESC's microcontroller?

Or are there any ESCs in the 50 amp range with a much better idle current?

The strange thing is the blower fan circuit is always on even when keys are not in the ignition. It's PWM controlled, so I was thinking about using a solid state relay, but the voltage drop (~1.3V) would be too much.
Last edited by UncleJoe; May 08, 2017 at 02:34 PM.
Sign up now
to remove ads between posts
May 07, 2017, 06:29 PM
Registered User
ESCs don't have off switches so as long as the ESC is plugged in the processor, the gate drive circuits, etc are all drawing power. Just because the power isn't being sent to the motor to spin it doesn't mean the ESC isn't just sitting there "thinking". Swapping out the BEC for your own regulator won't change this. There's also tiny leakage currents through the power MOSFETs themselves.

You might randomly find an ESC with slightly better idle current, but the fact is that it's simply not a design consideration for hobby ESCs. In their intended application, once they are plugged in they are probably running almost continuously and motor power consumption dwarfs anything by the ESC itself.

Can you just stick a large electromechanical relay in series with the ESC so it's completely de-powered when keys are not in the ignition? Too bad it's not an ESC without a BEC. Then you could just a small electromechanical relay to depower only the ESC's logic and then the only currents would be leakage currents through the power MOSFETs themselves which should be in the uA.

You might want to add bigger input decoupling capacitors right next to your ESC...the wire run to your ESC is longer than your ESC was designed for and may be degrading your ESC due to spikes. It will also help with the commutation noise.
Last edited by DKNguyen; May 07, 2017 at 06:42 PM.
May 08, 2017, 02:21 AM
Registered User
Thread OP
OK, thanks for the tip, but I couldn't find any suitable mechanical relays. I searched Digikey for relays >= 35A, but found only 3 choices and they're all probably too big to fit in the blower enclosure

Panasonic ADQM16024 38x29x17
TE PB2134-ND 37x30x17mm
K40P332 45.7x16.5mm diameter

I think the reason mechanical relays >= 35A are rare is because 35A can easily weld the contacts.

Here are pictures of the whole setup. As you can see, the electronics barely fit.

https://www.instructables.com/id/Mor...-Blower-Motor/

Quote:
the processor, the gate drive circuits, etc are all drawing power
I understand, but 60mA seems quite inefficient. The STM32F334 I use as a controller only draws 4mA @8MHz. I suspected the BEC was using a linear regulator, which would be < 50% efficient, so I thought by replacing it with a Texas Instruments simple switcher, I could bring the power down to 30mA. But even 30mA seems inefficient.

Sorry if I sound whiny. 31 days until the battery is drained isn't a big deal since I will almost certainly recharge before that happens. I'm just looking for peace of mind in case I go on an extended vacation.
Last edited by UncleJoe; May 08, 2017 at 02:40 AM.
May 08, 2017, 11:45 AM
homo ludens modellisticus
Ron van Sommeren's Avatar
Quote:
Originally Posted by UncleJoe
... I've souped up the blower motor in my Camry with a much more powerful SK 5045 450KV. ...
Have you considered the max.rpm of the fan? Stretching and/or disintegrating? Forces are proportional to rpm squared, i.e. proportional to voltage squared.
Quote:
Originally Posted by DKNguyen
... You might want to add bigger input decoupling capacitors right next to your ESC...the wire run to your ESC is longer than your ESC was designed for and may be degrading your ESC due to spikes. It will also help with the commutation noise.
too long battery wires will kill ESC over time: precautions, solutions & workarounds - RCG

Vriendelijke groeten Ron
May 08, 2017, 12:00 PM
Registered User
Quote:
Originally Posted by UncleJoe
OK, thanks for the tip, but I couldn't find any suitable mechanical relays. I searched Digikey for relays >= 35A, but found only 3 choices and they're all probably too big to fit in the blower enclosure
Hmmmm...well if form factor is an issue then solid-state relays wouldn't help you either since they only achieve their rated currents with a giant heatsink bolted onto them. Too bad you couldn't just modify the ESC firmware so it goes into sleep mode when throttle is zero for long enough.

The battery is your startup battery right? What are the user controls that you are making available?

Quote:
Originally Posted by UncleJoe
Sorry if I sound whiny. 31 days until the battery is drained isn't a big deal since I will almost certainly recharge before that happens. I'm just looking for peace of mind in case I go on an extended vacation.
No problems. We already figured out why idle current is a problem for your application.
May 08, 2017, 02:32 PM
Registered User
Thread OP
I just measured the idle current of a few other ESCs:

Hobbyking Blue Series 30A: 58ma
Hobbyking Blue Series 20A: 44ma
HobbyWing Justock club spec sensored: 30mA
Team Novak sensored ESC: 43ma (with 2 LEDs - desoldering might drop it to 20mA)

So if I get a decent plane ESC, I should get better idle power. Any suggestions?

Most car ESCs also have an on/off switch, but I'm not sure if that can be converted into electronic control. If it can, perfect.

Quote:
too long battery wires will kill ESC over time
I can imagine I will have a big problem then since the wires from the car battery to the blower under the glovebox will be very long presumably. I'll hook up an oscilloscope and test.

Quote:
Forces are proportional to rpm squared, i.e. proportional to voltage squared.
Intersting. I thought torque due to air resistance should be proportional to RPM. But apparently F = RPM^2 does match my measurements:

2860 RPM @ 267W
3900 RPM @ 650W power = torque * RPM ~= (3900/2860)^3 * 267W


Quote:
The battery is your startup battery right? What are the user controls that you are making available?
Correct. The blower speed is controlled from the front panel as before, which sends a PWM signal that my microcontroller detects and translates into the ESC control signal.
May 08, 2017, 02:36 PM
Registered User
Quote:
Originally Posted by UncleJoe
I just measured the idle current of a few other ESCs:

Hobbyking Blue Series 30A: 58ma
Hobbyking Blue Series 20A: 44ma
HobbyWing Justock club spec sensored: 30mA
Team Novak sensored ESC: 43ma (with 2 LEDs - desoldering might drop it to 20mA)

So if I get a decent plane ESC, I should get better idle power. Any suggestions?

Most car ESCs also have an on/off switch, but I'm not sure if that can be converted into electronic control. If it can, perfect.


I can imagine I will have a big problem then since the wires from the car battery to the blower under the glovebox will be very long presumably. I'll hook up an oscilloscope and test.



Intersting. I thought torque due to air resistance should be proportional to RPM. But apparently F = RPM^2 does match my measurements:

2860 RPM @ 267W
3900 RPM @ 650W power = torque * RPM ~= (3900/2860)^3 * 267W



Correct. The blower speed is controlled from the front panel as before, which sends a PWM signal that my microcontroller detects and translates into the ESC control signal.
Air resistance/drag increases to the squared of the velocity. Velocity is proportional to RPM and current is proportional to RPM. Most of the torque from a propeller is due to air drag.

I was thinking that if you had your own MCU interfacing between the front panel and the ESC, you could put a relay under the MCU's control such that the relay never actually interrupts current. That is, the MCU always ensures the ESC is throttled down to zero for a set amount of time before the relay is ever opened or closed. The lowest limit on relay current ratings is their interruption current, not the actual amount of current they can pass so you might be able to get away with a 35A relay for your 50A current if you do this. Is your fan actually drawing 50A to begin with?

Or maybe squeeze in a very large (1-2 mohm or less), heatsink power MOSFET to disconnect your ESC. Even at 5mohm, that's 1/4W power dissipation max so you shouldn't need too much heatsinking since there are no essentially zero switching losses for your application. You would have to run another line from your MCU to your blower fan enclosure to switch the MOSFET whenever the front panels tells the fan to not be off. Then your MCU can disconnect the ESC and put itself to whenever the front panel is turned off to not deplete your battery,

Instead of an extra line from the MCU to control the MOSFET, you could run the MOSFET standalone within the blower enclosure by driving it with a pulse stretching circuit to drive the MOSFET where the pulse stretching reads ESC command pulses from the MCU. The pulse stretcher needs to stretch any pulses longer than the maximum time expected between ESC command pulses (since it's 50Hz then it needs to be at least as long as 20ms. (I'd make it 100ms or even longer). So that when the first ESC pulse is detected the MOSFET turns on and stays on long enough for the next ESC pulse to arrive to "refresh" the MOSFET and keep it on. Obviously, this circuit would have to be designed to consume less power than the ESC. Probably just a flip-flop clocked by a very slow, power frugal oscillator. If it's a logic level MOSFET, you could probably drive the MOSFET gate directly from the flip flop without a gate driver since it's an application with very infrequent switching.

Also don't forget to consider flyback supression for the MOSFET...though if you do things right you shouldn't need any. MCU controlled means you can ensure the ESC is throttled down before the MOSFET is turned off and pulse stretcher controlled MOSFETs can do the same if the pulse stretching interval is made long enough so that the MOSFET continues to stay on for long enough time for currents to diminish after the last ESC pulse is received (another reason to make it 100ms instead of just 20ms though 20ms is probably already plenty).
Last edited by DKNguyen; May 08, 2017 at 03:03 PM.
May 08, 2017, 02:57 PM
homo ludens modellisticus
Ron van Sommeren's Avatar
Current is proportional to voltage², and proportional to Kv³.
Current = torque x Kv (edit: Kv in SI units! I.e. rad/s)

I meant the forces that are trying to rip a blower/prop/wheel apart, outwards, they are proporional to rpm².
Double rpm and those forces will quadruple. Don't forget to wear eye protection.
Centripetal_force (wikipedia)

Vriendelijke groeten Ron
Last edited by Ron van Sommeren; May 16, 2017 at 06:20 AM.
May 15, 2017, 11:56 PM
Registered User
vollrathd's Avatar
Switching high magnitude DC currents can really be rough on a common relay. You need a relay with real wide contacts, or even a magnetic $$$$ blowout relay. We used magnetic blowout relays at work.

One option would be a half dozen high current MosFets, all wired in parallel. With the MosFets now available, you can pick up something with six in parallel that will easily handle 50 Amps.

For automotive use, surge protection is required. Circuitry is quite simple. Just parallel the MosFets on a piece of perfboard, wire up the gates in parallel. The whole thing can be switched on or off with just the application of 12 Volts DC to the gates.

I've done it, works well. If interested, I can provide a schematic, along with a USA source for a MosFet that will do the job.

On the other hand, very high power MosFets are available as a single unit. This one costs $22, and is rated at 100 Volts DC and 350 Amps.

This unit uses terminal screw connections. Wiring it up only takes a resistor around 5000 Ohms, a resistor of around 270 Ohms, a 15 Volt Zener diode to protect the gate against voltage surges, and a toggle switch or 12 Volts DC to turn it on.

On resistance of this MosFet is only 0.002 ohms, so at 50 Amps, you only loose about 0.1 Volts or so. And at 50 Amps, a piece of aluminum plate perhaps 4 inches square will handle the heat issues. Or, maybe no heat sink at all???

https://www.digikey.com/product-deta...10T-ND/2116933
Last edited by vollrathd; May 16, 2017 at 12:28 AM.


Quick Reply
Message:

Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Discussion ESC getting really hot on idle Nanoc Beginner Multirotor Drones 2 Oct 03, 2016 03:05 AM
Discussion Power consumption at idle Larrythelathe Multirotor Drone Electronics 10 Aug 13, 2015 06:11 PM
Help! slow power loss... OS34... IDLE UP, help d1cky_kds 3D Fuel Heli Flying 3 Oct 25, 2011 12:24 PM
Discussion ESC with Idle function 785boats Scale Boats 6 Mar 30, 2008 02:23 PM
ESC - missing idle adjust screw theKAT Parkflyers 4 Mar 01, 2002 10:26 PM