when is an ESC being overworked?

[Snaaakeey]

The above is spot on but this may help to ( videos and pics always help)

NEW - ESC vs ESC - HeliFreak

and this vid is taken from later on in the thread. It should explain why ESCs heat up under parcial throttle and what active free wheeling is .

How BLDC Active Free-wheeling (Synchronous Rectification) works.. - YouTube

[trillian]

The Kontroniks and YGEs have active freewheeling. I was reading an article about his not long ago.

Many thanks to everyone for your input. At the moment I think I'll continue with the 12s setup at around 2100/2150 ish for idle-up1 and bump down the normal/hover mode a little, to maybe 1750 /1800 and I'll keep a close monitor on the ESC temps.

It certainly seems more than capable of handling the gearing just in power / torque terms.

A problem with ESCs is due to the freewheeling current.

The motor windings have significant inductance, so when the FETs turn off, the current can't just stop flowing.

What ESCs do is to remove the battery voltage, but still provide a path for the current to freewheel down to zero. By default, this is through a sizable diode in the FET.

This works - however, the diode forward biases at ~ 0.7V. 0.7V * 50A is 35W and there may be two diodes doing this (one in each FET). Hence, heat & inefficiency.

Now, this freewheeling happens each time the current is turned off; and this kind of equates to more at lower throttle settings. [There was some hand-waving there...] . Hence some ESCs can get hot at low throttle settings (possibly also at higher PWM frequencies).

There are various techniques to deal with this. One is to only turn one FET off, so the other doesn't drop 0.7V. I believe later CC ESCs do this, and alternate between the two FETs as to which is off (spreading the heat - and upsetting RPM sensors...).

I believe Kontronik have the option of Active Freewheeling. Rather than have current flow through a diode; the FET is carefully controlled to act like a diode. When the current wants to flow in the direction that a diode would allow, the FET is switched on; when it starts to reverse, the FET is switched off. ie the FET is like an active diode: more complicated, but doesn't drop 0.7V while passing big currents.

Is this any help..? Er, maybe...

I reckon that provided your throttle settings aren't silly, and your ESC is well specced (ie it's well capable of supplying the current you want); you may well be ok. Keep an eye on how hot it gets. [Maybe do some test runs in the garden, -ve CP on the floor ?]

And if you blow it up, you have an excuse to buy a Talon 90, which seems to be well-priced...

[Snaaakeey]

Your setup is very similar to a kit 6hv. Same esc, gearing and head speed and battery size. 300+ flights on mine and it hardly warm. I think you' ll fine too

[trillian]

Your setup is very similar to a kit 6hv. Same esc, gearing and head speed and battery size. 300+ flights on mine and it hardly warm. I think you' ll fine too

I hope so because it flies really nicely and I love the power! Next time out I'm trying a set of slightly lighter narrower chord blades. Even though it's heavier than the 6HV the power is amazing compared to what I'm used to. The HW ESC is so 'plug'n play' and just works without any fiddling and I think I paid something like £35.00 for it 2nd hand.

[cjcj1949]

The diodes used are fast recovery schottky and are 0.2V. 0.7V is a typical silicon value and varies with current. Active is better, if the active device is better than the diode at the current being used.

[cjcj1949]

Very good test that. It looks a though the fets are fully on for full throttle so less energy dissipated. Looks as though the answer is that it is very complicated and will depend on many factors. Looks like one that needs detailed knowledge of the esc, motor and load and needs the waveforms looking at to see what is happening. Practically checking your own installation by flying and measuring the temperature rise would tell if you are safe. Running shorter flights to begin with might be useful.

[scallybert]

The diodes used are fast recovery schottky and are 0.2V. 0.7V is a typical silicon value and varies with current. Active is better, if the active device is better than the diode at the current being used.

I can quite believe you on the details. I've just pieced things together from reading between the lines of various postings, and applying fundamental engineering prinicples...

0.2V means 10W @ 50A; compared to 35W. Still a fair amount. ~Corresponding to a resistance of 0.004 Ohms. I'm pretty sure that a bank of good FETs does much better than that, hence the advantage of 'active'.

[trillian]

Just to add to the discussion for those interested. I just setup the ESC for my Rave ENV and in setting up the governor I found this interesting. When you do the governor settings in the Castlelink software you put in all the pertinent details of motor KV, battery voltage (44.4 for 12s) gear ratio etc. You then ask for certain headspeeds and it tells you what throttle percentage to use for each headspeed.

If the speed you want is OK it gives you a percentage throttle to set in the TX and if you open the info box it will say something like (and I am paraphrasing) 'this headspeed is at 90% output blah blah' essentially saying this is good as far as the governor is concerned.

If the governor doesn't think it has enough headroom it gives you a warning saying insufficient headroom. The point being, you can see the boundary where it says it can deliver what you want. (you can actually trick it into accepting a slightly higher headspeed by temporarily upping the voltage a bit and it will then give you the throttle setting you need and you can put the voltage back).

Soooo, the point is I have determined that what the Castle Ice governor wants is a minimum of roughly 83% of the max theoretical possible rpm based on motor KV X voltage. Using this as a guide my setup on the Hirobo should be right on the money for a headspeed of about 2160. So if I was using a Castle Ice and wanted 2160 or slightly lower, it would be telling me that's perfect just as it is.

[cjcj1949]

I can quite believe you on the details. I've just pieced things together from reading between the lines of various postings, and applying fundamental engineering prinicples...

0.2V means 10W @ 50A; compared to 35W. Still a fair amount. ~Corresponding to a resistance of 0.004 Ohms. I'm pretty sure that a bank of good FETs does much better than that, hence the advantage of 'active'.

I'm doing much the same. I did digital logic design mainly and then went into software. The stepper motor drivers I inherited are etched in my memory as they kept blowing up. It cost a lot of money to replace the transistors. We were driving lathes and milling machines. In the end we found the designer had made a mistake with the value of a resistor. A transistor was being saturated when it had been meant to stay out of saturation. If he'd documented his work the problem would have been obvious. I agree that fets are better, I thought I'd said that, but I'm pretty sure this is off the topic of lower drive heat dissipation. It looks as though there are more switching operations at lower throttle settings, but that doesn't automatically mean more power loss. It depends on many factors and for any individual case there will be different results depending on the on resistance of the fets, their switching speed, the chopping frequency, the speed of the moter and its effective inductance at the load being considered. I don't think the recirculated current losses come into it directly, but they may make a differnce indirectly as the di/dt is proportional to V from V= -L di/dt and this would imply less switching being required as the current will take longer to decay to the lower level. However I don't think they measure current, as we did with the stepper drives, I think they switch at a predetermined frequency. It should all be in the documentation, if there was any provided.