If you need assistance, please send an email to forum at 4hv dot org. To ensure your email is not marked as spam, please include the phrase "4hv help" in the subject line. You can also find assistance via IRC, at irc.shadowworld.net, room #hvcomm.
Support 4hv.org!
Donate:
4hv.org is hosted on a dedicated server. Unfortunately, this server costs and we rely on the help of site members to keep 4hv.org running. Please consider donating. We will place your name on the thanks list and you'll be helping to keep 4hv.org alive and free for everyone. Members whose names appear in red bold have donated recently. Green bold denotes those who have recently donated to keep the server carbon neutral.
Special Thanks To:
Aaron Holmes
Aaron Wheeler
Adam Horden
Alan Scrimgeour
Andre
Andrew Haynes
Anonymous000
asabase
Austin Weil
barney
Barry
Bert Hickman
Bill Kukowski
Blitzorn
Brandon Paradelas
Bruce Bowling
BubeeMike
Byong Park
Cesiumsponge
Chris F.
Chris Hooper
Corey Worthington
Derek Woodroffe
Dalus
Dan Strother
Daniel Davis
Daniel Uhrenholt
datasheetarchive
Dave Billington
Dave Marshall
David F.
Dennis Rogers
drelectrix
Dr. John Gudenas
Dr. Spark
E.TexasTesla
eastvoltresearch
Eirik Taylor
Erik Dyakov
Erlend^SE
Finn Hammer
Firebug24k
GalliumMan
Gary Peterson
George Slade
GhostNull
Gordon Mcknight
Graham Armitage
Grant
GreySoul
Henry H
IamSmooth
In memory of Leo Powning
Jacob Cash
James Howells
James Pawson
Jeff Greenfield
Jeff Thomas
Jesse Frost
Jim Mitchell
jlr134
Joe Mastroianni
John Forcina
John Oberg
John Willcutt
Jon Newcomb
klugesmith
Leslie Wright
Lutz Hoffman
Mads Barnkob
Martin King
Mats Karlsson
Matt Gibson
Matthew Guidry
mbd
Michael D'Angelo
Mikkel
mileswaldron
mister_rf
Neil Foster
Nick de Smith
Nick Soroka
nicklenorp
Nik
Norman Stanley
Patrick Coleman
Paul Brodie
Paul Jordan
Paul Montgomery
Ped
Peter Krogen
Peter Terren
PhilGood
Richard Feldman
Robert Bush
Royce Bailey
Scott Fusare
Scott Newman
smiffy
Stella
Steven Busic
Steve Conner
Steve Jones
Steve Ward
Sulaiman
Thomas Coyle
Thomas A. Wallace
Thomas W
Timo
Torch
Ulf Jonsson
vasil
Vaxian
vladi mazzilli
wastehl
Weston
William Kim
William N.
William Stehl
Wesley Venis
The aforementioned have contributed financially to the continuing triumph of 4hv.org. They are deserving of my most heartfelt thanks.
Registered Member #56
Joined: Thu Feb 09 2006, 05:02AM
Location: Southern Califorina, USA
Posts: 2445
Something about power dissipated going as the square of current in a resistor... It it actually worse in a 3 phase drive since at any given point of time all of the motor current is going through 2 of the fets, so despite the fact that there are 6 devices on the board the total 'effective' on resistance for power calculations is 2x the on resistance. Then throw in the fact that the high side device is being PWMed at some non-100% duty cycle (so peak current is well above RMS), and the total power dissipation is even worse than the theoretical best of motor current * 2 * on resistance. Throw in some switching losses, copper losses, and 10w seems like a reasonable number for a device like that.
Your plan to add a little heat spreader seems like a good one, what I have seen done in the past is just sticking a piece of thermally conductive film (like that which you would use in lieu of thermal paste to go between a transistor and heatsink) and then an aluminum plate to act as a heat spreader. This helps equalize the temperature between the low/high side devices and helps conduct heat to the outside world better. Whatever you do I would recommend load testing the copter (maybe give it an extra big battery so it can't get airborne? and usual motor currents?) to avoid further damage. Also don't forget that the worst thermal loading may not be at 100% duty cycle due to the aforementioned reasons. IIRC 50% duty cycle is the worse for a motor which draws current directly proportional to voltage, and <1% is the worst for a motor which draws constant current. I would expect it to be around 70-90% for a propellor.
It wouldn't put it out of the question that you just got a 'bad' esc, in my experience with buying similar controllers (intended for ebikes instead of airplanes) even within the same model number and outward packaging you can get a completely different device. I once ordered a lot of 10 controllers and within the lot I got ones with completely different internal design, they did not even have the same microprocessor.
Registered Member #2906
Joined: Sun Jun 06 2010, 02:20AM
Location: Dresden, Germany
Posts: 727
so despite the fact that there are 6 devices on the board the total 'effective' on resistance for power calculations is 2x the on resistance
Good. Then 2*2mOhm*30A^2 = 3.6W Worst case.
(so peak current is well above RMS)
Its per definition not true: a 32A rated mosfet shall not be exposed to currents above 32A. Thats why i wrote
generally a 30A switching device will not drive a 30A motor if the owner is in anyway sane....
That still makes the worstcase 3.6W valid. (and overestimated!)
Then throw in the fact that the high side device is being PWMed at some non-100% duty cycle
Fine. So its less then 3.6W then.
and 10w seems like a reasonable number for a device like that.
.........didnt you just say the complete opposite? You actually argued yourself that the conduction losses are below 3.6W... some magic copper losses in short traces and filled copper areas which will not significantly contribute and switching losses blow that up to 10W? No Way. Yes, Switchginlosses are huge, Hen918 is totally right. But that should not exceed the conduction losses. So we are talking about around 5W which is less then 12W*0.5 which i tried to explain above. This PCB with all the copper area is well capable of disipating the heat....
The actual problem here seems that people have a 30A regulator and think that they can load it with 30Amps continous... thats totally INSANE. The 30A may only be peak currents, that means maximum acceleration and motorstartup but that might not be in any case the normal load since every acceleration event will then cause the current to rise way above the limit for the regulator which causes unavoidable component failure.
The flight above last about 4 continuous minute at 30 amps or so before it blew.
Any questions?? Really? The regulator PCB is totally fine. it just was abused
You can add a heatsink as much as you want, the Mosfets and the Bondwires arent made for that load. the whole setup is wrong and missdesigned and what even worse... the finger is pointed at the totally fine PCB.
Registered Member #230
Joined: Tue Feb 21 2006, 08:01PM
Location: Gracefield lower Hutt
Posts: 284
Patrick Take a close look at that data sheet heaps of watts capable at 25 degrees and only 3 watts of dissipation at 70 degrees Heatsink away my friend
Registered Member #2906
Joined: Sun Jun 06 2010, 02:20AM
Location: Dresden, Germany
Posts: 727
please give correct information if you do...
heaps of watts capable at 25 degrees and only 3 watts of dissipation at 70 degrees
Thats due to the thermal resistance from the die to the PCB so that the Die is below 150°C. The Powerdisipation is below 1W per Mosfet, so there is absolutely no problem with powerdisipation... the PCB can get as hot as 110°C and thats actually bound by the FR4 material not the thermal resistance of the Mosfets. (which get only as hot as 132°C in that case - that is actually ok if my power estimation is off by a factor of 2!)
What do you want to achieve with a Heatsink? The design is perfectly good engineered. It ran out of spec.. thats all. And a heatsink will not prevent damage, it will only take longer to be fatal given that systematic abuse when you use a 30A switching device to have a 30A continous load and the current demandingpowerspikes on top of it.
Registered Member #2431
Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639
OK. Good debate, good info to consider.
I love to find a topic where i pull the pin on a hand-grenade, stand back and watch. Ill be remmebering all these comments and issues when im designing tiny 1 amp motors and 6,000 amp submarine motors.
The below pics show practice work done on the failed ESC. I have one spare witch will now be heatsunk then flown with a in-flight data logger. Ive bought 2 new ESCs from china. Im wondering if i should buy better transistors or parralel them. But even then that presumes proper heat control.
holes and notches for bending.
Bent. Its 0.032 inches thick brass.
Didnt bother to bend it 90 degrees.
Testing the fit.
Silver epoxy, very little.
The "squish" test, to verify, no overflow.
As per DerAbi's point, ill use the data logger to mearsure the real battery current into the device. it will record peak too. the tail ESC always works abit harder then the others, as it responsible for yaw and lift. so a diagonal vector means more power.
Registered Member #580
Joined: Mon Mar 12 2007, 03:17PM
Location: Melbourne, Australia
Posts: 410
you didn't by any change upgrade the version of SimonK on the ESC did you? or change any settings like comp_pwm (active braking), with that setting on, i've had it blow the same mosfet on the same ESC twice, turns out the motor had 2 windings that were twice the resistance of the other. replacing the motor fixed it.
Registered Member #56
Joined: Thu Feb 09 2006, 05:02AM
Location: Southern Califorina, USA
Posts: 2445
It will be interesting to see the result of the logging test.
Since some seem skeptical of my comments I will explain my reasoning, which is based on simple power electronics theory. For a more theoretical discussion see the excellent whitepaper written by Fairchild at which talks about a few more sources of loss which I have left in this simple analysis. I also assume the motor commutation is slow compared to the PWM duty cycle so those switch cycles can be ignored. Consider the controller is running at a 16khz PWM frequency with 1us switching times, 12v input voltage (maximum rated for this controller is 16v), and 30amps out of the battery, and 50% PWM duty cycle. If we assume that the motor has sufficient inductance that the current is constant during the pwm 'on' periods (which is quite generous with modern airplane motors, in real life you will further increase your peak current through the transistors due to current variations during each pwm 'on' period) and appears resistive, then by necessity the current during the pwm 'on' periods will be 60amps. The conduction losses will then be (60amps)^2*4milliohms*50% duty cycle = 7.2 watts. This will then be spread equally across the 6 transistors for an average power dissipation of 1.2watts per transistor due to the conduction, for transistors at 25C (this number increases to 1.4w per transistor as the die temperature rises to 100c, as per figure 7 of the data sheet).
Now consider the switching losses. We assume the designers of the controller were smart and only PWMed the high side devices, since you do not need to PWM both. During the 1us switching time you will be dissipating worst case 60amps*12volts*1microseconds = 0.72millijoules. In real life you do a lot better than the theoretical worst case, I typically use half the theoretical worst case, which gives .36millijoules per switch, or 0.72millijoules per PWM cycle. Multiply by the 16khz pwm frequency and get 11 watts. This is shared across the 3 high side switches, and comes out to 3.6watts per transistor.
The copper losses are hard to estimate without knowing the board details but consider that for heavy 350um (so called 4oz) copper, the sheet resistance is about .12mR per square. I count about 4 squares from the + wire to the center of the bridge, which comes out to 0.48mR for the positive side, lets assume the bottom is the same. Going back to the full 60amps at 50% duty cycle we get 1.8watts for the copper losses. Note - because the input filter capacitor is on the far side of the traces the full 60amps peak flows through those traces, not the 30amps average.
So, for this calculation we get a total of 7.2+11+1.8=20 watts For a device that should only dissipate 3.6w "worst case". Things get better as you go to higher duty cycles, which shows the importance of getting the correct motor KV so you can keep the duty cycle as high as possible. Also, it is possible that they used slightly better gatedrive than on the devices I have worked with, or that they run at a lower switching frequency, which might reduce the switching losses down to the general target rule of thumb of 'make your switching losses equal to conduction losses'
The real saving grace for these suckers is that they are designed to be mounted in the prop wash so they have plenty of forced air cooling, but even with that using a device with 2mR transistors and rating for at 30amps is a bit optimistic without any sort of additional thermal management. In any case, considering that the device is running at 360w average input power, that comes out to 94% efficiency under these 'pessimistic' conditions, and quickly rises in the 97-98% range as the conditions get more favorable which is pretty good for a $14 device
Registered Member #2431
Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639
Yes the miracle of modern copper and silicon is amazing. And yes, 14 $ stunning. The day and age we live in !
ive got the heat sink on the good ESC, will fly tomorrow.
The 1250kv motor, 30 amp ESC and 14 inch props are all important. As im trying to increase lift and endurance. Ash and i are working on the 30" prop. but the ESC and motor remain an issue. Im thinking of desoldering the existing transsitors and put in 200 amp ones. (thats the super sized motors and props though.)
i dont understand motors, alternators and generators as well as i should.
Registered Member #2906
Joined: Sun Jun 06 2010, 02:20AM
Location: Dresden, Germany
Posts: 727
... your estimations are somewhat way off...
first of all: No one, absolutley no one uses 32A rated Mosfets to PWM a current of 60A. This is an assumption that needs to be prooven and actually if true, the load is then way too much for the Mosfets. As i said before: in constant overload case there will be nothing to prevent mosfet death in the long run. So in the end you quadrupled your conduction loss only by a) making false assumptions or b) running out of spec. Thats a bad argument, sry. For your information: if you have a switchingdevice rated for x Amps, and you have a dutycycle of 50% your max peak current is only allowed to be sqrt(2) times higher. Buy heeeey for the sake of argument and spreading fear and misinformation we do everything, right?
Switching losses:
During the 1us switching time you... 60amps*12volts*1microseconds
WTF. Are you really saying that the engineer goes for 1us switching time?? Thats plain bullshit, sry. The mosfet switches in about 20ns, (says the datasheet, not some magic imagination of overimpressed people who cant bevlieve what going on here) Thats not factor 2 off by yor estimation, thats factor 50! But ok. beeing faithfull, i wont half the estimation, i will double it (for what reason ever...) so your 11W shrink down to 0.88W and that will distribute over 6 mosfets which endup considerable less than the conduction loss. (Which it should, if the designer did their homework). So you still getting only 1.6W or so a the mosfet allowing 113°C ambient which is again bound by 110°C FR4 operating temperature. And again the design is totally ok. (Even when using your 60A idea which is also wrong. swichting losses are therefore even less if you stay in spec)
Gongrats, you made your self an argument by a) assuming out of spec operation b) making false assumptions c) assuming switching speed that are insanely slow... slower than most IGBTs btw . d) i refuse to even start about your assumptions with the copper loss You did not understand what resistance per square means.., sry And actually used 60A again... just to run out of spec and let your self having a 4 times higher number in the end without any reason. still getting negletible 1.8W)
You also could have said 10us switchgintime.... that would have beed way more in your flavour and equaly useless.
All you where saying is what "if you do it wrong it wont work", but the PCB - and i feeling like i repeat myself - is not doing anything wrong.
This site is powered by e107, which is released under the GNU GPL License. All work on this site, except where otherwise noted, is licensed under a Creative Commons Attribution-ShareAlike 2.5 License. By submitting any information to this site, you agree that anything submitted will be so licensed. Please read our Disclaimer and Policies page for information on your rights and responsibilities regarding this site.
Three Phase Electronic Speed Control In-Flight Blow-Out.
For a device that should only dissipate 3.6w "worst case". Things get better as you go to higher duty cycles, which shows the importance of getting the correct motor KV so you can keep the duty cycle as high as possible. Also, it is possible that they used slightly better gatedrive than on the devices I have worked with, or that they run at a lower switching frequency, which might reduce the switching losses down to the general target rule of thumb of 'make your switching losses equal to conduction losses' 
. 
And actually used 60A again... just to run out of spec and let your self having a 4 times higher number in the end without any reason. still getting negletible 1.8W)