Search

[Phil Lewis]

Hi Graeme,


Sorry to hear of your loss, I wasn't originally going to reply to your post until I read that you had tested the motor and found it to be fine which leads me to believe that you might try the same set up again in a new plane having invested a lot of time once again in building the airframe.


So, in light of that and especially in your last post welcoming constructive criticism please take this as that exactly that based on my knowledge and more importantly experience which is telling me that your powertrain set up is highly likely to fail you, potentially disastrously so.

What Dualsky have done here is to design a range of motors that will perform the same using different battery configurations, it is easy to see that they are aiming at a no-load motor RPM of approx 8-8.5K RPM, (which is why their 12 cell motor has half the KV of the 6 cell motor), that equates to approx 6,800 – 7,000 with a prop on (85%). They have chosen to produce motors for 6, 8 and 12 cell configurations by using different windings (producing different KV's) but it's then easy to work out that using the same formula if they had chosen to produce a motor for 10 cell use it would have a KV of between 210 and 230KV which is a fairly typical KV of a dedicated F3A motor.

That is a potential problem, more so as the manufacturer quotes a maximum prop size of 20X10 which is what is being used, but that is based on 8 cells not 10, by using 10 cells rather than the 8 it was designed for the motor is being asked to turn (2/8 = 25%) faster.

We all know that air resistance increases to the square not linearly, so the power required to turn the prop 25% faster isn't 25% more it's probably at least 50% more.

Putting more cells through a set up designed for less seldom has good outcomes, typically the ESC will fail first if it doesn't have a lot of headroom and if doesn't fail then the motor is likely to fail soon after as it overheats and shorts out the phases.In a matched system such as this the failure point is most likely to occur when transitioning from low/mid power to high power, such as pulling up into a vertical.

In this situation the motor is asked to accelerate the mass of the motor and prop very quickly and as it does so the prop is temporarily then partially stalled until the airframe picks up speed. That results in a massive increase in current drain as the motor struggles to achieve it's required RPM (37V X 270 X 85% = 8,500). A typical outcome to this is that the ESC will simply blow. The motor will be OK because the ESC blows so quickly that the current doesn't flow long enough to overheat the motor. This ESC (which is actually a Hobby Wing Platinum ESC) has over temperature safety settings but no overcurrent limiting setting. The over temperature setting doesn't get a chance to kick in as a large overcurrent flow to the FET's would mean they would blow straightaway before they even get a chance to get hot.

The likely follow on to this is that as the FETS blow the ESC goes short circuit across the power input wires, (this is quite common, just ask any F5B pilot) and in seconds the battery temperature rises to point of ignition probably blowing the cold soldered joints between the individual cells meaning you lose all servo and RX power as the individual cells are no longer connected. I'm not saying that this is what happened but that this is a likely scenario.

I appreciate what you have said about measuring the current draw on the ground and I don't doubt you but the above situation is what on paper and in theory is what I would anticipate could happen given the components used.

I also agree with Stuart, whilst I also agree with you that you are using a dual receiver set up you aren't using a dual power set up. Your source of power in effect doesn't come from two 5 cell batteries it comes from one 10 cell battery, and if you do use a secondary power source it needs to be isolated from the first or a short across one simply shorts the other out as well. You can either use a shotky diode to do this (very cheap and what Maplin's were good for) or use a sensor switch or its cheaper HK equivalent.

If you want a true redundancy set up RX system you can achieve this now at reasonable cost by using one of these (which use dual power inputs) and two of these (or of course any two Futaba S Bus RX's as S Bus is S Bus like USB is USB universally).

Lastly also please bear in mind that the potential prop RPM at 8,500 is beyond the maximum safe operating RPM for that propeller, so stand well clear when testing.

I really would urge caution with this set up, if it were mine I would try it on eight cells as 3,000 watts on eight cells is still more than enough power.   
Phil     

[Graeme Jones]

Phil - thanks for that, some interesting stuff there. A 120 Amp ESC should have been quite capable of handling 80 Amps, but one thing I didn't do was to retest the current draw after increasing the timing from 15 to 18. Even so, I'd had 3 successful problem free flights, using much more throttle than was applied at the time of failure, when it was only about half open. I'd used about threequarters testing vertical behaviour.


One point, do FET's always fail short circuit? I would have thought the usual failure mode would be open circuit, which is what happened to the only other ESC failure I've ever had. That was in a Hobbyking Vampire, where the ESC went bang upon connecting the battery to it. Fried most of the FET's but the BEC still worked.


On APC propellers, yes I'm aware of their RPM limit (150000/diam in inches). I already have a 17X10 and 18X10 to try instead.


Coffee break over, back to the workbench


Graeme

[Phil Lewis]

Hi Graeme,


In my experience (almost exclusively through others) there seems to be three ways an ESC can fail, the way you describe, another is where they only get to what seems like about half throttle and the way I described in my post where the ESC shorts the battery.


The first two can happen at any time and for no apparent reason but it appears from experience that when an ESC receives a severe current overload there is a good chance (but not always) that it will fail short circuit. I have no idea what dictates that it does or it doesn't.


An 18X10 prop gives 19% less airflow per RPM and a 17X10 gives 28% less if that helps.


I'm still intrigued though as to why you are getting such low current flow, assuming Dualsky figures are correct they are getting 100 amps on 8 cells and every theory in the world says that on 10 cells you should get considerably more current draw.


However as this is a 28 pole motor it should need high timing of about 25deg and that might release a bit more power (although I wouldn't expect the difference to be that great).


Unless you have a spare ESC to use if you need to get a new one you might want to consider a HK YEP ESC as they have over current protection as well as overheat protection, in theory no matter what you can't blow that ESC because the programming won't let you (a useful argument if invoking the guarantee).


Phil

[Mike Wood]

Hi Graeme

> ... Right hand side- Pack A 5000 5S with SBEC A powering receiver A driving right aileron servo, right elevator servo and ESC...

I think there would be an advantage to having one receiver driving opposite aileron and elevator halves i.e.right aileron with left
elevator on one receiver, which would result in a working surface on both sides of the fuselage, in the event that one receiver is lost.

Mike

[Graeme Jones]

Phil - I haven't seen Dualsky's test results for this motor apart from what's on their online shop. That doesn't show current for any specific setup, just a range of batteries (8S to 12S) and props (18X10 to 20X10). It does show the idle current on 10V as 1.8 amps. I assume that this is the current at full throttle but with no prop, rather than at idle RPM of, say, 1000. This raises the possibility that my wattmeter is reading low by a considerable amount. I tested it after the crash using a Kforce 70HV and on 6S only got 1.5A at full throttle with no prop. With a 17X6 it was around 40A on 6S. I'll check my wattmeter out later. I'll have a look at the YEP's but didn't I read a lot of bad reviews of them?


Mike - practical advantage or aerodynamic?


Graeme

[Phil Lewis]

Hi Graeme,


I was looking at these specs, that quotes 3,000 watts as max power at (roughly) 30 volts i.e. 100 amps and quotes 100 amps as max, looking at the graph below for the twelve cell motor that is clearly the max power figures for a 20X10 prop.


I have no idea of the real life relevance of the current draw at idle but I doubt idle can mean at full throttle, incidentally NEVER run an electric motor at full throttle without a prop, you are liable to damage it for no real gain at all. With no prop there is virtually no resistance to turning the motor faster so the current draw will be very low.


Assuming you are testing the 6 cell and 70 amp ESC with the sane motor I wouldn't necessarily say your meter is out, you are using a 3" smaller prop with 2" less pitch and using 6 cells your motor is only going to about 60% RPM, the squaring effect squares down so 40 amps doesn't sound immediately all that wrong to me. Only way to check is with another known quantity meter.


Phil

[Leslaw Przytocki]

Phil - I haven't seen Dualsky's test results for this motor apart from what's on their online shop. That doesn't show current for any specific setup, just a range of batteries (8S to 12S) and props (18X10 to 20X10).


Graeme

   Hi Graeme,


GA3000.7 where the spec is showing 8S TO 12S for the 270kv motor?

Why you used 10S for 8S motor?

For me is clearly max prop 20x10 but for 8S not to 12S... (for 270KV motor).


 

[Graeme Jones]

Phil - yes, I see what you mean, but the burst current of 100A is with 15S battery, prop not specified. What is wrong with running these motors with no prop? A motor of X RPM per volt is not going to turn any faster is it? It just won't pull the current that would be needed to turn a prop as well. Brushed motor, yes, since one that would turn at, say, 10000 RPM with a prop could easily rev at 2, 3 or even 4 times that with no load before something broke. I didn't test the GA3000 with no prop on 10 cells, but I have tested smaller motors on 6 cells to check how accurate the "KV" figure was. Using a cardboard disc with a black stripe to simulate a prop and give something for the tacho to read. Never had any problems and they don't over-rev.


Leslaw - sorry but I don't quite understand your point. The manufacturers spec. shows data for 8, 10 and 12 cells. Are you saying that it won't cope with more than 8?


I've now checked my wattmeter against a DMM and it is accurate, at least up to the maximum 10A of the DMM. My new ESC, from Bondaero, should be here later today so when I can find time I'll do some more extensive testing. I'm up to my elbows in glass/epoxy for the rest of the day, making the tubes taht the wing and tail joiners fit into.


Graeme

[Phil Lewis]

I'm interpreting their graph of the 12 cell motor and guessing they did the same with the 8 cell one. The graph shows the current attained with the largest prop on it is the same as in the table above at (as I recall) 75 amps. So my guess is that repeating the test with the same prop and 8 cells on the 8 cell motor the graph showed 100 amps (as in the table above).


I can't see why you shouldn't run a motor flat out without a prop but all manufacturers I know off tell you do not do it, my logic therefore says if there is nothing to gain by doing so why risk it?


Anyway when you get your new ESC you will be able to try various combinations (no of cells and different props) then see which works best for you.



Phil


(up my arms in the swimming pool and Spanish sunshine for the rest of the day).

[Graeme Jones]

I've just realised, while my brain was in neutral mode spreading epoxy, that the first part of my last post was wrong. The burst current is of course for 15 seconds, not 15 cells!!! Doesn't epoxy go off quick in this weather?

Phil - yes,I agree with you re. the graph, and yes I'll do a lot of ground testing before the new one flies, starting with a smaller prop.

By the way, I may not have described the problem on the last flight quite correctly. The grinding noise did not start when I opened the throttle. I'd done that gradually from the bottom of the previous manoeuvre, so it had been at about 3/4 power for a few seconds. This was different to the previous occasions, where it only happened when power was increased.

Graeme

[Leslaw Przytocki]

Leslaw - sorry but I don't quite understand your point. The manufacturers spec. shows data for 8, 10 and 12 cells. Are you saying that it won't cope with more than 8?

Graeme

here is specification and they are for 6,8 and 12S (not for 8,10 and 12):
http://shop.dualsky.com/ga3000-motor_p0147.html


Your motor is dedicated for 8S (GA3000.7) others are for 6 and 12S so something is wrong might
i have wrong specification?

[Mike Wood]

Hi Graeme

I was thinking aerodynamic advantage.

Mike

[Graeme Jones]

Leslaw - yes, the nominal cell counts are as you say. Nominal, not maximum or minimum. I don't think this is relevant to the fire, since the ESC was the apparent failure point, and it's maximum is 12 cells and 120 amps, 150 in short bursts. That's about 6.8 KW!


Graeme

[Leslaw Przytocki]

Leslaw - yes, the nominal cell counts are as you say. Nominal, not maximum or minimum. I don't think this is relevant to the fire, since the ESC was the apparent failure point, and it's maximum is 12 cells and 120 amps, 150 in short bursts. That's about 6.8 KW!


Graeme

Graeme,
motor for 12 cells it is not the same motor as yours -has different kv. They are not the same inside- outside is no different. It looks like you think- the specification  is showing the same motor with different setup cells/esc/props- NO.
There are 3 different motors,yours is the middle for 8S and max recommended prop size 20x10.
For the motor 8S you using 10S battery and 20x10 prop....
Sad story with Catfish- I really liked your design (have seen at Buckminster) and the highest build quality- respect:)

EOT for me and good luck with the next project.




   Best Regards


      Les

[Graeme Jones]

Just a short update on the rebuild. Target date for first flight is still 10 August. Progress so far is that the wings, winglets, fins, rudder and cowl are all built and ready to cover. The fuselage and tail are about half done, awaiting an order from HK with some 13mm aluminium tube for the AMT pivot mechanism (12mm carbon tube joiner). If I can find time I'll post some photos.


Graeme