/dist/images/branding/favicon

Crash cage or dome

Discussion in 'Cinestar 8' started by Mikko Kosonen, Jul 12, 2012.

  1. William Johnston

    Joined:
    Oct 2, 2012
    Messages:
    140
    Likes Received:
    5
    Hi Josh,

    Your problems and my problems sound a lot alike. (I hijacked your "How hot is too hot to fly" thread on that other forum.) Are you sure that you have a temperature problem rather than a current problem? It may be that the high temperatures you are seeing are just a symptom of the actual problem. Would you do me a favor? Would you take the camera mount off your Cinestar and hover the Cinestar for five minutes (or less if the BL limiting forces it down) and post the GPX file? I would like to take a look at it.

    --William
     
  2. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,383
    Likes Received:
    1,164
    Intriguing thought, William, but could you expand upon it a bit?
    What, apart from the current to the motors, would cause the Bl-Ctrls to run hot?

    Andy.
     
  3. William Johnston

    Joined:
    Oct 2, 2012
    Messages:
    140
    Likes Received:
    5

    Well I have a theory that I have been reluctant to say because if I am right it means that all the XL power boards have a problem. But since you asked this is how it goes.

    Both Josh and I have had the solder on our Deans plugs melt. We both replaced our Deans plugs with heaver plugs that allow us to fly, but we quickly overheat. We both stuck heat sinks on our ESC, but it didn’t really help. So what’s going on? When I do the test that I asked Josh to do, I get this for my average currents for each motor:

    16.4 14.7 13.5 14.9 14.8 17.2 16.9 16.6

    16.6 14.9 13.6 13.7 13.9 17.5 18.4 17.2

    I was using one 4S battery here. I didn’t realize I had a problem; the thing flew. So I loaded it up with two 5S batteries and a 2 axis gimbal with the GoPro. I got to fly a whole 16 seconds and here is that motor profile:

    21.0 19.4 21.2 21.7 23.4 22.7 23.7 23.5

    I now finally asked the guys at Quadrocopter the question I should have asked on day one, “How much current should a Cinestar 8 pull?” There answer was, “between 40 and 90 amps depending on load.”

    Wow I’m pulling a lot more current than I ought to, but why? But then I got a break. Literally, ESC 5 broke. I burned it up trying to take off on another flight right after that one with all the 20’s for the motor currents. Well, I replaced the ESC and ran my first test again. One 4S battery and no camera mount:

    18.0 19.8 14.9 12.7 5.4 13.2 14.3 19.0

    Will you look at that, the ESC I replaced is only pulling 5.4 amps. Other than the new ESC, everything else is exactly the same as before when ESC 5 was pulling 14 amps. And that 5.4 amps is right in line with what Quadrocopter tells me the Cinestar should be pulling. So let’s go with some theories:

    1) I did a miserable job soldering the motors lines the first time and bang up job the second time. Well I don’t think so, and actually the second time I redid all motor lines just to be certain. I would have had to somehow just accidently got it right on the one I happen to have replaced. It seems unlikely.

    2) There was a bad batch of ESCs and Josh and I just happen to get them. The one I put on could have come from a latter batch that was good. I suppose that could be it. But it’s probably the exact same design, coming from the exact same plant, and being manufactured in the exact same way. So maybe this is a little unlikely.

    Well I’m almost out of ideas. But I left out some information. When I was cleaning up the solder on the power distribution board I bungled it and lifted a pad. This forced me to solder the cap on the other side of the board; the side that would be in between the ESC and power distribution board. This little bit of solder kept me from sandwiching the two boards together tightly. There was a little air gap in between. And this was my lucky break, because I think this is why ESC 5 is pulling so much less current than the others. And here is where you say, “That’s preposterous.” Now bear with me and I shall try to explain.

    If you look at the Okto power distribution board, not the XL, but the previous one, you will see that the ESC sits beside the power distribution board in its little nook. The top of the ESC is the positive layer and the top of the power distribution board is the positive layer. When they designed the XL boards they simply moved the top of the ESC under the power distribution board. Now the positive side of the ESC, and not just the positive side, but a big positive copper power pad, is sandwiched up to the NEGATIVE side of the power distribution board. Uh oh, that doesn’t look good. But wait, what about the insulation layer on the power distribution board? OK so we have some insulation there. We don’t have a short circuit. But that’s a thin layer of insulation and it is probably designed for 5 volt printed circuit boards. But still system does not pull a lot of amps when the motors are off (0.45 amps if I recall correctly), so there is no current flowing across the boundary when the motors are off. But is it possible that when the ESCs start pulling a lot of current that current starts leaking through the boundary. Well maybe. It would explain a few things. Obviously it explains why I’m pulling so much fewer amps on ESC 5; the air gap prohibits current flow. But it also explains other things too. For example, motor power is related to the voltage multiplied by the current, so if you increase the voltage you can decrease the current and get the same power. This is why Quadrocopter is recommending going to a 5S battery for the Cinestars. But in another thread Josh says, “I am wanting to stay with a 4S setup since the 5s setup runs hotter...” I noticed that too, on the bench I turned up the voltage on my power supply and the current to the Cinestar went up. Isn’t that the opposite of what we expect? To keep the same rotor speed the current should go down when the voltage goes up, right? But that is exactly what we would see if current were leaking across the insulator. The higher the voltage the more likely the current is going to cross the insulator.

    One thing that bothered me, after I finished building my CS8 is that there weren’t any tests I could to do to check the correctness of my build. May I suggest a one; hover your Cinestar without the camera gimbal for five minutes then average the motor currents. You should get between 5 and 6 amps for each motor. For example, your average motor currents should look something like this:

    5.3 5.6 5.2 5.8 5.5 5.3 5.1 5.6

    At least, that is what I think they ought to be.

    And now you’re wishing you hadn’t asked.
    --William
     
    Morgan Friedland likes this.
  4. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,383
    Likes Received:
    1,164
    Thanks for the detailed answer, William. I'm still digesting it, but here's the Current/Temperature data for a short flight I did with a C8, 3-Axis gimbal, no camera mounted, with a 4S 6200.
    The left hand group is the current in units of 0.1A (so 37 is 3.7A). The right hand group is ESC temperatures in degrees C.
    The ambient air temperature was 23C. The wind was 5 mph, gusting to 8 mph, which may explain the disparity in the current for the motors -- boom 1 was pretty much pointed into wind.
    The C8 was hovering under manual control out of ground effect at about 8 feet. Each reading is one second after the previous.

    Let me digest the rest of your email and then see if I can comment on it.

    Andy.

    37/39/66/53/14/21/26/30/0/0/0/0/ 28/28/30/30/26/27/29/30/0/0/0/0/
    37/39/66/53/56/62/54/30/0/0/0/0/ 28/28/30/30/28/30/31/30/0/0/0/0/
    37/39/66/53/56/62/54/65/0/0/0/0/ 28/28/30/30/28/30/31/32/0/0/0/0/
    38/39/66/53/56/62/54/65/0/0/0/0/ 33/28/30/30/28/30/31/32/0/0/0/0/
    38/64/44/53/56/62/54/65/0/0/0/0/ 33/33/34/30/28/30/31/32/0/0/0/0/
    38/64/44/58/41/66/54/65/0/0/0/0/ 33/33/34/34/33/34/31/32/0/0/0/0/
    38/64/44/58/41/66/37/63/0/0/0/0/ 33/33/34/34/33/34/36/37/0/0/0/0/
    38/64/44/58/41/66/37/63/0/0/0/0/ 33/33/34/34/33/34/36/37/0/0/0/0/
    37/59/35/58/41/66/37/63/0/0/0/0/ 35/36/36/34/33/34/36/37/0/0/0/0/
    37/59/35/50/34/66/37/63/0/0/0/0/ 35/36/36/37/35/34/36/37/0/0/0/0/
    37/59/35/50/34/67/50/75/0/0/0/0/ 35/36/36/37/35/37/38/40/0/0/0/0/
    37/59/35/50/34/67/50/75/0/0/0/0/ 35/36/36/37/35/37/38/40/0/0/0/0/
    47/58/35/50/34/67/50/75/0/0/0/0/ 37/38/36/37/35/37/38/40/0/0/0/0/
    47/58/41/45/34/67/50/75/0/0/0/0/ 37/38/37/39/35/37/38/40/0/0/0/0/
    47/58/41/45/38/64/46/75/0/0/0/0/ 37/38/37/39/37/39/40/40/0/0/0/0/
    47/58/41/45/38/64/46/68/0/0/0/0/ 37/38/37/39/37/39/40/44/0/0/0/0/
    53/58/41/45/38/64/46/68/0/0/0/0/ 39/38/37/39/37/39/40/44/0/0/0/0/
    53/66/39/45/38/64/46/68/0/0/0/0/ 39/40/38/39/37/39/40/44/0/0/0/0/
    53/66/39/48/40/71/46/68/0/0/0/0/ 39/40/38/40/38/41/40/44/0/0/0/0/
    53/66/39/48/40/71/41/69/0/0/0/0/ 39/40/38/40/38/41/42/45/0/0/0/0/
    53/66/39/48/40/71/41/69/0/0/0/0/ 39/40/38/40/38/41/42/45/0/0/0/0/
    34/22/39/48/40/71/41/69/0/0/0/0/ 40/40/38/40/38/41/42/45/0/0/0/0/
     
  5. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,383
    Likes Received:
    1,164
    Ha! I blew a message limit when I tried to reply to your message.
    I'll split the reply across two messages.

    Tabb: The limit is 10,000 characters. Can this be increased a bit please? :)

    The only rational explanation for that is the current was so high that it, combined with the resistance of the connector, heated the connector up.

    I just measured (with a Fluke VOM) the resistance across a brand new male/female Deans connector from Pololu. It is was somewhere between 0.1 and 0.2 ohms.

    So let's imagine you've got 60 amps at 15 volts going through that connector.


    Whoa....that's 125 Amps total. That's waaay above the rated maximum for a Deans connector (which I understand to be 40-60 Amps). Little wonder the plugs melted. The Joule heating on that is 1,250. (125 x 125 x 0.1). That's 1,250 Watts per second!



    Yeah. That's the effect of the 5S -- increase the voltage across the resistive load and you'll square the heating effect.

    Right. That's what I had heard.



    That 5.4 does stick out a bit like a sore thumb doesn't it?



    As you say, that seems somewhat improbable.



    Agreed. Let's apply Occam's Razor for the time being.

    See next message....
     
  6. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,383
    Likes Received:
    1,164

    Hmm. There you have me. I don't know which board you're looking at.
    Do you mean the standard one that comes with the Cinestar 8?

    Can you post an image of this board? If you don't want to disassemble your Cinestar can you find an image on-line?

    Hmm. I don't understand why, electrically speaking, increasing the current flowing through the ESC would cause an electrical resistance to break down? If the ESC gets hotter, that might alter an electrical resistance....

    Also, without seeing the board, I can't be sure whether the negative side of the PDB has a ground plane there, or has some connections, or is actually just circuit board.
    Well, if you view the motor as a simple resistive load, then yes. But I'm not sure that is the case when you have a motor dragging a propeller threw the air and slowing it down. I suspect that going to the higher voltage allows it to draw more power and spin the propeller faster, but pulling more current -- which is why the conventional wisdom was that things run hotter (the heating being proportional to square of the current).

    I'll happily defer to any other forum members who are EEs and think otherwise.

    Oh. I didn't realize that QC was recommending going to a 5S battery.
    Yes. That's strong empirical evidence that your expectations might be at odds with reality! (Don't kill the messenger please! :) )

    But are you getting the same rotor speed? Motors are rated in units of Kv -- which I've always taken to mean the number of rpm per volt. More voltage, more rpm (moderated by those draggy propeller thingies, of course).



    That's true, with one modification: An insulator is an insulator. You're assuming (not unreasonably for your hypothesis) that the "insulator" is actually a resistance and subject to Ohm's Law.

    Why not measure the resistance across one of those areas that you suspect is behaving like a resistance instead of an insulator?



    I agree. That's more in line with what I'm seeing.


    Not at all. I think you might be on to something. The elevated current draw you are seeing is a surprising fact. Swapping out the ESC #5 and seeing the current draw drop dramatically is also undeniable.

    All that remains to be answered is why the current draw is so high for the other motors now -- and I don't think it's the motors, but perhaps it could be the firmware on the ESC that is different or something else about the ESCs?

    I'd start with a Volt-Ohm-Meter and try and getting a resistance reading in the areas on the PDB where you think there's a current leak.

    Andy.
     
  7. William Johnston

    Joined:
    Oct 2, 2012
    Messages:
    140
    Likes Received:
    5
    Well there are two power distribution boards the "Okto" and the "Okto XL." The "Okto" is the older one, and it's now on sale at Quadrocopter :
    http://www.quadrocopter.com/Okto-power-distribution-board_p_31.html

    The problem I imagine would not have manifested in this one as the ESCs do not overlap the board in this version, rather they dangle by their wires. A design that could only be thought of after a really good Oktoberfest.

    This page shows you how the ESCs were attached:

    http://www.mikrokopter.de/ucwiki/en/QuadroPowerDistribution?highlight=(power)|(distribution)
     
  8. Adam Paugh

    Adam Paugh Distributor

    Joined:
    Jun 29, 2012
    Messages:
    323
    Likes Received:
    78
    Fellas,
    The top of the distro ring is "+" and the bottom is "--". These layers must also be separated by an insulator similar to the solder mask that covers the entire PCB. The surface interaction of these two internal layers is in magnitudes, greater then the 32 vias that are on the Xl okto distribution ring that stitch the ESCs into the PCB via caps.
    5.4amp is a glaring anomaly in the data set and being that it was a replacement ESC makes this argument worth bringing to the manufacturer.

    I will see that I can C/P this thread and send it along to Holger Buss and will report back with the findings.

    Further, it is recommended that 25V capacitors are installed at C1,2,3,4 on the xl okto distribution ring for 5S applications.
    Greetings,
    Adam
     
  9. William Johnston

    Joined:
    Oct 2, 2012
    Messages:
    140
    Likes Received:
    5
    Yes the weak point of my argument. I'm pretty sure there isn't any power drain, other than the computer boards, when the engines are off. And there seems to be an undo amount of power drain when the engines are running. More power is being drained than is required by the engines. And that power drain is measured by the ESC so it has to be current flowing through the ESC and not someplace else. So I imagine there must be a leak in the ESC. My best candidate is current flowing from the ESC through the insulation (working like a resistor) to a negative ground plane on the power distribution board. My only evidence is that when there was an air gap there there was no undo power drain. Of course, that could be a coincidence and it could be something else entirely.
     
  10. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,383
    Likes Received:
    1,164
    Hi Adam: Actually, I suspect the 5.4 Amps is closer to normal. It's the other amperage readings that worry the hell out of me? That's a huge amount of current and no wonder the Deans connectors are melting.

    Thanks for offering to escalate this up to Holger. We'll await his response with bated breath....

    Andy.
     
  11. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,383
    Likes Received:
    1,164
    I see a drain of about 0.5 Amps at 16 volts when the motors are off. Not enough to arouse suspicion, though.

    I agree with you -- the current those ESC and motors are sinking is an order of magnitude higher than it should be. I'm not sure it is a "leak" in the ESC (check underneath the Copter for a puddle of electrons? :rolleyes: ) or that the ESC are simply sucking more power than they should.

    Your hypothesis is an interesting one. Do you have time and inclination to work on one of those suspect ESC and unsolder it enough to slip a piece of laser printer paper between it and the PDB?

    That would provide some pretty good insulation and would prove your hypothesis in a heartbeat. Or you could use thin plastic, a scrap of Mylar, etc. Anything that would act as an insulator. If that changes the current for that motor, your hypothesis is proven, you have your culprit. If it doesn't, then that would point to the ESC themselves working terribly inefficiently.

    Andy.
     
  12. William Johnston

    Joined:
    Oct 2, 2012
    Messages:
    140
    Likes Received:
    5
    Well they do tangentially. If you look at their 4S battery:
    http://www.quadrocopter.com/QuadroPower-6200-Lipo_p_367.html
    they recommend it for:
    XL HexaKopter
    XL OktoKopter
    Droidworx AD-6HL
    Droidworx AD-8HL

    While their 5S battery:
    http://www.quadrocopter.com/QuadroPower-6200-5S-Lipo-_p_541.html
    is recommended for:
    CineStar 6
    CineStar 8
    Droidworx AD-6HL
    Droidworx AD-8HL

    They also have a video showing them doing a comparison of the 4S and the 5S which sort of shows that using a 5S pulls less power on a Cinestar 6.
     
  13. William Johnston

    Joined:
    Oct 2, 2012
    Messages:
    140
    Likes Received:
    5
    Well I'm pretty sure the ESCs work to maintain a speed specified by the flight controller board. I believe that given a certain motor test value in the MK-Tool that the motor will spin at a particular speed regardless of the supplied voltage (within reason, not zero volts of course).

    I built a little device to measure the speed of the rotors. But I didn't play with it much. By then I had satisfied myself the problem was not caused by me. Now I wish I had run some more tests, but seven weeks a ago I was ready to be done with it so I RMAed the thing to Quadrocopter and tried to forget about it. I failed miserably on the forgetting about it part.

    You know over twenty years ago I wrote a brushless DC motor simulator for testing different control laws. It got published in some conference proceedings of something or other. I have been wondering what I did with the through this whole thing. Not that it would run on any computer I have now as it used Open Look for the GUI, but I would like to take a look at the code. It's a shame how things get lost. hmmm... Now I'm just an old man reminiscing.
     
  14. William Johnston

    Joined:
    Oct 2, 2012
    Messages:
    140
    Likes Received:
    5
    Except that my theory require that it behaves like a insulator until certain conditions are met and then it breaks down. And if that's the case then any ohm meter will measure infinite resistance since it won't meet the condition which cause the barrier to break down.
     
  15. William Johnston

    Joined:
    Oct 2, 2012
    Messages:
    140
    Likes Received:
    5
    I'm not saying that there is leakage through the PCB board from the positive side to the negative side. I'm saying there is leakage from the positive copper connector of the ESC to the negative side of the distribution board. The only thing separating those two is the solder mask on the distribution board.
     
  16. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,383
    Likes Received:
    1,164
    Hmm. If you take the props off, I wonder how much current the ESC will draw? Would that come close to replicating the phenomenon, I wonder? Maybe you could try it just one one motor (using the MK Tool motor test) and measure the resistance for that one specific ESC once you can see what the current draw is?

    Edit: "Resistance" in the previous sentence might really need to be "voltage drop" -- let's see what you find....

    If your theory is correct -- and I'm intrigued by the fact that it might be as something, somewhere seems very fishy -- then this should be a repeatable test for the other motors and their correspond ESC. For which read: I think you're on to something, but it's going to be tricky to prove it up.

    Andy.
     
  17. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,383
    Likes Received:
    1,164
    This too has intrigue. If that solder mask is all that's between the two, it might act as the dielectric in a capacitor -- which means it would block DC, but allow AC to pass -- just like the AC that the ESC's use to drive the motors.

    Andy.
     
  18. William Johnston

    Joined:
    Oct 2, 2012
    Messages:
    140
    Likes Received:
    5
    Well, I sent the board back seven weeks ago for repair so I can't do any more tests on it. I am really just hoping that the replacement will work fine and I don't have to fiddle with it anymore.
     
  19. William Johnston

    Joined:
    Oct 2, 2012
    Messages:
    140
    Likes Received:
    5
    Oh say, there's a good thought. I too thought it might act as a capacitor, but I was only thinking that there might be a break down in the dielectric. If an AC signal is showing up on the ESC positive connector then it might pass right through to the power distribution board.

    That would explain why there is no extra current flowing when the motors are off.
     
  20. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,383
    Likes Received:
    1,164
    Oh. That's too bad. I guess we'll never be able to put your theory to the test -- at least not on your board.

    Andy.
     

Share This Page