ROV Circuit Board

Discussion in 'General Electronics Chat' started by jlotz, Dec 28, 2010.

  1. jlotz

    Thread Starter New Member

    Dec 28, 2010
    9
    0
    Application Details: ROV board

    1)controls 2 DC motors for thrusters by L298 IC2,IC3 - motor stall current 2A

    2)controls 2 smaller dc servo motors by L298, IC6 - motor stall current 1A

    3) converts video to diff signal IC21

    4) Cap C10 is 470 uf for taking out large spikes caused by pulsing motors.

    5) J11 and J12 are ground headers I connect to the aluminiun shell of the ROV. Without this the camera picks up too much interference when the thrusters are being pulsed.

    6) J10 is the charge port. When charger is plugged in it flips relay1 over to connect batteries to these pins. When the charger is unplugged the relay1 reconnects the batteries to circuit and disconnects header J10 which will be bridged with water once the ROV is wet. The pins on J10 do not touch the aluminium shell, but are within 3mm of it. It should be noted that there could be a thin film of water between these pins and the shell if it is being charged as soon as the rov was removed from water (which is often the case when I use it).

    7) ROV is turned on via relay 3 by using the video lines from the surface to jumpstart the circuit and run the IC1 program. The IC1 program (Atmega 16) then locks on relay 3 and then switches relay 2 to connect video to surface. This is a little confusing but works very well. It is done this way because there is no on-off switch on the ROV, only the controller, and there are disadvantages of running more control lines down the long tether.

    8) System voltage is 19.2V for motors (battery power), 12VDC for camera and light (IC9), 5VDC for ICs (IC8)



    Problem:
    There seems to be something in this circuit that is taking out my ICs and most commonly the L298s. I have an auto-program set up on the controller that runs the ROV through a typical dive routine and I let it sit in a tank in my basement and drain the batteries. Once the batteries are dead I charge and re-run, just to get some hours on this thing to see where it will fail. At first I ran it about 9 times with no problem. Then after a 10th charge I turned it on and some weird stuff was happening. The traces where I circled on the board layout had burnt. So I totally re-built PCB using all new components and board. Ran it 3 more times with no issues. On the fourth day after the charge was complete I turned on ROV, it was fine, then turned off 10 seconds later. Then within 5 seconds I turned on again and there was a flash/spark on the board about were the IC1 microchip is but the opposite side, and IC6 went down. Everything else still worked. Took it apart and found a hole blown in the front side of IC6 at about the pin 4 location. De-soldered IC6 and hooked to batteries/controller. Fired it up and find that the 5VDC regulator is extremely hot, and the other two L298s are warm when they shouldn’t be, but still kind of working.

    There is no water inside the ROV causing this problem.

    Sorry for the long message, but I know more info is better then not enough. Any comments or ideas are appreciated.

    Thanks,
    Jeff
     
  2. thatoneguy

    AAC Fanatic!

    Feb 19, 2009
    6,357
    718
    1) Do you have a BIG heatsink on the L298 ICs? You are running at the limits.
    2) Are the diodes on the outputs Fast Recovery/Low Breakdown (Schottky)?
    3) Do you have access to an oscilloscope to look at the EMF on the motor lines?
    4) Is this for MATE (or similar?)
    5) Do the drivers work fine, right up until they quit, or do they fail immediately, if they work, how long do they function?
    6) Can you sink heat to water? Not much air circulation in an ROV.
     
  3. jlotz

    Thread Starter New Member

    Dec 28, 2010
    9
    0
    1) Yes I have heat sinks. They are a bar of aluminium 1/8" X 1.2" X 2.5" in size. In air I noted the L298s running at about 120C.

    2)Diodes are Fast Recovery/Low Breakdown, Schottky 3A on the thrusters and 2A on the servos.

    3) Yes I do. C10, 470uf really helps level out the spikes.
    Should I be looking for somthing else?

    4) No just a hardcore and expensive hobby.

    5) L298s appear to work well for 3-10 hours before they quit. I should note that on the second board it sparks and appeared to fail upon startup, before the L298s had a chance to warm up. It is a little unclear when the first board failed, but I notced the problems again on a startup.

    6) Not too easily. The shell is aluminium about 10" diameter, so there is alot of inner ROV air - shell area.


    Thanks,
    JL
     
  4. shortbus

    AAC Fanatic!

    Sep 30, 2009
    4,019
    1,542
    Why not use the outer aluminum shell as the heat sink?
     
  5. jlotz

    Thread Starter New Member

    Dec 28, 2010
    9
    0
    The PCB is moving relative to the shell, so a mechanical connection cannot be made. Do you feel that the L298s are running too hot? I thought they had an internal shutdown for overheating. Do you think that they overheated on the last test but still ran damaged/unnoticed, then after charging I noticed?

    Thanks,
    JL
     
  6. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    You're overheating your L298's.

    With a 2A load, each H-bridge will drop up to 4.9v across itself. 2A x 4.9v = 9.8 Watts per H-bridge. You wrote that the heat sinks are running about 120°C in air, but that's outside of the aluminum enclosure, right?

    Absolute maximum for the junction operating temperature is 130°C.
    Thermal resistance from junction to case is 3°C/Watt
    If you're dissipating 9.8w x 2 = 19.6W internally, then the junction is 19.6w x 3°C = 58.8 + what you measured on the heat sink, so 178.8°C.
     
  7. jlotz

    Thread Starter New Member

    Dec 28, 2010
    9
    0
    You are right I measured that in air with the thrusters in the water only. I will rig up some measuing device and place it inside the ROV. My Fluke meter/themrocouple has a shutdown feature after 5 min so that will not work.

    When you say "9.8w x 2 = 19.6W internally" what do you mean by the X2? I have 3 L298s inside. Two of them are wired in parrell for the thrusters drawing 2A max (L298), and the last drives 2 servo motors, one at 1A max current, and the other at maybe 200ma max.

    An overheating L298 does not seem like the problem, since they all went out on me at the same time on fauilure #2. Could one have blown, and somehow taken out the rest?

    Thanks,
    JL
     
  8. shortbus

    AAC Fanatic!

    Sep 30, 2009
    4,019
    1,542
    Did you get the L298's from a reparable dealer like Mouser or Digikey? Or off Ebay? There are a lot of counterfeit chips out there these days.

    Do you have a power-up routine in the micro? If all the inputs of the L298's turn on at the same time you can get a shoot through condition that may blow the chips.
    The same goes for reversing the thrusters, there has to be a short dead time before reversing, or you could get shoot through.
     
  9. jlotz

    Thread Starter New Member

    Dec 28, 2010
    9
    0
    Interesting comments. Chips came from Digikey.

    Start up routine is as follows.
    Relay is switched providing +20V to the board & L298s. Atmel Atmega16 micro sets thrusters to 50% duty cycle (which is neutral in this case), then flips the enable lines on. Enable lines will remain on until the thrusters exceed current of 2.1A. This only really happens with the propeller gets clogged. The routine for the L298 that drives my two servo motors is essentially the same. The thrusters are ramped and so there is a delay, between forward and reverse, however the one servo can flip back and forth alot using fuzzy logic in order to hold a proper position. There is no ramping on this, but the motor is small, and I am only sending 9V to it.

    Thanks,
    JL
     
  10. shortbus

    AAC Fanatic!

    Sep 30, 2009
    4,019
    1,542
    I know nothing about the micros and the coding, but if both thruster directions are enabled at the same time at start up, somethings got to give, usually the driver :)

    As far as the servos having only 9V, it's not the volts its the amps that matter. Most motor have a much higher start-up amp draw than the free running amp draw. So you might want to check that out too.

    The other thing, like SgtW. said is your heat sink. It seems kind of small, According to all I have read about the L298 in robots, they say to make your heat sinks huge. Also your sense resistors have to match the amp draw of the motor. I don't for sure know if using one sense resistor for both channels is( as in your schematic) kosher.
     
  11. jlotz

    Thread Starter New Member

    Dec 28, 2010
    9
    0
    Sense resistors are 0.1 ohm, 3W. I do not recall failing these, but at the same time I have never checked their temp either.

    As far as I understand the input on the L298 can be any combination and not blow it up, unlike normal H-bridges. However because I send a single PWM signal though a hex inverter for phase anti-lock control I don't think this is happening anyways.

    I will work to collect data on the temperatures, but first need to find a thermocouple small enough to fit inside this thing. My multimeter is too big.

    Thanks,
    JL
     
  12. jlotz

    Thread Starter New Member

    Dec 28, 2010
    9
    0
    OK, I ran a few tests to obtain some temps on the L298.

    If I run the thrusters full out for a whole battery (1 hour) toggling the duty cycle between 99% & 100%, I cannot get the L298 to get hotter then 107C. The temp in the ROV rises from 18C to 30C very slowly, but is fast to flush the heat, likely because it is a large aluminium shell.

    When I increase the surface area of my heat sink by 1.8X it tops out at 99C, so I will likely do that for a better factor of safety, but it does not explain why these things are blowing up on me.

    Any further help is highly appreciated.

    Thanks,
    JL
     
  13. jlotz

    Thread Starter New Member

    Dec 28, 2010
    9
    0
    I should have noted that the thermo couple wire was pinned between the heatsink and rear face of L298 for the test. I think this is the best spot...?
    And the ROV was closed up and submerged in saltwater for the test.

    JL
     
  14. Kermit2

    AAC Fanatic!

    Feb 5, 2010
    3,795
    951
    You're deep into troubleshooting territory, and that is hard to help someone with. Start suspecting simple things that you would never imagine needing to check. Use an eye loop and look into any mating connector holes for signs of good non oxidized metal surfaces.

    I think I see an SMD electrolytic on the trace that burned? Is that right? Any caps you are using for noise and spike suppresion on the motors should be substituted out with known good ones, or just name brand new ones.

    etc. etc.
     
  15. shortbus

    AAC Fanatic!

    Sep 30, 2009
    4,019
    1,542
    The L298 should be mounted directly to the heat sink, with nothing but some heat sink paste between the two. The more the contact the better. Your thermocouple can be mounted to the screw.
     
Loading...