Replacing a DPDT relay with solid state devices

Discussion in 'General Electronics Chat' started by MrSoftware, Jan 21, 2016.

  1. MrSoftware

    Thread Starter Active Member

    Oct 29, 2013
    504
    124
    I'm looking for tips on the best way to replace the DPDT relay in the system below with solid state devices (MOSFETs?) for the purposes of greater reliability and lower cost. Right now the DPDT relay runs me about $10/ea.

    It's a pair of H-bridges to control a DC motor that runs a mechanical lift. One is run by an automatic system, the other is a manual override (an actual physical 20A DPDT rocker switch). When the manual H-bridge is operated, it trips the relay allowing the manual H-bridge to override the automatic one. Typical duty cycle is about 6A @14v for about 1 minute, followed by about 20A @14v for about 3 seconds, followed by several minutes of rest.

    Any tips on how to replace that DPDT relay with solid state devices, that would possibly not exceed $10/relay in quantities of less than 100, would be greatly appreciated!


    h-bridges.jpg
     
  2. mcgyvr

    AAC Fanatic!

    Oct 15, 2009
    4,771
    971
    Whats the input signal/trigger to the automatic H bridge?
    Can you do something to eliminate that (manual switch just supplies the signal to the automatic h bridge when needed)?
    Thus removing the need for a relay/mosfet completely?

    I see no point in having 2 motor controllers (h-bridges) for a single motor..
     
  3. Stuntman

    Active Member

    Mar 28, 2011
    181
    47
    Perhaps they have intentionally installed redundant drivers so if either the manual or automatic bridge smokes, the other will work fine, and, in the event the automatic system hangs, the manual system can override via the relay.

    As mcgyvr hinted, if you can intercept the incoming signals to the auto driver you can either use a clever arrangement of resistors and diodes to allow the switch to "override" the automatic signals. Another option may be a DPDT signal relay (cheaper than your 20A + jobber) in a similar manner shown above to transfer between the manual input signal and auto input signal.
     
  4. MaxHeadRoom

    Expert

    Jul 18, 2013
    10,565
    2,379
    Incidentally what is the cause of 20A for 3 secs? Is this a turn off point with the motor stalled?
    If the relay were to switch off at this point arcing may shorten the life without a means of arc blow out, maybe this high hi current can be overcome?
    Are both drivers fed from the same supply?
    Max.
     
  5. MrSoftware

    Thread Starter Active Member

    Oct 29, 2013
    504
    124
    The automatic part is a retro-fit add-on to an existing manual system. It's actually a wireless remote control add-on, so the user can operate the lift from some distance away. I'm calling it automatic because the H-bridge is operated by logic driven by the remote control, where the pre-existing system is a high current manual rocker switch.

    There are multiple different lift designs, produced by different manufacturers, so the designs vary. BUT the one thing they all have in common is they all end with the manual rocker switch used as an H-bridge. So with this general design we have a retrofit that works for every model .

    This is a great idea, thanks! When the manual switch is operated, I can kill power to the automatic logic controls. A relay would work, but I could also do it with small MOSFETs. I'll just put a p-channel MOSFET in-line with the V+ power for the automatic system (only draws a few mA). When the manual switch operates I'll pull up the gate to turn off the MOSFET essentially turning off the control logic. hmm...how high do I have to pull the gate to turn off a p-channel MOSFET? Will a diode in-line with the gate cause a problem? This H-bridge is made of N-channel MOSFETs so I'll have to keep the V- connected so their gates can be pulled down and they turn off when the control voltage is removed. I'll post a schematic when I get time to draw it. Cool idea, this might work, depending on how fast the H-bridge turns off, so the 2 H-bridges don't short each other out.

    Yes the current spike is when the lift hits its limit and the motor almost stalls (there is a torque limiting clutch that prevents a full stall). I'm hoping to eliminate the relay entirely and replace it with solid state devices, hopefully eliminating this issue. Both drivers are fed from the same supply.
     
    Last edited: Jan 21, 2016
  6. Stuntman

    Active Member

    Mar 28, 2011
    181
    47
    Without seeing the parent schematic I cannot suggest if your PFET idea will work. One forseeable problem may be that even though you are cutting out V+ to the logic, the signals you are sending from the manual switch may backfeed into the automatic logic.

    Depending on the inputs to the bridge, there is still he possibility of using current limiting and signal diodes to allow simple override of the manual switch avoiding a relay and FETS altogether.
     
  7. MaxHeadRoom

    Expert

    Jul 18, 2013
    10,565
    2,379
    As with many posts here it is difficult to give any in depth informed resolution when all the facts and details of the whole system is mainly unknown.
    Without this, any answer to the original question may not fit or be the best solution.
    Max.
     
  8. MrSoftware

    Thread Starter Active Member

    Oct 29, 2013
    504
    124
    Can you point me to an example circuit or description? This sounds interesting and I would like to understand what you mean.


    I will put together a full schematic and post it. I'm just learning Eagle so it's going to take me a bit (why I hand drew the circuit in my first post), but this would be a great excuse to get some seat time with Eagle.
     
  9. Stuntman

    Active Member

    Mar 28, 2011
    181
    47

    Assuming 5V logic and that the AI can handle sourcing/sinking >5mA.

    upload_2016-1-22_11-52-29.png
    Note that the switch (I did not have a momentary DPDT schematic) will now override any logic signals from the automatic mode system. You will note, this requires knowing what kind of signal the HBridge is accepting, and the capabilities of the logic driver.

    ETA: Missing the 5 in "5mA."
     
    Last edited: Jan 25, 2016
  10. MrSoftware

    Thread Starter Active Member

    Oct 29, 2013
    504
    124
    @Stuntman - Thank you, I understand exactly what you mean now. I'm not sure if we can apply that in this case since we have to keep both H-bridges, but I will definitely keep this in mind in the event we get to change the design to have only 1 H-bridge. Thanks!
     
Loading...