LM5101 mosFET drivers and thermoelectric coolers

Discussion in 'The Projects Forum' started by leonhart88, Sep 13, 2012.

  1. leonhart88

    Thread Starter Senior Member

    Feb 23, 2007
    118
    1
    Hi All,

    I'm using two LM5101 mosfet drivers to implement an h bridge for thermoelectric coolers (TECs). Each LM5101 makes a half h bridge. I left my schematic at work, but I have attached a rough sketch of the circuit (will post the real schematic tomorrow). I've left out some resistors, diodes, caps, pins, etc.

    I was testing the circuit by applying the voltages as shown on the sketch to both LM5101s. Typically, I don't think you would want both high sides on at the same time, and they won't be when I implement the proper PWM. However, these voltages were actually what my microcontroller was outputting on reset.

    With these voltages, I would expect both HS lines to have 24V on them, but I am not seeing this. In fact, I am seeing around 10-11 volts on those lines, and they are not the same voltage (they are around 1V off). The voltages I see at the gates of the FETs are also around the same voltage and are slightly off as well. I'm suspecting that the FETs aren't turned on completely, which is why the HS lines do not have 24V. Similarly, because one of the gate voltages on one of the FETs is higher, one of the HS lines has a higher voltage as well.

    I have a 10 milli ohm current sense resistor on the top HS line and I noticed the voltage drop was 0.2V. I don't have any load attached, except for an LED and a 4.75K resistor so I'm not quite sure what's going on with the current draw as I was not expecting any.

    Am I correct in expecting an even 24V on both HS lines? Or am I totally misunderstanding something...


    Thanks!
     
  2. leonhart88

    Thread Starter Senior Member

    Feb 23, 2007
    118
    1
    I should also point out, that the circuit I'm using has already been used by my colleague and it's the exact same circuit (unless I drew something wrong in the schematic capture) with the same parts, so I know that it should work with the parts used. I'm just a little stumped and trying to make sure I understand how it works.

    Will post the schematic tomorrow for more detailed info, but was hoping that someone might already have some insight/advice for me.

    Thanks!
     
  3. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    The bootstrap capacitor precludes the device from working at DC. The cap basically functions as a battery, which will discharge if you don't keep "pumping it up". I don't see any mention of the value of the bootstrap cap in the datasheet. Maybe I missed it.

    You need to apply pwm to the HI input in order to keep the cap charged.
    And, of course, you have to control the timing of the inputs to avoid shoot-through in the two MOSFETs.
     
  4. leonhart88

    Thread Starter Senior Member

    Feb 23, 2007
    118
    1
    Hi Ron,

    Thanks. The thought that I might need PWM crossed my mind earlier but I'm not sure I understand what the bootstrap capacitor is for or how it works in an h-bridge configuration. I think the cap value might have been 10uF. I will research into the topic further to try and understand better.

    Just curious, but do most mosfet drivers have this bootstrap? I also notice the bootstrap is only for the highside FETs, so I assume it has something to do with the gate and source voltage difference.

    I guess lastly, if an application would like to have 100% duty cycle, then what other methods are there?

    Thanks!
     
  5. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    In order to use them as switches, most N-channel MOSFETs require that the gate voltage be 10V (logic-level devices require less) above the source voltage. When the source is grounded and the load is in the drain circuit, all that is required is for the gate to swing between 0V and ≈+10V.
    A high-side NMOS switch is a different animal. I started to write an explanation of how bootstrap drivers work, but didn't feel I could write it so you could understand it. I recommend you Google "high side bootstrap driver" and read some of the explanations.
     
  6. leonhart88

    Thread Starter Senior Member

    Feb 23, 2007
    118
    1
    Hi Ron,

    Thanks for your help. I've been researching the topic and I have a better understanding now. I'll just post what I've discovered in case this topic might help someone else in the future.

    As you said, the n-channel MOSFET gate needs to be around 10V higher than the source. I'm expecting the source to be 24V, so a charge pump circuit is necessary in order to drive this gate and turn the FET on.

    I also discovered that there are other methods of designing h bridges so a charge pump is not needed (eg. using PMOS). However, n-channel mosfets are much more efficient, but the high-side fets need to have a gate voltage greater than the source. As you mentioned, it's easy to drive the low side fets since the source is referenced to ground.

    Again, as you mentioned, since I wasn't using PWM at the time, and just supplying a DC (or 100% duty) signal, the gate voltages are not enough to turn the FET on completely which is why I am getting those strange voltages.

    I'm going to read up a bit on charge pumps to further my knowledge, then I will try getting the circuit to work again. Thanks again for your help!
     
  7. leonhart88

    Thread Starter Senior Member

    Feb 23, 2007
    118
    1
    I think I understand the bootstrap cap now. When the high side FET is off, the cap will charge to 12V. When we turn the FET on, the voltage at HB will be the source + the bootstrap capacitor voltage (because they are now in series?). This large voltage at HB is then used in the internal charge pump circuit to drive the MOSFET gate!

    Am I missing anything?
     
  8. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    I don't understand what you are saying. You may be right.:)
    The bootstrap (BS) cap becomes the power supply for the high side driver. It has about 11V across it. The low side of the driver is tied to the source (HS) of the high side MOSFET. As the driver output (HO) goes high, HS goes high. This reverse-biases the bootstrap diode, maintaining 11V across the BS cap. The entire high side driver (HB and HO) goes positive as HS goes positive, so, when HS reaches ≈24V, the gate will go 11V higher than that. When you bring HI low, HO initially goes back down to the HS voltage, turning the transistor off.
     
  9. leonhart88

    Thread Starter Senior Member

    Feb 23, 2007
    118
    1
    Thanks Ron. It makes sense now. However, I encountered another head scratching situation.

    I'm playing around with different HI/LI settings. With HI and LI both low, none of the mosfets should be on. However, I am still measuring around 12 and 13V on both HO lines. Is this expected?
     
  10. leonhart88

    Thread Starter Senior Member

    Feb 23, 2007
    118
    1
    I didn't realize I hadn't posted my schematic. Here it is.

    I can't seem to get any expected results...setting all the inputs to 0 does not result in a logic HIGH after the NAND gate. I just get 0 volts on both the HI/LI lines and ~12/13V on the HS lines as previously mentioned.

    I did try using PWM but was not getting 24V on the highside and the LEDs are flashing very brightly and making audible sounds. I may have set up the PWM incorrectly and maybe shorted something.

    I think I understand at least how the high-side MOSFETs switch now, but for some reason I still cannot get this thing working in the least bit. Ideally, I'd just like to confirm that I can set the lines to 0V and 24V and that the circuit is working before I actually put any load onto it.
     
  11. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    657
    Where?

    There are no NAND gates in the device.
    LEDs are audible? Do you have current limiting resistors in series with them?
     
  12. leonhart88

    Thread Starter Senior Member

    Feb 23, 2007
    118
    1
    Oops, I didn't attach the schematic in my last post. I don't have the full schematic with me at home (I should really bring it home), but here is a picture of part of the schematic, you can ignore the notes I've written on it.

    I have LEDs and a current limiting resistor further down in order to show the polarity (eg. LED1 will be on if operating in +/-, LED2 will be on if operating in -/+).

    Even if ignoring the NAND gate, with all inputs at 0, I don't expect to see any voltage on the HS lines because the capacitors will be essentially open circuit at DC steady state. But for some reason I'm still seeing a voltage on the HS lines.

    The LEDs may be audible and very bright because there is a shoot-through situation...but I have yet to verify this yet...
     
  13. leonhart88

    Thread Starter Senior Member

    Feb 23, 2007
    118
    1
    Here is the actual full schematic.
     
Loading...