Help needed with IR2110 driver & full bridge buck-boost.

Discussion in 'The Projects Forum' started by WayTooLate, Apr 12, 2010.

  1. WayTooLate

    Thread Starter New Member

    Mar 15, 2010
    4
    0
    This is my first time posting here. This problem has been killing me for a week now. I have looked at several posts relating to the IR2110 driver in a full-bridge setup but still can't figure out the problem. Any help is greatly appreciated.

    I'm building a small-scale version of a full bridge, buck-boost that will be later implemented for high voltages. I am using two IR2110 to drive the gates of the MOSFETs. Before implementing the full bridge, I tested each driver in a half bridge configuration successfully (with correct Vgs signals). The problem comes when I link the bridges together with the inductor. The converter is load dependent, but in the full bridge setup no current is being drawn from the power supply. I know these drivers are usually used in a H-bridge setup where Vsupply goes to the drain of both high side MOSFETs and this might have something to do with it. I'm thinking I need to add additional capacitors or resistors to the 'left side' of the bridge.

    Specs: fs = 100kHz, DC = 60% (for top-left/bottom-right MOSFET), Coss (MOSFET) = 1.28nF
    Attached:

    1. Simplified design (H.V. listed)

    2. Half-bridge implementation with the IR2110 (when the other side is linked the inductor is connected between both Vs pins and b/w the sources and drains of the MOSFETs on both sides)*

    *Mistake on the circuit diagram: Vcc = 13.5V; Vcc should be connected to anode of Schottky diode
     
    Last edited: Apr 12, 2010
  2. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Are you applying the same signal to input P07 as you are to P08? If so, you will wind up with "shoot-through" problems.

    You are assuming that the 74HC04 will invert the signal and that will be enough. However, the 74HC04 adds a propagation delay to the signal. Look at the datasheet for the particular manufacturer's IC that you are using to determine the propagation delay.

    The propagation delay, basically, is how long it takes for a change of signal on the input, to be reflected by a change in the output. The propagation delay might be anywhere from 4nS to 30nS, but check the datasheet to find out the actual range.

    MOSFETs do not turn on and off instantly. You need to look at a datasheet to see what the turn-on and turn-off times are. You will need to provide "dead time" where both the upper and lower MOSFETs are turned OFF before turning one back on.

    You are using a 330 Ohm pull-up resistors on the output of LM339's. LM339's have pretty wimpy open-collector outputs; 8mA is about the most that you can sink reliably. 13v/330=39.4mA, or about 4.9 times the maximum. You would be better off to use a max of 5 to 6mA, or around 2.2k.

    What is the lower side of C2 connected to? Doesn't look like it's connected to anything. It probably should be connected to Vs. You should probably have a 0.1uF metal poly or ceramic cap in parallel with a larger cap.

    You have 13v connected to VB. That won't work. Move D2 up and connect the cathode of it to the junction of C2/VB, and the anode to +13v. The Vcc pin needs to be connected to +13v

    I'd use a 0.1uF metal poly or ceramic cap in parallel with the 1uF C3.

    IRF510s are virtually antiques, and you're using them with a 20v supply. There are lots of more modern MOSFETs that would be a better match. Look for a Vdss rating of about 1.5x your source voltage, with a low total gate charge and low Rds(on).
     
  3. WayTooLate

    Thread Starter New Member

    Mar 15, 2010
    4
    0
    Thanks for the great info. The pull-up resistor was the main problem; I completely overlooked the max current out of the LM339. After changing the resistors, the converter works in full-bridge with a usual error of only 0.3% in the duty cycle/output voltage.

    I had considered the propagation delay and shoot-thru but haven't implemented a solution yet. The prop. delay of the inverter is only 8ns and with my switching frequency of 100kHz (10us period) I didn't think it would be a huge deal starting off.

    Using IR's An978a app notes, the 'worst case mismatch' of the IR2110 is 10ns' with the turn-on prop. delay 25ns longer than the turn-off. "This, by itself, should insure that no conduction overlap of the power devices would occur, even if the on and off input command coincide. As an added safety margin a resistor diode network can be added to the gate...the purpose of this network is to further delay the turn-on, w/o affecting the turn-off, thereby inserting some additional dead-time."

    I have the resistor on the gate already and plan to add a 1N4148 (anode connected to HO/resistor) as suggested. The app. note makes it sound like this is all the deadtime I should need; should I be worried about this?

    C2 is connected between VB and VS. And Vcc is connected to 13.5V and the anode of the diode (only a diagram mistake) along with C3.
    And I will try adding a cermaic cap (100nF?) in parallel with C3.

    Also I will switching to the IXFK80N50P MOSFET which has a Rdson = 65mOhms and Coss=1.28nF

    Thanks
     
  4. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    It helps if you update your schematic rather than try to describe your changes in words; it's much easier to spot mistakes that way.

    If you are following the app note's recommendations, you should be OK.

    100nF=0.1uF = 1,000,000pF

    I don't know what you are talking about with Coss=1.28nF.
    Look at Total Gate Charge; it's given in nC's (nanoCoulombs)
     
  5. WayTooLate

    Thread Starter New Member

    Mar 15, 2010
    4
    0
    Yeah I don't know why I gave Coss. I used Qg(on) = 197nC for the bootstrap calculation.

    I attached the full-bridge implementation with the described updates. Hopefully the shift from Vin = 20-40V to 150-200V will not cause many problems. I also am going to try to add current programmed control to the converter (already have Hall-effect current sensor). I have a power electronics book that has all the transfer equations I need for a 2 Mosfet, inverting buck-boost but not a full bridge, non-inverting one. So this may take some time unless I find an appropriate example.

    *I forgot to change one of the 13V to 13.5V on the schematic.
     
  6. djaitly

    New Member

    Aug 9, 2012
    1
    0
    Hello, I have a question, is this circuit will work with 24Volts/20A motor for speed and direaction control by changing pwm timing? Any suggestions are highly appreciated.
    Thanks
    Djaitly
     
Loading...