IR2113, cant make HIGH SIDE work

[nadre25]

Hi guys, we are trying to use ir2113 to drive the IGBT for our project. We plan to make an H-bridge using the IGBTs but we cant drive the IGBT using our IC. When we tried placing our PWM as input in the HIN pin of the IC, the HO pin only outputs a straight DC voltage.

When we tried on the LOW side, it worked fine.

Our PWM is 0 to 5V
VCC = 15V
we followed the schematic in the datasheet of ir2113

can anyone suggest a specific configuration that we need to do to activate the HIGH SIDE.

some of our questions.

1. We are expecting to see a squarewave output in the HO pin for a PWM input in the HIN pin, is that correct?

2. If the IR2113 is a driver, how exactly is it able to drive the IGBT, is it because it increases the current? /voltage?

Thanks!

 

[alexan_e]

In order to activate the high side, the bootstrap capacitor has to be charged from a path to the ground, how exactly did you make the test?
Can you post a schematic?

Alex

 

[nadre25]

This image is the schematic what we used



we used 0.1uF and 10uF capacitors

VCC and VDD = 15V

COM,VSS,SD = Grounded

HIN = 0 to 5V input

LIN = 0 to 5V input

we are just checking the output at HO pin and what we get is a DC voltage, which is about near VCC.

 

[alexan_e]

Vs must have a path to ground so that bootstraping can work properly, do you leave the output floating?

Alex

 

[Orson Cart]

OK, to make it work needs the presence of a low side fet (Vs must be connected to the drain of the low side mosfet/igbt), when this fet is turned on it allows current to flow into the bootstrap cap which then allows the high side output to work, so no low side device = no high side operation. The only other way to get the high sdie going is to put a fully isolated supply (e.g. a 12V battery) across the high side bootstrap cap - this will power the high side output - hope this is all clear - regards, Orson Cart.

 

[alexan_e]

OK, to make it work needs the presence of a low side fet (Vs must be connected to the drain of the low side mosfet/igbt)

The datasheet schematic shows that Vs should be connected to the source of the high side mosfet so using just a high side mosfet with a resistor load would be enough to make it work.

Alex

 

[nadre25]

When you say "leave the output floating" do you mean that the output is not placed in the gate of the IGBT / any device? if so then YES, we left the output floating and tried to see the output w/ an oscilloscope.

---------- Post added at 14:39 ---------- Previous post was at 14:33 ----------

OK, to make it work needs the presence of a low side fet (Vs must be connected to the drain of the low side mosfet/igbt), when this fet is turned on it allows current to flow into the bootstrap cap which then allows the high side output to work, so no low side device = no high side operation. The only other way to get the high sdie going is to put a fully isolated supply (e.g. a 12V battery) across the high side bootstrap cap - this will power the high side output - hope this is all clear - regards, Orson Cart.

Ok, thanks Orson Cart. I will try this, i might be able to update you next week since the equipments are in our lab. So basically, i have to use 2 IGBTS one for low side and the other for the high side and connect the VS to the drain of the low side IGBT? I'll connect them like a half h-bridge ckt?

---------- Post added at 14:40 ---------- Previous post was at 14:39 ----------

The datasheet schematic shows that Vs should be connected to the source of the high side mosfet so using just a high side mosfet with a resistor load would be enough to make it work.

Alex

i'll also try what you said alexan_e. Thanks! i might be able to make an update in 2 days.

 

[alexan_e]

Yes by floating I meant without any connection and no it will not work this way, you have to connect at least a mosfet/igbt in the high side with a resistor going to the ground as load.
The bootstrap capacitor would probably be able to charge through that resistor, if not then you need the low side device too.

Alex

 

[Orson Cart]

Unfortunately the below is not completely correct

The datasheet schematic shows that Vs should be connected to the source of the high side mosfet so using just a high side mosfet with a resistor load would be enough to make it work.

Alex

- without a low side mosfet/igbt - the bootstrap cap may not charge up enough - hence the high side output can not operate, Regards, Orson Cart.

 

[alexan_e]

I haven't tried it but have you?
Why wouldn't the bootstrap capacitor charge from a resistor (lets say 100 ohm) as load connected to the gnd?

Alex

 

[alexan_e]

I can only simulate a similar bootstrap circuit with discrete components and it seems to be fine (50KHz input, duty 50%)



Alex

 

[Orson Cart]

Good question - yes we have seen this effect (albeit long ago), typically the duty ratio will be 50% max, hence charging via an RC will give 50% of the available volts (say7.5 out of 15, less the diode drop, say 0.7 volt) giving only 7V or less on the bootstrap cap - which may not be enough for the UVLO on the high side electronics, using a low side fet/igbt guarantees that the bootstrp ckt works completely, giving 14.3 volts on the high side supply, and allowing a user to see the drive on the high side output.

There is no load on the bootstrap cap in your simulation - which affects your result.

Hope this all makes sense, Regards, Orson Cart.

 

[alexan_e]

There is no load on the bootstrap cap in your simulation - which affects your result.

What kind of load should I put?
The bootstrap capacitor provides the power to the level translator mosfet and the totem pole driver, what is the load that is missing?

Alex

 

[FvM]

typically the duty ratio will be 50% max, hence charging via an RC will give 50% of the available volts

The presented circuit actually does not average the supply voltage according to the duty cycle. It charges the bootstrap capacitor to full supply voltage - one diode foward voltage. In addition, the RC time constant must be sufficient low.

It should be also noted, that the problem of the original post was an inappropriate test setup for the push-pull circuit - operating the high-side switch without previously activating the low-side.

If the discussion is about high-side switch only, e.g. for a buck converter, we should define the load voltage and current waveform to decide about the feasibility of a bootstrap driver.

 

[Orson Cart]

What kind of load should I put?
The bootstrap capacitor provides the power to the level translator mosfet and the totem pole driver, what is the load that is missing?
Alex

The load on the high side bootstrap cap will depend on the operating freq and size of the device you are driving (and if you have a G-S or G-E resistor present) - a best guess (worst case) is a 1k-ohm load.

Regards, Orson Cart.

 

[alexan_e]

You mean a physical resistor connected from the gate of the mosfet to the source?
I don't see a resistor like that in the application data of the device, what is the purpose of using that resistor?

The device I'm driving for testing purpose is 100 ohm resistor, the same load will be connected to IRF2113 for test of the high side operation.

Alex

 

[Orson Cart]

Yes a physical resistor from G-S, for the IR2113 you may not need it, but it helps keep the mosfet off at power up, the 100 ohm resistor will dissipate approx 5W for your 22V supply if the duty cycle reaches 90% say.

(The 2113 loads the bootstrap cap by using power from it to drive the mosfet)

If you have a 1k from G-S and a 100 ohm load, then you will get 10/11ths of the 22 volt available x 1/(duty ratio), so at 50% you will get 10V on the bootstrap cap and at 90% out you will get 2V. The 2113 is really designed to operate with an half bridge of mosfets where the bootstrap cap is fully charged every cycle very quickly when the lower device is turned on.

Also for a 100n cap and a 1k load operating at 50%, 20kHz say, the volts on the cap will droop 5V (very approximately) from a 20V starting point every cycle, and this will be worse at 90% ON time, so a bigger cap is in order (470nF say).

Regards, Orson Cart.

 

[alexan_e]

My point was that the IR2113 should be able to work just with the high side mosfet and a 100 ohm resistor to test the operation of the chip.
There is no need to add the GS resistor or change the pulse ratio to 90% to do that test.

Alex

 

[Orson Cart]

Agreed, but as you didn't state how you were going to test it, it was best to explain that at 90% duty cycle things will not be as good as at 50% or lower, and it may be that your final requirement may have required hi duty cycle...
Regards, Orson Cart.

 

[nadre25]

Hi Guys. i tried to connect the VS to the source of my high side IGBT. This is my schematic. Sorry if the drawing is not that good, i dont have an IR2113 on multisim. For clarifications, just ask me.





The HO was 0 to 30V our VCC was 15V, we also had a 20V supply on the drain of the IGBT. The waveform of HO was almost a squarewave but there is a distortion... what could possible be done for that?