Electronics > Projects, Designs, and Technical Stuff
Using IR2108 for three phase inverter
<< < (2/2)
jbb:

--- Quote from: hussamaldean on November 03, 2018, 07:21:00 pm ---aha
but the IC requires both inputs to be same phase otherwise it won't work since I tried with not logic gate and function generator and the results didn't seem ok at all

--- End quote ---

So... I checked the datasheet.  The Low side INput (LIN) is inverting.  (This is so that you can tie HIN and LIN together and drive them with a single input signal.  You'd then rely on the chip to produce dead time for you.  However, you then wouldn't have an easy way to turn off both MOSFETs.)  Can you set a bit in your micro to invert the PWM outputs?

I would suggest a change of chip for 3 reasons (see feature comparison on the front page of the IR2108 datasheet):

* The chips with a single COM input can be susceptible to noise between the microcontroller.  The chips with a split VSS/COM can help with this a lot
* The HIN and inverted LIN of the IR2108 (or IR21084) seem like a pain to deal with.  I suggest moving to HIL / (non inverting) LIN or IN / SD inputs
* Adjustable dead time could be nice.Therefore I suggest the IR21094 might be better for you.  Or the IR21064 if you need < 500ns dead time and you're confident of your microcontroller outputs never generating a shoot-through (remember to think about microcontroller default pin states during programming / boot loader / boot up).

FYI:

* Running MOSFETs in a totem pole / half bridge with Vdc > 200V or so may blow up due to reverse recovery losses in the MOSFET body diodes (Qrr).  The MOSFET body diode is (as the name suggests) built-in, and usually really crappy because there's a competition between MOSFET performance and body diode performance.   IGBTs with anti-parallel diodes are likely better here because the diodes are built separately and have much better performance.
* I think LIN is wrong in U1 on your circuit diagram.
* 1k gate resistors are likely way too big (I guess they're placeholders)
* You might want to include some fast turn-off diodes and resistors for each gate to allow harder turn off.  This helps with switching losses.
hussamaldean:

--- Quote from: jbb on November 03, 2018, 09:08:12 pm ---
--- Quote from: hussamaldean on November 03, 2018, 07:21:00 pm ---aha
but the IC requires both inputs to be same phase otherwise it won't work since I tried with not logic gate and function generator and the results didn't seem ok at all

--- End quote ---

So... I checked the datasheet.  The Low side INput (LIN) is inverting.  (This is so that you can tie HIN and LIN together and drive them with a single input signal.  You'd then rely on the chip to produce dead time for you.  However, you then wouldn't have an easy way to turn off both MOSFETs.)  Can you set a bit in your micro to invert the PWM outputs?

I would suggest a change of chip for 3 reasons (see feature comparison on the front page of the IR2108 datasheet):

* The chips with a single COM input can be susceptible to noise between the microcontroller.  The chips with a split VSS/COM can help with this a lot
* The HIN and inverted LIN of the IR2108 (or IR21084) seem like a pain to deal with.  I suggest moving to HIL / (non inverting) LIN or IN / SD inputs
* Adjustable dead time could be nice.Therefore I suggest the IR21094 might be better for you.  Or the IR21064 if you need < 500ns dead time and you're confident of your microcontroller outputs never generating a shoot-through (remember to think about microcontroller default pin states during programming / boot loader / boot up).

FYI:

* Running MOSFETs in a totem pole / half bridge with Vdc > 200V or so may blow up due to reverse recovery losses in the MOSFET body diodes (Qrr).  The MOSFET body diode is (as the name suggests) built-in, and usually really crappy because there's a competition between MOSFET performance and body diode performance.   IGBTs with anti-parallel diodes are likely better here because the diodes are built separately and have much better performance.
* I think LIN is wrong in U1 on your circuit diagram.
* 1k gate resistors are likely way too big (I guess they're placeholders)
* You might want to include some fast turn-off diodes and resistors for each gate to allow harder turn off.  This helps with switching losses.
--- End quote ---
for now, I am stuck with IR2108 since the any other one is not available here in Iraq
langwadt:

--- Quote from: hussamaldean on November 03, 2018, 07:21:00 pm ---
--- Quote from: langwadt on November 03, 2018, 07:17:52 pm ---
--- Quote from: hussamaldean on November 03, 2018, 06:36:37 pm ---
--- Quote from: langwadt on November 03, 2018, 06:30:35 pm ---yes you can use them, the complementary outputs on timer1 is for drivers like that (haven't checked if you got the right pinout)

--- End quote ---
could you please elaborate ?

--- End quote ---

the advanced control timer1 has complementary outputs on ch1,ch2, and ch3 (ch1n,ch2n and ch3n) so when setup correctly chXn is the opposite of chX and there is a programmable dead time every time they change

--- End quote ---
aha
but the IC requires both inputs to be same phase otherwise it won't work since I tried with not logic gate and function generator and the results didn't seem ok at all

--- End quote ---

then just tie HIN and /LIN  together, the chip has build in dead-time

hussamaldean:

--- Quote from: langwadt on November 03, 2018, 09:56:45 pm ---
--- Quote from: hussamaldean on November 03, 2018, 07:21:00 pm ---
--- Quote from: langwadt on November 03, 2018, 07:17:52 pm ---
--- Quote from: hussamaldean on November 03, 2018, 06:36:37 pm ---
--- Quote from: langwadt on November 03, 2018, 06:30:35 pm ---yes you can use them, the complementary outputs on timer1 is for drivers like that (haven't checked if you got the right pinout)

--- End quote ---
could you please elaborate ?

--- End quote ---

the advanced control timer1 has complementary outputs on ch1,ch2, and ch3 (ch1n,ch2n and ch3n) so when setup correctly chXn is the opposite of chX and there is a programmable dead time every time they change

--- End quote ---
aha
but the IC requires both inputs to be same phase otherwise it won't work since I tried with not logic gate and function generator and the results didn't seem ok at all

--- End quote ---

then just tie HIN and /LIN  together, the chip has build in dead-time

--- End quote ---
yup
it does have dead-time
hussamaldean:
I can confirm that the LIN is the shurdown pin, when the pin is pulled HIGH, the output is turn on
weird though
Navigation
Message Index
Previous page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod