Delay matching in IR2xxx low and high side gate drivers

[Eva]

I have been analyzing the delay matching between channels in the following IR gate drivers. Test conditions are 22ºC, no load except 70pf (or so) from oscilloscope probes, Vcc=12V with 220nF decoupling applied to all three sections of the IC and both inputs driven in parallel from a fast square wave generator.

- IR2112: I have 10 pieces of these, all from the same batch. The high side output turns on 5ns later than the low side, and it turns off 30ns later. Those delays are consistent among all the devices tested.

- IR2010: I have 9 pieces of these, all from the same batch. Both outputs turn on almost at the same time, but turn off of the high side is delayed 30ns with respect to the low side. The delay is consistent among all the devices tested. I have also noticed that these delays drift a lot with temperature changes.

- IR2113: I have 9 pieces of these, 4 from one batch with white letters and another 5 from other batch with yellowish letters. The low side seems to be delayed by 5ns with respect to the high side, both at turn on and at turn off. The delay is very consistent among all the devices from the two batches.

[Eva]

Edit:

IR2113 seem to suffer very little drift with temperature, the delay was almost the same when the chip case was at 60ºC (due to self heating, measured with IR thermometer).

[Hartono]

Hi Eva,

very nice testing, thanks for sharing

Hartono

[BWRX]

Hi Eva.

What switching frequency were your tests performed at?

[luka]

Hi

I think she said that at 4MHz if not she will correct me

[Eva]

Originally posted by BWRX
Hi Eva.

What switching frequency were your tests performed at?

I tried 50Khz, 1Mhz, 2Mhz and 4Mhz. The delay figures mentioned were obtained at 50Khz and verified at 1Mhz and 2Mhz but not at 4Mhz, where strange things happen...

[Hartono]

Hi Eva,

"where strange things happen..."

What happened ?


Hartono

[Eva]

At 4Mhz with IR2010 the high side channel is further delayed, particularly as temperature increases. Both turn on and turn off may be delayed by up to 100ns at 60ºC case temperature with respect to the other channel. This does not happen with IR2113.

[sawreyrw]

The data sheet on the IR2010 shows graphs of Tj as a function of frequency and load. Depending on the load Tj, starts to rise above a frequency of 100 KHz. At 4 Mhz I suspect Tj will be high even with light loading.