TIDA-010059: LMG3411R050 fault output error

Hi Higa,

Thank you for reaching out to us on testing this reference design and asking the question.

In fact, the design do have integrated both high-side and low-side 'Fault' signals of all three phases. Please refer to Figure 7 in page 12 of this document: Three-phase, 1.25-kW, 200-VAC small form factor GaN inverter reference design (Rev. A) (ti.com).

Please also note this in page 11 of the design guide:  In-phase current sense reference design for 230-VAC motor drives using hall-effe (ti.com) that:

'The only difference is that the TIDA-00915 reference design uses the 150-mΩ GaN power module (LMG3411R150) that has a 6-A RMS continuous current rating, while the TIDA-010059 reference design uses the 50-mΩ GaN power module (LMG3411R050) that has a 12-A RMS continuous current rating.' So we can also refer to the test waveforms in this document, too.

For the problem of motor runs slow in open loop control, it may possibly be related to the parameters in the control. Could you check it up with aother MOSFET/IGBT drive?

As for the waveform oscillation, the probe setting may impacts much. Try to make the GND connector as short as possible.

Best regards,

Jerome Shan

Jerome-san

Thank you for answering.

However your answer does not answer my question.

My question is why only the low side GaN outputs a fault signal.

I understand that the fault signal is integrated.

The FAULT signal in the waveform shown is directly measured at pin 32 of Q6 in the schematic.

The same was true for the other low-side GaNs, Q4 and Q2.

Could you please give us your opinion on the fact that the FAULT signal is being output rather than the measurement environment problem as you say?

Regards,

Higa

Hi Higa-san,

I am sorry I did not catch your points earlier. 

So you have measured fault outputs on all three low-side switches, but have not captured one on the high-side? 

LMG3411 can report a fault on OCP, OTP and UVLO. Do you know the cause of the fault for the low-side now?

Is it possible to make a test on the high-side for this fault purposely?

Best regards,

Jerome

Hi Jerome-san

you said:

So you have measured fault outputs on all three low-side switches, but have not captured one on the high-side? 

-> Yes.

you said:

 Do you know the cause of the fault for the low-side now?

-> I measured the current, but no abnormal current was flowing.
(The current probe I'm using can only measure up to 20MHz, so it may simply not be possible to measure it.)

you said:

Is it possible to make a test on the high-side for this fault purposely?

-> OK. I'll let me check with only Hi-side.

Best regards,

Higa

Hi Higa-san,

If it still has problem, try to increase R21 to 150K (reduce dV/dt to 25V/ns) and increase C45 for better noise rejection on LDO5V.

Best regards,

Jerome