LM5170EVM-BIDIR: ISETD Not Working as Expected

Can you show me your PWM signal waveform?

Hi,

Yes, please see attached.

I verified the duty cycle using the oscilloscope to be 0.1% as well.

Hi Youhao,

Could it be that the PWM signal frequency I'm giving to the ISETD pin on the EVM is too low?

According to Equation 4 on page 21 of the LM5170 datasheet, frequency of ISETD signal should be at least 1/ (2pi*1 kOhm * CisetA).

Looking at the Figure 28 of the LM5170 EVM User guide on page 20, I thought that CisetA = C13 = 100pF. Using the equation mentioned above, this translates to a minimum frequency of 1.6 Mhz. This seems to high and I'm not even sure if I can reach this frequency on a PWM signal using a C2000 MCU.

What would your advice be in terms of resolving this ISETD issue I'm experiencing?

Hi Kartavya,

Instead of increasing the ISETD frequency, you can increase CISETA to a higher value.   The purpose is to convert that PWM pulse to a dc signal without much ripple.  The filter is 100k internal and the CISETA.

Thanks,

Youhao

So to clarify your response, are you agreeing that the PWM signal frequency is the reason for this ISETD issue?

Moreover, are you further suggesting that the only way I can get current control to work on the EVM is by swapping capacitor C13, which is CisetA, on the EVM with an appropriate value?

This does not make sense, because this would mean that the EVM would never work in the configuration setup by TI. I'm very confused by your response towards how to solve the ISETD problem I am facing. Could you please clarify if you don't mind?

Sorry Kartavya for the confusion which may cause you.  The EVM includes a lot of flexibilities to allow the user to reconfigure or modify, but the default setting is what the user's guide mainly addresses, which include the use of the analog voltage loop control. The smaller CISETA helps reduce the delay for the loop control.  Since you would like to use ISETD, you will need to replace CISETA if you don't have a fast enough PWM signal. 

Hope this clarifies.

Thanks,

Youhao

Sorry what error do you refer to? 

The programming error may be larger at light signal because the IC accuracy is trimmed at V(ISETA)=50mV, and at light signal level the internal offset becomes more significant as a relative error, but the absolute error should still be acceptable. 

So, when I set D = 1%, the expected current in the channel should be 20*3.125*0.01 = 0.625A. However, I only see approximately 0.38A in reality.

Sorry I mistakenly clicked TI think resolved.

I see what you mean by the error.  J6 should be open and you can tell from the schematic drawing.  Sorry for the confusion and we will update the user guide.

Could you start to raise the duty cycle and see if the accuracy get better?  The EVM is set for 32A each phase and please do not focus on the accuracy at 0.625A for now.  I also wish to see your PWM signal waveform and see if the wave shape needs to be improved. Could you share?

Thanks,

Youhao

Ok got it. Here it is:

The D is very low so you can't see it clearly here. Also, I'm experiencing some issue with Boost mode and I created a separate question for it on the E2E forums. Would you please be able to comment on that right now as well? Question title: LM5170EVM-BIDIR No Output Power in Boost Mode

You need to zoom in to show the rising and falling edges. Also, the frequency seems too slow in my opinion. What the CISETA value do you have now?

On the other hand, the error is ~0.3A. You most probably see similar abs error up to 32A per phase.  Please continue to raise the duty and you can easily find.

Got it, thank you. CisetA = 100 pF as of now. This is C13 according to schematic.

Please change the CISETA per the datasheet equation, and you will get better results.

Good luck and let us know if we can be of further assistance.

Thanks,

Youhao

Got it, thanks

Also, could you please check out a post about another issue I'm experiencing with the EVM. Title: LM5170EVM-BIDIR No Output Power in Boost Mode

I got it and just replied.  

Thank you. Also, can I disable output voltage control by just removing jumpers J36 and J37? Because, I still want to have access to V48 and V12 sense pins to measure voltages

Can I assume the issue has been resolved hence I can close this thread?  You are welcome to re-open it by adding a new post.

Thanks,

Youhao