TIEVM-HV-1PH-DCAC: Current Monitoring - Active Leg

Part Number: TIEVM-HV-1PH-DCAC

Hi There,

We are looking at the EVM and was wondering what MCUs the inverter will work on and whether there is any porting necessary from the original TMS320F28379DZWTT MCU?

e.g. are there many/any code changes necessary for the baseline code to operate on a single core MCU, and whether there will be changes to operate with 176 or 100 pin package versions?

Also, if you had a choice, which TMS320F2837x D/S MCU would you choose and why?

Thanks in advance.

Robin

Robin,

1. Given the requirement is to only control the inverter, the F2837xS is sufficient

2. We only use one core on the existing design and hence porting should not be a big issue.

3. For changes between pin packages, the interface on this software is not as clean as some of our newer solutions but should not be too bad. Unfortunately we may have some hard coded parameters that you may have to dig into, not everything is in settings.h file

4. F2837x9 has CLB and if you are not using that then it is same as F2837xD, porting between should be very minimal

5. I checked the old schematic and there is onlt HCT1, there are two relays but only one current sensor? can you provide a snapshot where you see two current meas? 

Hi Manish,
1. MCU F2837xS - Confirmed, this MCU's solely role is inverter control, nothing more; I have just reviewed the datasheet as well as checking online and have down selected the TMS320F28377SPZPQ as a candidate; I chose this particular MCU based upon AEQ, instock at digikey and had the smallest footprint available; Please advise if you perceive any issues with this choice.

Side Note: There was a suggestion earlier on about the TMS320F28075, out of curiosity I did a quick check and noted that TMS320F28075PZPQ was also a suitable candidate; Is there any benefits if we were to use this particular MCU? I note it has only 512KB of FLASH, whereas the Delfino above has 1MB; Would the code port easily enough and would it operate on this MCU?

2. Thankyou; I am a little worried about porting between architectures, mainly from ignorance tho, once I am comfortable with the new MCU it shouldn't pose a problem;

3. As per point 1 above, I am hopeful for the smallest package, 100 pin HTQFP;
We considered the BGA package, however were not sure if the IC came with high-lead solder ball, are you able to confirm this?

4. Not being overly familiar with the code, but I take it does not require nor does it use the 2nd CLA (as per your point 2)?

5. The Transphorm EVM equivalent of the old design has a current shunt in both power rails, which differed from the old design; review of code showed it didn't seem to act upon this source of information tho;
I am not sure whether the change to bipolar full bridge as to whether the new code would benefit from a second current shunt in its control mechanism or not;

Note as of late yesterday we made the decision to not implement SDFM, just ADC; purely as a risk mitigator activity;

Thanks for your assistance

Robin

Apologies Manish, I realized I didn't answer point 5 properly...

5. Current sensor - (simplified)
Q1/Q2 output feeds inductor L1, then passes through current sensor U1 before exiting as AC_LIVE;
Q3/Q4 output feeds inductor L2, then passes through current sensor U2 before exiting as AC_NEUTRAL;
Sitting between AC_LIVE and AC_NEUTRAL is capacitor C1;

Cheers
Robin

Robin,

1. I will suggest you keep with F28377xS/D because of the qualification needs of silicon and the timeline.

2. No problem.. 

3. For the solder question, I have asked my colleague to comment. 

4. It does not use any core except CPU1 C28x core.

5. Unfortunately, I cannot answer why Transform has two current sensors  :) as I was not involved with that development, it may be good for you to reach out to them. Our old design uses 1 CT. 

 

 

Hi Manish,

5 - Thanks, so I can remove the second current sensor as the new code relies on only the one current sensor, regardless if I have a unipolar or bipolar switching topology?

PS Should I raise new issues for every enquiry, or perpetuate this one?

Cheers
Robin

Robin,

I can only advise what we did, as it was sufficient. There might be a functional safety aspect to the second sensor that I do not know and hence cannot advise, you should consult with the original designer.

It helps to have a unique thread for the related question, but for example, the solder question could have been a separate thread as different expertise is required to answer the question in detail.

Robin,

Regarding the BGA package, I think that it has no-lead balls, but I will confirm and let you know tomorrow.

REgards,
PEter

I just re-reviewed the Transphorm code now and confirmed that there is no code content that processes the second current sensor, e.g. ADC content is read in, but that is it, no further references to it in either the ASM or C files;

I will discuss internally and move forward from there.

Thanks for your advice

PS Should I mark this as resolved or wait for the BGA question to be answered?