InstaSpin SVM Questions

Hi Hayden,

Welcome to the wonderful world of motor control!  As you have already discovered, motor control can be challenging at times, but few fields in engineering can be as rewarding!

 

To answer your question, an IGBT's ON time can be longer than the switching period, especially at higher speeds when you need to utilize over-modulation.  If this is a concern, you can of course limit your MAX voltage value to produce a duty cycle less than 100% to prevent this from happening.  But this will also limit the highest speed possible from your drive.

To better understand how we do SVM in InstaSPIN, I have attached a document which may help.  When you get to the page which shows the system block diagram, click on the SVM block to see an explanation of the technique, as well as the code that we use.

Also, to help you better understand motor control in general, you might find some of my seminars to be of use:

www.youtube.com/watch

Regards,

Dave

https://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/171/InstaSPIN_2D00_FOC-Secret-Decoder-Ring.ppsx

 

 

Thanks Dave! I have started watching the youtube videos and have been looking through the file attached. The videos have made my long commute to work a lot more interesting!

I was mainly curious about the switching technique used in your SVM module. I have read that there are SVM solutions that reduce the switching thus reducing the switching losses. Or is that a property that is inherent to SVM as a whole? The reason I ask is that I am actually in the process of developing a drive for a PMSM and will be using the Instaspin FOC solution and I was trying to size a bootstrap capacitor. Any advice?

Thanks Dave,
Hayden

Ha!  Hopefully you aren't driving while you are watching the videos!  :-)

 

The SVM technique we use with InstaSPIN is sometimes called the "alternating reverse sequence" technique.  This gets its name from the fact that in the first half of the center-aligned PWM period, two voltage vectors are generated followed by one of the null vectors.  Then during the last half of the PWM period, the voltage vectors are generated in reverse order, and it ends with the other null vector.  However, this technique does NOT reduce switching losses compared to standard sinusoidal modulation.  But it does result in 15% greater modulation voltage capability compared to sinusoidal modulation.

 

One SVM technique which does reduce switching losses is the "NULL V0" technique.  This is discussed in my video in the section on SVM.  With this technique, everytime you have an opportunity to use a null vector, you always use null vector V0.  With this technique, one phase voltage is always low for 120 degrees.  This means that one inverter phase is not switching at any given time, resulting in a 33% reduction in switching losses.  But at low speeds, this may present a problem, as any given transistor may stay on for a long time and possibly overheat.

 

Sizing up the bootstrap capacitor depends on several factors, which are listed below:

1.  How much charge from the bootstrap capacitor does each switching event require?

2.  What is the time required to charge the high-side bootstrap capacitor when the bottom transistor goes low?

3.  What type of SVM technique is used?

4.  Is overmodulation required?  This may require the top transistors to remain ON for a prolonged period of time.  During this time, it is the quiesent current of the high-side gate driver circuit that bleeds the capacitor, not the switching current itself.

So the cap size is related to all of these issues.  I would strongly recommend doing a Spice simulation for one phase of your inverter circuit so that all of these factors can be observed.

 

Best Regards,

Dave

 

 

 

Dear Dave,

Very thank you for the documents that you uploaded on this topic. I spent two days to watch all 5 parts of motor control series that TI posted on Youtube. It's so interesting and very useful for me, an engineer who want to manage advanced motor control techniques.

However, in point of view of a Motorware's user, I would like to suggest you have more training videos in which:
1. You can guide motorware users to understand the working mechanism of every modules of motor control that is coding in Motorware.
2. If you can, please share or public all presentation that you used in the motor control training series, so that I (and many Motorware's users) can review in detail and deeply understand more about your lectures.
3. I really like the InstaSPIN-FOC Secret Decoder Ring.ppsx of yours, and I hope that you can share more the similar documents in order to Motorware's users can track easier the structure of sotfware as well as the related codes for every modules/blocks when developing their own projects with Motorware,

Your job is great, and thank you so much again.

Tran,

Hi Tran,

 

I am very happy to hear that the video series and the Decoder Ring tool were helpful to you.  Thanks for your feedback, and also for your suggestions for future training ideas!

 

Warmest Regards,

Dave