Intelligent Power Modules 1

[ShamilaWije]

Hi,

I'm designing a 3 phase motor drive for a 3 phase BLDC drive. The drive should operate from a 440V 3 phase system and should be capable of handling 5kW to 7kW.

I've been looking for a package which integrates all the IGBTs and the drive electronics from manufacturers such as IOR and Fairchild. They all offer very good solutions but only upto 400V working voltage.

The best solution I found yet is from Mitsubishi semiconductor (Powerex). The part no is PS12036. It integrates all the IGBTs, drive ccts and provides protection such as over current protection, control side under voltage shutdown, shoot through lock out etc.

Do you guys have any suggestions?

 

[blacksea]

Maybe better way would be on-the-shelf servo amp including motion controller - for example Elmo (

http://www.elmomc.com)?

 

[itsmoked]

http://www.ixyspower.com/

Keith Cress
kcress -

http://www.flaminsystems.com

 

[ShamilaWije]

Blacksea,

I considered that option but 2 points

The motor is a custom design and run at a very low speed generating high torque and short operating time. Couldn't find a suitable controller.

Ultimately the drive and the motor will be packaged as a single unit. So I thought it's better to tryout our own design.

 

[ShamilaWije]

itsmoked

IXSY has CBI IGBT modules but they do not integrate the gate drive ICs. I was looking to see if IPMs are available with the HV gate drive ICs integrated as well

 

[jraef]

"They all offer very good solutions but only up to 400V working voltage. "

What most drive manufacturers do is use the 400V "class" devices and test to higher voltages, i.e. 480VAC input. By the way, use 480V, not 440V as a design spec. Only legacy systems still have 440V any more.


"If I had eight hours to chop down a tree, I'd spend six sharpening my axe." -- Abraham Lincoln
For the best use of Eng-Tips, please click here -> faq731-376

 

[blacksea]

ShamilaWije,
First, I hope designed by you motor would be as standard one, i.e. 3-phases wye-connected with 4...8 pairs of poles. BTW what's about position sensor?
Second, Elmo amp is very compact and have integtated with MPC motors.
And finally, current sense and sampling, IGBT drivers interface, filters for current / velocity / position loops, PCB design for high voltage and current - too many problems for such design...

 

[sreid]

IMHO the on module protections (especially Short Circuit Protection) is worth ever penny. If you try to design your own, expect fire, smoke, and big bangs a number of times.

 

[electricuwe]

1. Very often IPM modules are choosen for the design assuming that this would allow an easy design without much of power electronics knowledge.

This is definetly not the case

2. The IPM you mentioned contain driver circuits, but they do not contain the signal isolation and the driver power supply to SELV level.

3. In such a direct drive application as you describe, take care to consider power cycling. With low output frequency the junction temperature of the die
- will reach the same value as having the peak current of the sinewave as DC
- will vary over time
- the variation will cause degradation of the bond finally leading to failure. Read manufacturers on this and consult a technical representative. Considering that, Ihave doubts that the module you have choosen provides sufficient current capability.


4. A further pitfall of this drives is that parasitic capacitances of the motor winding are significantly higher than usual. This may cause
- EMC-Problems
- Unintended SC-tripping
- increased swicthing loss

 

[sreid]

Electricuwe;

1. Absolutely true. Designing Power Electronics without some thought results in "Big Bangs." This does not happen with 3.3V logic circuits. i.e., the consequencies of design erors in Power Electronics has more dire results.

2.Yes, fast optical Isolators and power supplies for the Gate Drive signals must be provided. Other than that the supplies must be floating, this is not a really big problem.

3. The Worst Case Values must be sought and respected. Otherwise see "Big Booms." Thermal cyclying and wire bond failure is a feature of all power electronics.

4.Motor winding capacitance-use small series inductors in the motor leads (on the circuit board).

EMI. A life time of study (and disappontment).

SC tripping-see series inductors.

Increased switching losses? Where when using an IPM?

 

[ShamilaWije]

Thanks guys for the response.

electricuwe

2. The IPM you mentioned contain driver circuits, but they do not contain the signal isolation and the driver power supply to SELV level.

Yes i understand that the signal isolation is not provided. But if need be I can optically couple the signals as needed while isolating the microcontroller supply from the IPMs control cct supply. Although this is strictly not since I can use the same supply for the microcontroller while isolating all the signals to the outside. Such as comm lines and Digital I/O etc.
By the way what is SELV level?

3. In such a direct drive application as you describe, take care to consider power cycling. With low output frequency the junction temperature of the die
- will reach the same value as having the peak current of the sinewave as DC
- will vary over time
- the variation will cause degradation of the bond finally leading to failure. Read manufacturers on this and consult a technical representative. Considering that, Ihave doubts that the module you have choosen provides sufficient current capability.

Yes I understand the current capability is not quite enough for the power rating I mentioned(about 5Arms according to the datasheet) but lets assume its enough for this discussion.

4. A further pitfall of this drives is that parasitic capacitances of the motor winding are significantly higher than usual. This may cause
- EMC-Problems
- Unintended SC-tripping
- increased swicthing loss

Can you please elaborate on the parasitic capacitance and how the IPM effect this.

 

[electricuwe]

ShamilaWije,

let me go more into details.

3. Even the RMS current rating of 5 A does not take into consideration effects of power cycling. Low frequency operation may lead to further derating if long lifetime is expected.

4. Switching losses of IPM (as well as for standard IGBT modules and discete IGBTs) are measured assuming an inductive load. A standard electric motor connected directly to the module comes quite close to representing such a load. However, if a long cable is involved or the motor is designed for low operating frequency there is significant parasitic capacitance connected to the module unless you connect further components in between as suggested by sreid.
At each switching of the module this charging and discharging of the capacitances leads to high peak currents and most of the energy stored in the parasitic capacitances has to be dissipated by the IGBTs inside the module additional to the conduction losses and the swithing losses for inductive load.

 

[sreid]

Even worse than the motor winding capacitance, the motor Iron has the permiability of air to fast switching edges. Initially it appears that you are driving an air core inductor until the iron "catches up." The series inductors minimize this effect which, along with the capacitance, can cause some truly impressive high current spikes.

 

[ShamilaWije]

Thanks guys for the info.

Can you give me some guidance on either measuring or estimating these capacitances and permiability effects. Also how do I choose the required series inductance assuming I know the above data.

 

[sreid]

As a general rule, the bigger the inductor the better. Ferrite or powered iron code (for high frequency performance). The inductor has to handle the peak motor current without saturating so for PC board area considerations you will be limited to 10s to 100s of microhenrys.

 

[ShamilaWije]

Thanks sreid. I'll see if I can get some actual measurements on the motor I'm working on.

 

[sreid]

For the Opto Isolators, Avago (HP spun off Agilent who spun off Avago).

http://www.avagotech.com/pages/home/

For current sensing, LEM Closed Loop Hall Current Sensors.

http://www.lem.com/

 

[LionelHutz]

Look at Skiip packs from

http://www.semikron.com

 

[ShamilaWije]

Thanks for the info. I'm already using avago opto-isolators and LEM current transducers. So I seem to be on the rught track.

 

[electricuwe]

Even the smallest SkiiP is severly oversized for this application.