[zwmaster]

Hello.

I'm new to this community and would like to get some opinions on critical part's selection.

Plan is to build all in one controller with DC-DC converter for board electronics, BMS and controller for servo pumps. I want to keep all car electronics as is and use ECU information to drive electric motor.
Basically take information’s from CAN network in use it for driving the electric motor.
This part is already done and it’s working well (Pedal position, many other signals from brake, motor speed, vehicle speed, shifter etc.).
I have Kostov 11” 250V motor.

On start I have few questions:
- Has anyone done regen with Kostov 11”? Perhaps Plamenator could give us official statement?
- IGBT selection. 4-6pcs of IXGN400N30 or 1-2 pcs of older Gen2 modules? Like CM300DY-12NF, FZ1200R16KF4 or 1MBI800U4B-120
- Parallel IGBT drive recommendations. Plan to use +-15V and few TI or similar drive chips.
- Snubber cap’s, large Elko’s or many small ones or metalized polypropylene film cap’s with low esl
- Cap combination experience?
- Has anyone tested design with mounting the diode to motor not inside the controller?

I hope this will be a constructive forum other can use for guidance on controller construction.

Thnx in advance.

 

[Techie]

I've been looking for a dc controller that'll do regen, seems most of the common manifacturers are afraid to impliment it... I'm still deciding between ac and dc (mainly on that regen bit).

 

[zwmaster]

Hi Techie.

It’s controller mfg. nightmare, but in legal aspect. The problem is in the DC motors. As far as I have learned till today, many can’t be used for regenerative braking and are destroyed if used in this way.
I only hope Kostov 11” can be used like that.

How to control regenerative braking is my other question.
Could the stator and rotor coils be controlled separately? Not be conected in series by default but using IGBT modules. I will know exact position of the motor axle (komutator) and have a lot of processor power in my DSP controller.

There are also other problems from here.
Storage of large amount’s of energy during regen.
Maybe intermediate storage in large capacitors is the way and also burning excess of energy in electric heaters and AC drive.

Happy 2010

 

[Qer]

This is not an AC versus DC issue, but more a series wound versus the rest of the world (well, almost at least). For example, a SepEx brushed DC-motor can regenerate just fine and for example Renault has built some SepEx-based EV's that can regen just fine.

The problem with series wound that makes them tricky to use for regenerative braking is that the field and armature are connected in series. This means that if you induce a magnetic field in the field your armature will generate current that will feed the field that will create more magnetic field in the armature which will increase the current that will cause a bigger magnetic field, etc. A series wound motor is, simply put, instable in regen which makes it very tricky to create a braking force that won't shake your fillings out or blow either the motor or controller due to over current.

Then there's an additional problem that's pretty much self inflicted in the DIY-community and that's the habit of advancing the brushes to increase the voltage the motor can handle. While advancing the brushes works just fine to make a construction originally meant for 72 Volts (I believe?) handle twice that voltage and even more, it also means that running the motor in regeneration or reversing the direction might easily damage the motor due to excessive arcing.

An interpoled motor won't have the problem with advanced brushes since it's neutrally timed, but you still have the problem with the rather instable nature of the series wound motor in regen mode. Then there's the other aspect of it all and that's if it's even worth bothering with. Unless you live in a hilly place or do a lot of stop and go regen is probably not even worth considering and even if there might be benefits from regen with your daily driving habits most of the regenerated energy will probably go to waste anyway.

When you convert the kinetic energy to electric energy and then back to kinetic again it means all the losses are doubled. The energy has to go twice through the transmission, twice through the motor, twice through the controller and first charge (with losses) and discharge (more losses) the battery pack. Battery chemistry also matters, with for example lead acid regen is pretty much worthless since Peukert will rob you of most of the regenerated energy.

Also, when you regen you must keep track of SoC too. If you, for example, first charge the pack to 100% and then regen down a hill you might damage your pack unless the controller is intelligent enough to count Ah or listen to feedback from the BMS. So is struggling with regen worth a few more percent range? Or should you just stack a few more cells to your pack? The latter is definitely simpler.

 

[DavidDymaxion]

I have done regen with an 11 inch sepex wound Kostov motor. It's not really all that exciting, though, I have just hooked up the field and armature in parallel. On 48V the motor idles at 1570 rpm. Push it faster than that (down hill or towing with a truck) and it does regen.

Regen is pretty trivial with 48V. It would be very challenging to do at high voltages, you would have to buck or boost (or whatever it is) the voltage.

Kostov said something on their website about going parallel/series on their motor, that might help. Something I didn't get an answer on yet, does the regular motor put all the fields in series? Are the brush pairs in series or parallel?

... On start I have few questions:
- Has anyone done regen with Kostov 11”? Perhaps Plamenator could give us official statement? ...

 

[Techie]

I've seen the Warp11 HV is advertised as being regen capable.

 

[Plamenator]

All interpoled motors (i.e. the Kostovs and the advertised WarP 11" HV) are capable of regen as their brushes are not advanced and interpoles in addition improve the heavier arcing present during regen.
The problem, as QeR described it, is in the controller, which has to ensure safe operation.
Should one find a series regen controller, the interpoled motors should work fine.
Note that a non interpoled motor, even with neutral brushes, will probaly arce too much in regen mode.

David, in his post, refers to the series/parallel FIELD mode of the Kostov 11" 250V (http://kostov-motors.com/tractionmo.../seriesdcmotorsforelectricvehicleconversions/)- it only changes the way the 4 stator windings are connected to each other and not the basic rotor-stator connection which is always series.
The series/parallel FIELD modes switch the 4 main windings from parallel (default) to series connection. When in series, rpm of the motor fall down which in turns allows it to go to 250V while maintaining rpm below 6000.
For racing purposes it is better to use the parallel mode as the windings can take much more amps there (the stator windings of the 11" 250V are far superior than those of the 11" 192V in terms of overload).

 

[jackbauer]

I used the 1MBI800U4B-120 in my design. So far no problems. Also added a 1.5uf snubber between c and e.

 

[zwmaster]

Thank you all for your answers. I will respect your comments as you are a lot more experienced than me.

I plan to use ½H Bridge design.

jackbauer thank you for your input.
two 1MBI800U4B-120 should be enough to get maximum performance out of 250V kostov. How many modules do you use?
I also have one FZ1200R16KF4 but when we look at spec’s I think dual 1MBI800U4B is better selection.

br

 

[jackbauer]

I used two modules. Pumped 400amps through them for about 10 seconds on one occasion. Heatsink peaked out at 28c. I could almost hear em saying " You're gonna have to do better than that!"

 

[zwmaster]

jackbauer:
How is your gate driver holding up?
What peak current did you designed it for?

I will opto-insulate the drive part from CPU. I will try to use VO3120 from Vishay as pre-driver. In correlation to maximum current thru modules I think much more peak current must be provided then 2.5A or 9A drivers on the market are offering.

Does someone remember the post on link between max. current required for the motor and size of capacitors in controller? Maybe short rewrite of the formula?

thnx

 

[jackbauer]

Gate driver seems fine its a 9amp part. Basically i'm using a standard control board from the open source controller project and just driving the igbt with a 10R gate resistor. I'll be doing more extensive tests this month when my batteries turn up.

 

[zwmaster]

jackbauer are you using one 9A driver for two IGBT's? Is this enough to get full potential of design?

And one other info for people looking for cooper bar's

Today I found a really low cost source in my neighborhood.
And guess who it is.

It’s my local roofer!

They use allot of 30mm x 5mm, 20mm x 3mm and other's of 6m length
The bar length's EV enthusiasts use is just scrap for them...
It’s really clean cooper

So I’m lucky

 

[DavidDymaxion]

Plamenator, thanks for the great details, it's exciting to see what Kostov has been doing of late.

Is that true for the sepex motor, too, that the fields are wired in parallel? I see I have four brushes, are the pairs wired in series or parallel? Many thanks for the education, I'd rather ask than pull the motor apart just to learn!

All interpoled motors (i.e. the Kostovs and the advertised WarP 11" HV) are capable of regen as their brushes are not advanced and interpoles in addition improve the heavier arcing present during regen.
The problem, as QeR described it, is in the controller, which has to ensure safe operation.
Should one find a series regen controller, the interpoled motors should work fine.
Note that a non interpoled motor, even with neutral brushes, will probaly arce too much in regen mode.

David, in his post, refers to the series/parallel FIELD mode of the Kostov 11" 250V (http://kostov-motors.com/tractionmo.../seriesdcmotorsforelectricvehicleconversions/)- it only changes the way the 4 stator windings are connected to each other and not the basic rotor-stator connection which is always series.
The series/parallel FIELD modes switch the 4 main windings from parallel (default) to series connection. When in series, rpm of the motor fall down which in turns allows it to go to 250V while maintaining rpm below 6000.
For racing purposes it is better to use the parallel mode as the windings can take much more amps there (the stator windings of the 11" 250V are far superior than those of the 11" 192V in terms of overload).

 

[Qer]

Is that true for the sepex motor, too, that the fields are wired in parallel? I see I have four brushes, are the pairs wired in series or parallel? Many thanks for the education, I'd rather ask than pull the motor apart just to learn!

Neither. A SepEx has two fields but they're not wired in parallel (nor in series). Instead you need a dual controller, one for the (static) field and one for the armature where the field typically runs at approximately 1/10 of the armature current (but that's depending on the motor specs). This gives several advantages over series wound motors, such as:

• You can get a high level of torque by applying a high field current
• You can get a high rpm, as far as I've understood higher than a series wound motor for a given battery voltage, by lowering the field current (aka field weakening).
• Regen is as simple as for a PM-motor (and MUCH simpler than for a series wound motor).

The cost is that the controller gets more complicated and that the controller and motor has to be matched with each other. This can usually be achieved by configuring the controller to comply to the motor parameters, but it's still more complicated than the series wound motor which is more or less just plug and play.

 

[zwmaster]

Hi. I'm just preparing some tests.
Is someone using "dual IGBT" design instead of standard IGBT + freewheel diode?

Is this bad idea?
What would happen if we try to use two igbt's driven with pair of PWM signals with 180deg phase difference and dead-band between them?

br

 

[major]

I see I have four brushes, are the pairs wired in series or parallel?

Hey David,

These are 4 pole motors. They have 4 brush holders spaced at 90 degrees about the commutator. Brushes at 180 degrees apart must be the same polarity. Brushes at 90 degrees apart must be of opposite polarity. Otherwise you short the armature. All the motors I've seen have internal "cross connectors" which connect brushes at 180 degrees to A1 and the other brushes (which 180 degrees from each other) to A2. So this would be parallel connected.

Regards,

major

 

[major]

What would happen if we try to use two igbt's driven with pair of PWM signals with 180deg phase difference and dead-band between them?

Hi br,

I think they call this synchronous rectification. Yes, it is done this way in some controllers.

Regards,

major

 

[Tesseract]

I think they call this synchronous rectification. Yes, it is done this way in some controllers.

It would be synchronous rectification, if the active semis were MOSFETs, which conduct bidirectionally, and not IGBTs, which don't... the C-E junction IS a diode, you know

 

[zwmaster]

It would be synchronous rectification, if the active semis were MOSFETs, which conduct bidirectionally, and not IGBTs, which don't... the C-E junction IS a diode, you know

thnx.
Is my conclusion correct?
There no use of second IGBT in drive mode other than it's diode.
It can be used for regen. and braking.
Combining 2nd IGBT during OFF stage of PWM would mean too much stress on the motor or have no effect.