HVKit with ACIM 220VAC Marathon 48T17D2000K No Load Speed

1. I'm not sure that the motor shipping as HVACIMTR is the same as it originally was. Several people have commented that the values are quite a bit different than what is in user.h for the 5Kxxxx

2. use proj_lab10 with the modulation changed to 0.66 and see what speed you can reach. This should be suitable.

3. the magnetizing current could certainly be too high, the motor ID process for ACI motors is not nearly as accurate/repeatable as for synchronous motors. one of the ways to test this is to enable powerwarp while running at rated speed. The magnetizing current will drop very low. Then start applying some load slowly and watch magnetizing current rise. Once you are at rated load and rated speed look at the magnetizing current. This is what should be used in user.h

4. there could also be some element of speed controller tuning changes required for the high speed issues you are seeing. to take this variable out of the equation you could use InstaSPIN-MOTION and the SpinTAC controller. Much easier to get a very well tuned velocity controller which is stable over the entire operating range.

Chris, thanks for your quick reply.

Specific followup questions...

1. Can TI please verify that the motor shipping in the sold dev kit is the expected motor, and confirm what the motor params should be?

2. Can TI / others please verify / confirm what operating environment must be in place on hvkit to achieve standard no-load speed (just as if motor were directly connected to 3phase power)?  Example...Once motor params are defined...DC Link voltage minimum with sv mod at 0.5, PFC must be used??, you must always use PowerWarp mode to expect to reach no load speed??? Stuff like that.  This is the HVACIM sold by TI with the HVKit, so I would think TI would have this motor fully characterized.

3. I am working on running lab10.  thx.

4. I am attempting again to post an image inline with my post...It is a cut and paste of section 4.8.1.3 discussing Getting the Rated Magnetizing Current of an ACIM. It states that the ACIM magnetizing current USER_MOTOR_MAGNETIZING_CURRENT should approximately be the no load current (peak).

My motor nameplate and measurements have shown the no load current to be roughly 1.2amps RMS.    So calculated value for USER_MOTOR_MAGNETIZING_CURRENT would be roughly 1.69. 

The instaspin ID process came up with 1.53 peak.

It seems like the Magnetizing current value should be right according to these rules...The user guide value is even higher than the value generated by motor ID process. 

5. I will look at running InstaSPIN-MOTION and the SpinTAC controller.  thx.

6. Can TI please verify that I should be able to reach standard no load speed with the ACIM sold in the kit from TI with standard magnetizing current and no PowerWarp, and no PFC active? (Just as a directly connected motor would do)

7. Can someone please review/correct my brief guess / analysis regarding svgen voltage behavior possible impact on my issue.  Please correct me where I am wrong.  BTW, is the multiplier by 2/3 because there is always two low sides active, putting two low side impedances in parallel with each other and causing the high side phase to have 2/3 of the l-l voltage drop across it?

8. Why aren't inline pasted images getting posted even though I submit with images? :)

Thx again.

1. I've requested a HVACIMTR be shipped to my attention so I can see. From a few similar posts I have the feeling that the vendor started shipping us a slightly different motor that they (or the manufacturer) felt was equivalent. This is going to take at least until the end of next week before I can reply though.

2. Minimum bus voltage (this should be about 165Vdc).  You will need to use over-modulation (proj_lab10a) to utilize the bus fully like line driven. With the previous proj_lab## you are using about 25% less voltage.  We don't really characterize this motor though...not like it seems you think we do. It's just a motor to demonstrate some of the control techniques. 

4. yes, that is correct. 

6. Yes, using proj_lab10 with appropriate USER_MAGNETIZING_CURRENT and tuned speed loop (or using torque mode)

7. I'll have to review later

Chris,

 

Thanks for the update/reply.

 

Here are my findings regarding overmodulation when studying no load speed (Target is 1.792krpm based on 230VAC direct 3ph line connection behavior):

 

1. this could be from the speed controller (but if you didn't change the speed gains in lab10 it's not) or from noisy current sampling

2. the overmodulation includes an algorithm for choosing the current sampling set-point and which current readings are used. it looks to me like this helps quite a bit once you reach the 0.45 modulation

3. It sounds to me like this is probably what I call voltage hunting. Instead of commanding 1800 RPM, try commanding 2000 RPM. It won't reach this value, but it should stop the hunting effect. If you are still seeing some of it than I think most of the issues are from getting poor current measurements into the system.


4. I haven't used MOTION on this motor in a few years, so I can't comment of the top of my head. But they are higher than what is in the default USER_MOTOR for Marathon_5K33GN2A

5/6. sorry for your troubles with MOTION. Others have had good success. Sounds like there are some documentation items to clean up though...

7. this is the behavior expected with PowerWarp. It reduces the magnetizing current which lets you achieve higher speeds with much lower current. The trade off is less torque. When you need the torque PowerWarp will increase the magnetizing current, but you do lose some dynamic capability (meaning the torque has to build) so it shouldn't be used for an application where a load could be applied very strongly.

8. I would look at adjusting your USER_MAGNETIZING_CURRENT based on enabling PowerWarp and then slowly providing a load (if you can do this with your set-up). You can follow the torque / speed curve for a motor and plot the Id required for a given speed and load.

well, I did confirm that this is definitely a different motor than the original. changed happened in early 2014.
and we just heard from the manufacturer that THIS model is being discontinued. Hopefully they are sending the recommended replacement and we can take a look...

1. Glad you heard something. I appreciate you pursuing this.

2. I would post a picture of the motor I have, but when I paste a picture in here it doesn't make it in to the submitted post.  What am I doing wrong?

3. I looked at the voltage ratings on the parts on the HVKit rectifier section.  The line to line cap C1 is rated at 250VAC, but it is not populated on my kit.  The varistor was the next limiter with a rated voltage of 275VAC.  So, I went ahead and ran the rectifier line voltage up through about 270VAC.  Sure enough, I was able to reach the no load speed observed when the motor is directly connected to 230VAC 3phase60Hz.

Calc DCLink voltage...

265VAC*sqrt(2) = 374.4VDC.

270VAC*sqrt(2) = 381.8VDC.

Testing...

At 265VAC I achieved 1.77krpm with reported DCLink @ 360VDC

At 270VAC I was able to reach my target ~1.8krpm with reported DCLink @ 367.7VDC.

Motor params as documented earlier in this thread, Using Lab10a, gMotorVars.OverModulation=0.5 (Can't go any higher than this otherwise motor starts sounding noisy/rough, need to figure out what is happening here).

 

4. Excited to see what results you get for motor params, basic performance/behavior etc.. when your motor comes in.

5. I am still going to run the slowly loaded PowerWarp magnetizing current experiment you suggested.

 

Thx

2. I got a picture from your TI sales/apps. You have the currently shipping version, not the original that I am more familiar with.

looks like you are making some good progress

I receive the motor which is shipping as HVACIMTR
and ID'd as:

#elif (USER_MOTOR == Marathon_48T17D2000K)
#define USER_MOTOR_TYPE MOTOR_Type_Induction
#define USER_MOTOR_NUM_POLE_PAIRS (2)
#define USER_MOTOR_Rr (6.73)
#define USER_MOTOR_Rs (9.7)
#define USER_MOTOR_Ls_d (0.0051)
#define USER_MOTOR_Ls_q USER_MOTOR_Ls_d
#define USER_MOTOR_RATED_FLUX (0.8165*230.0/60.0)
#define USER_MOTOR_MAGNETIZING_CURRENT (1.75)
#define USER_MOTOR_RES_EST_CURRENT (1.0)
#define USER_MOTOR_IND_EST_CURRENT (NULL)
#define USER_MOTOR_MAX_CURRENT (5.0)
#define USER_MOTOR_FLUX_EST_FREQ_Hz (5.0)

I am only using 110 Vac = 160Vdc bus

With this voltage it runs fine using proj_lab10a to
715 RPM, 0.5 Vs
835 RPM, 0.57 Vs
~880 RPM, >0.6 Vs it starts breaking down

there are certainly some bad samples getting in the system at this overmodulation range.

I tried increasing gCmpOffset with same/worse results.

with powerwarp (Id reduced to minimum) I can run to
1800 RPM, 0.35 Vs

I lowered the MAG_CURRENT to (1.5) A and I was able to run a little faster, but it still broke down over 900 RPM.

I think this is primarily a problem with the over-modulation / current sampling.

increasing the svgencurrent.minwidth helps some. Try setting to 220 - 270 range in the expressions view or the GUI

Chris,

1 . Glad to hear you have received the marathon 48t17d2000k hvac devkit motor. Thanks for posting the params that your motor ID process came up with. As you saw, my values are similar. That is good to know.

2. Definitely both of our Ls_d/Ls_q values are different than the 5K33GN2A motor that had been previously mentioned.

3. When TI runs / (Can you please consider running one) a manual no-load/locked rotor/DC resistance test to check/verify the motor id results, can you please share those with this thread?

Thx

DanC,
I tested both the 48T17D and the original 5K33GN2A using both MotorWare v12 and 16. My results were reasonably consistent on both for each version. Both versions showed similar issues in the over-modulation region, though 16 was slightly better.

When you run ID are you using a 208Vac+ line? Or 360Vdc bus?

I think one of the issues I'm seeing is using 110Vac line (~160Vdc) for ID.

For 48T17D

If I lower
USER_MOTOR_RATED_FLUX (0.8165*110.0/60.0)

then I only require
USER_MOTOR_MAGNETIZING_CURRENT (0.64)
to make this flux, which is actually much more reasonable for a 1.2A rated motor

and the inductance is then ID'd as
USER_MOTOR_Ls_d (0.035)
which is more reasonable IMO.

The Rr ID's slightly lower, but still in the same 6ohm range.

I can then run this motor to 1849 RPM with a modulation index of 0.60 (MW 16). I have to lower the speed gains to 2.0/0.02 and the minimum width to 160. Modulation index over 0.6 becomes unstable unless I lower the magnetizing current further.

Chris,

In response to your line voltage used during motor ID question, I have been using 230VAC. 

Thanks for sharing your experimenting results.

I recently re-ran motor ID with this 48T17d2000k marathon motor.  Just for reference debugger view shown below...

With overmodulation set to 0.5 I can reach 1.5krpm max.

You mentioned reducing the magnetizing current...If I lower the rotor flux value, the required mag current to generate that flux would go down.  Less flux I can run the motor faster (up to and beyond natural line voltage sync speed).   I get that.  However, doesn't that reduce the max torque I can get from the motor?  Maybe I'm confusing magnetizing current (i_d)  with the "torque producing" current i_q??

My best results so far with the calculated rated motor current value (1.5-1.7 amps peak) have been to ensure that the DC link voltage is high enough to support the svgen section reaching the line-line voltages that the motor would normally see when directly connected to 3ph power.

I just need to understand/review the impact of reducing mag current on generated torque capability a bit more.

Thx

reducing the magnetizing current will reduce the flux, which should reduce the torque. but there is also a tradeoff in current which can be used for torque (Iq) vs. flux (Id) as you are limited to Is as set in MAX_MOTOR_CURRENT. If you use all your current to produce flux (Id) you don't have enough left for Iq.

At 230Vac I believe your 1.5A of Id is correct for rated torque, and being limited to about 1500 RPM @ 0.5 makes sense. You should be able to use overmodulation to get to 0.6 which should get you right to 1800 RPM. I do believe that beyond this modulation we are seeing some issues, and you certainly need to maintain enough bus voltage to use the over-modulation appropriately.