Variable Frequency Drive Capacitor Issue

Thread Starter

qrb14143

Joined Mar 6, 2017
112
Hi all,

I am building an AC Induction Motor drive using a six switch inverter. I have a 540V / 5.5A supply, feeding a DC link capacitor of 30uF which in turn feeds the six switch inverter. The inverter outputs are connected to the motor which is rated at 550W, power factor 0.8. My switching frequency is 20kHz.

When I power on the circuit, the motor begins to spin, the power supply then goes into current limit mode and shuts off, then it recovers and the same thing happens again. If I connect a larger 235uF capacitance in parallel with the existing DC link capacitor, everything runs beautifully.

Does anyone have any suggestions as to why the system won't run at the lower capacitance value? I have worked through the procedure in this paper https://www.kalbeck.com/asset/630/Whitepaper SalconeBond.pdf and it suggests that I should see only 105mV of ripple on the DC rail and hardly any ripple current.

Thanks
 

kubeek

Joined Sep 20, 2005
5,623
Does anyone have any suggestions as to why the system won't run at the lower capacitance value? I have worked through the procedure in this paper https://www.kalbeck.com/asset/630/Whitepaper SalconeBond.pdf and it suggests that I should see only 105mV of ripple on the DC rail and hardly any ripple current.
Which part of the procedure exactly do you mean? Ans what value of inductance did you use? Did you measure the actual ripple current and ripple voltage on the capacitor?

Also, the paper assumes the source inductance to be very large, so that the ripple current peaks dont flow through it and the source provides only the average current. However in your test setup I would expect short wires betweent the PSU and your board, which will have litlle inductance and the PSU will be paritally covering the current peaks that would normally be covered by the bulk cap.
 

Thread Starter

qrb14143

Joined Mar 6, 2017
112
If I apply a very crude simplification, the motor is rated at 550W, so assuming Pin=Pout, then the power supplied from my high voltage supply should also be 550W. At 540V, that equates to around 1A, so I am operating my HV supply at well below what it's capable of.

When I power on the rig, the motor spins briefly, then the power supply "overcurrent" warning LED comes on and kills the supply voltage at which point the motor slows down. When the motor slows down, the overcurrent warning goes away and the supply voltage is restored. This repeats over and over again and things never settle down. However, if I add a larger DC link capacitance, the issue goes away and the motor runs nicely at its rated speed.

As you say, I have short thick wires between the PSU and the DC link capacitors and even shorter and thicker wires into the inverter so I would expect my source inductance to be very low, allowing the source to pick up some of the current ripple, hence why I don't understand why it needs a larger capacitance to run correctly.
 

Thread Starter

qrb14143

Joined Mar 6, 2017
112
To obtain a value for the inductance, I assumed the motor was in steady state, drawing 550W at a power factor of 0.8 lagging and I knew the line voltage so I went from there. I arrived at a per-phase inductance of 0.915H
 

kubeek

Joined Sep 20, 2005
5,623
To obtain a value for the inductance, I assumed the motor was in steady state, drawing 550W at a power factor of 0.8 lagging and I knew the line voltage so I went from there. I arrived at a per-phase inductance of 0.915H
I have little practical experience with AC motors, but I would expect the actual inductance to be masked by the back EMF of the motor in steady running state, so the inductance could be a lot less.
As you say, I have short thick wires between the PSU and the DC link capacitors and even shorter and thicker wires into the inverter so I would expect my source inductance to be very low, allowing the source to pick up some of the current ripple, hence why I don't understand why it needs a larger capacitance to run correctly.
The ratio of impedance of the wiring indcutance to the impedance of the dc link capacitor will dictate how the peak current is shared, and which of the two energy sources provides the majority of the ripple current. If you increase the capacitance you skew the ratio in favor of the capacitor, so the capacitor provides more of the peak current and the psu provides less.
Another factor can be that the psu can have some sort of integration on the OCP, and even very short current peaks over a long enough time can trigger the overcurrent protection.

Also, you didn´t answer my question, what did you measure? Are your calculated results even in the ballpark of what is actually happening?
 

Thread Starter

qrb14143

Joined Mar 6, 2017
112
I have little practical experience with AC motors, but I would expect the actual inductance to be masked by the back EMF of the motor in steady running state, so the inductance could be a lot less.
The ratio of impedance of the wiring indcutance to the impedance of the dc link capacitor will dictate how the peak current is shared, and which of the two energy sources provides the majority of the ripple current. If you increase the capacitance you skew the ratio in favor of the capacitor, so the capacitor provides more of the peak current and the psu provides less.
Another factor can be that the psu can have some sort of integration on the OCP, and even very short current peaks over a long enough time can trigger the overcurrent protection.

Also, you didn´t answer my question, what did you measure? Are your calculated results even in the ballpark of what is actually happening?
Thank you for your help so far. Your suggestion about the OCP applying integration is very interesting and may well be the source of my problems.

In answer to the question of inductance, I am not sure how I would actually measure the inductance of a running motor. Calculating the inductance as I described above was the only way I could think of to get a value for the inductance.

How would you recommend obtaining an inductance value?
 

kubeek

Joined Sep 20, 2005
5,623
I think you can simply measure it without the motor running, that would be the value I´d use in the calculations.
 

Thread Starter

qrb14143

Joined Mar 6, 2017
112
I think you can simply measure it without the motor running, that would be the value I´d use in the calculations.
I will certainly give that a go. I had assumed that it would be influenced by a myriad of factors when the motor was spinning and that a reading at standstill would be meaningless.
 

MisterBill2

Joined Jan 23, 2018
4,025
I suggest running the 540 volt DC supply in a constant current mode, and setting the current at a level that is within the limits of both your switches and your motor. It seems that the smaller value capacitor is not able to provide the peak currents needed by the motor. Trying a constant current mode will allow you to see how close the smaller capacitor is to what you need. BUT are you certain that the capacitor is actually 30MFD? And I am wondering about what speed the motor is supposed to be running at.
 

Thread Starter

qrb14143

Joined Mar 6, 2017
112
The motor is currently running unloaded, neglecting frictional losses of course. This is my final year project and I only had a week left so I just went with the larger capacitors for the duration of my testing which seemed to do the trick. The capacitor is most definitely 30uF since it is a parallel combination of 3 x 10uF caps so I would expect a very low ESL and ESR from this arrangement as well. This has been replaced with a series/parallel arrangement of electrolytic caps giving 470uF in total.

For a three phase application, I am unsure as to why the DC link capacitor is needed since the motor should present a reasonably balanced load meaning that the DC current should be fairly constant, should it not?

I did not have time to perform a locked rotor and no load test to deduce the parameters of the machine. I have now stopped work on the project and am preparing my final report.
 
Top