Overcurrent Danfoss

[pattiane]

Hi, I'm using a dSPACE 1104 to control a Danfoss drive VLT5000.
The DC bus of the inverter(Danfoss) is fed by a power feed back unit to maintain stable the DC voltage.
I put a dump load in AC side(Danfoss inverter side).
And when I tried sending a PWM signal in one leg, I can see a overcurrent in my Interface Protection Card.
Someone can help me? Is it due to dead time? How can check?

 

[DickDV]

pattiane, there is a lot about your post that seems unclear.

What do you mean be a "power feedback device to keep the DC voltage stable"? Could this possibly be a brake snubber or brake chopper?

And what is a "dump load"? Could this maybe be an overhauling load pushing the motor or is the motor not connected?

And, finally, how are you putting a "PWM signal into one leg"? Is this maybe a PWM signal on one motor lead and if so, how are you doing that, and why?

 

[pattiane]

Hi Dick, I knew my explanation was not clear...

Actually "the power feedback device" is a REVCON Unit(german brand).I feed the DC bus via the rectifier from the Danfoss.
The only function of the Revcon unit is to maintain the voltage stable.The measurement done indicates 550V.

I measure via transducers(LEM)Vdc,Idc, and 3 phases Vac and Iac from the output of Danfoss inverter.

I have an external brake chopper, that I can control via the dSPACE PWM.The brake chopper is used when the REvcon unit doesn't work correctly to limit VDC.

The dump load is just a load which consumes active power.
Basically I used this load instead of the motor.

My final objective is to implement in the Danfoss unit some controllers(the power for instance) via the dSPACE system.But before implementing whatever, I just tried testing whether my systems works.
So after fedding the systems (e.g. DC Volatge is present in the bus), I just tried to send one PWM signal in one leg.Just to check.Normally the overcurrent shouldn't occur.
I don't know whether it's the deadtime or something else.
But normally the deadtime is already implemented in the IPC board.

Hope it helps you to find my problem.
Cheers.
pattiane

 

[pattiane]

I forget something.
I tried using without the dump load.But the overcurrent is still present.

 

[DickDV]

Sorry, pattiane, I'm not the one to help you on this. Sounds like you are designing or redesigning an inverter.

 

[Skogsgurra]

Is it possible that you mean "dummy load" when you say "dump load"? That would make some sense.

Gunnar Englund

http://www.gke.org

 

[ozmosis]

pattiane
have you checked out the transistors within the Danfoss VLT5000 inverter section? (without power connected). If you are getting an overcurrent then it is possible you have an open-circuit IGBT.
what is the VLT5000 connected to if it is not a 3phase motor?

 

[pattiane]

No, the dump load is a resistive load designed to dissipate the excess power generated by the generation systems when the batteries are fully charged.

 

[pattiane]

hi sed2developer,
Actaully I'm not using the VLT5000 for a motor control, but for a power control.
Initially I'm trying with the dump load(active power consumer).

But my problem is only a hardware problem, since there is any complex control.
When I send a PWM signal to one leg of the inverter, I have a overcurrent with or without a load.

 

[ozmosis]

but have you checked the power transistors within the VLT5000? A quick diode check with a decent multimeter will indicate if you have a problem there.

 

[pattiane]

Hi Sed2developer,
No, I haven't check the power transistors yet.
I will.

 

[pattiane]

Interesting remark:
When I send PWM signals to the 3 legs, there's no overcurrent!
But sending PWM to one leg or two legs occur the overcurrent.
If anyone can explain that...

 

[CJCPE]

How are you determining that you have an overcurrent? Is that a fault message from the VLT5000? Could it be that the VLT5000 is detecting that there there are not balanced currents in the three phases but does not have a separate fault message for that problem?

 

[ozmosis]

I agree with CJPE. Here is an extract from the VLT5000 service manual:
"The VLT detects over current conditions by monitoring the
current output of all three phases. An over current condition exists when the current in an individual phase exceeds 165% of the units maximum overload rating. At that point the IGBT's in that phase are turned off. Shortly after turning off, the IGBT's will once again be gated on as the current in that phase will have dropped. This turning on and off of the IGBT's will continue for up to 1.5 seconds after which time the VLT will trip on Over current.

The inverter section is looking for a balanced 3ph output and any deviation on this will create an overcurrent trip. The drive thinks that you are running a motor on a 1ph.
Still not sure how you are bypassing the gate driver circuit of the drive and just runing "1 leg".

 

[pattiane]

Actually I'm not using the original card of the VLT, I'm using and other one:

http://ftp-dee.poliba.it:8000/Liserre/23-24 Lezione 07 06 2004/IPC5manual.pdf

Ive done some other measurements, the conclusion is :
If a put in PWMa = alpha;
PWMb and PWMc must be alpha+0.03 or alpha-0.03
For instance: alpha = 0.5;
PWMb = 0.52 => Ive no problem but if I put PWMb=0.55
There's an overcurrent.

Could be my LC filter in the inverter output which creates the overcurrent? L=0.45mH C=6.9uF, switching frequency could be 5kHz, for instance.

 

[pattiane]

Hi All,
The cause of the overcuurent was the wrong connection of the LC filter in inverter output (in CL filter)!!! Electrician's problem!

I wonder now why I don't have ant sinusoidal voltages at the filter output. It'd be normally sinusoidal, should it?

I wonder whether it's insignifiant to get 60Volts variation of voltage in DC bus for a normal work of the inverter.

 

[Skogsgurra]

Good to have the feedback. Interesting board, why is it being used? Is the standard board not good enough?

Gunnar Englund

http://www.gke.org

 

[ozmosis]

I'm suprised that the system worked with the 3-legs operating then if the LC filter was wired as a CL filter.

 

[Skogsgurra]

Me too. But it didn't work very well. Did it?

Running with the C first is a little like running a long motor cable without reactors. It puts extra switching load on the IGBTs when they charge the cable capacitance.

Also, most LC filters do not produce a sine-wave as pattiane expected it to do. And it is also dependent on the carrier (swiching) frequency. Turn carrier frequency up and you will probably see something similar to a sine. But, do read the manual for the inverter and filter first. Most inverters and filter combinations are designed for a certain carrier frequency - ususally 2.5 kHz (in the VLT 5000 case) and going above that often results in either an overload message or very hot filter components.

Gunnar Englund

http://www.gke.org