"Braking" with GE VFD

Tom Jenkins

Lifetime Supporting Member
Join Date
Apr 2002
Location
Milwaukee, WI
Posts
6,299
I just got back from a start-up trip with GE VFDs supplied by others. This is on a multistage centrifugal blower application, of a type and size I have done successfully many times before. These were two 350 hp and two 200 hp VFDs, GE Model 6KGME4___DSL1001 "clean power" VFDs. The blowers are a fairly high inertia load.

I normally set the accelleration and deceleration on the VFDs for 15 to 30 seconds. This is to prevent over current trips on accelleration and over voltage trips on decelleration. In rare cases I may have to go to 40 seconds. I cannot recall ever needing more than 45 seconds. I also have the same time for both in all projects that I can recall.

On this job the GE starty-up technician was not as experienced as I might have hoped. (See - I can be diplomatic!) We started at 30 seconds for both accell and decell times. Accell was fine, but on decell we had overvoltage trips during a normal stop cycle until we got the decell time up to a minute. The GE start-up technician indicated he really thought we should be using braking resistors. We did not see any elevated temperatures on the VFD or have any temperature faults.

Finally, the questions:
1) I generally expect a VFD to have similar decelleration torque capability as accelleration torque capability. I expect to use braking resistors only if I have either a short decel time requiring more torque, or repeated braking requiring more heat dissipation. Is this assumption incorrect?

2) On A-B and C-H and Magnetek VFDs I haven't had any problems with decell. Is the GE drive generally less capable in this area than some of their competitors?
 
I can't give you much insight for question 2, but for question 1 unless there is a safety reason or repeated start/stop (heavy duty cycle) you don't need dynamic braking. Unless there is some reason where 60 seconds to stop the motor is unacceptable, there should not be any damamge or heavy wear to the drive without braking resistors. Would be curious to see at what voltage on the Bus the drive trips out on. Don't quote me, but I think for AB it's around 700V for the 1336II drives.
 
Typically all VFD's have about the same braking ability. The primary difference in how much braking you need is load driven.

It is possible the load is being pulled along preventing a faster decel than 30 seconds. I have set as much as 90 to 120 seconds decel on ID fans. Usually most HVAC fans I deal with are factory set 60 seconds.
 
a thought

Something else to check is the incoming power. I don`t know what state you`re working in Tom but in Ok and Tx they set the voltage around 510 volts on a 480 volt system. This is with no load and we`ve had a few problems with over buss faults. At one mill where they where using AB 1336 +`s and the power company put in a bank of capacitors on the primary line we had to have them lower the incoming voltage and the over voltage faults went away. I don`t know what they set the volage at on 4160 and above, but this can cause a problem in the old days it was always low voltage problems, but i haven`t seen that in a long time. You don`t by any chance have dampers that close on these fans when you are trying to stop them do you? They might be harder to stop without a load than with one since it`s a fan? Just thoughts.
And yes we had set the par. for buss voltage as high as possible on the AB drives. They also had a Danfoss it was on a 500 hp chipper and it only tripped once during this time and had the same fault. So may be some drives are more susceptible to these faults. Also they usually faulted while stopping.
 
Last edited:
Tom, I think your basic assumption that a drive will have about the same accel and decel capabilities is not right.

To illustrate, if you have a heavy-duty (that would be constant torque) rated drive, you will get about 150% of the motors rated torque for accel purposes if you ask for it. However, you have essentially NO braking torque capability. Whatever the load coast time is--that's very nearly what your shortest possible decel time is.

Now, if you put a full-capacity rated snubber brake package on that same drive, you will have 100% of the motor rated torque available for decel purposes--maybe a little more. So, even with a full-capacity brake package, you still have more accel torque available than decel torque.

ABB drives (and maybe some others) have a feature called flux braking. This allows a standard drive to develop about 10-15% of motor rated torque as a brake without any additional braking equipment. They use the braking energy to somehow saturate the motor so the braking energy is wasted as heat in the motor. This sometimes gives you just enough braking to avoid adding a braking package but it will not come anywhere near giving your equal accel and decel torque. To get that, you would have to go to a four-quadrant drive.

As to the second question, drives built for the US market only will typically have an input voltage rating of 460 +/- 10% which comes out to 414 to 506V. Outside of these levels, you get either a low or high voltage trip. These drives will typically fault on High DC Bus at about 780V.

Drives built for the world market will have a much broader input voltage range, typically 380-10% to 500+10% which comes out to 342 to 550V. These drives will fault on High DC Bus at around 870V. It gets downright hard to get these drives to fault on Low or High Input Voltage.
 
Tom

I agree with DickDV. Without braking resistors, you will get very little braking torque from a standard drive.
On your previous installations, did you actually time the stopping or just assume that it was stopping in the time set for the decel? Most drives will automatically adjust the decel rate to keep the drive from tripping on Bus Overvoltage. As a result, the decel time may be much longer than the time programmed into the drive. Some drives do this by default, others have to be programmed to do it. The GE drive may not have been set up to modify the decel rate when the Bus Voltage gets near the tripping point.
 
Vic said:
Tom

I agree with DickDV. Without braking resistors, you will get very little braking torque from a standard drive.
On your previous installations, did you actually time the stopping or just assume that it was stopping in the time set for the decel? Most drives will automatically adjust the decel rate to keep the drive from tripping on Bus Overvoltage. As a result, the decel time may be much longer than the time programmed into the drive. Some drives do this by default, others have to be programmed to do it. The GE drive may not have been set up to modify the decel rate when the Bus Voltage gets near the tripping point.

Thanks for the input, guys.

The coast time for these blowers is actually close to five minutes for a complete stop. That isn't a factor. The coast time on these machines is more a function of good bearings than high inertia.

I was basing my question on past experience, where accell and decell were set the same. For example, we did a job recently with A-B VFDs on about the same size machines, and the accell and decell times were both around 15 seconds. On other similar jobs with C-H VFDs we had 30 seconds accell and decell.

After reading your explanation, Dick, it makes sense that braking torque isn't the same. The WK² is 102. This calculates to a braking torque for 30 seconds decell from 3600 rpm of 40 lb-ft. This is less than 10% of the motor rated torque. I didn't think that is asking too much of the drive for a stop once every few hours. The blowers are open inlet and direct discharge to a submerged air header. There is a check valve to prevent back flow into the blower, so the air system isn't trying to free wheel the blower and won't add to the decell torque required.

Vic, you may be right about the VFD adjusting the decell rate as voltage climbs. I've never had to do this in the past, and it didn't occur to me.

Thomas Sullens may be on to something with the high input voltage. I'll have to check into that.

This is my first experience with GE drives in quite a few years. I was less than impressed with them, and I guess I was quick to question the VFD capabilities. I just don't think they are as robust as the VFDs we supply from C-H and A-B.
 
I've had large inertial loads that require MINUTES to coast to a stop. As long as you're allowing current to flow (such as an IGBT inverter back end with a diode rectifier in parallel [OK, they're for inverse voltage protection, but they act as a rectifier for the motor's regenerated energy]), you're going to keep re-gen'ing energy from the motor as long as you can maintain the motor flux. If you break the cycle, that is, interrupt the flow from the motor, the rotor flux will dissipate, and you can elminate the need for braking altogether.

As far as flux braking is concerned, I believe it works on Lenz's Law such as other eddy current brakes use. Make a magnetic field, and rotate a conductor in the field, and the conductor will develop a magnetic field that will oppose the magnetic field inducing the rotating conductor's current. This energy becomes heat in the rotating conductor.

I'd suggest you simply bring the fan to full speed, then disconnect the drive altogether, assuming an output contactor, to see what the natural coast-to-stop time is. Set the decel ramp for this time (most decel ramps are set for the time from full speed to zero speed if time-based).

The short answer is that as long as the motor is regenerating more power than the DC link can absorb (voltage-sharing resistors across the capactiors or braking resistors as dissipation sinks), voltage will rise on the bus. If you reach the voltage level where trips are triggered, you must either dissipate the energy faster (braking resistors) or slow down the deceleration rate until the overvoltage point is not reached. Large inertial loads, such as fans, are typical for this type of problem.
 
DonsDaMan said:
... I'd suggest you simply bring the fan to full speed, then disconnect the drive altogether, assuming an output contactor, to see what the natural coast-to-stop time is. Set the decel ramp for this time (most decel ramps are set for the time from full speed to zero speed if time-based). ...

... Large inertial loads, such as fans, are typical for this type of problem.

The coast time is about five minutes to a full stop. If you set the decell time to this, it kind of eliminates the advantage of having the VFD decellerate the blower, doesn't it?

Again, this isn't an application problem - I've done this dozens of times before on similar installations. In fact, this system is an upgrade, and they have torn out old Robicon VFDs that we were controlling and replaced them with the GE VFDs. The decell and accell time on the Robicon was around 30 seconds, which is typical, and it worked fine for years.

I'm trying to find out the typical parameters for the VFDs. What is the typical deceleration torque available?

Also, I'd like to know if the GE VFDs have a known design difference or reputation for being weak in this function.

One last little question that bothered me. All of the VFDs were purchased from GE, and the phase shift auto-transformers and breakers and so on (clean power VFDs) are mounted in identical enclosures. Four of the actual inverter units were labeled GE, and one one was labeled Saftronics!! Who builds what for whom in these units?
 
Last edited:
I have a manual for a GE VFD that's a couple of years old, so I can't say fro sure whether it applies to your situation.

There is a parameter F41 that establishes the braking torque limit from 20 to 200% of nameplate. Default value for the parameter is for torque limiting to be inactive.

There is a note that if torque limiting is activated, accel and decel times may vary from the settings established by parameters F07 and F08.

I thought GE VFDs were made by Fuji.
 
From what I could tell the GE AF300 series drives don't have any internal DB resistors. All the resistance needs to be connected externally. So you are just left with the resistance needed for the baseline DC bus design, which isn't much.

I didn't think any of the drive manufacturers used internal DB resistors once you got to the power levels Tom is using. So it surprizes me a little that any of the drives can decel that quickly. Unless they were set up to use flux braking and just chewed the energy up as motor heat.

Tom, if you have an application you can look back at see if the drive in a successful application was set up to use flux braking. If so, set the GE drive to use that and you may be in better shape.

Keith
 
I was going to post exactly what Steve mentioned about the torque limit setting. My manual is for an AF-300 P11 drive (I'm not familiar with the model # you mention). I had a similar problem a couple months ago with a 200HP GE drive (Fuji). I would get the overvoltage fault on Decel only.

I must also mention that there is a setting on that F41 parameter to "Automatically prvent OU2 trip due to power regeneration effect". That is how it is stated in the manual.

I had my decel time to well over a minute before adding this setting. After doing this, I could lower the decel back to "normal". I personally could not notice much increase in the actual decel time when this setting was in effect.
 
You can't decelerate a motor with a VFD unless you have either a braking resistor or some other method of removing the rotational (mechanical/kinetic) energy from the motor. The energy conversion (from mechanical to electrical) is done in the motor by having a rotating magnetic field (the rotor) and a stationary set of conductors (stator). The energy that must be dissipated is dependent on the inertia (mass times velocity) and the friction (such as bearing drag, windage, etc.) and other losses (IR loss, PN drop, etc.).

In all cases that I'm aware of (except regen braking which most diode-front end drives cannot do) then the energy is converted to heat either in the motor or the braking resistor. The first law of thermodynamics tells me you can't just destroy the energy.

Flux braking has already been discussed.

If you have a properly-sized braking resistor, then there should not be an overvoltage problem on deceleration. If you don't have a properly-sized braking resistor, then that's the problem. If you don't have one at all, then you're depending on the losses to provide the sink for the electrical-from-mechanical energy.

Your problem can be summed up as "I'm getting more energy from the motor during deceleration than the drive can handle, so it keeps tripping on overvoltage." You need to either dissipate the energy more quickly, or allow the load to slow for a longer period of time.

Edit:
As far as torque limiting is concerned (I've never used this parameter; I always use DC link voltage to determine braking need), then it must time-limit or duty-cylce-limit the current flowing to the braking resistor. With this in effect, you'll allow the DC voltage to rise when the torque limit has been reached, creating a chance for overvoltage. I can only see this being used when one is concerned with couplings or other torsional stresses that can be detrimental when braking too quickly.
 
Last edited:
Tom, basically it comes down to this. An inverter with no snubber braking and no flux braking will have essential NO braking capacity---maybe 1% of motor torque. Your braking rate will be entirely determined by rotating inertia and friction losses in the load.

As for brand relabeling, as far as I know, GE drives are made by Fuji and are definitely on the low performance end of the spectrum. Saftronics relabels Yaskawa and even ABB in some cases. Omron, Magnetek, and ElectroMotive are all Yaskawa. And the list goes on.

Some modern drives have the ability to extend the decel ramp when the DC Bus is charging up close to the Overvoltage trip. You can usually turn this feature on and off in the software someplace. This is nice for avoiding High DC Bus Trips on decel but, of course, it makes your decel time a bit unpredictable. That might not be acceptable depending on the application.
 

Similar Topics

I need some clarifications from VFD experts. I have a project with a vibratory conveyor that must be stopped relatively quickly. Stopping is once...
Replies
2
Views
760
Hi. I have a FR-V520 3.7Kw VFD that seems to have lost it braking. I used to run the machine it is driving and when the motor stopped the...
Replies
9
Views
3,510
Howdy, I was called onsite this afternoon to take a look at a Lenze SMV drive with a braking unit and resistor. Whoever had installed the braking...
Replies
4
Views
2,624
Hello! What would be the proper value for the DC Brake time on VFDs for the DC injection to the motor? Is there a way to calculate? Having it...
Replies
7
Views
4,391
Am teaching a class Oct 18 and 25 Basic VFD and 3 phase motor applied to electric vehicle. Have VFD and motor setup made by Fuji for GE field...
Replies
12
Views
6,784
Back
Top Bottom