What's new
What's new

400VAC 50Hz Cost Effectively

I'llFixIt

Plastic
Joined
Feb 19, 2014
Location
Michigan, USA
Hi all,

The company I work for has machines that will be shipped to Europe, the power system at the destination is a 5-wire system, 400VAC, 50Hz, 3-phase. Our customer is requiring us to run the machine off on our floor in this configuration. We have 120, 240, and 480V 60hz in our shop. We are trying to figure out the cheapest way to get to the target voltage and frequency. We found special generators for this purpose, but they are fairly expensive to rent. My thought was to use a Variable Frequency Drive configured to output 400VAC/50Hz. But I'm not certain I can command BOTH frequency and voltage of the drive.

Any thoughts? Better ideas?

FWIW, the searches I've done for this come back to things like motors will run a little faster/hotter when changing voltage nominally, say 380V/50Hz to 460V/60Hz, but that's not my circumstance. I HAVE to supply the machine with 400VAC/50Hz, only for about 8 hours. The machine is very simple, no motors... Everything on it is ran on 24VDC, and my VAC to VDC power supply is fine with running at 400VAC 3-ph so that's not a problem.

Thanks.
 
You didnt say what kva but, rent a 480 60 hz diesel genset and have the shop turn the gov down to 1500 rpm (50 hz) and set the volt reg to 400 volt, no big deal 1 day rental and a little labor...Phil
 
If your load only consists of a DC power supply, all power is rectified, then the frequency will not matter.

I tested 5 large 3Φ 400V DC power supplies for a customer last week, connected to a transformer. They all worked just fine.

I have a 5KVA 400V 3Φ transformer, if you need something like that, for your testing. PM me if interested with your location.

SAF Ω
 
I know a guy in St Louis that has a 50 Hz electric-to-electric motor-generator type genset at about 7.5 kVA. At least he had it the last time I saw him, and he usually holds onto stuff.

I know the actual genset, it was used at one of my prior employers. No idea what he wants for it, or what shipping would be, but I know he has not been using it, and he did not pay a lot for it.

BTW, rectified systems MAY find the frequency important, the filters, etc can behave differently. Even inrush suppression may act differently. Anything with a PFC (that would be most EC stuff) needs to be capable of 50 Hz.
 
I would agree, all of the power supplies I tested, were rated for 47-63Hz 400/450V. Intended for the EU market.
His may be different, or not. Only he can tell.

SAF Ω
 
Biscuits, I just created a new thread because I couldn't find this one. If a mod posts the other, I'll attempt to delete it.

So, I now have an AB powerflex 755 11amp VFD, running in voltage adjust mode which is supposed to be much cleaner power than a standard VFD. It gets an immediate hardware overcurrent fault when the power supply is connected. VFD is verified outputting 400VAC@50Hz. Power supply is rated 380-480VAC 50/60Hz. The drive is having no problems running a computer programming outlet or a couple of small operator cooling fans, but it just can't do the power supply.

Power Supply, IFM EFector DN4033 is rated at 4amps peak in rush current, drive can handle 26 Amps peak, 11 amps continuous. There is NO WAY the power supply is pulling even 11 Amps on boot up. Power supply is verified as good on another piece of equipment using 480VAC@60Hz. Rockwell Automation tech support hasn't been able to help me yet either.

SAF, I understand where you are coming from with the power being rectified and frequency not mattering, but I have to supply the machine master disconnect with the 400VAC@50Hz, it is a customer requirement I can't get around. I have to prove the machine capable of running like this as it is going Europe.
 
BTW, sorry for long period of absence. Between coronavirus shutdown and all sorts of problems, we are just getting this worked out in practice instead of just theory.

I know I made another post here, odd that the moderators would approve one but not both, hopefully the other one comes through.
 
Cheapest method will be a 480v vfd with a line load filter on the output and delta connected capacitors to clean up the waveform. Collectively all 3 phases will still have a square wave of 340volts peak,at the pwm frequency of the drive from line to earth ground. But line to line will be clean sine wave.

The pwm output of a vfd will blow up most electronics down stream. Your 480v vfd will deliver 700 volt square waves without a filter.

Another option is a 240v to 480v transformer. Drive it from a 240v vfd configured for 200v 50hz. The leakage inductance of the transformer may be enough you can put y connected capacitors on the output and you can earth ground the neutral formed by the caps. This will get you clean line to line and line to ground voltages.
 
Thanks for the reply Johansen.

Seems my other post isn't going to make it so here goes again.

I currently have a PowerFlex 755 drive with voltage adjust mode, it's an 11 amp continuous drive. Input of the drive is 480VAC@60Hz, programmed output is 400VAC@50Hz. Supposedly this drive is special with this voltage adjust mode, it's intended purpose is not for non-motor specific voltage applications. The drive is going to an IFM Efector DN4033 DC power supply, 380-480VAC in and 24VDC 10Amp out.

With Lo-Pass enabled on a Fluke 87V, I see I am getting 400VAC@50Hz from the drive. When the output of the drive is turned on, and the circuit breaker between it and the power supply is on, the drive immediately craps out with a hardware overcurrent fault. The drive is rated to 26 Amps peak, the power supply is rated at peak startup inrush current of 4 Amps. I'm reasonably certain the power supply is incapable of pulling enough amps to shut the drive down, power supply should blow up first.

The drive output is perfectly happy with powering a small computer programming outlet which is single phase, also a couple of small fans which are also single phase, but it REFUSES to run the power supply.

I don't currently have any caps or filter on either side of the drive. I do have a large AB filter on a shelf, Rockwell/Allen Bradley part number 2090-XXLF-X330B, might that help?

The DC power supply output has a circuit breaker, so no downstream electronics have been effected yet. Downstream of the power supply is a PLC, HMI, industrial barcode readers, prox sensors, ethernet switches, etc. If the DC power supply can turn on, I'd trust that it's switching circuit cleans up the power enough that all DC components would be safe.
 
That filter seems suspicious to me. It is probably a CLC configuration for noise removal.

Go to home desperate and buy 3, 500 foot spools of 12 guage wire and you've got 3 inductors. Any ordinary motor run caps will work, you probably want about 20uf delta caps.
 
Johansen,
Thanks for the reply. I have plenty of large spools of wire in the shop. Someone here told me not to wind, say a ferrite toroid ring, as a choke. How would you suggest winding my load wires through the inductor(s)?

And the filter I have, you're correct it's for EMC/RFI, I just thought the guts might have the same effect as a toroid/inductor.
 
I got away with as little as 500uH for a 400watt 240v vfd. (A 9mh 3 phase line load reactor had only 480uH common mode inductance)


I found that 65 feet of 12/3 wire coiled up as tightly as i could only measured 120uH.

500 feet of 12 or 14 guauge should be plenty. Inductance follows turns squared.

I found that i could use y connected caps and connect the y point to the dc bus without blowing up the vfd. Bug there are issues specific to various vfds that prevent me from recommending people do that.
 

Had some corespondance with a man who tried a cheap automation direct line load reactor for the sine wave filter applications i have spoke of, and while it worked it got very hot. He found the laminations to be .020" thick and did not appear to be insulated. The MTE reactors i have used are .014 and the extra heat loss from adding delta connected caps is minimal
 
Johansen, I am going to go with an MTE unit, it's only a bit more than the automation direct option.

Rockwell Automation recommended a true sinewave filter, which is $$$. The way I understand them, they do pretty much what a line load reactor does just far more cleanly. Am I mistaken? I'm thinking the load reactor should work, if I could pay for a sinewave filter, I'd might as well just rent a fancy generator.
In our shop it's generally acccepted as a bad thing to have capacitors on the outputs of VFDs. Do you think the delta wired caps are really necessary with the load reactor?

Thanks all, really.
 
the pure sine wave filter should be a special core that has a very high common mode inductance to filter out the common mode voltage, which is what destroys motor bearings.
for example: Sine Wave Filters | Products | KEB does not filter common mode voltage, its a 3 leg core with delta caps. (only 2 required)

this reactor core, https://base.imgix.net/files/base/e...xus_motor_protection_filter.5e67b3815c949.png the air gap in the middle provides the energy storage needed to filter the line to line voltage.
the circular path through all 3 coils has no air gap so it has a very high inductance (but because of that, nearly no energy storage capacity (not needed anyways)) to filter the common mode voltage.

each leg of the vfd is a square wave of varying duration and the 3 leg core of a line reactor can filter that into a sine wave without capacitors provided the load is "stiff" at the pwm frequency. so, a 5% motor reactor might get rid of half of the pwm at the motor leads, and that's enough. yes you still need capacitors after the line reactor, depending on what you're driving. a small 24v 10 amp power supply may work find without capacitors. or, you could blow up the capacitors in the emi filter from the current flowing. or, you could create a standing wave between the reactor and some capacitor somewhere and the voltage generated could be problematic.

The 3 phase voltage output of a vfd is not steady relative to ground but collectively all three phases float up and down at +170v, -170vdc at the pwm frequency of the vfd, which is typically 4-16 Khz. this common mode voltage cannot be filtered by a 3 leg reactor. (double those voltages for a 480v drive)



if instead of buying a line reactor (which has no common mode inductance) you just use 3 separate inductors.. you can the use y connected capacitors and ground the center point.


additional problem: don't ground the y center point directly. ground it through a 120v 500 watt lightbulb.

if the light lights up when you start the vfd up, then you have a problem. here's why: the vfd can arbitrarily make the common mode voltage delivered by the vfd.. anything it wants it to be within limits, at whatever frequency it wants it to be. the programming is bad on some of them. this is why i stress Delta connected capacitors when recommending people just hook up a line/load reactor.

when the vfd is producing on the order of >80% of its maximum voltage or more, there isn't room for the vfd to be delivering any funny stuff in its common mode voltages, so you can ground the y point directly. in my experience the weird stuff is produced below 20% voltage. i think it has to do with poor programming associated with the "Boost" voltage which is normally on the order of 10%.


anyhow, i got away with 600uH of inductance on a 400 watt vfd which seemed to me like there should have been enough ripple current to blow the drive. but it worked fine.

an isolation transformer will remove the common mode voltage completely. you can then ground the Y point of the secondary and have a legitamate 5 wire system. delta or Y caps on the transformer will filter the line to line voltage and my guess is most transformers (sized comparably for the vfd) will have enough leakage inductance you won't need a line/load reactor.
 
I don't have any motor bearings to worry about messing up. The most sensitive equipment on here is a single phase electric torque driver, which has it's own power supply outputting DC voltage to the driver. And a fan for a worker. Everything else is 24VDC, so I only need to worry about getting that 24VDC power supply going, without frying it over time. Ordered a load reactor from MTE, RLW-001403. Will give that a go. If that doesn't work, the idea will be scrapped altogether in the interest of time vs. cost. Will report back.
 
Update, load reactor allows the power supply to power up instead of overloading the VFD as before, so the reactor must be filtering. However, the power supply squealed like a pig and sounded like it might blow up. So I guess this idea is dead at this point.
 








 
Back
Top