PDA

View Full Version : DC power supply for larger sized VFD's



Mcgyver
02-15-2013, 08:56 AM
This thread linked below has some good content on why you'd use a larger VFD if inputing single phase vs 3. On large vfd's this gets really expensive, and as pointed out in the thread the drop can be huge. I just looked at a 7.5 hp baldor drive that is only rated for 2hp with single phase input. Ugh

http://bbs.homeshopmachinist.net/threads/55479-where-are-we-at-with-VFD-s-how-good-are-new-ones


One solution is putting a DC power supply before the VFD and powering the VFD that way. I think that's what I need to do to get the desired output without spending a fortune.

My question is, on building the DC power supply. I've built small power supplies and view it as a fairly simple thing .......switch--->fuse--->rectifier---->capacitor. throw in a resistor to drain the cap and probably a thermistor so I don't pop the breaker with the inrush current and you're done. Supply would be 240 single phase in and up 7.5 hp.

Assuming an appropriate steel enclosure and grounding, is there anything I need to be aware of in scaling this up quite a bit? Just wondering what the practical and technical challenges are, things I should be aware of etc in building such a supply...how much capacitance, do i need special caps for that amount of power, does it need cooling, etc

Thanks!

Duffy
02-15-2013, 10:35 AM
Mcgyver, wont this be very closely related to the rectifier in a welder? I know a welder is only running under 100 volts, but the circuitry would be similar. I suspect that even the rectifiers, or bridge, will have a PIV rating as high as you would need. In a PERFECT world, you might even be able to just buy, (scrounge,) a replacement welding rectifier and drop it in. have to check the voltage rating on any capacitors though.

Mcgyver
02-15-2013, 10:44 AM
Mcgyver, wont this be very closely related to the rectifier in a welder? .

i'm asking because I don't know what I don't know....but is the rectifier much of an issue? a quick check of digikey has 400V 50amp rectifiers for under $5.00

MaxHeadRoom
02-15-2013, 10:52 AM
All the VFD's I have stripped for parts have had a 3ph bridge rectifier and large capacitor bank at the supply input, the main difference between the 1ph rated and 3ph only is the Cap. bank size, you could up the energy capability by just upping or adding extra capacitors after the bridge.?
The diagram in the link shown is a bit misleading as the bridge is not made up of SCR's in a VFD, the output is usually IGBT's that synthetically produce a High freq switching variable freq. variable amplitude 3 phase sine wave
Max.

Forrest Addy
02-15-2013, 11:07 AM
If a VFD was used to drive a sewage pump or screw air compressor 24/7 then a conservative de-rate for the VFD driving it from single phase would only be prudent.

But consider the end use and likely duty cycle of the load motor. If a manuial machine tool is in the hands of home shop guy in no hurry chances are there will never be a moment when the motor is close to full load amps. In this situation the de-rating of a VFD running from single phase can be relaxed to say 1 1/2 times the motor HP. The same machine in a small commercial shop where efficency counts and the operator skilled and aggressive there may be many high HP cuts per hour and the VFD should be sized accordingly like 2X the motor HP. If I were a civil engineer sizing a VFD for a sewage station located in the boondocks where only tsingle phase was available, I'd be sure to spec the VFD 3X the motor HP.

The weak point of a three phase rated VFD when running from single phase is the bridge rectifier and especially the filter capacitors. The capacitors charge and discharge to flywheel the pulsating DC from the rectifier so the VFD's output can efficiently switch it into sine coded three phase AC. This charge/discharge cycle is ecpressed as "ripple currect" and its effect on capacitors is heat. Too much ripple current and the capacitors fail, sometimes loudly. An off-board DC power supply with the beef to handle single phase input would require less of a de-rate of the VFD.

If yuu separately supply DC to the VFD another problem comes when you have to isolate the VFD's on-board DC logic power supply that powers the VFD's control circuitry. If you can isolate the VFD's power supply and separately power it or somehow power the logoc separately you can efficiently run a normally rated VFD from single phase - but you have to jump through those puwer supply hoops: the main supply and the logic supply.

becksmachine
02-15-2013, 11:16 AM
One solution is putting a DC power supply before the VFD and powering the VFD that way. I think that's what I need to do to get the desired output without spending a fortune.

Thanks!

You can buy vfds already configured to install this way, multiple sections with no power supply that use a common dc buss power supply. I don't know if you could find specifications for this power supply online?

Failing this, it is sometimes possible to augment the existing diode bridge and then replace existing filter caps with some larger values to keep the ripple currents in line.

Dave

Bruce Griffing
02-15-2013, 06:34 PM
Peak rectifier current and ripple current in the capacitors are the issues you face in designing a dc supply for this application. Why? If the capacitors hold the voltage up as they should, then the rectifier will only conduct at the peaks of the sine wave input. This has a couple of implications - peak currents in the rectifier are 10-30 times the average current. Most large bridge rectifiers have a peak as well as an average current rating. When the rectifier is conducting, it is mostly charging the capacitor. That means that the capacitor has to handle the peak current as well. This calls for a very low ESR (equivalent series resistance) capacitor(s). All of that said, you will be fine if your parts are rated for the peak and ripple currents. The only other problem to handle is the start up surge - that can be handled with an inrush current limiter.

J Tiers
02-15-2013, 08:16 PM
Yes, you CAN add your OWN rectifiers and capacitor setup, and uprate the VFD to full duty at nominal rating.

The problem is doing that in some cases......

Here is a typical 3HP VFD semiconductor module made by FUJI. The 5 HP and even 10 HP or larger may use similar modules. Obviously, it would be rather difficult to replace the rectifiers without significant board surgery to disconnect these and swap in your own.

http://img.photobucket.com/albums/0803/jstanley/Fujimodule2.jpg

If you can find a larger unit that has the DC bus brought out, which usually occurs with units that do NOT have an internal brake transistor, but instead get an optional external "brake module" wired in.......... THOSE you can often directly wire to the brake terminals with good DC and get full duty. Some have an internal DC choke which they want for protection as well as ripple and inrush suppression, and you may need to connect before that.

You also are responsible for your own inrush suppression..... a big cap bank draws a big surge when you flip the switch, so usually there is a resistor to cut the surge, which is shorted by a relay or SCR after the voltage reaches 70% or so of normal. That you would need to put in your external rectifier system.

For most purposes, all you are really needing is a heavy rectifier, and good sized capacitors, with inrush protection.

The ones another poster referred to are "common bus" VFDs, which are made to operate off DC, and have no rectifiers. They are usually associated with a "front end" that may be either an SCR rectifier unit, OR a bidirectional VFD which can return power to the line if the DC bus voltage goes up.


A new deal coming in is 3 phase with SMALL capacitors (physically and electrically small), like only 68 uF or so, of a "film" type which handles large currents and has no lifetime problems. I am currently designing a 3 HP very small and compact unit of that type for one of our clients.

They recently hit us with a requirement for "some" single phase input applications. The only way that can work is with a good-sized external capacitor bank that is optionally connected. If that isn't done, it is not the ripple current, it is the fact that the DC bus voltage will "follow" the AC line..... from nearly zero volts to peak, and back down, because the small film cap cannot hold up the voltage with heavy current. With 3 phase, even with NO filter, the voltage never goes below 86% of the peak, so on 208V, the minimum voltage with NO filter is near 250VDC.

That is ANOTHER limit on the single phase usage, but is easily fixed by external capacitors and rectifiers for commercial units which follow this design style.

Lew Hartswick
02-15-2013, 09:37 PM
Don't any of these folks know about choke input filters??????
...lew...

lakeside53
02-15-2013, 10:09 PM
Not if they are less than 50+ and never dealt with vacuum tubes :)

Bruce Griffing
02-15-2013, 10:46 PM
Choke input filters are more expensive. Think about the current involved. To JT's point - the DC bus does not have to be external. You can simply power the VFD with DC. You are wasting an extra diode voltage drop or two, but no reason is should not work.

J Tiers
02-15-2013, 11:55 PM
Don't any of these folks know about choke input filters??????
...lew...

Of course..... MANY units already have internal chokes, or require an external one in certain circumstances.

However, a choke is not a cure in every case. Along with reducing the inrush, the choke ALSO reduces the bus voltage, and a choke large enough to suppress excessive current surges may ALSO reduce the bus voltage to a point that it is not useful. If you have a circuit simulator, you may have some fun trying it out.

A choke and capacitor filter tends to allow through only the average of the voltage presented. Since the average of half sines is 0.66x the peak, that is the voltage which a heavy inductor/capacitor filter will output. A lighter filter will output something between peak and average, while a rectifier and capacitor is a "peak hold" and tends to charge the capacitor to the peak voltage of the sine, but only at the cost of a large peak current.

You pays your money and you takes your choice.

ldbent
02-16-2013, 12:38 PM
How about making your own rotary phase converter to drive the VFD. Here in southern Canada(Minnesota) a 10 HP 3 phase motor can be had for about 100 Loonies. A used contactor, some run caps and momentary contact switches and your in business. On second thought this probably isn't for your hobby shop. Never mind. If a RPC is a possibility the Practical Machinist board has a forum dedicated to RPCs and VFDs.

lakeside53
02-16-2013, 01:01 PM
Assuming you can feed the vfd from an external dc source, you can take this single phase conversion one step further. Make the dc front end "voltage doubling" and use surplus 480 volt VFD's. These tend to sell much cheaper then their 240 volt equivalents.

But... not for the unknowing or inexperienced. The voltages exposed can most certainly kill you... Even the type and condition of the probes on your voltmeter need to be taken into consideration.


There are few variations on the theme, but here's a common form : http://www.coolcircuit.com/circuit/voltage/vd.GIF

This shows how it works :http://www.talkingelectronics.com/projects/ThePowerSupply/imagesP79/VoltageDoubler-1.gif

macona
02-16-2013, 01:03 PM
I would just add extra buss capacitance.

lakeside53
02-16-2013, 01:47 PM
Yes, but aren't you roling the dice as to whether your diodes can take the added peak current?

Similarly, there was an earlier comment about creating an external dc power supply then feeding the dc to the three phase inputs in parallel. That of course assumes that the VFD doesn't "need" the ac for anything else, BUT, only half of the vfd rectifiers will conduct, and the average current through them will be doubled. I think...

EVguru
02-16-2013, 03:14 PM
Placing a battery bank on the DC bus is becoming common for VFDs used in Lift applications.

Determining the spec of rectifiers and capacitors usually isn't that hard, so you can get an idea whether they're adequate for single phase operation. Monitoring the DC bus voltage and capacitor/rectifier temperature tells you a lot too.

Mcgyver
02-16-2013, 04:09 PM
How about making your own rotary phase converter to drive the VFD.

I've got a 10hp RPC, works perfectly...BUT I have to put a VFD on this drive for speed control and if I have to put a vfd in, I would really really really like to not use the RPC. Shop time is lot more Zen like without its drone in the background


But... not for the unknowing or inexperienced. The voltages exposed can most certainly kill you... Even the type and condition of the probes on your voltmeter need to be taken into consideration.

not 100% sure what you mean - doing 240V house wiring can kill you as well? I'm far from an expert but do whatever needs doing, all my house wiring, lots 240 and 600 3P in the shop, built a rpc, rewired my DSG 600 V control panel etc. ....is there something about this type of project that makes it more risky, or are you just giving friendly reminder that this isn't hooking up the car stereo? :D

Mcgyver
02-16-2013, 04:11 PM
I would just add extra buss capacitance.

to where - hacking in side the VFD? how does that overcome Roberts point that the main issue you're only using one let of the bridge diodes?

Mcgyver
02-16-2013, 04:15 PM
If yuu separately supply DC to the VFD another problem comes when you have to isolate the VFD's on-board DC logic power supply that powers the VFD's control circuitry. If you can isolate the VFD's power supply and separately power it or somehow power the logoc separately you can efficiently run a normally rated VFD from single phase - but you have to jump through those puwer supply hoops: the main supply and the logic supply.

I thought, from that other thread i linked to, that you could get power to the VFD va AC 1P, AC 3P or DC and the VFD figures it out.....is that not the case? do you have to supply power seperately to the VFD for its own uses if you're supply DC instead of AC as the main input?

J Tiers
02-16-2013, 05:00 PM
I thought, from that other thread i linked to, that you could get power to the VFD va AC 1P, AC 3P or DC and the VFD figures it out.....is that not the case? do you have to supply power seperately to the VFD for its own uses if you're supply DC instead of AC as the main input?

MOST recent VFDs use the DC to get internal control power and the isolated "user control" power for switches etc.

SOME VFDS, more typically larger ones, will check to see that 3 accurately phased input lines exist before they will agree to operate. Those may not be happy with DC or single phase.

SOME that check only look for voltage, and can be fooled by strapping L2 to L3 or the like. That obviously does not help the diode current issue if there is one.

SOME VFDs are rated for full power on single phase. Most of the smaller Invertek VFDs are, for instance, and those are widely available under other names, SEW, TBWoods, Baldor, etc. (for certain models, others may be from a different source)

MaxHeadRoom
02-16-2013, 05:14 PM
There are some, Mitsubishi is an example, where the internal control power is fed off of two of the dedicated phases, so if using single phase, these two have to be used, in fact they have a jumper where you can leave the internal control on and remove the power to the main rectifier for standby purposes, if you so wish.
Max.

lakeside53
02-16-2013, 05:39 PM
not 100% sure what you mean - doing 240V house wiring can kill you as well? I'm far from an expert but do whatever needs doing, all my house wiring, lots 240 and 600 3P in the shop, built a rpc, rewired my DSG 600 V control panel etc. ....is there something about this type of project that makes it more risky, or are you just giving friendly reminder that this isn't hooking up the car stereo? :D

Just friendly advice. It wasn't aimed at anyone in a particular, but there are a lot of noobs reading this that have no idea just how nasty voltage doubled 240 (480) can be... there is roughly 679 DC volts across the capacitors. You can get nibbed by 120, bitten by 240, but 480 will really get you.

I work around a lot of 480 3P. If the panel or junction box is live the electricans have to put on flash protection before removing the front, and they do. Their "live work" 600 volt AC rated screwdrivers, wrenches, gloves, meters and probes are kept in a separate padded containers and are in pristine condition - not hanging on a toolbelt or banging around in a tool box. Wish I was a careful every time, but I try to be most of the time ;)

J Tiers
02-16-2013, 05:54 PM
Just do not work with "live" circuits.....

Hook up voltmeter before applying power.

Keep one hand in your back pocket if the thing HAS to be live while you do something.

You already have safety glasses on for any shop work, but you need 'em for poking in live electrical stuff for sure.

The really serious flash burn issues probably don't apply to the limited power we usually have around. But for 480 at 800A, the effect is like being a cockroach standing on the part someone is welding. The arc is the same sort as welding, but considerably worse.

Don't be the guy who has to think, after the fact, about how easy it would have been to protect against "what got you".

lakeside53
02-16-2013, 06:00 PM
Very good advice.

macona
02-16-2013, 06:10 PM
Usually the input bridges are rated pretty high. I wouldn't worry about it. You could always look it up and see what the rating is, it may be an integrated power module or it could be a separate 3 phase bridge. You could also use an external bridge to supply power to the external caps as well if you are that worried about it.

On one I had I just ran heavy wires from the cap traces outside.

Mcgyver
03-04-2013, 12:19 PM
I was looking at this Baldor unit in 7.5 hp model....

http://www.baldor.com/downloads/manuals/_downloads/737-702.pdf

there are DC terminals in the illustrations, but I'm not seeing text about DC input....there's also mention of the resistor brake hookup and its unlcear to me if this uses the DC connections.

Do you guys think this unit will accept DC input? I'm thinking fuse, rectifier, thermistor and caps. Should i use the oil filled ones like on the phase converter? any guess as to required capicitance?

Or should I just hook it up figuring i'll get 5 hp or so and that that'll be enough

thanks

J Tiers
03-04-2013, 10:34 PM
The brake resistor uses the plus and the collector of the brake IGBT, NOT the minus. Look at the photo I posted in one of the threads of the de-canned Fuji module..... I point out the brake IGBT and diode.

There appear to be plus and minus DC terminals, which MAY be OK to use. However the DC- is marked "no user connection" on page 3-11, which leads to a suspicion that it isn't present, and that the marking is just a standard marking molded-into the case, and is connected in some sizes and ratings, but not all.