EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: viper on December 09, 2023, 04:37:42 pm
-
Hoping to spot a guru with VFD experience here! I am mostly familiar with VFDs having used them for decades and even repaired them. But I have a question as I want to solidify this in my brain. As is known, a VFD has a full wave rectifier on the front end to make DC to be sent back out via the IGBT pack. What is common, and most drives will do is accept single phase as input. Where this gets tricky is you are now limited on input phases thus you can't get full power from the drive.....so they are commonly derated for the application. I am trying to assess the math to ensure I am correct.
My brain....a VFD has a rated input current. I think protective circuits such as CTs on all phases is a gamble. Some have more protection than others. Some are happy to self destruct. If I take for example a drive with a rated current of 15A, that is 15A for each phase and assume 200V, that is 200*15*1.732 = 5.2kw. However, as you run on single phase, you STILL must not exceed the 15A limit for any input conductor unless you want to make smoke. So on single phase that is 15*200= 3kw. So the derate is NOT simply assuming derate by 1/3 due to the missing phase, it is factored by the square root of 3 or 1.732?
If I am on target so far, I am simply investigating any possibility to better optimize the front end of a VFD for a single phase application? I had vaguely thought about adding capacitors over to the missing leg to get all 3 phases powered, but I don't think it can work. AC current through a cap is limited by capacitance and frequency and I am not even sure what the voltage and phase shift would look like on L3.
I had thought about any internal mods that could be done to a 3P VFD to better utilize 1P power, but it would seem only a redesigned 1P full wave rectifier could work as you would have to turn up the current with a factor of 1.732. Once the DC bank is loaded, nothing in the drive would know any different, but still seems like quite a challenge, especially if the front end has protections such as CTs.
-
There's also ripple to consider. The unsmoothed output of a single phase bridge rectifier drops to zero twice per cycle. That of a three phase bridge rectifier doesn't, as the next phase takes over when the preceding one has dropped to half the peak (cos(60)=0.5), which happens six times a cycle, so much less reservoir capacitance is required.
You'd need to know the min. acceptable DC bus voltage, total reservoir capacitance on the DC bus, and possibly the ripple current rating of said capacitors to calculate the actual derating factor, and know whether its diode current, ripple troughs or ripple current that limits it.
-
Some ruminations on the problem:
I work with VFDs & Servo drives and designed one from the ground up in the mid 80s for a product. That one was single phase input and I had the luxury of being able to use a series inductor/choke/reactor to limit peak current. Many VFDs and Servo drives often suggest using series reactors on line inputs to reduce peak current and improve power factor.
The 1.73 factor is about right - 1/1.73 is 0.578 which is close to the 60% derated power limit that many VFD manufacturers quote. I suspect it's based on average rectifier currents for single vs three phase. Unless we can get the specs on the module used in the VFD or, even better, the characteristics of the diode chips in the module and run a complete analysis or simulation it's pretty much a rule of thumb.
I have a personal project using a 5 HP Lenze VFD to drive a 3 HP motor for a belt grinder. With metal, faster is better and I can crank up the RPM to 150% of base speed, so concevably It can put out 4.5 HP. The VFD is 3 phase in but Lenze does not rate it for single phase input, period. I tested out the VFD with single phase but I was lucky enough to have a 2.5 mH 20 A reactor from the project above that I used in series with the line. I loaded it down quite well grinding some big pieces of steel and it at least didn't expire.
I'm not happy with abusing the VFD this way and have been mulling over various solutions:
1.) the most practical is to build a complete rectifier and capacitor bank that bypasses the VFD's rectifier and augments the DC link capacitance. Not all VFDs have the DC link brought out though. I've got the parts, I just need to get a round tuit. (joke, joke)
2.) I thought of designing a passive phase shift network that synthesizes the proper three phase input. It occured to me that the proper angular separation between the phases is only important at high loading. At low load, the required capacitance is reduced and the inductance is increased. I also thought that a center tapped transformer would be needed to balance the synthesized phases. I made a few experiments using parts at hand and it sort of worked but, if memory serves, a capacitor in series with one of the VFD line inputs forms a voltage doubler that causes the DC link voltage to rise at low load. I didn't confirm the mechanism but it seems reasonable. It could also have been a resonance with the inductance.
3.) the Phase Perfect single phase to three phase converter synthesizes the third phase using some electronics and a PWM stage. Here's a link to their brochure: https://www.phasetechnologies.com/downloads/phase_technologies_phase_perfect_brochure_102a5cdea0.pdf (https://www.phasetechnologies.com/downloads/phase_technologies_phase_perfect_brochure_102a5cdea0.pdf) They used to have a white paper that went into the concept but I don't see it on their website. If memory serves, the single phase input power lines are passed through with minimal processing. These lines also drive a rectifier and capacitor bank to provide the DC link voltage that is chopped by a PWM to generate the 3rd phase. I thought of doing something similar but solution 1. above is easier.
In North America, two things work in our favor:
- we generally have 220 to 240 V single phase line whereas three phase motors (NEMA anyway) require 208 V. This means that the VFD can be less efficient and can tolerate more ripple on the DC link.
- we have 60 Hz line frequency which means smaller DC link capacitors than for 50 Hz. Most VFDs these days are sized for 50-60 Hz.
Hope this gives you more insight into this,
Cheers,
-
Great points mentioned guys, regarding the ripple running on single phase. My general 'assumption' on industrial drives is they don't tend to put things on the edge, but provide a little headroom. Most drives I have been around, Yaskawa, Mits, ABB, etc, will have an OL capacity of 150-200% for brief periods. And even pushing harder, usually (nearly always) there is a CT on at least two output phases to monitor load conditions and shut the drive down before total destruction occurs. One reason I am a big fan of VFDs is all the protections you get!
And as Duak mentioned, many of them are 50hz capable. So I have to assume the front end is pretty robust, and since many drives allow for turning off phase loss detection, I would think the ripple would be within range on single phase. Never tested that though. I do know they work. However, I could not imagine turning off phase loss in a real 3P application as they would seem a recipe to burn out a drive.
It would seem that contacting an OEM or doing a deep dive on the front end of a particular drive would be the only 'proper' thing....
Duak mentions something that is even deeper, that I forget about, that many drives have DC buss taps, in which I guess you could inject more DC, or maybe even run the drive through those means. Probably not worth the effort but in theory, maybe.
I am familiar with the phase perfect units and they are indeed pass-thru on the single phase legs, and they digitally synthesis the 3rd. I think that is where they did some things right because one issue we have faced is dynamic braking. The decel of a load can push current back to the line in which the IGBTs cannot accept the return current. Usually either the power is burned off through a resistor (cheap way), or a line regen module is used to feed power back.
I know on one CNC application, a brand new machine was blowing up everything. It was under warranty but the guys they send are doing good to even use a meter correctly! In that machine, they were using nothing but an oven resistor element for braking power absorption, and a wire was loose. This caused very high DC buss voltage. On that machine, they tap the DC buss at the spindle VFD to run all other servo drives, so when that voltage went high, it really did some damage! A small rant ,that OEM should be slapped for not having proper DC buss voltage protections! It was quite a miss IMO! I know on good Fanuc drives, they will shut down if the DC buss gets too high.
-
An auto transformer producing extra volts, as many as possible without over volting the drive.. then put a cap in series with the third rectifier terminal.
This will force current through the cap into the rectifier at 90 degree phase shift.