Phase converter voltage deviation.

Status
Not open for further replies.
I

icy78

Member
Location
US
Hello first time posting here, that I can remember at least. Looong post.
I am an hvacr Tech and I have a question on a 3-phase Chiller and pump being powered by a roto phase. What I've always understood as far as supplied voltage and voltage under load, I should be within or less than 2% deviation. I usually get this by adding the three voltages dividing by 3 and then calculating how far off the farthest one is from that. So what I have on this phase converter is 275 254 + 240 volts when the 208-230 3 phase pump motor starts at 3 horsepower. All phase converter capacitors connected. I will read 9.8 amps 7.7 and 7.7. F L A is 8.4.

Now my understanding is that never should I have a motor running with any leg over FLA , but the phase converter manufacturer says that is fine, the motor supplier says that is fine. However if I read information from Fluke or US Motors, or another phase converter manufacturer, they all say that is not o.k.

So the phase converter manufacturer said to disconnect a few of the Run capacitors , to drop the voltage , and that would steady out the amps and bring them lower. The amps did not drop they actually went to 5, 8, 9.8. At new voltages of 235, 235, 246. 3% deviation.
The chiller itself, (when running one of the 4 compressors) the amps on it run 30-27 and 13 at 3% voltage deviation.
Another thing that I have read is a 3% voltage deviation can result in 50% reduction of motor life due to insulation over-heat.
Both amp draws deviate by 40%! 10% is allowable.

I'll bet you want more specific equipment info but I don't have that right now. Site is 3 hours away.
I could get more from others tho.


I'd love to know what you guys say about this. Maybe just more generic input about deviation.
Thanks

PS....the chiller startup manual calls for less than 2% voltage deviation.

Sent from my SM-G930V using Tapatalk
 
First off, understand that people who have a horse in the race will tell you whatever it takes to go away. The RPC supplier was likely selected by the HVAC equipment supplier, so they agree because it's in their collective best interest to agree and to obfuscate the issue until you give up and go away. Then they get to sell replacement equipment once their warranty expires.

Your instincts are correct, a small voltage imbalance creates a more sever current imbalance, and a current imbalance results in excessive heating of the motor windings. When the currents don't match, the rotor generates what are called "negative sequence currents" that counter rotate and create negative torque. It's much less than the normal torque, so it has little effect on that, but what it does is become heat in the motor. So if you look at motor heat produced PER UNIT of current, it means that at a given amount of current in the motor, the windings run hotter. Overload protection of the motor is going to be based strictly on the amount of current actually flowing, but is ASSUMING that it is relatively balanced. So when it isn't, the OL relay may not trip, but the motor insulation is being damaged by that extra heat, hence the issue of reduced motor life; Heat x Time = Failure.

But the other question is, how was the motor sized? Because if the chiller mfr purposely over sized the motor KNOWING that this would happen because it was going to be run on an RPC, it may be fine with it. So for example if the chiller mfr put in a 5HP motor and re-labeled it as a 3HP motor, it can run like this forever.
 
Note also that to some extent the three sets of windings in the motor share a common heat sink and heat dissipation mechanism. That allows the motor to tolerate a small excess over FLA on one winding (without even having to take advantage of any Service Factor the motor might have.)
But this allowable excess is nowhere near allowing you to act as if the total heat production from the three winding can still be tolerated if two windings are "under budget." One reason for this is that the increase of the temperature of the center of the winding over the outer parts of the winding will not be reduced just because the other windings are lightly loaded.
 
Thankyou guys!
Can you explain what you mean, Goldigger, by 2 phases being under budget?

Sent from my SM-G930V using Tapatalk
 
icy78,

You didn't mention the size of your RPC (rotary phase converter) or the size of your 4 compressors, only the 3HP circulator pump.

Right now, your worried about the circulator current balance. When your 4 compressors and condenser fans start coming on line, your problem may go in the opposite direction.

A RPC can only be accurately balanced at one particular load, it's the nature of how they are constructed. The capacitors are selected to resonate the voltage on the generated leg. Usually selected for the maximum rating of the RPC output. At outputs below maximum, the voltage will be higher, and the voltage balance will become a bigger spread. At the full load output, the voltage should be relatively well balanced.

Running small motors on large RPC's can cause unbalanced operation and excessive heating. Most RPC manufacturers will list what is the smallest motor that can be run, on a particular model, to prevent excessive heating.

With a single large load, RPC's can perform well. With multiple smaller loads that come and go, things can get more complicated. If I understand correctly, you have one circulator pump motor, four hermetic compressors, and several condenser fans, all staged to come on line as the load requires. So your balance will be all over the place, depending on the load conditions. High voltages and motor currents when lightly loaded and low voltages and currents at the maximum loads.

A couple of suggestions that can help mitigate the problem. If the circulator runs continuously you might consider swapping out the pump for a single phase model, so that it runs under all conditions of chiller loading without overheating. For the four stages of compressors and condenser fans, it is possible to balance them individually as a zone, by moving the capacitors from the RPC, to the load side of the contactor for each individual zone.

In this way, each compressor and associated fans, can be balanced as an individual unit, and will have minimal effect on the overall system balance when it is switched in and out. Its more complicated and time consuming but can be a benefit to system reliability.

When removing caps from the RPC, to use for zone balancing, can affect the starting of the RPC idler. Both start and run capacitors are paralleled (additive) in the RPC for the phase shift required to get the RPC idler started, at which time the start caps are switched off. This loss of start capacitance, by removing the run capacitance, can be compensated for by adding additional start capacitance, that gets switched off.

As to the unbalance operation and affects. It will never be perfect, you want to shoot for the happy medium, of the typical running load. Any time you run over the nameplate FLA, you are potentially overheating one phase winding, and reducing lifespan. Once you get over the service factor rating, if any, then your getting in the quick degradation area.

In comparing the voltage and current symmetries, I would use the difference between the utility supplied phase and the highest manufactured phase reading. Comparing the average to the highest reading, is not the difference that the motor is working with.

MTW
 
Quote......For the four stages of compressors and condenser fans, it is possible to balance them individually as a zone, by moving the capacitors from the RPC, to the load side of the contactor for each individual zone.

Thanks MTW. That's a lot of great info.

I I am concerned about the chiller and its various loads. I was just wanting them to address the pump, if they get that right , likely,the chiller will fall into line. However, I really like that bit about putting capacitors on the load side of the contactors. I would have to see how that is calculated. It may take somebody with a lot more knowledge of this stuff than me.
I did see on Ronks site, they dicussed that setup, but I didn't really have time to get into it.

Anyhow, I'm really appreciating the input from you guys.

Sent from my SM-G930V using Tapatalk
 
icy78 said:
BTW. How do I quote, or "like" others posts?

Sent from my SM-G930V using Tapatalk
Click to expand...

In the bottom right there is a box that says Reply With Quote, click that instead of Post Reply. You can quote multiple posts by clicking the icon that looks like a speech bubble on any additional post you want and finish with same Reply With Quote on the last post you want to include.
 
ActionDave said:
In the bottom right there is a box that says Reply With Quote, click that instead of Post Reply. You can quote multiple posts by clicking the icon that looks like a speech bubble on any additional post you want and finish with same Reply With Quote on the last post you want to include.
Click to expand...
Thankyou.
I'm on Tapatalk so it must be different. I tap on the post to highlight it and on the top of the page I tap the arrow.
I don't see a way to " like" the post tho. Maybe not an option on this forum?

Sent from my SM-G930V using Tapatalk
 
icy78 said:
Quote......For the four stages of compressors and condenser fans, it is possible to balance them individually as a zone, by moving the capacitors from the RPC, to the load side of the contactor for each individual zone.

Thanks MTW. That's a lot of great info.

I I am concerned about the chiller and its various loads. I was just wanting them to address the pump, if they get that right , likely,the chiller will fall into line. However, I really like that bit about putting capacitors on the load side of the contactors. I would have to see how that is calculated. It may take somebody with a lot more knowledge of this stuff than me.
I did see on Ronks site, they dicussed that setup, but I didn't really have time to get into it.

Anyhow, I'm really appreciating the input from you guys.

Sent from my SM-G930V using Tapatalk
Click to expand...
Problem with balancing the compressor is the load likely varies at times, whenever load varies from the level you balanced it at, the balance will be off again. The fans and pumps are more of a constant load level.
 
kwired said:
Problem with balancing the compressor is the load likely varies at times, whenever load varies from the level you balanced it at, the balance will be off again. The fans and pumps are more of a constant load level.
Click to expand...
Well there's a point too.
Hmmm. Is a CNC RPC what should've been installed?
Something that will automatically add or subtract capacitors based on load, per design?
(At least that's my understanding of it.)

Sent from my SM-G930V using Tapatalk
 
Wow...been in the industry 50 yrs and I have learned alot today. Have several farms in area with 1 P and use RPC. When running < 5 HP motors and running close to FLA for extended tengths of time, we have experenced short life on the motors ( say less than 5 years). I am in fact installing 3 - 7 1/2 grain aeration fans (grain bins) this spring on an exiting RPC system. Currently the RPC is used on a small grinder / mixer mill which has 2 - 7 1/2 motors plus a 3 HP Leg used when unloading truck in over head bin. The RPC and grinder mixer are 30+ yrs old with original motors still going strong, running about 2-3 days /week, 8 hrs per day year around, except during planting and harvest. The RPC is rated to start 10 HP motor and max system HP of 30. They will not be able to run both grinder/ mixer and aeration fans at same time. What problems if any do you RPC experts forcast for our expanded system?
 
icy78 said:
Well there's a point too.
Hmmm. Is a CNC RPC what should've been installed?
Something that will automatically add or subtract capacitors based on load, per design?
(At least that's my understanding of it.)

Sent from my SM-G930V using Tapatalk
Click to expand...
Rotary converter isn't as bad as a "static converter" for the most part, but both typicaly have imbalances from my experiences.

I had someone ask me to look at an air compressor once they were having troubles tripping overload. Someone put a static phase converter on it for them - they bought a used three phase unit. The converter had enough extra capacitors to just move leads around to get different values out of them. I could get reasonable balance- until it started to approach "cut out" pressure. then it was loaded more and off balance again. Don't remember where we left things set, I know we did turn the motor overload up a little more then I would have liked to see - but I guess it is what it is. I told them when that motor burned out that they should just get a three phase motor for it. Something tells me the compressor maybe gets shelled before then anyway - they were operating it without cover on, and it sure looked to me like it needed to be on for cooling air to flow properly through the unit, on top of that it did have an oil cooling coil, but not necessarily getting proper air flow over it.
 
icy78 said:
Well there's a point too.
Hmmm. Is a CNC RPC what should've been installed?
Something that will automatically add or subtract capacitors based on load, per design?
(At least that's my understanding of it.)
Click to expand...

A CNC RPC is normally one that is 2X the size of the load, and voltage balanced within 5%. And best when used with a single CNC, to help maintain that balance under varying load.

There are several methods of constructing and balancing RPC's, some use caps from a single utility line to the manufactured leg, to resonate the voltage up to the required level. Others use caps from both supply legs to the manufactured leg.

Many RPC builders compensate on the high end, +10%, of the line voltage, to the manufactured line, to allow for starting and running the largest motor advertised. The manufactured leg is weak compared to the utility lines, it has a lot more impedance and hence the voltage and current on the generated phase falls off as load increases. This is where it can becomes like partially single phasing. The two utility lines are taking the brunt of the load, and the manufactured line may be doing very little contribution to balance the system.

With small loads placed on a large converter, as a purchased unit, typically the manufactured voltage and current will be high while the utility legs are more normal. In this region you have one phase leg that is running hotter than it should, especially if the utility voltage is in the high range to start with. The closer you are to that +/- 10% specification on the supply, the quicker you can exceed the limit with the manufactured leg.

A single large load on the other hand, can be enough to pull down on the manufactured leg just enough to get the balance near the sweet spot. Where the phase angle, the voltage and currents are as symmetrical as possible. There is only the one sweet spot, per balance job, above and below that point, you move further away from symmetry. Whenever the load varies, the balance point will vary.

If you run a 3Φ RPC idler with minimal balance caps, it will never have a high voltage and current on the manufactured line. Therefore you wont get much power from the manufactured line, but it won't overheat either. Then the question is, in this region, do you have enough power at the motor to serve the load. Oftentimes the motor could be oversized, to compensate for a weak leg and still provide enough load power to satisfy the load.

In the multi motor situation, you need to consider the characteristics of the load, what loads come on and stay on first, then what is the typical way that the loads will be used, and what variation the system will have in use. Only after considering the the real world load application, what would be the best way to go about balancing for a mid point sweet spot.

Switching caps in and out of the RPC circuit while running can be a bad idea, some builders do it, but it can produce some serious voltage spikes in the system during the switching. If you want or need to switch them in and out, it's best if you can do it before you start the system and apply load. One builder spent years designing, building, testing and patenting an automatic solid state switcher control, but discontinued it because in order to suppress the spikes to prevent problems it proved non economical.

That's why I suggested splitting them up with the loads, if that is even possible, depending on system design and function. Switching them with the loads will have lesser spikes when combined with the inductance of the load motor. When placed close to the load, most of the current through the capacitor is reactive, and flows between the cap and the motor windings, not down the supply leads. This would affect the system balance less, when the load was switched in and out. Switching caps with the load will make additional inrush current on the contactor or starter that is supplying it, so that needs to be considered as well.

With the chiller in question, it really depends on what process it serves and how is it designed to run. Does all the fans and pumps run continuously, and the compressors cycle online as required. Or does the entire system cycle on and off as the load controller call for, and the fans are switched with the compressors they serve. It would be a different balancing approach depending on either case.

I have done a utility supplied 3Φ chiller with 4 zones, where each compressor was variable with a VFD for each zone and one for the circulator pump, fans were switched across the line. I would think it could be possible to obtain a manufactured unit with oversize drives that could do the phase conversion as well as the motor control. More expensive on the purchase, but a lot more workable in the field with 1Φ supply.

If you look back at the numbers you gave in the first post, I think you can see what I'm alluding to. But you need to keep track of the numbers in a chart form so that you keep the appropriate values and phases in the proper order, for comparison, a column for unloaded conditions and loaded conditions. Without that you will loose track quickly of what your test condition readings mean.

MTW
 
Yet another option in some cases would be to have a multiple idler system. You switch on more idlers, and parallel them as the system requirements dictate. Another way to keep relative balance with more diverse load conditions. Say a small and a large unit, for little loads to prevent over voltage and excess heating, and a big unit for when you bring on major loads. The utility lines will already be in phase, so long as you wire up the supply phase the same on each, then you just need to get the rotation in sync with phasing of the manufactured leg. These can be switched in and out without too much disturbance on a running system.

Rotary RPC's are the ones that give a bit more variance capability in the output , due to the manufactured phase supplied by the back EMF the motor rotor provides to the unpowered idler winding. But there are many variations of static converters as well, some better than others, and some even better than a RPC. Some designs go back about a hundred years, others are new solid state design that keep the balance tighter than utility supplied 3Φ power. There is a use case for each. It really depends on the load, the design and of course the budget.

PhasePerfect is the solid state design and is also gangable (parallel) for more output than standard units. Nice units that can run any type of load, and stay in complete symmetry under any condition of load, but hit hard in the pocket.

A 60 Year old induction converter design that uses a step up transformer and capacitor bank, can also be effectively used on loads that are steady, and only need one balance point. Many times used for oilfield service, these units can deliver full HP and torque for a fixed load, and require no spinning idler. Just transformer losses until the load is started. ADD-A-PHASE is a manufacturer of these type units.

Then at the bottom of the heap are start capacitor only static converters. The lowest cost unit with just a start cap, cap starts the 3Φ load motor, then switches off the cap and runs it in single phase mode. Reduced HP and torque, but acceptable for small individual loads that start unloaded and can operate the load with less than 2/3 of nameplate. Such as drills and mills, used in a home shop. There are many builders of this low cost/performance units.

One step up are static cap units, that include start and run caps. Both caps used for starting, then the start cap switches off, and just the run caps remain connected to the load. Better than the previous model type, is not running just on 1Φ power, so it has better HP output. But still not generally suited for demanding type loads and full HP output.

And lastly there is the Steelman method, it uses a 12 lead Wye Load motor and re-configures the windings in a different way for operation from 1Φ power, with a start and run cap. With these unit more than nameplate torque can be achieved if set up properly.

MTW
 
Last edited:
MTW said:
Yet another option in some cases would be to have a multiple idler system. You switch on more idlers, and parallel them as the system requirements dictate. Another way to keep relative balance with more diverse load conditions. Say a small and a large unit, for little loads to prevent over voltage and excess heating, and a big unit for when you bring on major loads. The utility lines will already be in phase, so long as you wire up the supply phase the same on each, then you just need to get the rotation in sync with phasing of the manufactured leg. These can be switched in and out without too much disturbance on a running system.

Rotary RPC's are the ones that give a bit more variance capability in the output , due to the manufactured phase supplied by the back EMF the motor rotor provides to the unpowered idler winding. But there are many variations of static converters as well, some better than others, and some even better than a RPC. Some designs go back about a hundred years, others are new solid state design that keep the balance tighter than utility supplied 3Φ power. There is a use case for each. It really depends on the load, the design and of course the budget.

PhasePerfect is the solid state design and is also gangable (parallel) for more output than standard units. Nice units that can run any type of load, and stay in complete symmetry under any condition of load, but hit hard in the pocket.

A 60 Year old induction converter design that uses a step up transformer and capacitor bank, can also be effectively used on loads that are steady, and only need one balance point. Many times used for oilfield service, these units can deliver full HP and torque for a fixed load, and require no spinning idler. Just transformer losses until the load is started. ADD-A-PHASE is a manufacturer of these type units.

Then at the bottom of the heap are start capacitor only static converters. The lowest cost unit with just a start cap, cap starts the 3Φ load motor, then switches off the cap and runs it in single phase mode. Reduced HP and torque, but acceptable for small individual loads that start unloaded and can operate the load with less than 2/3 of nameplate. Such as drills and mills, used in a home shop. There are many builders of this low cost/performance units.

One step up are static cap units, that include start and run caps. Both caps used for starting, then the start cap switches off, and just the run caps remain connected to the load. Better than the previous model type, is not running just on 1Φ power, so it has better HP output. But still not generally suited for demanding type loads and full HP output.

And lastly there is the Steelman method, it uses a 12 lead Wye Load motor and re-configures the windings in a different way for operation from 1Φ power, with a start and run cap. With these unit more than nameplate torque can be achieved if set up properly.

MTW
Click to expand...
MTW said:
A CNC RPC is normally one that is 2X the size of the load, and voltage balanced within 5%. And best when used with a single CNC, to help maintain that balance under varying load.

There are several methods of constructing and balancing RPC's, some use caps from a single utility line to the manufactured leg, to resonate the voltage up to the required level. Others use caps from both supply legs to the manufactured leg.

Many RPC builders compensate on the high end, +10%, of the line voltage, to the manufactured line, to allow for starting and running the largest motor advertised. The manufactured leg is weak compared to the utility lines, it has a lot more impedance and hence the voltage and current on the generated phase falls off as load increases. This is where it can becomes like partially single phasing. The two utility lines are taking the brunt of the load, and the manufactured line may be doing very little contribution to balance the system.

With small loads placed on a large converter, as a purchased unit, typically the manufactured voltage and current will be high while the utility legs are more normal. In this region you have one phase leg that is running hotter than it should, especially if the utility voltage is in the high range to start with. The closer you are to that +/- 10% specification on the supply, the quicker you can exceed the limit with the manufactured leg.

A single large load on the other hand, can be enough to pull down on the manufactured leg just enough to get the balance near the sweet spot. Where the phase angle, the voltage and currents are as symmetrical as possible. There is only the one sweet spot, per balance job, above and below that point, you move further away from symmetry. Whenever the load varies, the balance point will vary.

If you run a 3Φ RPC idler with minimal balance caps, it will never have a high voltage and current on the manufactured line. Therefore you wont get much power from the manufactured line, but it won't overheat either. Then the question is, in this region, do you have enough power at the motor to serve the load. Oftentimes the motor could be oversized, to compensate for a weak leg and still provide enough load power to satisfy the load.

In the multi motor situation, you need to consider the characteristics of the load, what loads come on and stay on first, then what is the typical way that the loads will be used, and what variation the system will have in use. Only after considering the the real world load application, what would be the best way to go about balancing for a mid point sweet spot.

Switching caps in and out of the RPC circuit while running can be a bad idea, some builders do it, but it can produce some serious voltage spikes in the system during the switching. If you want or need to switch them in and out, it's best if you can do it before you start the system and apply load. One builder spent years designing, building, testing and patenting an automatic solid state switcher control, but discontinued it because in order to suppress the spikes to prevent problems it proved non economical.

That's why I suggested splitting them up with the loads, if that is even possible, depending on system design and function. Switching them with the loads will have lesser spikes when combined with the inductance of the load motor. When placed close to the load, most of the current through the capacitor is reactive, and flows between the cap and the motor windings, not down the supply leads. This would affect the system balance less, when the load was switched in and out. Switching caps with the load will make additional inrush current on the contactor or starter that is supplying it, so that needs to be considered as well.

With the chiller in question, it really depends on what process it serves and how is it designed to run. Does all the fans and pumps run continuously, and the compressors cycle online as required. Or does the entire system cycle on and off as the load controller call for, and the fans are switched with the compressors they serve. It would be a different balancing approach depending on either case.

I have done a utility supplied 3Φ chiller with 4 zones, where each compressor was variable with a VFD for each zone and one for the circulator pump, fans were switched across the line. I would think it could be possible to obtain a manufactured unit with oversize drives that could do the phase conversion as well as the motor control. More expensive on the purchase, but a lot more workable in the field with 1Φ supply.

If you look back at the numbers you gave in the first post, I think you can see what I'm alluding to. But you need to keep track of the numbers in a chart form so that you keep the appropriate values and phases in the proper order, for comparison, a column for unloaded conditions and loaded conditions. Without that you will loose track quickly of what your test condition readings mean.

MTW
Click to expand...
MTW, I don't know how to respond to that wealth of information. But allow me to say :
THANKYOU!
For your time and knowledge.


Sent from my SM-G930V using Tapatalk
 
Update....
The factory tech came out to verify voltages and amperage is and give his yay or nay blessing. He had at that time a 6% deviation on voltage. After discussion with the factory they said that that was fine it just can't go to 10%. I don't understand that. Two of the compressor amp readings recorded by him were approximately 30 26 + 14 and 30, 27 + 11. To me that's just insane but the factory said that was okay. Anybody want to comment on that? I'm a little confused by it, especially since the manual spells out 2%. Only thing that I can think of is that they are confident that it will last beyond the one-year warranty and after that it's not their problem anymore. Thats the only thing that makes sense to me. I don't agree with it but it makes sense.

Sent from my SM-G930V using Tapatalk
 
Icy78,

For the part that you cant believe, and don't understand, go back and read post #2. Jeff spelled it out pretty succinctly.

The chiller specifications calls the voltage deviation very tightly to cover their behind, an ideal condition not always achievable in the field, the tight spec makes it your problem if something should go pear shaped.

You would never be able to hold that, with any rotary converter, with the type of variable load you are running. The only converter that could hold that spec in your situation, would be the Phase Perfect mentioned earlier.

As you see, the voltage deviation is the least of the concern. The current deviation is normally multiples of the voltage differences. This is the major problem, the current is what does the real work. The main thing to watch for is that none of the line currents run at above the marked FLA rating, of each respective motor. If the currents remain at or lower than the FLA, then burnout is not too likely. As Jeff mentioned, with non symmetrical currents, you might get some circulating currents producing heat, but that's not usually a major factor.

Then the next consideration is, do you have enough torque to start and run the load with the reduced voltages and currents. It appears that you do, cause it runs OK. Not as strong as it could be if well balanced, but good enough to get er done. The heat of summer will be the acid test to see how well she hold's up to extreme loading. Having hermetic compressors as the main loads is beneficial, as the refrigerant helps keep the temps of the windings in check.

MTW
 
Status
Not open for further replies.
Top