xfmr stiff enough... Elevator soft starts - lights dim

[wyreman]

ok, finally got some progress from the poco ... and the [old thread of the same name] old thread is closed

re: transformer / secondary wire sizing for a big inrush elevator current.


quick background
Elevator soft starts - lights dim every time
156a 208v 3p pretty new elevator - soft start has been optimized. 312a instantaneous inrush
Isee 108kVA instantaneous demand for the elevator inrush alone and 200' of run

175a cb on the elevator disco with less~50' 2/0 Cu

600a service HAD [3] 50kVA TX at 200' run w1000mcm AL
poco replaced them with
3-100KVA banked TX's, they are showing about 150Kva load between them so they're 50% loaded

hope someone here is more knowledgable than me about how 3 round xfmrs at 100kva are half loaded at 150kva.
When 1 big dry 150kva 3ph transformer would be fully loaded
Separate mag fields more efficient than one combined?

In total they've quadrupled the secondary size from the smallest section 1-1/0Cu to a continuous 500Cu run, including a #5 Box to accommodate that. Of course doubling the service run as well from 1-1000Al QPX to include 1-500Cu QPX. Loosely speaking they’ve doubled the size of everything. The only solution that remains is adding a TX right outside the building

I still think the xfmr is too small! But they have doubled everything and say that's it.
Report from the users is that the lights are still dimming, worse than before, but that may just be because they are sensitized to it now.

My question is:
Are those three oil filled round transformers
enough to drive 156 amp three phase elevator
load w a 312 amp inrush current...
Along with the 8 other apartment General use Meters

what about the I R drop over 200'
should this xfmr/secondary combo be able to drive the 156a elevator and keep VD below 5%?
We had 8-10%

I'm going in with a FLIR camera and heat test everything.
I suggested they come in with a 100+kVA load bank and "superbeast" it at both sides of the meter

 

[rlundsrud]

To answer your question about why 3 100 kVa transformers are only half loaded at 150 kVa, 3x100 kVa= 300 kVa. If the total load is only 150 kVa then it is at 50%.

As to your other questions, perhaps you could provide a one line diagram. Include conductor sizes and voltage where it is transformed. Also, do you have any load calcs for the services that are also on this circuit. Could you also include the manufacturer specs for minimum circuit ampacity on the elevator.

I would suspect that the 2/0 conductors might be two small and that is causing your issue, but that borders on a WAG.

 

[wyreman]

To answer your question about why 3 100 kVa transformers are only half loaded at 150 kVa, 3x100 kVa= 300 kVa. If the total load is only 150 kVa then it is at 50%.

.

so then a typical Single
Three Phase Transformer
rated 150kVA is good for 150kVA


but a bank of [3] Single Phase Transformers
is additive

so a Single Three Phase Transformer
rated 150kVA is
really like a bank of [3] 50kVA Single Phase Transformers

thanks, that's kind of what I inferred, but need to read up on why.


the other questions are mostly shown in the photo journal

https://goo.gl/photos/7WyN66YHxiYCnmHh6


Utilize 80% of existing xfmr
The transformer is equal to or less than its rated kVA and rated voltage.
The average temperature of the cooling air during a 24-hour period is 86 F.
The temperature of the cooling air at no time exceeds 104 F.

 

[wyreman]

even if the elevator feeder conductors were undersize

how would that dim the building distrib
since that is before the feeder

plus we never clamp more than 300a VD at the main?

The only place i can think of to IR gun [or push and pull]
for the high impedance connection would be at the main buss, after the meter but before the elevator feeder

the whole building dims 7-10% when the elevator goes up.
300A flow at the meter- same at the elevator

600a 208v service with the wires as described above

 

[wyreman]

i uploaded a crude single line to the photo set

 

[MyCleveland]

Never seen an elevator in a eight suite building.
I would start looking at whether the apartment feeders were sized properly for voltage drop.

When sizing the service, assuming you took advantage of all the derating calc's available, you may be left with a design where your house load is close to half of the service size.

If close to half your load is one large motor....then you are going to experience a voltage dip on starting.

 

[wyreman]

Never seen an elevator in a eight suite building.
I would start looking at whether the apartment feeders were sized properly for voltage drop.

When sizing the service, assuming you took advantage of all the derating calc's available, you may be left with a design where your house load is close to half of the service size.

If close to half your load is one large motor....then you are going to experience a voltage dip on starting.

i remember doing the bldg calcs and came up w 565a for the bldg including the big motor - have to go back and check -
great call that the feeders to the units might be have the ir losses.

but they really should be very lightly loaded - i have some datalogging graphs here that show the voltage dip coming when the main is flowing 300a
- pretty much just the inrush thats all there is-
but who knows, there might be some elec heaters running now!

great point, i will have to check that!

 

[gar]

161228-2054 EST

When dealing with linear components absolute load has no effect on the magnitude of flicker when a load change occurs.

A 50 % change in load where % is in terms of full rating, then you get essentially the same change in voltage from the load change whether the initial steady load was 0 or 100 %.

You need to measure the change in voltage to the lights (I assume they are incandescent) from just before the elevator starts to during the starting time. After acceleration of the elevator and motor the voltage drop will be less. This may be short enough in time that you need to use a Fluke in min-max mode.

Make the measurement at bulbs that typically show flicker, and at ( I really mean on the wires ) the hot to neutral at the main disconnect. You could also do this on the output side of the main breaker.

Using 120 V bulbs ( a 15 W incandescent, a Cree 9.5 W, and a GE 13 W CFL ) and a voltage change at the bulb of about 4 V, then:
(1) I see a visible change but not objectionable with the indandescent. In a moderately controlled experiment I can detect about an 0.5 V change with an incandescent.
(2) With the Cree I detect a short flick, but possibly not a sustained intensity change.
(3) Same as the Cree with the GE.

I did not try to measure light intensity. These were just human reactions.

In your whole chain of components I don't know how much of the impedance is common to lights and the elevator. To minimize this common impedance the lights should come from their own main panel that does not include the elevator's main panel disconnect. This would give you the least common impedance.

.

 

[Bugman1400]

I would rather see the voltage waveforms at the bus instead of at the incand bulb. Can you post those? Please post voltage, current, and angle between. This sounds like a typical flicker problem that can be solved in a variety of ways but, I would check the settings on the Siemens controller first. Otherwise, it would be best to know what the volt dip is and for how long.

Can you list the existing settings in the controller?

 

[wyreman]

?Give me a couple days for the controller settings
Remember it's 200% , don't recollect the ramp interval
We tried default, quickest and slowest (without dogging)
Think we settled on default

I have data logging charts I can dig up
Mostly shows all good but voltage dogs when amps bump just over 300
It had a bad scale for the amps
Where you couldn't see it vary too much

 

[Bugman1400]

It would be best to see waveforms that have a fast sampling rate. It would also be good to see what the source impedance (R+jX) is at the bus. If the bus is weak, there may be little you can do until the incoming source is stiffened. An option is to add capacitors to help stiffen the bus; however, caps wouldn't do anything if the majority of the weakness is from resistance-based impedance (R). Knowing the bus reactance (jX) will help determine this and how much capacitance should be added to help the flicker.

 

[NewtonLaw]

[Partial/QUOTE=wyreman;1793304]ok, finally got some progress from the poco ... and the [old thread of the same name] old thread is closed

re: transformer / secondary wire sizing for a big inrush elevator current.

My question is:
Are those three oil filled round transformers
enough to drive 156 amp three phase elevator
load w a 312 amp inrush current...
Along with the 8 other apartment General use Meters

what about the I R drop over 200'
should this xfmr/secondary combo be able to drive the 156a elevator and keep VD below 5%?
We had 8-10%
[Partial/QUOTE]

I tried to do a couple of back of the envelope calculations based on what I see you posted as the approximate one-line diagram. I made the following assumptions: 1) The transformer bank used transformers with 2.5% impedance (Typical impedance for this size distribution transformer) and then again with a transformer impedance of 5.0%
2) I assumed a 208Y/120 grounded Wye service supplied by an infinite primary bus.

Based on your one line we have transformer bank impedance, service conductor impedance (200' of 1000 MCM Al in parallel with 200' of 500 MCM CU) in air, 50' of #2/0 CU in conduit.

I calculated the final impedance at the end of the 50' run of #2/0 CU. For the 2.5% transformers I get %Z = 7.63. For the 5.% transformers I get %Z = 9.9%. Since I assumed all the calculations on the secondary side of the transformers I used V-base as 208 volts, kVA-base as 300 kVA, Z-base as 14.42 ohms and I-base as 832.

Converting the current inrush of 312 amps to per unit I get I-pu = 312/832 = 0.375 amps. Based on this current flow through the above two values I got for impedances to the motor I get 2.9% voltage drop for the 2.5% transformers and 3.7% voltage drop for the 5.0% transformers. This does not take into account the phase angles which result in an X/R ratio of 1.45, nor the power factor at the motor during start-up which is typically 18 to 35%.

So these voltage drops would be worst case conditions and I would expect actuals to be less. I could not come up with the 8 to 10% voltage drop you speak of in your first post. Was this a measured voltage drop at the motor terminals or soft start terminals? Are you sure you do not have a connection problem?

The 3 100 kVA transformers should be more than adequate and unless they are the customer's responsibility, the utility should have done all the necessary calculations to ensure the voltage drop to your service point of connection would be within the State Utility Commissions regulations, usually +/- 5% for residential service and +/- 10% for commercial/Industrial service.

I just do not see a problem with your service unless there is a really bad connection somewhere. The IR camera may not find anything due to the thermal lag time of the wire/connectors. You would need to place load on it for about 7 minutes, 15 minutes would be better. Last, make a resistance check across each connection using a low resistance ohmmeter or an alternative would be to measure the voltage drops across each connection supplying the motor during start-up or place a load across each connection looking for the same high voltage drop.

Hope this helps,

Newton Law

 

[Bugman1400]

[Partial/QUOTE=wyreman;1793304]ok, finally got some progress from the poco ... and the [old thread of the same name] old thread is closed

re: transformer / secondary wire sizing for a big inrush elevator current.

My question is:
Are those three oil filled round transformers
enough to drive 156 amp three phase elevator
load w a 312 amp inrush current...
Along with the 8 other apartment General use Meters

what about the I R drop over 200'
should this xfmr/secondary combo be able to drive the 156a elevator and keep VD below 5%?
We had 8-10%
[Partial/QUOTE]

I tried to do a couple of back of the envelope calculations based on what I see you posted as the approximate one-line diagram. I made the following assumptions: 1) The transformer bank used transformers with 2.5% impedance (Typical impedance for this size distribution transformer) and then again with a transformer impedance of 5.0%
2) I assumed a 208Y/120 grounded Wye service supplied by an infinite primary bus.

Based on your one line we have transformer bank impedance, service conductor impedance (200' of 1000 MCM Al in parallel with 200' of 500 MCM CU) in air, 50' of #2/0 CU in conduit.

I calculated the final impedance at the end of the 50' run of #2/0 CU. For the 2.5% transformers I get %Z = 7.63. For the 5.% transformers I get %Z = 9.9%. Since I assumed all the calculations on the secondary side of the transformers I used V-base as 208 volts, kVA-base as 300 kVA, Z-base as 14.42 ohms and I-base as 832.

Converting the current inrush of 312 amps to per unit I get I-pu = 312/832 = 0.375 amps. Based on this current flow through the above two values I got for impedances to the motor I get 2.9% voltage drop for the 2.5% transformers and 3.7% voltage drop for the 5.0% transformers. This does not take into account the phase angles which result in an X/R ratio of 1.45, nor the power factor at the motor during start-up which is typically 18 to 35%.

So these voltage drops would be worst case conditions and I would expect actuals to be less. I could not come up with the 8 to 10% voltage drop you speak of in your first post. Was this a measured voltage drop at the motor terminals or soft start terminals? Are you sure you do not have a connection problem?

The 3 100 kVA transformers should be more than adequate and unless they are the customer's responsibility, the utility should have done all the necessary calculations to ensure the voltage drop to your service point of connection would be within the State Utility Commissions regulations, usually +/- 5% for residential service and +/- 10% for commercial/Industrial service.

I just do not see a problem with your service unless there is a really bad connection somewhere. The IR camera may not find anything due to the thermal lag time of the wire/connectors. You would need to place load on it for about 7 minutes, 15 minutes would be better. Last, make a resistance check across each connection using a low resistance ohmmeter or an alternative would be to measure the voltage drops across each connection supplying the motor during start-up or place a load across each connection looking for the same high voltage drop.

Hope this helps,

Newton Law

Most of the posts on this thread seem to concentrate on steady state parameters (the design). The design could be correct but, if the source is weak, you'll have issues like this. So, I think it is incorrect to assume an infinite bus. To find the solution of how to fix this, I think more data is needed. Perhaps a flicker calc would also help.

 

[NewtonLaw]

Most of the posts on this thread seem to concentrate on steady state parameters (the design). The design could be correct but, if the source is weak, you'll have issues like this. So, I think it is incorrect to assume an infinite bus. To find the solution of how to fix this, I think more data is needed. Perhaps a flicker calc would also help.

I agree that this is possible. An easy solution to this is to request a voltage survey be accomplished by the supplying Utility since they are responsible for ensuring that the voltage drop due to load is within the regulatory limits. In PA the +/-5% or =/-10% based on Residential (or Office space) or Commercial/Industrial.

 

[wyreman]

I agree that this is possible. An easy solution to this is to request a voltage survey be accomplished by the supplying Utility since they are responsible for ensuring that the voltage drop due to load is within the regulatory limits. In PA the +/-5% or =/-10% based on Residential (or Office space) or Commercial/Industrial.

that's how i got them to upgrade the xfmrs from [3] 50kva to [3] 100kva and double the size of the secondary...

they are coming back with some powersight eq to see if that gave them compliant readings at the meter
they are hoping that the people have just been sensitized to the issue
that why they complain
and that they are under 5% and compliant! hmy:

a second issue... oh i'll start another little thread

 

[NewtonLaw]

You have a partial answer as to where to look ...

You have a partial answer as to where to look ...

even if the elevator feeder conductors were undersize

how would that dim the building distrib
since that is before the feeder

plus we never clamp more than 300a VD at the main?

The only place i can think of to IR gun [or push and pull]
for the high impedance connection would be at the main buss, after the meter but before the elevator feeder

the whole building dims 7-10% when the elevator goes up.
300A flow at the meter- same at the elevator

600a 208v service with the wires as described above

From my previous post on the thread that is now closed, and based on your starting amps of 312 amps inrush during the elevator startup, I did not get the 7 to 10% voltage drop. I calculated a much lower value closer to about 3% at the elevator. You say the entire building sees about a 7 to 10% voltage drop during start up, is that correct? If so, look at the point of common service for a bad connection, corroded connection or wire. If the entire building sees the voltage drop, the problem has to be at or ahead of the common point that serves all the building load.

Looking at the photos you supplied, what I see as common is the following, moving back from the Elevator Shunt Trip Breaker, I see a tap from the House main and mouse meter, then to the bus bars common to an 8 meter panel, then bus from the meter panel to the service disconnect, then cable to the Utility company's meter and point of connection. At this point, I assume you pick up the service conductors (1000 MCM AL & 500 MCM CU) back to the 300 kVA transformer bank) which may have a lose connection point, splice, etc.

If you choose to use the load bank you spoke of, I would remove any and all panel covers, connect the load bank at or ahead of the Elevator Shunt Trip Breaker, and load the system. Take voltage readings before turning on the load bank. Turn on the load bank, read the currents and voltages (Include all three phases and neutral). Allow it to run for the 7 to 15 minutes I suggested and then use your Infrared camera to identify any hot spots following the same chain I suggested above back to the transformer terminals.

This should locate the common problem. At the least it will give you a measured steady state voltage drop at the top of the Elevator Shunt Trip Breaker. I would be looking for heating between the 8 gang meter panel to the 600 amp disconnect continuing back to the Utility company transformers.

I hope this helps,

Newton Law

 

[wyreman]

Based on your one line we have transformer bank impedance, service conductor impedance (200' of 1000 MCM Al in parallel with 200' of 500 MCM CU) in air, 50' of #2/0 CU in conduit.

... Are you sure you do not have a connection problem?

The 3 100 kVA transformers should be more than adequate...

I just do not see a problem with your service unless there is a really bad connection somewhere.

Hope this helps,

Newton Law

first of all,
thank you for your well considered reply

the only assumption that i must not have explained well enough - i believe the entire secondary run is 200' of parallel conductors in conduit - going thru a number of manholes, presumably with splices in each.
200' each of the 1000mcm and the 2/0Cu

further, i believe the voltage drop shows up before the meter - after all, thats where he measures, thats all he cares about.

that's Why i can't see how poco can even have the idea that the problem is on my side of the meter.
As you clearly stated, the trouble must be at a common point before the high impedance connection.

They replaced the source and the long run... but the issue -seems to me- must be a bad splice in the 1000mcm run.
Thats all thats left of the old source
But that also doesnt make sense because doubling up with the new parallel run of 2/0s should not have made it worse!
unless we have much higher aggregate draw => each unit has electric heat.

I got to go back and take some readings, see the new values.

Keep an accessible log
12 columns ABC&N

1before meter
2load side of elevator shunt
3load side of elevator disco​

time for the rows

see if there are any +5% moves any more

the 8 meters are probably heavier loaded now because I seem to recall that each unit had electric heat, they may be 100a subfeeds.
Its been cold here. Total bldg load may have gone way up.

 

[wyreman]

hot work permit

hot work permit

going back today with [2] sets of powersight 3000 loggers and some low resistance ohm meters
I'm going to comb thru the buss over the next few days looking for the high impedance connection on my side of the meter, or proving it is on poco side
If I can't find the loose bolt on the buss, I'll have to load it to prove the bad connection is on poco side

Checking the size of feeders to the units is also on the menu, because the complaints of dimming are from the units

I was about to bring in a Mosebach 100kw loadbank w 5kw $tep$.
I do have some 480v 30kw heaters available, and could make something easily from them
A heater like that, run instantaneously at 208 after a 480v 100a disco, should be a decent substitute.
only concern is the fans must also be 480v

I'll start with the main closed in, all distribution off
Keeping one set of the PS3000 on the utility wires and one set on the line side of the test point, I'll bump it at: the elevator breaker, the elevator disco if needed.

• general flir
• visual inspection
• comb thru buss at meter for high impedance connections
• compare voltages ,under normal loading, at points of interest on customer side of meter - compare these to the PS300 parked on poco side
• bring in 30kw 480v heaters at ~85a/ea @208 as needed to get an instantaneous dip to compare
• min permissable voltage 208 = 197

 

[Sahib]

Wyreman: Per your statement the POCO transformer is 50% loaded already. So to overload it your building steady load should be greater than 50% ie more than 150 KVA, which may be ascertained from service meter readings.

 

[Sahib]

Wyreman: As others told, the problem is due to loose connection. And not due to overloaded POCO transformer. Check voltage drop at service disco during elevator operation. If it is excessive, fault is upstream. Otherwise, downstream.