Could someone give me a quick overview of what to be on the lookout for when a transformer is used as a wye-delta step up transformer?
Background: I am a customer of a PCB fabrication house that has a large solder reflow oven. While I was there the oven kept tripping out seemingly do to power quality issues, and the owner said that this happens frequently. I was not there as an electrician, and it would be inappropriate for me to attempt to do any work on this equipment or in this facility. But as a favor to the owner, I would like to understand what to be on the lookout for, so that I can explain it to him and he can ask the right questions of his EC.
I noticed that the feed to this oven was supplied by a dry type transformer (didn't catch the KVA rating, looked to be at least 100KVA) of the sort normally used as a step down transformer. However the transformer is being used step-up, to supply 480V to the oven. The primary is thus wye connected and the secondary is delta connected. I do not know if the primary neutral is connected to the supply neutral or ground.
Given this state of affairs, what are the issues?
My short list: The ungrounded secondary can cause problems for neutral connected transient voltage suppression, especially if there is capacitive neutral current from an adjustable speed drive. If there are ungrounded control transformers internal to the apparatus, they may require a grounded supply for safety (there is what seems to be a 120V computer system running the oven...possibly from a separate supply?)
I don't believe that ground fault detection is required for 480V ungrounded delta supplies...but would it be a good idea? (Duh.)
(Note: I posted this question to electrialmatters...silence )
Dispelling some vicious rumours, regardless of which way power flows through it, by ANSI convention the higher-voltage windings remain with “H” terminals and lower-voltage windings remain having “X” terminal designations.
A problem with 480V ungrounded-delta systems can be potentially disruptive ground-overvoltage occurrences. If 480Y/277V service is needed, there are 208∆ - 480Y/277 drytypes available in 15-750kVA that are intended for stepup service. As opposed to solidly-grounded 480Y, a high-resistance-ground system may also be employed, but it may not serve and ø-n loads, and temporary excursion to 480V on unfaulted phases must not cause undesirable operation in the served equipment. It is crucial that surge-protective device ratings accommodate neutral shift that may occur for insulation degradation at any place in the local 480V system. Remember that transient or resonant ground overvoltages can cause temporary misoperation and sustained insulation damage if allowed to persist. Ther is a very good reason why 480Y/277V systems have been recommended in the US since 1955.
[Note in ∆Y-stepup operation—delta “X” terminals and wye “H” terminals.]
If a 208Y winding is used on the transformer ‘source’ side as with a more common 3ø drytype, the XO terminal is usually best left unconnected to minimize undesirable neutral circulating current.
[This message has been edited by Bjarney (edited 03-10-2004).]
Backfeeding transformers is very common. I know of no manufacturer that advises against the practice. The commercially available step-up transformers exist because of the need to create 480Y/277V systems.
Some additional notes, some one else has put together, on backfeeding transformers.
Backfeeding is not recommended for transformers smaller than 3kVA, and should not be done on a control transformers of any size, because windings are compensated. Backfeeding will result in lower than expected output voltage.
Backfeeding causes very high excitation inrush, making coordination to breakers or fuses difficult. Avoid backfeeding wherever possible for this reason.
If a Delta-Wye transformer is to be backfed so that the Wye side is the input, DO NOT connect the neutral terminal (X0)to the primary system neutral, nor should the neutral terminal be connected to ground.
When backfed, transformer taps will not help compensate for poor (other than nominal) source voltage. They do not provide the appropriate magnetic flux levels in the core as they are designed to do.
I will pass this information on when I bring in the next batch of PCBs.
My bet: a slight bit of leakage (and there is always something) is causing a voltage excursion up to the full 480V, causing some protection relay intended for 277V to trip out. The main load is presumably line to line resistive heaters, with a control transformer or two providing 120V or 240V.
Winnie, could you elaborate on “…protection relay intended for 277V”?
If there is anything that needs a ‘neutral’ in the 480V circuit, you are out of luck with the 480V 3-wire ∆ winding. There is no “277V” available with the described configuration, as would be for ordinary 480Y/277V 4-wire service.
By 'protection relay intended for 277V' I mean _ground referenced_ overvoltage protection rather than a _neutral referenced_ load served.
I encountered this with an inverter based VSD that we temporarily set up in our lab. There was no neutral supply terminal, only the three hot supply terminals and a grounding terminal. The loads were all line to line loads (a 480V to 120V control transformer, and a line to line three phase rectifier supplying the DC bus). However there was a set of MOVs intended for transient voltage suppression which were connected wye to ground.
With no detailed knowledge of the oven, it would not surprise me if all the loads in the over were line to line loads, but that there were some _ground_ referenced, normally high impedance, overvoltage protection. Extremely small leakage currents would cause the line to ground voltages to float anywhere from 0 to 480V, rather than remaining solidly at 277V.
I can imagine several scenarios where there would be an 'impedance derived neutral' inside the oven, where things would work most of the time, but where transient problems could show up. For example, if the oven heaters were 277V and wye connected, then the heater neutral would pretty well be at 277V to any phase conductor...until that heater contacter opened or closed, especially if the poles didn't actuate at exactly the same time. I could envision a system that was designed for a 480/277V 4 wire supply, where the loads were well enough balanced that it would work (although imperfectly) with a 480V 3 wire supply.
Winnie, it sounds like you have the situation under control. You are not "mixing" 277V components with an ungroudned 480V system.
To give an idea of how crazy this ground-transient overvoltage stuff gets, Allen-Bradley recommends as an alternative to THREE ø-g MOVs, using FOUR line-voltage MOVs all tied together on one side, with three terminals in turn connected to three phases, and the fourth MOV lead connected to ground. This implies the MOV set is intended to conduct at the expected double ø-ø clamping voltage, BUT ALSO, double the ø-g clamping voltage.
Although in this case the ungrounded-delta system in quite small, one needs to be well aware of the peculiarities of ungrounded systems.
I was on a job several years ago where the poco 3 phase 120/208 service was stepped up to 480v. the foreman had the engineer explain what he was supposed to do with the neutral and the xo connection. He was told to leave it alone and not to land the neutral. The next day the ground was smoking from a red hot bonding electrode he installed from the xfmr to the building service. the xfmr winding laminations were dripping off the coils. during the resulting finger pointing. the poco stated that the neutral not being landed on the xo and bonded to earth ground had been the cause. Any comments would be greatly appreciated.