I have a problem that, for now, has me stumped.

The building is 3 years old; 5 years before, it was a patch of dirt. There has been extensive utility work done in that area. The building is served by it's own pad-mounted transformer, 208 3 phase.

The building itself has 6 tenant meters, and a 'house' meter. My problem is with circuits served by the house meter; other areas seem to not have the problem I am addressing.

While the other services have 3 phase HVAC loads, the house meter only serves nine single-phase 120 circuits. Five of these are lighting circuits, and one is the lighting controls.

My problem? I an losing electronic HID ballasts. 17 of 39 so far. The ballast maker tells me they look as if they received some sort of voltage spike. Lighting circuits are loaded to about 8 amps each.

I have also lost a photocell - 25 amp rated, controlling about 8 amps of HID pole lights.

Finally, I had a new Intermatic electronic timer, with battery back-up, fail several times to keep time. That is, it would suddenly drop an hour or two. This timer works with a photocell to control the two lighting contactors. These controls have their own neutral to the panel.

All panel connections are tight.

None of the tenant spaces are losing fluorescent ballasts, or experiencing light flicker.

Any thoughts?

Check the Neutral, could be the POCO neutral at fault here or high resistance.

Problems with voltages will occur when the total load from the TX becomes unbalanced.

Fit a datalogger on the circuit affected which can record spikes.

Thank you. PoCo has added a data logger today.

I'm interested to see what the data logger picks up. Please keep us posted Reno...

Seeing as how you're burning out control devices, I would suggest putting a contactor in the control circuit. Electronic timers can be pretty pricey (intermatics, anyway) and the 25A photcells are about $22 too, so you can save yourself some service calls and waranty work by taking the HID load off of those items. I go through a lot of 25A cells whenever they are running without a contactor, especially on ballasted fixtures.

I want to say harmonics sound like a possible culprit, based on the HID factor, but the logger should bring the problem to the surface.

Good Luck!

Zapped ,,, I have a pair of contactors. The 'control circuit" has the photocell, the timer, and the pontactor coils on it. Period. It also has its' own dedicated neutral.

I'm also losing ballasts ... nearly half of them ... $230 each, electronic HID ballasts only 3 years old. Now, these are spread over several circuits, and I agree that a loose neutral connection might cause voltage surges. Yet, even in those instances, the loads are really well balanced - again, mitigating the effect of a loose neutral. Overall, the sparky who did the place did a very good job.

Now, the pole lights are on their own photocell, and no contactor ... but a 25 amp rated photocell ought to be plenty for 800 watts of MH!

So, all I can do is wait and see what the datalogger shows - though I might take apart some of the switchgear, just to see if I can spot anything not 'right.'

Reno,
Another source of spikes is an occasional lightning storm in the area. You mentioned parking lot lights. I have seen damage in parking lot fixtures and also several hundred feet inside buildings when poles attract the big sparks. That is why some engineers require a ground rod in the concrete base of metal parking lot poles.

Another possibility could be faulty underground 12.5kv utility wiring coupling to parking lot wiring.

Have you checked the impedance of the ground and neutral for the house panel? Ideal makes a fairly inexpensive tester ($200) that can be used to test branch circuit and feeder ground impedance.

On the subject of switching those outside ballasts, in my experience a contactor or heavy duty lighting relay seems to work more reliably than a photocell only. If the photocell is not conducting the full sine wave, there could be harmonics created by the photocell. If the photocell has a relay, the contacts are usually pretty small. Some photocells use solid state relays which could be source of your problem if they are acting like a dimmer-only conducting for a percentage of the sine wave.

And a final suggestion, what is the voltage drop on the circuits at the actual load?

Why not try a non electronic ballast the next replacement. Maybe those newfangled ballasts are really the problem?

I'm sure you have thought of this but just in case: If there are no complaints of a similar nature from the tenants (None of the tenant spaces are losing fluorescent ballasts, or experiencing light flicker.), then I would suspect a poor connection somewhere between the final point of connection of the transformer and the house panel. If it were a utility problem one would think it would affect the tenants as well??

You should also keep in mind that many, if not most, spikes can be generated within the facility. Motor loads, especially refrigeration/cooling seem to be big offenders. If the problem does end up being spikes, you'll likely end up putting some protection on the house panel, but I wouldn't spend a nickel until you see the results from the POCO.

Do the tenant spaces have the same type of lighting & ballasts? You mentioned the ballasts were for HID fixtures, which I will assume means something other than fluorescent. If that is the case, I don't feel that the technology for electronic HID ballasts is anywhere near being mature. Early on in the T5HO technology rollout, I ended up replacing 90% of the ballasts in one installation.

Also, if all the ballasts were from one lot, there is the possibility of a manufacturing defect that the manufacturer may not have detected as yet.

I'm definitely interested in seeing the outcome, keep us posted.

Since Electronic / SMPS Ballasts for HID Lighting are relatively new to the Industry, I would expect some early failure issues.

The Ballasts _SHOULD_ have some sort of Factory Installed TVSS Protection on the Line Input side of the Ballast. The level of TVSS may not be as beefy as needed, or might be built too sensitive to TVS/S.

Just before the Line "PI" Filter - if installed, there should be - at the least - two MOVs.

Just out of curiousity, where any discrete components visually "fried"?
If so, what components lost smoke, or where are their locations on the Ballast's PCB?

The source of the Spikes / Surges would be from something generating trash back into the system, at "Off Hours" - i.e., when it is dark, and the Site Lighting is active.
Could be late night, could be early AM (pre-dawn).

Things to consider:

1: Are there any other Customers fed from the same Transformer's Secondary, besides the ones fed from this Service Section?

Other Customers may be creating issues with their equipment, and doing so during off hours.

2: Are there any Customers fed from this Service Section, which might be working late night (dusk to maybe 12:00 AM), or early morning (2:00 AM to dawn)?
They might be running noisy equipment, causing the TVS/S issues.

3: Might be some Tenants on this Service Section with noisy HVAC equipment, which automatically starts in pre-dawn hours.

4: Utility may have a Power factor Correcting Capacitor bank, which "Cuts-Out" at an off peak time (like 8:00 PM), or "Cuts-In" prior to peak time (like 4:00 AM), and when the Cap's are Cut-In or Cut-Out, creates TVS/S issues, or an overvoltage scenario.

5: May be heavy use, 24 Hour shift Customers on the same Primary Circuit, which are creating noise issues during night time hours.

Verify the time frame & duration of the noise (Spikes / Surges) with the Datalogger, to determine if there is an issue with TVS/S (Transient Voltage Spikes/Surges).

If there is an issue with TVS/S, may need to fit the House Panelboard with an SPD Panel, or query the Ballast Manufacturer for Point-Of-Use TVSS devices.

BTW, who made the Ballasts, and what are the ratings?

If the costs and maintenance factor prohibit an extensive TVSS application, the Fixtures may need to be "Retro Engineered" back to Reactor based Ballasts (Magnetic Ballasts).
Could use the "F-Can" enclosure types, or the "Pole-Based" round ones.

Will see what arises from your next reply(s).

Scott

I can only provide partial answers.

The ballasts are made by Advance, and many were sent back to Advance for examination.

It is Advance that tells me that components failed on the input side, suggesting exposure to high voltage.

Ballasts are rated for the nominal voltage of the circuit; that is, a 120v rating on a circuit that measures 124 - 125v.

The pad mounted transformer serves seven customers, all in the same building. 4 are retail spaces (mini mart, fast food, butcher shop, novelty shop), 2 offices, and the house meter. Each tenant has their own HVAC (house does not have HVAC).

All problems so far seen are on circuits served by the house meter, and on the lighting circuits. 10 circuits, and seven involve the outside lights - leaving the alarm panel, and two I can't recall offhand.

Any chance the taps on your transformer need bumped down a step or two? I wonder if the peak voltage is excessive in the wee hours of the morning or if there are automatic tap changers that effect your area.
Joe

My guess is these Ballasts are being spiked / surged, from one (or more) of the Tenants connected to the same Service Section / Transformer.

The noise is bursting outwards, and affecting the HID Ballasts at Night when the Site / Parameter / Area Lighting is on - which is, of course, fed from the House Panel, through the House Meter, which is tapped from the Meter Section, common to the Tenants'.

As stated before, the "Electronic HID Ballasts" are relatively new in the Industry, and are still going through a kind of "Public Sector, In-Use Beta Test" thing, in order to "Find The Bugs".
Expect a more stable version in a few more years.

The same was experienced when Electronic Fluorescent Ballasts for Linear fluorescent Lamps, were first introduced to the Industry.

Failure rates were high in the first few years (some were higher than ever expected!).
After time, the Manufacturers gathered failure data, tried newer approaches, and eventually came up with stable versions used today.

If the failures are not _DIRECTLY_ resulting from TVS/S generated by the Tenants, or other Customers, the problems may be from the operation of the Lamp, applied to an erroneous Ballast design - or due to unstable Line Voltages.

To be specific;
* What are the Lamps' Wattage + Type?
I kind of think they are 55 Watt Metal Halide, Pulse Start Lamps.

* Is the Power Factor > 0.95?
If you are not sure, just list the Line Input Amps, and the Lamp's rated Wattage, for a given Ballast.

* The Measured 124-125VAC: was this taken at daytime, or night; and were the Lights on or off; and was the test done at the first fixture, last fixture, or at the Panelboard?

I am really curious about this scenario! Want to find out what is going on, not only because I want to help get this problem under control, but also because we have a few Clients requesting HIDs driven by Electronic Ballastry.

I have been explaining the "Learning Curve" for failure rates experienced with newer Technologies, and suggesting them to wait a few more years for the Technology to improve, so their investments are worthy - not a significant loss from maintenance nightmares due to large failure rates.

Lastly, the issues of failure may be from the Lamp's operation, and / or the Ballast's counter-action with the Lamp.

Simply, the Lamp may be causing the Ballast to "Push-Back" Spikes and/or Surges of Voltage - either as high levels of Harmonic Distortion (and the "Line Input" components are suffering from this noise), or the overall Power Factor is leading too much (< 0.7), resulting in a High Voltage being "Pushed-Out" of the Ballast, which damages the Line Input Components.

Please let me know what Lamps are used, the Wattage, and the Ballast information, so I can research this.

Also, if you have the opportunity to open one or two damaged ballasts, get some pictures of the fried components, plus mark down the Identifier marked on the PCB / Component; then post the information / pictures here.
Alternately, just make note of the smoked Components (ID, location on board, description of the Component, etc.), and post the info. here.

Good luck.

Scott

Just curious if there is any update info?? Did the Data Logger reveal anything?

I have not received any feedback from the PoCo ... but I have continued to have problems with one 'fancy' thermostat. While one would expect the 120/24v transformer to have protected the thermostat, at this point I'm not ruling anything out.

Failures of other components seem to have ended. The replacement ballasts, photocell, and timer are doing well.

Thanks for letting us know up to the date with the info what is going on there.

I am tempting to suggest the idea about the low voltage set up I don't know if it worth the money or not but get auto reguated transfomer it may able cure up some of the surges.

Merci,Marc

Absent either proof of a regular problem ... or further equipment failures ... I am tempted to let it all slide.

If I see other problems arise, my plan is to instal some manner of surge suppressor at the panel. I note that they are available in all manner of sizes ... any suggestions as to how to size them?

what size the exsting service is there now ? and yeah there are few diffrent size you can use some you will actally use in place of breaker space if have room otherwise a remote mount will work.

Merci,Marc