strobing lights..................

[Toneyz]

Can anyone help me on this at a Hyundai showroom/garage a couple of years ago in the workshop Havells Sylvania 4 tube HF Sylrack T8 fittings with 80w high output tube ( these are correct) have been fitted. The tubes are wired internally 2 tubes from one h/f electronic ballast both internal two then the outer two on another ballast. The problem was on two of the fittings the two internal tubes kept on flashing we thought that being they are on all day and installed over a year new tubes so got well-known brand and replaced, a couple of days later started again got in touch with Sylvania who replaced both F.O.C. now another fitting is doing the same with replaced tubes fitted in this in a strobing effect always mid-afternoon.Any ideas

 

[westward10]

Have Sylvania taken the old fittings to check them out. They only do this in the afternoon but come on fine in the morning?

 

[Toneyz]

All ok until 2.30, 3, 4 pm then the internal two strobe flash fast until turned on and back on again so I'm told. Sylvania came back and said could not find anything wrong only could be tubes but in two fittings and replacement tubes fitted.

 

[marconi]

This could be something very simple or more difficult to fathom. I can only offer the following advice at the moment. By strobing I assume you mean going on/off completely or do you mean changing rapidly in brightness but not actually going off completely?

1. Attend in the afternoon and find out what is happening in the garage and what electrical equipment is being started and stopped - car lifts, power tools, high pressure washers. Using an analogue voltmeter, observe the L-N voltage at the flashing fittings. Compare this voltage to L-N at the distribution board for that phase. Electronic ballasts don't operate well if mains is down by 10% or more.

2. Find out whether these fittings are at the end of the feed and how the fittings are connected to the supply eg: daisy chained, starred, ring, comb, ...

3. HF florries must be solidly connected to Earth and within the fitting the ballasts must have a very good Galvanic connection to the metal casing. In addition the metal reflector or cover immediately above the tubes must have a very good Galvanic connection to the metal casing of the fitting. You may have to use some emery paper to ensure that paint is not reducing the conductivity of these connections and that screws have 'star' washers to cut into the metalwork. Take R1+R2 measurements of the strobing and non-strobing feeds to confirm all is well at dc - 50Hz. Confirm that the incoming cpc is making a good low impedance connection to the casing of the fitting.

4. The tubes must obviously be properly connected at either end but they must also be within 10-15 mm of the adjacent metalwork or reflector for a good capacitative linkage with the tube all along its length. It's important for reliable ignition of the tube.

5. I think maybe, para 4 might explain why it always the inner two tubes that are affected because the linkage for them is only along one quadrant of their cross-section whereas the outer ones have a linkage along two adjacent quadrants.

6. Bring along some 6mm earth cable and the earth pin from a 13A plug. Connect one end of the cable to this pin and attach a croc clip or make a fish tail at the other end. If you observe a strobing light fitting, after first confirming it is earthed (Para3) plug in the earth pin to a nearby 13A outlet and attach the earth to the fitting temporarily. I am wondering if skin effect along a long supply run means the strobing fittings do not have a low impedance path to the MET/Means of earthing, remembering that HF ballasts/tunes operate at around 20kHz or higher.

7. Check (even better remake) all connections (L, N and E) along the run to the strobing fittings, to the mcb and mcb to busbar.

8. Finally (for now) is there any reason why these fittings would be running hotter or colder than the rest? if running hotter I am wondering if a ballast has a thermal cut-out which is cycling on/off, this cut-out being a thermal switch or some electronic mechanism.

 

[marconi]

Another experiment to attempt to discover if the problem is supply voltage related is to take along some 1.5mm2 3 core flex and a plug with a 13A fuse to the garage.

Disconnect the strobing fitting from its installed supply and wire in a temporary 1.5mm2 supply. Find a socket which is on the same phase as the fitting's normal supply and another one that is not.

Try supplying the fitting from the same phase for a day or more or until strobes, and then swap over to the other socket/phase.

I am assuming you can use an FRC which is lightly loaded (less than 1kW) and not supplying pulsed loads because I want to avoid voltage drops other than caused by the strobing fitting. Even better would be the cleaner's set of sockets if there are any.

Keep the 1.5mm2 run as short as possible. And if you can, use 2.5mm2 flex instead - I know it is hard to connect 2.5 mm2 into 13A plug but I have achieved it with an MK tough plug - to keep the supply impedance low.

Use flex so all cores have the same csa.

This experiment is related to point 1 in my earlier post.

 

[westward10]

I know electronic ballasts don't like high output tubes but they have been okay for a year.

 

[Toneyz]

Thanks for all the replies,
Westy, the fitting is designed for h/o tubes sold as such.
marconi , read your reply there are two lighting circuits both on the same phase these lights that have been affected have been to date on the same circuit the two replaced have been fine up to now it is another fitting now. there is not a great loading any inductive loads would be ramps & compressor for odd air tool. why not any effect on other lights. The circuit is wired in 2.5 with 2.5 CPC the fitting has a good earth also via metal to metal contact of fitting with lighting trunking back to D.B.

 

[marconi]

Maybe then the ballasts and its components are ageing overly quickly from running hot and causing its circuit characteristics to change.

Its more theorising from me I'm afraid without more information to go on.

The best I can come up with is in my earlier posts, especially paras 3, 4 and 5 in #4.

One way HF ballasts make some saving on the level of high voltage required to strike an arc reliably and quickly in the tubes is to rely on the distributed capacitance between the lamp and the earthed reflector/back shield. There has to be a good galvanic path to the casing of the HF ballast. It's all to do with preventing potential gradients along the tube which hinder the electrons emitted by the hot cathodes and accelerated along the tubes by the high electric field between the ends.

So my concern about the lowered mains voltage was about the ballast HV dropping as a result and by too much for the distributed capacitance trick to make up, and made worse by damp or dirty tubes...

Do the tubes get dirty or damp because such contaminents on the surface of the tubes, especially water can affect the potential gradients and ease of ignition? One 'cure' is to clean the tubes with soapy water and then spray them with silicone to replace the thin coating they are manufactured with. I have read this in the distant past ( - when researching for a post from someone in the EF about a florrie which would not start or did and later turned off for no reason) but so far can only find this mention of it:

See what ZEN1 says in :

101 uses of Silicone Spray - Home Improvement | DSLReports Forums - http://www.dslreports.com/forum/r21520532-101-uses-of-Silicone-Spray

 

[marconi]

You said the fittings are supplied by 2.5mm2 singles in trunking. Along the runs have you checked the temperature of these and other conductors in the trunking? Have all the cable selection factors been considered to arrive at 2.5mm2 as satisfactory? Maybe during the day the conductors feeding the florries are warming up in the trunking and thus increasing their voltdrop/metre as the day progresses.

Are the fittings at the end of a long/longish run?

 

[Squid]

Single or 3 phase?

 

[marconi]

The OP tells me in a PM that the florries in question are all on the Blue phase.

 

[Toneyz]

You said the fittings are supplied by 2.5mm2 singles in trunking. Along the runs have you checked the temperature of these and other conductors in the trunking? Have all the cable selection factors been considered to arrive at 2.5mm2 as satisfactory? Maybe during the day the conductors feeding the florries are warming up in the trunking and thus increasing their voltdrop/metre as the day progresses.

Are the fittings at the end of a long/longish run?

There are only 5 fittings on the circuit. The run max is only 20 meters max to end and this fitting is only 8 meters from the D.B. there is only one other circuit in the lighting trunking. I will have to wait until can return but since replacing the other two fittings that were faulty no problem except for this fitting. We did a school about a year ago with specified Sylvania fittings and have had to replace a few faulty fittings.

 

[Lucien Nunes]

Voltage and temperature are the likely factors to trigger the effect at a particular time. A simple max/min reading voltmeter will show if the voltage at the fittings is dropping unusually low at the crucial time, and a digital thermometer stuck to the ballast will show if that hits an abnormal peak. The fact that it's the innermost tubes may suggest temperature, but only with a larger number of fittings (e.g 10 doing the same) would it be statistically significant. At the moment it is annoying but there haven't been enough failures to show a clear pattern.

If the voltage is marginal, then one or other ballast or tube pair that is hotter or colder might react to the low volts before the other, and this also might mean that the fitting furthest from the board is not the worst affected. Connecting a variac into the DB so that you can vary the circuit voltage would give you a chance to provoke, or avoid the effect. Even taking a fitting down and plugging it into a variac might rule out voltage effects, if you can prove that it works reliably well below the lowest voltage seen on the circuit. Sylvania technical might be able to tell you how the ballast is supposed to behave if it sees out-of-spec conditions, e.g. shut down until power-cycled, retart after 5 seconds etc.

If the voltage is not particularly low, and from your description it doesn't sound as though it will be, there might be some disturbance on the supply that these fittings are more sensitive to than anything else in the building. Only a power quality analyser will reveal that, and again there might have to be a coincidence e.g. hottest fitting + oldest tubes + power disturbance, before anything shows up.

If you can swap ballasts and tubes, it might be worth interchanging a good set with a bad one to discover whether the problem moves or remains at the same fitting. That doesn't help the customer immediately but at least gives a clue as to what step to take next. If the ballasts are ageing prematurely, and they are all from the same batch, they might all start to do it at random times over the next year or two. Sylvania have replaced the failed ones so far, probably for their own peace of mind, but may not have done a detailed analysis of the ballast internals. They might have said they can't find a problem, but there could still be a subtle failure mode that will creep up on all the fittings, such as a rogue batch of capacitors that lose performance rapidly.

 

[marconi]

^^^^ - a typically top-notch post from Doctor Nunes.

 

[Toneyz]

Thank for all the replies, they are telling me that they just want to replace the fitting and see how it goes for now

 

[marconi]

Will you be taking the fitting and continuing some investigations on it? If not, could you send it to me and I will reimburse you for the postage and packing? It will keep me amused over winter.

I can send you bubble wrap and packaging.