External sodium floodlights - weird circuit issue

Some time ago I posted asking for help with our external lighting circuit (see thread here https://www.diynot.com/diy/threads/external-lights-floods-bulkhead-circuit-issues.531240/ )

Long story short, I've replaced the 3 sodium floodlights and all lamps, checked the circuit and fuses, and it is now all working. The circuit looks something like this:


The photocell is the switch to the two relays/contactors, which in turn switch the upper sodium lights and lower bulkhead lights around the outside of the property. Power to the upper and lower lights comes from the upstairs and downstairs light circuits respectively.

The whole system is now working as it should, with the exception of the 3rd sodium light. This spends its whole time switching on and off randomly. It is a completely new unit with switchgear incorporated and new lamp. I had replaced it once before and believe it has been replaced by the previous house owner too as it didn't match the other two originals.

So, it appears there is something up with the circuit but I don't know what I should be looking for. Could a voltage drop cause such behaviour? What should I be looking for with my multi-meter? Is it just a coincidence that the unit failing has the longest cable run?

Any advice/suggestions welcome.

wizbongre said:

This spends its whole time switching on and off randomly

Click to expand...

When you say that do you mean:
A Without the other lights on? or
B When the other lights are on?

Firstly, many an hour has been wasted on faulty new products.

I'm a bit confused by your diagram.
Two circuits - upstairs and downstairs - on one wire?

Being picky.... the device you show on the upper left is a photocell - a dusk til dawn switch. Is this what you mean?

A PIR is a movement sensor - sodium lights dont work well with a PIR because of the long run up time.

Are the fittings high pressure sodium (SON - a sort of golden color light) or low pressure (SOX - the traditional orange/yellow 'streetlight')

If SON, are they internal ignitor or external ignitor??

EFLImpudence said:

Firstly, many an hour has been wasted on faulty new products.

I'm a bit confused by your diagram.
Two circuits - upstairs and downstairs - on one wire?

Click to expand...

I guess he means that the power for the load side of the contactors comes from the upstairs lighting circuit for the sodiums & the power for the bulkheads comes from the downstairs lighting cicruit?

I guess he does - so the diagram is confusing.

Check the voltage at the third fitting, sodium ballasts have different tappings, you may need to move the wire to the 220 volt tapping

When you say going on and off randomly is it a reasonably consistent cycle eg on for 2 minutes off for 5 minutes on for another 2 minutes etc?

chivers67 said:

When you say that do you mean:
A Without the other lights on? or
B When the other lights are on?

Click to expand...

When the whole circuit energises and the lights come on, all 7 lamps light up and then after a bit of time the last sodium on the run switches off for a little bit, then re-ignites. It does this randomly at undetermined intervals, whilst the other 6 lamps stay on.

EFLImpudence said:

Firstly, many an hour has been wasted on faulty new products.

I'm a bit confused by your diagram.
Two circuits - upstairs and downstairs - on one wire?

Click to expand...

Sorry, PowerPoint isn't the best for drawing circuits, and I'm no spark so wasn't sure how best to show it. There is a circuit from the CU for upstairs lights, and a circuit for downstairs lights. The switched output from the photocell runs in parallel to the 2 relays. I've updated the image slightly to (hopefully) clarify a bit. Very simply, blue lines are power, orange the switched live. In my simple head, the relay is acting in place of a simple lightswitch as the photocell couldn't handle the load of all 7 lights on its own.

AdrianUK said:

Being picky.... the device you show on the upper left is a photocell - a dusk til dawn switch. Is this what you mean?

A PIR is a movement sensor - sodium lights dont work well with a PIR because of the long run up time.

Are the fittings high pressure sodium (SON - a sort of golden color light) or low pressure (SOX - the traditional orange/yellow 'streetlight')

If SON, are they internal ignitor or external ignitor??

Click to expand...

Not picky at all - good spot and a lesson to me not to do 2 things at once. I do indeed mean photocell - the circuit is designed to offer dusk till dawn lighting. The fittings are SON with internal ignitor.

AdrianUK said:

I guess he means that the power for the load side of the contactors comes from the upstairs lighting circuit for the sodiums & the power for the bulkheads comes from the downstairs lighting cicruit?

Click to expand...

I did indeed - hopefully a bit clearer now!

EFLImpudence said:

I guess he does - so the diagram is confusing.

Click to expand...

How could I have done it better?

333rocky333 said:

Check the voltage at the third fitting, sodium ballasts have different tappings, you may need to move the wire to the 220 volt tapping

Click to expand...

Didn't know this - very useful. Would it be feasible for the voltage to drop sufficiently along the circuit to the last sodium?

RF Lighting said:

When you say going on and off randomly is it a reasonably consistent cycle eg on for 2 minutes off for 5 minutes on for another 2 minutes etc?

Click to expand...

Doesn't appear to be. The cycle is longer than that I think, so I've not managed to measure it accurately. If it was constant, what might that suggest?

RF Lighting said:

When you say going on and off randomly is it a reasonably consistent cycle eg on for 2 minutes off for 5 minutes on for another 2 minutes etc?

Click to expand...

wizbongre said:

Doesn't appear to be. The cycle is longer than that I think, so I've not managed to measure it accurately. If it was constant, what might that suggest?

Click to expand...

I think were RF is going here is that SON lamps tend to cycle as they reach the end of their life.... they will burn for a while then extinguish, cool off, re-strike & repeat the whole cycle infinitum. This cycle often takes around 10-20min depending on the temperature. Could this be happening?

Also, if a SON lamp with an internal ignitor (shown by an 'I' within a triangle on the lamp) is installed into a fitting which also has a built in ignitor then strange things can happen.

SON lamps intended for use with an external ignitor (shown by 'E' withing a triangle) dont usually start at all if used in a fitting without an ignitor.

If you had to make an educated guess.... how long would you say the cable length is between the relay & the 3rd SON fitting?

Do you think they are looped from one fitting to the next or do they fan out from some central junction box, say in the loft?

These are brand new SON lamps in brand new fittings - I don't believe it is a lamp approaching end of life. As it has happened with this particular unit previously suggests there is something underlying. The cycle takes longer than 10-20min too. I have a spare lamp so can swap around to prove one way or another.

The lamps and units were bought together as compatible lamps & fittings, and the other 2 lamp/fittings work perfectly.

The wiring runs for 1m from the relay to a junction box; here it splits with a single run going to the fitting at the back of the house (about 5m) and another run going to the two fittings at the front. This runs for say 8-10m into another junction box, where it splits off for the 2nd fitting and then continues on for another 2-3m to the final fitting. So the fitting that is failing is on the end of a shared loop, farthest away from the supply. (It's worth bearing in mind there is about 4m between the CU and the relay too.)

So 1 fitting is off its own run, the other two share a loop.

I’ve had duff lamps and fittings straight out the box. Just because it’s new to you it doesn’t mean that it hasn’t been returned to the wholesalers by someone else and put back into stock.

A really simple test you could try is swap the lamp that’s playing silly devils with one of the good lamps from the other fittings and see if the fault moves with the lamp or stays with the fitting.