What????? This doesn't make any sense! Justin....HELP!!!!!!

OK - let me explain.....

I have the magnificent beast known as the Retro Roadshow.....16 feet of light screens burning just shy of 3Kw of incandescent lamps. It gets HOT, and occasionally smells a little "too" hot (some of the chipboard actually shows signs of charring in one of the screens).

So....a good idea methinks.....replace the 25w filaments with 4w LEDs - what could be wrong with that?

In theory....absolutely nothing. The LED units are these

On the surface, a simple replacement (well they work in a standard ceiling fitting anyway in that they turn on and off as expected!).

However....I've spent this evening going out of my tiny little mind.....to the extent I've actually ripped ALL wiring out of one of the light screens bar one lamp....just to prove a point!

A great quantity of Rioja isn't helping I might add ;) (don't worry - I stopped doing practical and switched to theory once I'd uncorked the bottle for those who are concerned!)

Once the electricity is applied via a bulgin 8 pin connector....the lamp lights.....even with the controller TURNED OFF! (and I've tried this with multiple controllers including a switch pack with a Frankenstein bulgin connector), and multiple cables.

Start a chase, and it burns brighter when it's "On", and then returns to a dimmer light when it's "Off"....even though these LED lamps aren't actually dimmable :daft:

Replace the LED bulb with a standard 25w filament...no problem.

Now - here's where it gets REALLY weird!

Reconnect the chain and replace just ONE lamp in the chain with a filament instead of an LED (keeping the rest of the chain LED)...and it works! How is this possible?


I'm totally confused. I'd like to convert to LED for sheer sanity (they're brighter and obviously run a lot cooler.....plus less of a fire hazard).
I have absolutely no problem with proper electricity (i.e. filament lamps), but LED obviously goes into the dark side of "electronics"...and I'm not even a novice in that field.

Any ideas?

* Do I need to rewire in a different way to live one side and neutral on the other? (I've tried reversing the polarity by the way)
* Do I need a completely different type of controller? I'm currently using a QTX dimmer pack as the main controller (with 4xIEC fudged into an 8 pin bulgin via a Frankenstein lead...although my PAT tester OK'd this on Saturday as being safe). The secondary controller is an old Transcension CC4....same results for both (using different leads)
* Do I absolutely HAVE to keep a single filament in the chain just to make it work or do I need something like a resistor to make it behave?

Quote:

---

Originally Posted by Imagine

OK - let me explain.....

I have the magnificent beast known as the Retro Roadshow.....16 feet of light screens burning just shy of 3Kw of incandescent lamps. It gets HOT, and occasionally smells a little "too" hot (some of the chipboard actually shows signs of charring in one of the screens).

So....a good idea methinks.....replace the 25w filaments with 4w LEDs - what could be wrong with that?

In theory....absolutely nothing. The LED units are these

On the surface, a simple replacement (well they work in a standard ceiling fitting anyway in that they turn on and off as expected!).

However....I've spent this evening going out of my tiny little mind.....to the extent I've actually ripped ALL wiring out of one of the light screens bar one lamp....just to prove a point!

A great quantity of Rioja isn't helping I might add ;) (don't worry - I stopped doing practical and switched to theory once I'd uncorked the bottle for those who are concerned!)

Once the electricity is applied via a bulgin 8 pin connector....the lamp lights.....even with the controller TURNED OFF! (and I've tried this with multiple controllers including a switch pack with a Frankenstein bulgin connector), and multiple cables.

Start a chase, and it burns brighter when it's "On", and then returns to a dimmer light when it's "Off"....even though these LED lamps aren't actually dimmable :daft:

Replace the LED bulb with a standard 25w filament...no problem.

Now - here's where it gets REALLY weird!

Reconnect the chain and replace just ONE lamp in the chain with a filament instead of an LED (keeping the rest of the chain LED)...and it works! How is this possible?


I'm totally confused. I'd like to convert to LED for sheer sanity (they're brighter and obviously run a lot cooler.....plus less of a fire hazard).
I have absolutely no problem with proper electricity (i.e. filament lamps), but LED obviously goes into the dark side of "electronics"...and I'm not even a novice in that field.

Any ideas?

* Do I need to rewire in a different way to live one side and neutral on the other? (I've tried reversing the polarity by the way)
* Do I need a completely different type of controller? I'm currently using a QTX dimmer pack as the main controller (with 4xIEC fudged into an 8 pin bulgin via a Frankenstein lead...although my PAT tester OK'd this on Saturday as being safe). The secondary controller is an old Transcension CC4....same results for both (using different leads)
* Do I absolutely HAVE to keep a single filament in the chain just to make it work or do I need something like a resistor to make it behave?

---

Ok calm down Wayne. Chillax, as duh yoof will say.. .

The reason the LED lamp is lighting up is that these modern LED thingies require but a very minimal amount of currants to operate. As such, a small amount of current leaking through an electrical circuit (nothing of the level that a switch closing will allow BTW) will let the LED light up albeit quite dimly.

What I reckon you need is a dummy load on that lighting channel - i.e. a real bulb in parallel with the LED type. Or just more of the LED lamps all in parallel. The leakage current might not be enough to light more of them at the same time.

I believe BigClive has an explanation here: https://www.youtube.com/watch?v=tzWz_guJHvY

The actual reason a low current draw lamp like these LED thingies can light up is due to parasitic capacitance. That is to say, capacitance between terminals (say of a switch, or adjacent cables) is allowing a tiny amount of current to flow. All an LED needs is a tiny amount of current, and so will glow.

According to a thread I just read on an IEEE forum, you can use a 100 nanofarad capacitor (class X rated!) in series with a 100R resistor (0.5 watt will be OK I think) across line & neutral on the lamp holder(s). Aka something Maplin sell (used to sell?) https://www.maplin.co.uk/p/rc-contact-suppressor-rg22y or Farnell.. http://uk.farnell.com/roxburgh/re120...t=rc%20network

While I'm not the Oracle in these things, I had a similar problem at 12v. Having replaced the filament lights on my trailer with LED ones, the indicators were constantly on. I solved the problem by putting an additional filament bulb in the circuit, which cured it.

What kind of strange sorcery is this?????

OK - this conversation actually continued offline last night and several solutions have been suggested - the easiest as suggested in my original post and indeed by Peter above being to keep a tungsten bulb in the circuit.
So...most of the boxes will be converted, with just the two outside ones (which have one of each colour as a filament bulb) will remain as real lamps (those boxes don't actually get too hot anyway).

Just the small matter of putting all the wires back into the box I dismantled last night now :o