Questionable Florescent LED replacement Efficiency ?

[Stueeee]

LED tubes I have seen only require power feed at one end. It's been a while since I opened a strip light fitting but, if you can, more easily just feed one end, that will be fine.

You can get blank light fittings for them. Like a normal strip light but with no gubbins inside. They have only one end wired up (marked) and yes, must be fitted the right way round. Old strip lights have power at both ends.

Wiring L & N to one end will be OK until someone puts a tube in the wrong way round. This will result in a dead short and blown fuse/tripped MCB. The 'other' end of these LED tubes have the two pins looped. A simple solution is to wire the live to one end, the neutral to the other, and a wire that connects the 'spare' connection at each end of the fitting. Then the tube will light whichever way round its fitted.

 

[awemawson]

Well at least the manufacturer is taking this seriously - they are sending their engineer out tomorrow afternoon to obverse the issue. Hopefully his measurements will be close to mine, but there's a fair difference between their claimed 24 watts and my measured 80 plus watts

 

[bloomi]

I think you might be comparing apples to oranges. They are quoting a real power of 24W, you are measuring an apparent power of >80W. If the system with the magnetic ballast fitted has a power factor of ~0.25 (which sounds like a realistic number for this system) then the two are equivalent. A clamp meter cannot be used to measure power in AC circuits unless the device is purely resistive (heater / incandescent bulb). You need to use a Kill A Watt style meter to measure the real power consumption (which has to be wired in series with the fitting, not using a clamp meter). Without measuring the real power you cannot claim the tubes are drawing excess power.

In practice what I believe you are seeing is the ballast (inductor) shifting the current out of phase with the voltage (and hence reducing the power factor below 1). This means that the light effectively sees a lower voltage and hence draws more current to deliver the same power. However as that current is being drawn while the voltage is lower the actual power consumption remains unchanged (less the loss of a few Watts in the ballast).

If you are on a domestic tariff then they are billed on real power only even if the apparent power is >80W you will only be billed for the 24W real power so that is fine. Some commercial tariffs bill on apparent power, in which case you do need to remove the ballasts, however I still do not believe the lights are missold as their real power will be correct (which is what everyone talks about when they talk about power). I would however still expect the loss of a few Watts in the ballast.

 

[MetalMonkey]

If you are on a domestic tariff then they are billed on real power only even if the apparent power is >80W you will only be billed for the 24W real power so that is fine. Some commercial tariffs bill on apparent power, in which case you do need to remove the ballasts, however I still do not believe the lights are missold as their real power will be correct (which is what everyone talks about when they talk about power). I would however still expect the loss of a few Watts in the ballast.

I have a feeling this will be set to change in the future for domestic customers, as the majority of the loads in a property move towards electronics with poor power factors and customers are put on smart meters which can measure reactive power.
At that point it will just be a simple billing change.

 

[awemawson]

Well the Independent Engineer rang up to ask if he could come earlier - so he's been - observed, and I'm delighted to say exactly confirmed my measurements. He has discounted Power Factor issues telling me that the choke PF is no where near low enough to give the observed results.

He has to report back to the Importer and TLC and I will be fascinated to hear what their proposal is

 

[Guineafowl]

Well the Independent Engineer rang up to ask if he could come earlier - so he's been - observed, and I'm delighted to say exactly confirmed my measurements. He has discounted Power Factor issues telling me that the choke PF is no where near low enough to give the observed results.

He has to report back to the Importer and TLC and I will be fascinated to hear what their proposal is

Watching with interest. How did he measure the power consumption?

 

[awemawson]

@Guineafowl - a clamp meter !

 

[Guineafowl]

Ok, so as we’ve established, he can only measure apparent power like that, and he’s making the assumption that the power factor can’t be a significant, er, factor.

What I’ve done, since this is interesting and I’ve got some spare time, is drawn the original fluo circuit (left) with the LED unit fitted (right). I’ve assumed the starter is removed. Ballast and cap values are typical. For the moment, I’ve modelled the LED itself as a 24W resistive load, just to see what the other components do.

What we then have, is an RLC parallel-resonant circuit, with a resonant frequency of 67Hz, very close to mains frequency. You can get quite surprisingly large currents near resonance.

Yes, it’s a limited model, but it does suggest that reactive effects shouldn’t be ignored in this case, as your engineer has done. The simplified circuit shows an apparent 40+W load dissipating only 24W, and that’s with the LED behaving as a nice, simple resistor.


Note that the power factor of 0.5 here, only needs to drop to 0.3 to show 24W appearing to be 80W.

 

[brewdexta]

Interesting stuff. I swapped out 4 fluorescent tubes for LED tubes a couple of years ago in the upstairs workshop. I just replaced the starter with a dummy. I never measured the benefit. I have a load more to change, probably about a dozen as they are double fittings in the downstairs barn. I could possibly monitor now as I fitted one of these recently while installing the solar http://lechacal.com/wiki/index.php/RPICT3V1

It sits on an RPI and I have current sensors on the solar, barn and main incomer plus a voltage sensor. This reports into an Influxdb time series database and I then query it via Grafana, it looks like this. Most of the time that the PF is low or variable is that the current is very low so mainly noise. The big blue spike is when we plug the EV in, you can seen the incomer PF goes to one. I haven't worked out why the barn apparent power goes positive and the real goes negative yet, they use the same sensor, it may be me and my influx database query, I'll get around to looking at it at some point.

So what I could do, is turn off all the other stuff in the barn, turn on all the tubes and take a measurement. Then do the same once I've swapped out the tubes, then again after ripping out the ballast. May take a while as I have sheds loads of other stuff to do though.

 

[awemawson]

@Guineafowl he's NOT "My Engineer" but the engineer appointed by the supplier & importer so he must know what he's doing . . . musn't he . . .

 

[Guineafowl]

@Guineafowl he's NOT "My Engineer" but the engineer appointed by the supplier & importer so he must know what he's doing . . . musn't he . . .

Here’s a story, while we await developments…

We have three phase here, and after a windy night the 3ph boiler was complaining of phase loss. All the 1ph stuff, distributed among the phases, was working fine.

Called out SSE, and the young lad measured at the service head, and showed me each phase measuring 240V to neutral, and concluded the supply was fine, must be a problem with the boiler. Was packing up to go.

Now, I’d already been in there just before and found two phases at 0V to each other, not 415. Not wishing to reveal that or be a know-it-all Nelly, I said ‘my electrician mate’ had suggested measuring between phases as well.

Grumbling that he’d left that meter in the van, he went off to get it, to humour me. He jabbed the probes in, ’there’s your 415, there’s your 415, there’s your… Oh. Zero. Looks like your mate was right.’

Two phase wires had twisted together in the wind, blowing one transformer fuse, so we had two phase via three lines.

Moral: dunno. Be systematic and thorough, I suppose, and even engineers can make mistakes.

 

[Guineafowl]

I’ve highlighted this thread to an electrical engineering forum. One of the advisers has come back with this answer:

“The inductors were there to keep current flowing through the tubes for a greater angle, not just at the peak of the voltage cycle.
The capacitor was there to neutralise the inductance, to partly restore the PF.

The clamp current being measured now is the reactive capacitor current. That will be of the same magnitude as the original tube current through the inductor. That explains why the clamp current shows the VA of the original tubes. Depending on the type of metering used and tariff, they may actually be paying for the reactive current.

Removing both the inductor and the capacitor will reduce the reactive current, and the circulating energy. It will minimise the bill for lighting, by an unknown amount, depending on the metering.”

Power factor of a retrofitted LED unit in a fluorescent fitting

A poster on another forum has fitted a large number of 230V 50Hz 24W nominal LED tubes, apparently designed to be plugged straight into the old fluorescent fittings. He has found, by clamp meter measurement, that the new LED lights appear to be consuming 80W, not the claimed 24W. A number of...

www.physicsforums.com

 

[Seadog]

The choke is there, primarily, to provide "ignition" of the plasma. For no other reason.

 

[awemawson]

I thought that the choke was there to limit the current through the original tubes, hence the colloquial name 'choke' for what is actually an iron cored inductor. I accept that at switch on it will also provide a back emf spike to strike the tube.

 

[Guineafowl]

I thought that the choke was there to limit the current through the original tubes, hence the colloquial name 'choke' for what is actually an iron cored inductor. I accept that at switch on it will also provide a back emf spike to strike the tube.

As far as I know, you’re right, it’s both. The inductive kick at start-up strikes the arc when the starter contacts open, then it limits the current since the arc is a bit like a short circuit.

 

[Seadog]

Think of an ignition coil, points, and condenser. That's exactly what you have in a traditional fluorescent fitting. The starter, a bi-metallic switch, opens after a brief warm-up period, which then causes the magnetic field in the choke to collapse, which in turn induces a voltage spike that causes breakdown of the warmed, mercury loaded, vapour in the tube.
Omit the choke and the light won't work. Full stop.

 

[awemawson]

Well we have a RESULT!

The importer of these LED replacement tubes, who authored the leaflets previously posted (that specifically say no rewiring is required) has paid for two visits by their specified electrician who has:

A/ exactly reproduced my measurements on his first visit and
B/ on his second visit proved that the majority of the excess power is drawn by not removing the capacitor from the fitting, and a small extra amount by leaving the chokes in circuit contrary to their instruction leaflet.

In view of these findings they are additionally paying for two electricians to come and rewire all 31 fittings in my main workshop.

Now legally my dispute should be with the retailer - and I feel that I am owed some compensation for the considerable exess power drawn in the last year that these fittings have been in use. Needless to say the importer doesn't want play this game but I am pressing the retailer for a substantial discount against the further 28 tubes that I was about to buy that triggered my questioning a 'real' power savings !

. . .we shall see . . . and I will report !

 

[Shox Dr]

RESULT

 

[Guineafowl]

Well done for persisting.

I guess the rewiring is to be done because, as fitted according to instructions, the LED power factor is below what’s required. The capacitor is still producing leading VARs that aren’t needed.

As for claiming back the excess power, you could only really claim for the I2R losses from the extra reactive current.

 

[awemawson]

Well the distributor has come back with the AMAZINGLY GENEROUS offer of a 10% discount on the next batch of 28 tubes. I've not quite decided how to respond to this overwelming offer. My extra energy spend due to this error is in the order of several hundreds of pounds, and I've been a regular customer for probably 20 years spending thousands of pounds with them.

I think I need to polish up my sarcasm and draft a letter