Electronic low voltage lighting transformer - argh!

[PetrosA]

Lots of the new 12V halogen fixtures are using individual electronic transformers nowadays. Many of the instruction sheets say you can't read the voltage accurately because of the high frequency output unless you have the correct DMM, and it turns out you have to put a bulb in to get the transformer to come on in the first place. Ok, I lost a lot of time on a service call and had to eat a $40 transformer today because I didn't know that and I can deal with all that but...

Now that I'm home and playing around with the (still) good transformer that I replaced because I didn't know how to get a reading from it, I'm trying to understand my DMMs and their limitations. I have an Agilent U1211A (clamp meter) and a Greenlee DM-810 (DMM), both of which are true RMS. I'm not sure how to interpret the specs and the readings I get are way different - same source, hooked up simultaneously. The transformer label claims an output of 11.7V (AC or DC not specified) @ 120V nominal and both meters give me reading only in AC mode. Here's what I'm getting with a line voltage of 122.7V:

DM-810 - 13.15V AC @ 58.35 kHz
U1211A - 9.7V AC @ 25.6 kHz

The only clue I have about the actual frequency output of the transformer is a vague statement from the manufacturer's website that says most of their transformers operate between 20 and 40 kHz and a meter with a minimum of 40kHz is needed. That would suggest the Agilent is correct. OTOH, the voltage showing on the Greenlee seems closer to what I would expect.

Any ideas from someone who can understand the meter specs better than I do as to which is showing me the true values would be very very welcome. Also, is it possible that neither meter is correct and that this is a case where a meter that can read AC+DC is needed? Thanks

[quantumfall]

I'm not sure if this will help but I suspect both are inaccurate with the output of these PSU's being so noisy.  Could you try it under load with a smoothing cap across the output ?

A thort I had to see if you are getting any ware near the effective voltage might be to run two lamps in series measure the voltage and get a visual idea of the lamp brightness, then try to match the same lamp brightness using a bench variable supply and taking a voltage reading.  It may give an idea of the effective voltage. I have not tried this so may be just daft. if you had four lamps you could match the brightness side to side.

[PetrosA]

I could try a cap, not sure what size to use though. I don't have a bench power supply to use (although I do have a light meter, so matching output wouldn't be so random). Part of the problem is that none of these manufacturers are giving frequency specs so test results could vary from one brand to another. One thing's for sure - the introduction of electronics to electrical systems is still in its early stages and frustration will abound

[Fraser]

This web page may be of interest to you.

http://sound.westhost.com/lamps/elect-trans.html

Others are having the same problem as you when trying to measure these things. I personally would use a portable Oscilloscope to view the waveform. This may seem total overkill but in order to know what is going on in a PSU circuit you sometimes need to see the shape and behaviour of the AC waveforms. A multimeter is sometimes not enough as it can miss-measure a complex waveform. Modern 'low-energy' lamp drivers are even worse ... with timed start up voltages, running voltages and current monitoring by a clever IC...but thank goodness they are throw away items !

[quantumfall]

This web page may be of interest to you.

http://sound.westhost.com/lamps/elect-trans.html

Others are having the same problem as you when trying to measure these things. I personally would use a portable Oscilloscope to view the waveform. This may seem total overkill but in order to know what is going on in a PSU circuit you sometimes need to see the shape and behaviour of the AC waveforms. A multimeter is sometimes not enough as it can miss-measure a complex waveform. Modern 'low-energy' lamp drivers are even worse ... with timed start up voltages, running voltages and current monitoring by a clever IC...but thank goodness they are throw away items !

Looking at the waveform in fig 3 in your link You are looking at a high frequency and modulated output that is not very uniform so I'm not surprised that a DMM is having trouble taking a reading.

You would need to rectify and smooth the output to take a DC reading and you will need to use fast diodes etc, thats more than I would know how to do or would be sure I have an accurate reading.

These "Electronic lamp PSU's" are often hidden in building structures and have limited if any air flow, they only have a short life if covered with fibre or building insulation I think.

An electrician would just swop out a bad unit no voltage test needed its the most cost effective way.  I have seen 70 to 100 of these put in a large 5 bed house and as a maintenance procedure it would probably be better to change the lot when they start to fail, otherwise you will be back many times as each unit reaches it end of life.

A large house that had halogen 12V lamps installed had large toroids to power the lamps in the hall, with a long run to the basement location of the toroidal transformers. It needs large high current cabling as well on the load side, a better job IMO and more reliable but still expensive.

[Kiriakos-GR]

Every one deserves to have an opinion, so here is my. 

In such comparisons, its best to always have an load.
An equal value resistor , as the halogen lamp.

( Rule No1 , never let the inner impedance of your DMM , to mess-up  the output of your circuit ) 

With the load on, redo the measurements .
And post back your findings.

[quantumfall]

I have been out of the electrical game for a few years now so I forgot that we have the mains halogen lamps now which I would think a better solution if the lamp life is good.

[Kiriakos-GR]

the mains halogen lamps

If this reference are about the new ones called as " Green technology " ,
they do die fast , and in them there is mercury .. 

Save the planet ( less energy ), but kill the planet with mercury ..

[quantumfall]

the mains halogen lamps

If this reference are about the new ones called as " Green technology " ,
they do die fast , and in them there is mercury .. 

Save the planet ( less energy ), but kill the planet with mercury ..

The ones I'm thinking of are Tungsten filament Halogen technology, just run at line voltage rather than 12V
http://www.screwfix.com/cats/A354590/Lighting-Lamps/Lamps/240V-Halogen-Lamps .

[Kiriakos-GR]

In my "village" we call those as spot-light bulbs .    

[gobblegobble]

In such comparisons, its best to always have an load.
An equal value resistor , as the halogen lamp.

And in transformers intended for x to y lamps have load that equals the minimum amount of lamps, not just one. I'd say it was bit confusing to see the single lamp become brighter when second one was added to the circuit...

[cybergibbons]

I spent a long time replacing almost a thousand of these in a large building - about 30 failed within a month of each other, one starting a small fire in a roofspace, and several others causing evacuations. Although they were all to be replaced, I had to check how many didn't work. I tested them using a bulb and two test probes formed out of sharpened welding rods.

I took some to play with as well though.

The output is horrible - I think it was 40kHz signal modulated at 50Hz with some nasty random spikes and discontinuities producing wideband noise. Cold solder joints, cracked insulation, broken screw terminals, poor heatsinking, loads wrong with them. They were replaced with 240V fittings.

The owner of the building recovered the cost of the transformers + replacement cost as they only lasted about 4 years.

[PetrosA]

Kiriakos, the readings I posted were from a 60W transformer with a 20W bulb attached. With no load, there would be no voltage at all.

I had just installed these two hanging lights on Saturday, and got an email on Monday that one of them had already stopped working.  I haven't had any failures yet with similar lights I've installed, so I really haven't had to troubleshoot them before. That's why I didn't know that the transformer wouldn't produce voltage without a load. The only time I had a similar job was with small spot lights that were easy to swap from one canopy to the other (they plug in with a small cylindrical post) but these hanging lights have a cable system and glass shades - not the kind of thing you want to keep moving around. The design of the canopy makes it difficult to test without a purpose made rig with a bulb and leads for attachment to the transformer and second leads for attaching probes. I will be making a test rig with a 50W bulb and test lead attachments for future use. Maybe one of these days Dave will get a portable o-scope for review and I'll have a chance to win it later

[Kiriakos-GR]

Maybe one of these days Dave will get a portable o-scope for review and I'll have a chance to win it later

Get in the line      I am aiming the 100$ UNI-T to be released . 

Oh and an Tip, next time post some pictures of what you do ..
The Delphic oracle, had close down the business, since long time ago.   

[david77]

Those SMPS 12V halogen supplies are quite nasty, especially the cheap chinese ones you usually get with
cheap lamps from DIY stores. I avoid them like the plague.
Usually they shouldn't catch fire when failing but I have seen some that did, some haven't even got a
mains fuse or over temp protection. Dangerous stuff.

At work we only use these, maybe more expensive (about 25 Euros) but we never had any problems with
them.

http://www.basislighting.com/stock/accessories/35---105va-240v-electronic-transformer,-dimmable/

[PetrosA]

Oh and an Tip, next time post some pictures of what you do ..
The Delphic oracle, had close down the business, since long time ago.   

Sorry about that. I probably didn't write it clearly enough for non-native English speakers to understand that there had to be a bulb connected for it to work in the first place, and I should have written that I used a 20W bulb - it would have been easier to understand...

[Zero999]

If this reference are about the new ones called as " Green technology " ,
they do die fast , and in them there is mercury .. 

Save the planet ( less energy ), but kill the planet with mercury ..

There shouldn't be any mercury in halogen lighting. I think you're confusing it with fluorescent.

Yes, electronic halogen lamp transformers are just an oscillator connected to a transformer. Most multimeters are designed for 50Hz to 60Hz usage and are not true RMS. To measure the output voltage from an electronic transformer, you need a true sine wave meter. I'd suggest using one rated to measure a higher frequency than that produced by the transformer because it will produce harmonics.

If they're covered in insulation, they should be derated, just like any other piece of current carrying equipment such as cable or an old toroidal transformer.

I wonder if they can be used to make a DC power supply?

Add a rectifier, a smoothing capacitor and a buck to the output and you might get a decent PSU. The minimum load requirement and inrush required to charge the filter capacitor might be a deal breaker though.

[Kiriakos-GR]

Yes you are correct.
Actually they called as  Compact Fluorescent Lightbulbs ..