While I did salvage the lights in the last part, that was not all! There were also downlight transformers and luminaires to be salvaged too. Using a similar strategy to the last, I decided to salvage one of each sort to look at (and play with).
iGuzzini Tiltable Gimbal Fitting
All of the lamps removed from the lecture room were probably used to light the edges of the room and were installed using iGuzzini fittings which featured an one-axis gimbal, meaning that it could be either mounted with the beam pointing straight down, or tilted around to face a wall. The purpose of the fitting is to hold the lamp safely, make the electrical connections and dissipate the heat as necessary.
To that end, the bi-pin connection at the rear is made by a firm “wiper” based connection, and the front seems to be covered in a mylar insulation.
The rear of the connector uses a ceramic block to handle the heat, screwed to the aluminium frame with a few fins to keep the temperature limited. This is connected to silicone wiring of a short length, which is then joined to ‘regular’ cable using a terminal connector block. Of course, being fitted for many years, many of these had some corrosion and lots of dust on them.
ATCO LVL24A-2 Transformer
The oldest transformer in pile seems to be this locally-manufactured ATCO LVL24A-2. This one features a fixed input lead which cannot be replaced, and is terminated in a regular two-pin GPO plug. This transformer is rated for 12v AC output at 4A (for a total of 48w), and has a 23A short circuit current. This particular one is “safety isolating” by nature of the insulated separated windings, meaning the output is considered SELV.
All of them seem to have been operated long enough that the heat has caused the potting compound, used to mechanically secure the transformer plates, to crack. This looks bad cosmetically, but is unlikely to affect the performance of the transformer itself.
ATCO TM50A-2 Transformer
The next transformer, age wise, seems to be another locally produced unit. There seems to be a bit of a change to the design, with replaceable input wiring to terminal connector block underneath the hinged cover on the left. The transformer seems to be made for 11.4v output instead, with the lower output voltage resulting in slightly longer lamp life but at the cost of reduced output. The short circuit current seems to have been slightly reduced accordingly.
Unlike the above unit, this unit is slightly smaller and fully enclosed in black plastic without visible potting in the rear.
I’m not sure what this marking is, but I assume this could mean that this was made in week 39 of 1997. Lets just say that a regular iron-core transformer isn’t anything particularly special, so I tried to take it apart. Prying at the bottom released the plastic cover plate, which revealed a metal-base, and a potted nightmare. Underneath the plastic, the transformer was potted into the plastic case. It took some heavy prying to get the core out.
Not really worth it, I’d say, but you can see that the left white wire is used to “series up” the thermal fuse to the input primary, so as to cut power permanently to the transformer if it were to overheat.
ATCO STE50A-2 Transformer
It seems this one is even newer, with a slightly more slim design and cable entry/exit all on the same side with push down retention clips. It’s still a regular magnetic transformer, but the casing has gotten a little more intelligent – as it raises itself up off the surface it is mounted on for some additional cooling.
OSRAM ET-REDBACK 60VA/230-240 20-60W Electronic Transformer
Now we get to the really exciting part. All the transformers above were all “heavy” magnetic transformers which give you roughly 12v AC at 50Hz on the output (as you put in 240v 50Hz in the primary). This unit is different – it’s compact, and it’s very light, as it’s an electronic transformer. An electronic transformer is like a switchmode power supply, but since this one is intended for lighting purposes, the output waveform and voltage isn’t as tightly regulated or specified.
There weren’t many of these units in the pile, which means that it’s likely a later install or a replacement for previously failed transformers.
This unit seems to support dimming by leading/trailing edge control, a high power factor of 0.95 and is double-insulated with SELV output. The output, for a lamp, is rated at 11.5V AC, with a 2m maximum cable rating (maybe for RFI reasons). Interestingly, the cable cap also features two different holes which act as cable retainers for different cable thicknesses.
To disassemble it, you merely have to pry at the bottom plastic tray to release it from the rest of the frame.
The main PCB is very small, and doesn’t have that many components.
You can see the input seems to pass through a large inductor or isolating transformer to reduce the RFI generated by this switching unit. There is a series thermal fuse on one of the heatsinks, with insulated wires reaching back into the PCB.There is also a regular fuse, covered in heatshrink to ensure no explosions in case of catastrophic failure.
The output transformer can be seen to be toroidal, which is efficient, but is made specially with a plastic cover to insulate primary from secondary in addition to the enameled copper wire. The feedback transformer is also separate and insulated from the main toroid. This is necessary to fulfill the requirements for double-insulation and SELV.
The underside shows a large gap between primary and secondary, with solder resist all over. Four surface mount diodes seem to form a full wave bridge rectifier for the input. A few assorted resistors and capacitors are also scattered on this side.
Electronic Transformer Output
This was the first electronic transformer I’ve come to own, so measuring its output was definitely of interest to me. So, I decided to power it up without any load on the output and see what came out.
Surprisingly, unloaded, the transformer didn’t give out anything like what I would imagine. It peaked over 23v positive over ground, and its AC RMS value was only 3.2v. It seems that the electronic transformer cannot operate below a certain amount of load which explains why replacing lower-powered single LED globes on some transformers just won’t work. The other thing is that the voltage regulation seems to contain high peaks, and lots of ripples, which can be stressful on LED globes – especially with those for poor regulation, as that can force excessive current peaks through the emitters.
The 50Hz seems to be existent in the output, so it seems that no work has been done to filter this out, and it becomes an advertising “feature” as being compatible with dimming, as the output will be duty-cycle modulated for dimming purposes.
Looking at the “ringy” pulse-train output, it seems to have a frequency of 3.54khz when unloaded.
I suppose none of this makes good sense when operating unloaded, so I threw in a regular 50w halogen downlight to see what its normal operating condition looks like.
The ballast seems to perform much more regularly, with a “flattened” top waveform that seems to correspond with the flattened top AC waveform that’s common here. It seems the transformer doesn’t operate below certain AC voltages resulting in “dead time” near the zero crossings of the AC input waveform. The average output voltage AC RMS is 11.76v, which is as expected.
The ballast seems to put out a roughly square wave voltage at 29.84khz which is actually a pretty high frequency, so maybe some lamps might not have fast enough diodes (due to recovery time) on the input stage to properly rectify the voltage coming out of these electronic transformers which may result in energy waste and diode overheating.
Putting on the Philips 10W lamp seemed to have it operate successfully.
The voltage delivered was 12.05v AC RMS, which seems quite suitable for the globe.
However, the peak to peak average is 40.8v, so it has to deal with about 20v “peaks”, when designed for 12v operation.
The ripple seems to be comprised of two components – a larger square wave component at 60.85khz and a smaller square wave component superimposed at 344.8khz. That’s pretty rough on the lamps, and could be a cause of RFI.
A look at the voltage spectrum seems to show quite a few spikes above the noise up to about 4Mhz, which isn’t entirely surprising. Shortwave listeners and HF DXers generally don’t like anything with switching supplies because of the inevitability of this due to the harsh switching waveforms.
Rather surprisingly, all of these various LED globes and transformers were salvaged from a single lecture theatre. It’s very surprising to see this many different models used, all attached to the same gimballed light fitting. It’s the first electronic transformer I’ve actually come to use, and its behaviour is quite interesting – its output is actually pretty lousy. But I suppose it is lighter, and potentially more efficient than an iron based transformer. It explains why not all LED bulbs are suitable for use with electronic transformers though.