Using Transformer with Separate 16V Secondaries in Split-Rail Supply

[van-c]

I’m working on a project that requires a power supply with +/-15 Volt rails.  The positive rail must supply around 500 mA and the negative rail must supply about 50 mA.  I’d like to build the power supply around a junk-box transformer.

As shown in the photo, the transformer has two pairs of 16V secondary windings.  Should I use two separate rectifier bridges to provide the two power rails, or can I can safely join the two 16V secondaries in some fashion and use only one bridge  (as though the secondary was 32V center tapped)?  I have plenty of rectifiers in my parts bin, but I’d like to know whether using one bridge is a safe and practical option to consider and how to go about it if so.

--Van

[SteveyG]

You can use the centre tap as the 0V rail and a bridge rectifier on the two outer ends of the windings for the rectified + and - supplies.

[van-c]

You can use the centre tap as the 0V rail and a bridge rectifier on the two outer ends of the windings for the rectified + and - supplies.

Problem is, the center tap (yellow wire) is electrically connected to only one 16V secondary via the brown wire.  The other 16V secondary (two blue wires) is isolated from the center tap.

[electr_peter]

Check phase (polarity) of 16V from blue lines (vs other 16V line) and connect one the blue wires to yellow.

[Richard Crowley]

No, electr_peter is not suggesting using the existing center-tap wire.  As you say, that is only the center-tap for ONE of the secondary windings.
electr_peter is saying to connect the two secondary windings IN SERIES. The point where the blue wire and the yellow wire connect would be the center-tap of the COMBINED secondary windings.

Yes, you could also use separate bridge rectifiers on each of the secondaries, and then wire the outputs into a dual rail, +15 and -15V output arrangement.  There are some "audiophool" circuits that use that topology.  They think that having MORE components somehow makes the circuit "better". Of course there is no merit in that fantasy, but audiophools are very stubborn and dislike facts.

[mariush]

Looks to me like the transformer has two secondary windings  :  16v (blue - blue)  and 19v ( brown - orange) with a tap somewhere at 3v, the yellow wire.

it could be   brown ----16v---- yellow -- 3v-- orange-blue ----16v---- blue
it could be  blue ----16v---- blue-brown ----16v---- yellow -- 3v-- orange

it could also be no connection between the two large secondary windings

One of the blue wires may be connected directly to the orange wire, I would check the resistance between blue wires and orange , or blue wires and brown wire, to figure out how the windings are layed out internally. 
two bridge rectifiers, 2  capacitors, one 7815 and one 7915 and you're done.

Note though.. not sure the transformer is big enough for 10-15va.

[SteveyG]

It can easily be tested, but the blue wires are unlikely to be connected to the other secondary winding. Connect blue to yellow, which is your new centre tap, and use brown and the other blue as the two ends of the winding. Make sure your phasing is correct.

[van-c]

The secondary containing the blue wires is totally isolated from the secondary with the yellow-brown-orange wires.  I wouldn't know how to determine the phase relationship of the blue wires with respect to the other secondary. I did measure resistance values, however.

Resistance between the red and black primary wires measures 15.8 Ohms. Resistance between the two blue wires (denoted “16V”) measures .5 Ohms, and between the yellow (0V) and brown (16V) wires measures 5.8 Ohms.  Resistance between both blue wires and the yellow lead is infinite.  (Why is the resistance between the two blue wires lower than that between the yellow and brown one, since both pairs supply 16V?  Different current specs?)

[mariush]

Yeah, different amount of wire results in different resistance.

[van-c]

Yeah, different amount of wire results in different resistance.

So that tells me to use the blue-wire pair for the 500 mA positive rail and the yellow-brown pair for the 50mA negative rail.

Something I'm not clear on:  If, per electr_peter and Richard Crowley, I wire the blue-blue pair in series with the yellow-brown pair, does it matter which blue wire is connected to the yellow (denoted "0V" in the photo)?  I.e.,  does the phase need to be measured somehow with the transformer energized to determine which blue wire to connect to the yellow one?

Once the secondaries are wired this way, I would then have a blue wire supplying the positive and brown wire supplying the negative with the yellow/blue "center tap" at gnd.  Right?

[IanB]

I.e.,  does the phase need to be measured somehow with the transformer energized to determine which blue wire to connect to the yellow one?

Yes. Connect one of the blue wires to the yellow one with no loads present and measure the AC voltage across the outer pair (other blue to brown). If the series voltage is ~35 V you have the right blue wire. If the voltage is ~0 V you have the wrong wire and need to swap.

The rest of your statements are correct.

[van-c]

Okay, I think I've got it now.  Thanks to everyone who contributed to the discussion.  Hopefully, someone else will also benefit from the shared knowledge.

--Van

[Mechatrommer]

Should I use two separate rectifier bridges to provide the two power rails, or can I can safely join the two 16V secondaries in some fashion and use only one bridge  (as though the secondary was 32V center tapped)?  I have plenty of rectifiers in my parts bin, but I’d like to know whether using one bridge is a safe and practical option to consider and how to go about it if so.

if i'm lucky as you with many parts in the bin with twin secondaries winding, i will... join 2 secondaries (make center tap as everybody suggested) and also make two separate full bridge (if i also have the luxury of diode/rectifier parts. single bridge you only utilize half the ac circle (50Hz), with full bridge i can utilize both ie 100Hz hence i have more power in my PSU.