# 4 channel +/- from dual secondary

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.

#### Jamh

Is it possible to provide four clean +/- from dual secondaries?

I have a bunch of 2x 30V toroidals. Would be great if I can somehow divide the load to power 4 tpa3118 btl amps with one. I think 4x15V would be Ok. Their VA rating is high enough.

What do you think? Good or bad idea?

I can think of 2 other possibilities:

1. Use 2 toroids, open them up, unwind enough wire to drop the secondaries to around 18V, so that with 1.4x rectifying I'll end up with 24V. This will be a fair amount of work I think.

2. Use 2 toroids with Step Down DC-DC circuits from 30 to 18.

Of course, minimal noise and cross talk, etc.. are preferable.

#### DF96

The basic problem you face is that you can't have two things simultaneously both connected to each other and disconnected from each other.

Dual secondaries can give you a single + and - supply via full-wave rectification (or two separate single polarity supplies) or two separate + and - supplies via half-wave rectification. You can't get four separate + and - supplies.

#### jan.didden

Paid Member
He could use four sets of rectifiers and caps. Not really separated, but visually, sort of

Jan

#### CharlieLaub

Paid Member
The missing piece of the puzzle that you didn't explicitly state but you discover if you look up the TPA3118 is that the power supply is single ended and can be from 8-24V DC.

You have transformers with separate 30V AC secondaries. You didn't mention the type of primary winding.

One possibility would be that you have dual primary (two separate 115/120V windings). Since you live in the USA, if you connect the primaries in SERIES instead of the usual parallel you have the voltage on the secondaries. This is because each primary now has 115/2 VAC on it instead of both having 115VAC if they are connected in parallel. The result would be that your secondaries would have HALF the voltage but the SAME CURRENT capability, which has the downside of also having the "VA" rating. 15VAC on each secondary would result (via rectifiers and smoothing caps) in two separate ~21V DC (unloaded) rails, which seems to be what you want. You could then connect two amp boards to one of these and the other two amp boards to the other - there is no need to have a separate supply for each board...

If that doesn't work, just buy a more suitable transformer new or via Ebay. It really would not be worth it to regulate the rails down to 24V because you need at least 3A current capability per amplifier and implementing a good regulator or regulators would be approaching the cost of a new transformer. Here is a link to a good quality one that you could use for these amp boards (cost is \$32 plus shipping):
AS-2215 - 200VA 15V Transformer - AnTek Products Corp

#### Jamh

The missing piece of the puzzle that you didn't explicitly state but you discover if you look up the TPA3118 is that the power supply is single ended and can be from 8-24V DC.

You have transformers with separate 30V AC secondaries. You didn't mention the type of primary winding.

Right. The TPA3118 boards do not need +/- power, just single ended. Ideally I'd like to end up close to 24V DC or even slightly higher.

The transformers have dual primary winding and dual secondary winding. I need to measure them again, I can't remember if they were 30V each on the secondaries or 35V. Something in that vicinity.

Using the + and ground, then - and ground will not work (the input ground to each of the amps will screw things up). I was thinking of 4 rectifiers.

#### Jamh

if you connect the primaries in SERIES instead of the usual parallel you have the voltage on the secondaries. This is because each primary now has 115/2 VAC on it instead of both having 115VAC if they are connected in parallel. The result would be that your secondaries would have HALF the voltage but the SAME CURRENT capability, which has the downside of also having the "VA" rating. 15VAC on each secondary would result (via rectifiers and smoothing caps) in two separate ~21V DC (unloaded) rails, which seems to be what you want.

Good idea! Didn't think of that. I'll need to check to see if they do indeed have dual primaries.

#### CharlieLaub

Paid Member
Right. The TPA3118 boards do not need +/- power, just single ended. Ideally I'd like to end up close to 24V DC or even slightly higher.

The transformers have dual primary winding and dual secondary winding. I need to measure them again, I can't remember if they were 30V each on the secondaries or 35V. Something in that vicinity.

Using the + and ground, then - and ground will not work (the input ground to each of the amps will screw things up). I was thinking of 4 rectifiers.

I believe you are slightly misunderstanding how a transformer works.... there is not necessarily a "ground" unless you connect one. By "not necessarily" I mean that the magnetic coupling between the primary and secondary in the transformer provides isolation and does not pass DC. For example, it would not matter whether you connect gound-centere-100VAC (think of a sine wave where the positive and negative peaks are at the same magnitude but opposite sign) or a sine wave voltage oscillating between 1000V and 1280V (1280 = 1000V + 2*1.4*100) the result is the same on the secondaries.

On the secondary side, there is also no "ground" per se - it's floating. Only AC voltage is induced into the secondary. For example, given ONE secondary (two wires), after rectification and smoothing (via one or more capacitors) you can connect either wire to ground to form a positive or negative voltage with the other terminal.

What can happen, and what might lead to confusion, is with a center tapped transformer the two "ends" of the secondary winding are moving current in opposite "directions" but with the same magnitude (they are just ends of the same wire after all) and the middle point is therefore at a midway potential. Because the magnetic coupling does not pass DC (only AC) this results in a center tap voltage that is 0V (ignoring unequal loading, etc), but you could just as well connect it to some voltage reference and it will be happy to sit at that voltage as well.

Your transformer has TWO secondaries. If you connect a rectifier and smoothing cap to each you have TWO independent supplies, however, they don't know about each other or about "ground". You can connect these together in different ways, for instance to make a dual polarity supply, or you can use them separately without any connection between the two. They are electrically isolated unless you connect them.

One way to make use of two identical (identical voltage) secondaries that might work well for you is to connect the secondaries in parallel. This can be done in two ways - one will give the same voltage as one secondary alone, but can provide twice the current. The other is a useless "backwards" connection in which the voltage induced in the first primary is cancelling the voltage induced in the second primary. Sometimes the secondaries are marked so that it is clear which wires should be paired up. If not, you can measure the AC voltage on the open secondaries before connecting them to determine the correct way to do it.

In your case, I would just connect the primaries in series and run your 115VAC main through them. This will give you half the rated voltage on the secondaries (15V or 17V AC). I would connect the two secondaries in parallel. This gives you one single ended supply. Then connect a rectifier and smoothing cap(s). This gives you a one single-ended smoothed DC supply. Run four separate pairs of wires to the power supply terminals of each of your four amp boards.

Don't bother connecting any of the PS to earth-ground anywhere - it's just not necessary and can cause more problems then it solves. Just measure the voltage across the two wires and connect the one that measures the "positive" voltage to the amp + supply and the other to the amp "-" or "ground". Remember, the multi-meter is NOT measuring the voltage with respect to earth-ground when you do this, it's making a relative measurement. DO connect the enclosure metal case to safety earth with a toothed washer, machine screw, lock washer, and nut and use a 3-prong plug.

#### Chris Daly

Is it possible to provide four clean +/- from dual secondaries?

I have a bunch of 2x 30V toroidals. Would be great if I can somehow divide the load to power 4 tpa3118 btl amps with one. I think 4x15V would be Ok. Their VA rating is high enough.

What do you think? Good or bad idea?

I can think of 2 other possibilities:

1. Use 2 toroids, open them up, unwind enough wire to drop the secondaries to around 18V, so that with 1.4x rectifying I'll end up with 24V. This will be a fair amount of work I think.

2. Use 2 toroids with Step Down DC-DC circuits from 30 to 18.

Of course, minimal noise and cross talk, etc.. are preferable.

Why not do your 4 x 15v + and - as DC, rather than all of this sweat with
deriving multiple AC taps.

Using DC regulators, ie 7815, 7915 you can have 4 x DC 15v + and -
from your 2x 30v transformers. if you need more current following regulation
than the 1 amp available from 78xx, and 79xx , use pass transistors.
or regulators with more current like LM338.

There is a LM338 neg and positive supply described here.
http://www.diyaudio.com/forums/chip-amps/56106-lm338-regulated-snubberized-psu-audio-amplifiers.html

Note 1 transformer for positive and 1 for negative, so your 4x positive
derived from one transformer, and 4x negative from the other.
Would this be a solution for you ? If not, then 8 regulators
4x for positive and 4x for negative.

Cheers / Chris

#### Jamh

Thanks guys. Food for thought. I like the LM338 route. So it should be fine to just drop in that Snubberized PSU schematic the C7 to C8 and C9 to C10 ground connections? Essentially have them as dual mono supplies?

Just to be clear, I don't need - (minus) V. Just GRND and +V.

Last edited:

#### Chris Daly

Thanks guys. Food for thought. I like the LM338 route. So it should be fine to just drop in that Snubberized PSU schematic the C7 to C8 and C9 to C10 ground connections? Essentially have them as dual mono supplies?

Just to be clear, I don't need - (minus) V. Just GRND and +V.

Hi
If you do not need V- , you would use the top half of the carlosfm schematic
and repeat that same top half for the other channel to provide 30v DC
on each .

Note V in to V out difference is up to 12 volts in his implementation suggesting the T03
cased devices must be on a suitable heat sink and be insulated ( T03 case hardware ) due to Vout of the LM338 also being the case of that device.

Following the bridge rectifier 39-42 Volts DC In the LM338 is providing 30V DC out.
91R, his schematic R3 and R5 2k2 is establishing the voltage Out figure.
He has allowed for diodes in his calculations. The diodes are necessary due to the values of C5 and C7 Note voltage values of capacitors must always be higher than
anticipated voltage. Minimum suggested would be 50v & higher like 63v better.
Diodes must be orientated exactly as shown

The formula substituting his resistance values Vref ( 1.25) x , R5 (2k2R) / R3 (91R)
+1 applies. ( Normally this is known 1.25 x R2/R1 +1 )