Transformer Selection

[mnemneth]

I have had the thought that a switching power supply would be nice, however my experience with switching power supplies is that you end up with one end tied to neutral... so you cannot just take two of them and connect them together to get a split power supply.. the one supplying -V ends up getting shorted out through the neutral. Doing some searches I have not been able to find something like a +25v and -25v switching power supply, I know they could be made, because computer switching power supplies output +5 -5 +12 -12 and other voltages as well. I suppose I could learn to build my own switching power supply, but it seems dangerous to experiment with something running directly off the AC line, so I decided to stick with transformers.

Got to be honest and say I've never had that experience.

Every SMPS that I've worked with, other than a PC unit has the output totally isolated from the mains input.

The supplies just "float" and can be connected as a + 0 - supply quite easilly.

Just buzz the output plug to the mails input to check.

Sandy.

 

[Zaaphod]

I don't have a lot of experience with SMPS, I just had one that was tied to neutral on one side, so I thought they all were. Thank you for the information. That would be great if they were floating. I did see some inexpensive smps that seemed to be in range. I will make sure to ask if they have floating outputs and test it myself before I try to hook 2 of them together.

 

[AndrewT]

Sorry, but no.

The lightly loaded supply for that "normal" 18Vac was ~27Vdc.
As the load increases the average output drops.
Firstly due to ripple voltage appearing on the output.
Secondly due to the reduced output as the transformer regulation comes into play.

At full power output you could expect the rail voltage to end up anywhere between 20Vdc and 25Vdc depending on how well you assembled your PSU.
Let's assume 23Vdc, allowing ~4V of sag under full load.

Use that +-23Vdc = 46Vcc and find what the datasheet tells you as a prediction for maximum output power into your nominal speaker load.

All the information that I have put up in this thread is already posted many times in this Forum and is freely available on other web sites.

 

[simon7000]

US AC line voltages are slowly being stepped up to improve efficiency. Normal is 127 but can go up to 132 or down to 117 and still be "Normal." Transformers are rated for output voltage at rated current. Regulation of 10% means it may drift 10% above that rating without a load. When you put a diode/capacitor load on a transformer it only produces current when the voltage stored on the capacitor is less than than the transformer's output voltage. That is for a smaller amount of time than the full cycle so the peak current goes up when charging is shortened. That means transformer losses also go up.

As transformers are not all the same efficiency or even construction there are rules of thumb for AC amperage ratings based on DC draw. It is typically 1.8 x DC amps for AC amps. In a power amplifier you rarely draw full current, but that number is still useful.

The AC output on some transformers is rated for 110 volts in. So using 130/110 plus 10% or so for regulation will give you the maximum voltage to expect. The capacitors and circuitry should be able to handle that. The rated output voltage should be your design target but allow for things to still work even 10% below that.

Toroidal transformers couple in more line noise than EI cores.

So pick for maximum voltage not to damage stuff. Select the current to be 1.8 times what you expect to drive into a pair of loudspeakers with 1/2 of their rated impedance. Rated impedance can be twice the minimum! EI core is first choice, but toroids are more common in small sizes. Toroids tend to have better regulation.