Talkin' PSUs

Meanwhile, at "new Babel," this they began to do...

They began to harmonize the World's voltages. The elephant in the room was Britain. Continental Europe had 220V, and Britain must be brought into line, but faced with replacing every substation transformer in the land, Britain refused.

Having caused a log-jam in the law-making machinery, they did what they always do - it was fudged. It was decided to "adjust" by statutory instrument all voltages in the EU to 230V.

The voltages in continental Europe remained at 220V, and the voltages in Britain remained at 240V, but the "230V adjustment" included an "accuracy annex," which allowed 230V the greater latitude of +/-10%.

In olden times the latitude was set to the achievable tolerance, which was +/-6%, and you could buy transformers with taps for 220V or 240V. Now it's 230V, period.

Therefore the off-the-peg transformer primary is rated for 230V, but as there is no such rated supply voltage, we must design for a supply voltage that may be 207V, or it may be 253V or anything in between.

Deciding on the transformer, part 2

In voltage, you can relate RMS to Peak (or DC) by √■ 2 and do the inverse using √■ 2 = χ-¹

Now you learned it on the calculator; we can start thinking of it as √2 and 1/√2 (it's easier for me to explain).

With current, there is no equivalent authoritative ratio (maybe they've done one since?). There is, however, a rule of thumb ratio, which happens to coincide with the Fibonacci ratio. The important thing about it is it works!

There are no calculator keys for the current ratio, so you'll have to memorise it. IAC to IDC = 0.62; IDC to IAC = 1.62.

If you want 1-amp DC, you will require 1.62-amps AC.

In the last post, the transformer secondary voltage was decided at 24V. If we multiply that by the AC amps, we get 24 x 1.62 ≈ 39.

39 is the required power, but modern convention says we must call it VA and not watts, which is how we used to think of it.

Looking in the RS online catalogue, the nearest VA rating is 50VA for an E-I laminated clamp mount transformer. It costs £23.66 including VAT. There is a toroidal transformer available for £62.65.

If the power supply is to live in its own cabinet, I can see no reason for a toroidal transformer, so I would choose the laminated one and save almost £40.

My findings from my 1970s design indulgence blog regarding circulating currents in paralleled secondaries sway my choice to a dual 12V output, which I can wire in series for the required 24V.

Looking at the RS page: https://uk.rs-online.com/web/p/chassis-mounting-transformers/0504672/, we can click on the datasheet link and find the information for part number 504672.

It's a lot bigger than the 12VA transformer of the "teaser circuit" on page 1, and it won't fit in a PSU1 case, but we decided we wanted 1-amp.

The datasheet shows a no-load voltage of 13.3V and 13.3/12 ≈ 1.11 - it has 11% regulation. We can now work out the maximum input voltage the 7824 will contend with.

The maximum mains voltage allowable in the GB is 253V, 230V plus 10%. It used to be 240V plus 6%, and it still is! Therefore the maximum mains voltage is 255V.

255/230 ≈ 1.11. We also have 11% regulation (for when the power supply is left on with no load), and so 1.11 x 1.11 ≈ 1.23. Therefore the 24V AC can go as high as 29.5V AC.

Let's see what it could be after the rectifier. We first remove the "entry-fee" of two diode-drops, but this is on no-load, and the diodes only need to pass quiescent current for the 7824 and the discharge resistor R1. Assuming we use regular silicon rectifier diodes, the diode-drops will be 0.6V. We take way 1.2V from the 29.5V and are left with 28.3V AC.

We have to take the peak voltage, which we do by multiplying 28.3V by √2.

The answer is 40.05V DC.

Another idiot question incoming (and might also be answered in your 70s page?)

Changing the ground wiring (yes, the ground) on the '70s project recently, so it didn't encircle the transformer (just one turn), the S/N improved by 6dB (the noise halved). Might you have noticed the diagram shows a round transformer?