Mains Transformer Sizing

[keithsterling]

Hi Tony,

Right now I only own one Yamaha PA20 which I have to switch it between my Oakley delay units, so I can only use one at a time. I would like to run them all from one transformer.

In the ADR30 BG you state:

"For powering two ADR30 boards, then the mains transformer's secondaries should be rated 18V-0V-18V, power: 15VA. This will give you a power supply that should be theoretically capable of providing just over 230mA to each rail."

How do you get from 36VCT and 230mA to 15VA? I’m having trouble making sense of the numbers. I’d love a short explanation of the calculations.

I would like to run 2 ADR30s, and 1 SRE330 and one DDR320 from one transformer. What VA rating should it be?

[Synthbuilder]

How do you get from 36VCT and 230mA to 15VA?

It is not easy to calculate the exact maximum current draw from just the VA rating. It is a complex calculation that has a great many variables. Indeed, the VA rating of the transformer will be tied to a certain temperature increase, eg. the iron core rises by 60 degrees Celsius (which is heck of a lot to have inside a sealed case) when the maximum current is being drawn.

So a rule of thumb is used. And a little bit of a overhead allowed.

36V at 15VA appears to allow a 0.42A. But this is with a load that is a pure resistance like a very long length of wire. The current flowing in this resistance is proportional to the voltage across it and both the current and voltage waveforms will be close to perfect sine waves. It's standard ohm's law stuff. The moment one connects the transformer's secondary coil to a rectifier, smoothing capacitors and linear regulators, it gets way more complicated. And I mean really complex. The peak currents are huge, but only last for a short time, so no more tidy sine waves. And the voltage waveform coming out of the transformer is also no longer a sine wave as the transformer struggles to supply these bursts of current.

Now you can model this in a circuit simulator with the various devices you have connected and what you'll find is that the root mean square (RMS) current (this is the equivalent current that is heating the transformer secondary coil) is approximately 1.8 times that the current drawn from the regulator. In our example above; 0.42A divided by 1.8 is around 0.231A. This rule of thumb is just that. It's not exact. But it's pretty robust so you can normally use with confidence.

But note taking the maximum current will heat your transformer core up by whatever the manufacturer stated, eg. 60 degrees C above ambient temperature. So you need to think about making sure that heat will be allowed to go some where. So I tend to use a rule of thumb that you need a factor of two between maximum secondary current output and DC load current. This will allow the transformer to run cooler - but it'll still run warm.

I would like to run 2 ADR30s, and 1 SRE330 and one DCR320 from one transformer.

Personally speaking, I'd not run three separate devices from one transformer. You'd have to distribute the AC output of the transformer to all the devices. And those leads have to carry those horribly large current spikes. It's usually best to 'divide and conquer' - keep each power supply local to the device.

Tony

[keithsterling]

Personally speaking, I'd not run three separate devices from one transformer. You'd have to distribute the AC output of the transformer to all the devices. And those leads have to carry those horribly large current spikes. It's usually best to 'divide and conquer' - keep each power supply local to the device.

I'm going to run separate leads from the transformer to each device, not daisy chain them, so each lead will see the same current spike as it would with separate transformers, correct? Also, I'll always use these things in the same rack, mounted in consecutive spaces, so no need for really long leads.

Ok, please humor me Tony so let's go ahead with one transformer:

Current draw from single phase AC supply:
DCR330: .6A
2xADR30: .6A
SRE320: .65A

Total from single phase supply: 1.85

Since we are working from a 2 phase supply we should then halve this current?

1.85 x 2 = .925A

Then we double the transformer capacity so that it’s not working near it’s maximum current:
.925A x 2 = 1.85A

Next multiply by 1.8 to find the equivalent current:
1.85Ax1.8 = 3.33A

Find transformer VA rating by multiplying 36V by this current rating:
36V x 3.33A = 119.88

Let’s round that to 120VA

This transformer seems to fit the specs:

https://www.mouser.com/ProductDetail/Ha ... 4psA%3D%3D

Am I waaaaaaaaaay off base, or will this transformer work?

Thanks Tony, I really appreciate your time helping with this.

[Synthbuilder]

Since I don't recommend any Oakley project be powered by a component level transformer I will not comment on any particular transformer.

However, the total current draw from the positive rails of the SRE330, DCR320 and two ADR30s is 575mA DC. The negative rail will be less but let's assume the negative rail is similar. Any transformer would need to be able to supply twice that, ie. 1.15A (RMS), into each rail. With a 36V centre tapped transformer, you would need at least a 41.4VA device. 50VA or 60VA would do.

But, the lead length may yet be a problem. The centre tap of the transformer will have to cope with some very odd current waveforms. Where the transformer is local to the power supply, the centre tap carries only the current to that power supply. The ground reference, ie. module 0V and mains earth, are tied to the 0V output of the power supply. Not the centre tap of the transformer. There will be a small but significant voltage difference between the centre tap and the 0V output of the PSU. Even though they are tied together with PCB track. This does not matter in any single Oakley project since the centre tap does not directly form part of the analogue circuitry.

But connecting all four modules together with common centre tap wiring and multiple connections to 0V via the audio wiring may lead to problems with low level hum in the audio outputs.

If you would like a common power supply to feed all four modules, perhaps a better way is to make a +/-20V DC output power supply. That is, a single mains transformer driving a single +/-20V DC regulated linear power supply which can then be used to power all the modules. You'd not need the big smoothing capacitors on each RPSU module as these can be replaced with smaller 47uF, 35V devices. You would still have to connect a common 0V between the modules but at least it'll not have to carry the spiky return currents of the smoothing capacitors.

[keithsterling]

Thanks Tony. It took me a while to understand where these “spiky currents” come from, but I think I have a bit of understanding now. This diagram helped me grasp what I think you are talking about:
“When the capacitor charges through the diodes in the full wave (or half wave) rectifier, that is done with a short, but of high amplitude, current spike coming from the transformer or from the AC mains, and the larger the capacitor is, the higher that current spike is. It is a current surge, or current inrush. A picture below illustrates this issue”

ripple.jpg (39.94 KiB) Viewed 433 times

It’s amazing how much one must know to do this right. I am going to follow your recommendation to build a DC supply to distribute to the modules. Is it ok to skip the voltage regulator in this supply and just use full wave rectified, capacitor filtered supply, since there are regulators on the RPSU boards?

[Synthbuilder]

Is it ok to skip the voltage regulator in this supply and just use full wave rectified, capacitor filtered supply, since there are regulators on the RPSU boards?

Yes. Smoothed unregulated DC is certainly better than low voltage AC in this case. They'll still be a bit of 120Hz (and its harmonics) on the common 0V lead but it should be small enough. A regulated supply will be quieter though.

But separate smaller mains transformers for each module would be better still.

In all three scenarios it is absolutely essential to earth your power supplies. For the unregulated smoothed power supply and regulated power supply, the output 0V line must be connected to earth at the point it leaves the case. Remember too that all external metal parts of your power supply case should be also connected to earth.

[D.Tilbury]

Best thread in forum at the moment.

[KSS]

Personally speaking, I'd not run three separate devices from one transformer. You'd have to distribute the AC output of the transformer to all the devices. And those leads have to carry those horribly large current spikes. It's usually best to 'divide and conquer' - keep each power supply local to the device.

I'm going to run separate leads from the transformer to each device, not daisy chain them, so each lead will see the same current spike as it would with separate transformers, correct? Also, I'll always use these things in the same rack, mounted in consecutive spaces, so no need for really long leads.

Ok, please humor me Tony so let's go ahead with one transformer:

I am sincerely trying to understand the reluctance to just make yourself two more PA-20's or PA-30's ?

After all they are only a CT mains transformer in a plastic box with a GX connector. There is nothing special about them in any way save that many similar looking 'line lumps' do not have a center tap connection. Flatpack style PCB mount transformers are available, the plastic boxes are available -but in your expressed use would not be required- and the connectors are available. You gain flexibility and redundancy having three plugged into a power strip rather than the path you're pursuing.

When you decide to go with a higher voltage unregulated supply, you are also choosing to accept worse performance. And potential problems that may end up worse than the current problem you're attempting to address. Gotta love how potential and current always insert themselves into power threads as puns!

Tony is being his usual kind and careful self, but i will say it straight, What are you thinking here? What is the actual problem you are trying to solve that demands you keep avoiding the advice of one of the most technically proficient and respected creators of synthesizers? And synth music too!

I urge you to re-think the answer to the problem in front of you.

Generic PA-20s and PA-30's used to be available at a *much* better price than Yamaha but that supply has dried up or changed enough that I can no longer source them. But as long as its your own DIY -and you don't require the certs- they can still be made for far less than the Yamaha. and probably less than your alternative also. In both time and cost.