I always hesitate when dimensioning a transformer for a DC power supply, and I guess I sometimes overdimension them. My confusion stems from the difference in transformer and DC voltage.
If I need 10V DC @ 1A, will a 10VA transformer do? My gut feeling says I need more, so I use a bigger transformer, but how much bigger does it have to be?

- 9,658
- 17
- 68
- 117
3 Answers
First, 10VA is not enough if you ultimately want 10V DC at 1A out. Theoretically the transformer is capable of putting out 10W, but only at a power factor of 1. If you only needed to heat a resistor with this transformer, then its spec would be just on the edge. For anything else it's not enough. Then any good engineer will add some margin anyway.
Second, even 15VA or 20VA is not enough to know the transformer can do what you want. You need a specific voltage out that has to be able to deliver 10W, not just any combination of voltage and current that comes to 10W.
Since you seem to be asking about a power line transformer, I'm guessing that you intend to only put a full wave bridge, cap, and maybe a linear regulator on the output. You need the peaks of the AC waveform after the full wave bridge to be a few volts above the target output voltage. This gives room for droop at high load and for the linear regulator to do its job. Figure the full wave bridge will drop 1.5 volts under the full 1A load and maybe 2V for the linear regulator. From this alone the AC peaks need to be at least 10V + 1.5V + 2V = 13.5V. Accounting for droop under high load is more tricky. In theory, the transformer output voltage rating is under full load, but often not specified for the worst case line voltage input. This is where you have to look at the transformer datasheet carefully. Then there will be a voltage drop between peaks as the current is drawn from the storage cap instead of directly from the transformer. So far we need a minimum of 13.5V / sqrt(2) = 9.5V AC sine out before accounting for the drop due to low line voltage and droop between line cycles. It sounds like a 12V transformer is probably the minimum, assuming a reasonbly sized storage cap.
For 60 Hz power line frequency, the storage cap will be charged up at a 120 Hz rate, or every 8.3 ms. Let's say we've budgeted for 2V droop at the full output current of 1A. That means the minimum storage cap is 1A * 8.3ms / 2V = 4.2mF. That's quite a lot, but doable. You can go with that or start with a higher voltage to allow more droop, which would allow for a smaller cap.
So to make a concrete recommendation, something like a 12V 1.5A transformer will likely do it with a big enough storage cap. Keep in mind this kind of power supply will be rather inefficient. The full wave bridge alone will dissipate about 1.5W, and the linear regulator more.
The above tradeoffs are good reasons you don't see direct power line transformers with "dumb" rectifier and linear regulator supplies much anymore. Even in North America, the power line is only 60 Hz, so the transformer will be big, heavy, expensive and the result rather inefficient. Nowadays you put the full wave bridge directly on the AC line, then chop that at high frequency thru a much smaller transformer to make the low voltage on the isolated side. Opto feedback to the chopper can allow the final output voltage to be regulated. This is much more efficient and can use a smaller, cheaper, lighter, and more efficient transformer because it will be operating at 100s of KHz. This is exactly what switching wall-wart type power supplies do.

- 310,974
- 36
- 428
- 915
-
I love your long detailed answers. I wish I could byte them off in chunks as I work. If you had a few headers for different sections it allows any reader to get what they need immediately and it makes a great placeholder for those of us that have 20-30 seconds at a time to browse the site. Maybe a section on deciding sizing, maybe another for explaining efficiency requirements of your regulation. Maybe a section on sag and the bulk capacitors(filtering). Maybe a sum up section. For those of us that want the lessons we could have learned without learning them. – Kortuk Jul 03 '11 at 07:37
-
@Kortuk: You have mentioned this before. While I respect your right to your own opinion, I disagree. This post is only 6 paragraphs and takes up only 2/3 of my screen vertically. I believe headings and extra formatting would clutter up and distract in a post of this short size. I will try to keep this in mind if I find myself writing longer posts. – Olin Lathrop Jul 03 '11 at 13:54
-
@Kortuk, @Olin - I agree with Olin. Also, he always takes the trouble to give detailed answers, which are worth reading completely. It's all useful information. I try to do the same. It would be seriously demotivating if I knew people were too lazy too read them. – stevenvh Jul 03 '11 at 16:34
-
@stevenvh (and on olin's answer it should motivate him), There is not an easy way to grade this. In the past I did some testing. I would write an answer, just like Olin's. I would get a few upvotes. I would edit in section headings, suddenly you get another 8 or 9. I could do this with a rather long delay and it still seemed to function. As I was trying to say for myself, I work in an environment were I constantly get a moment to do something but have to stop almost right away, it is only barely enough time to read a paragraph maybe 2, with section headings I can go from heading to heading. – Kortuk Jul 03 '11 at 23:33
It's always a good practice to buy and use the one which is twice of your requirements,
for example, you need 10 V, 1 A rated transformer and your application or equipment consumes 1 A current but the maximum offered by transformer, now the transformer heats heated up as it becomes saturated and more heat pumps out thereby reduced efficiency of transformer. So always buy the double rated.
And one more thing, I don't know about your location where you live, in my home place, India Transformers are available in terms if ratings rated like these 10 V , 500 mA 10 V , 1 A 10 V , 2 A represented like this
not in VA rating, if only VA rating is mentioned its hard, 20 VA can represnet '10 V 2 A' or '20 V 1 A' along with VA if, Voltage is also mention its fine..
Then as per your requirement 10 V, 20 VA is fine.
Hope this helps.
-
Another thing which should be taken into account is the power supply efficiency. Depending on the rectification and regulation system, it could have considerable impact on the supply design. The Pavan's recommendation of using twice the amount may seem excessive at first, but if you for example need 10 V 1 A DC at the output, and use the simplest circuit for that (say KA7810 regulator and KBL406 rectifier), you'll need more than two volts for regulator, two times 1.1 V for the rectifier bringing the amount to 14.2 V absolute minimum for stable operation. – AndrejaKo Jul 02 '11 at 12:37
-
When mains power voltage drops are taken into account, that quickly goes up to 20 V under load, bringing us to the 20 VA figure. – AndrejaKo Jul 02 '11 at 12:38
-
A filtering capacitor after rectifier and before regulator could help increase the voltage regulator sees though. – AndrejaKo Jul 02 '11 at 12:41
-
1@Pavan - *all* the transformers I've ever used were specified with VA rating and secondary voltage(s), occasionally also current, but not always. The reason for giving the VA rating is that it determines the size. Usually all 10VA types from one manufacturer will refer to a single mechanical drawing. – stevenvh Jul 02 '11 at 14:47
-
@AndrejaKo: Not only will a filtering capacitor increase the average voltage to the rectifier, it is necessary else the input to the regulator drops to zero at least at the line cycle zero crossings. – Olin Lathrop Jul 02 '11 at 17:58
-
When sizing a transformer for a DC load, the AC current rating of the transformer is almost always going to need to be larger than the DC current that's being supplied to the load, sometimes significantly so. How much larger depends on the topology of the power supply used - you can find a handy guide here.

- 2,926
- 1
- 22
- 23
-
That is one handy cheat sheet. Close to what I remember, but I'm under the impression that *peak voltage*s suggested are *no load* for most configurations and *5 to 10 % load* for some (the "choke before capacitor" ones to be specific). While it mentions rectifier voltage drop, maybe emphasise that here lest it gets ignored. – greybeard Dec 31 '22 at 09:59