Couple of transformer questions

Question

I'm looking at building a bench power supply following Peter's excellent tutorials on this and have a couple of questions on sourcing a transformer (maybe 2):

1) I want to be able to provide a 30V DC rail and a 12V DC rail and I'd assumed that it must be possible to find a transformer that could provide different tappings to achieve this. I suspect I'm either not looking in the right place or asking the wrong questions of the search engines. I have find a couple that would provide, say, 0 - 6/13/18/26Vac but these seem to be a single secondary, i.e. it isn't tapped to provide both a 13V (say) and 26V. Could someone point me in the right direction in tracking something down - I'm not asking you to do the work for me of course. If I'm asking the impossible, I suppose a suitable approach would be to build it with two transformers?

2) Based on your experience, would I be right in thinking that something like a 25Vac secondary would provide me with enough headroom for 30Vdc. My thinking is that 25Vac is RMS so pre-rectification I could be looking at 25x1.414 = 35.35v; post-rectification with a voltage drop of, say, 1v = 34.35v, then as long as further voltage drops through transistors etc is controlled (in that, I don't believe I will be pushing through enough of them to cause 4v drop) I could achieve 30V?

I would be happy to build this up in stages and do the necessary testing at each stage but I am hoping to not waste bucks on buying something and finding that I needed something a smidgen more powerful!

Thanks,

Andrew

michaelkellett · Answer

The situation with transformers is a bit more extreme than even John suggests: Take a 25V AC nominal winding, this will be at nominal input voltage and full load. The regulation of the transformer will be stated somewhere, that's the difference between no load and full load - for the Farnell 1785741 (just an example - not suitable for this job) it's 9% The mains voltage range in the UK is -6 -> + 10%. So the maximum output at no load from the transformer is 25 * 1.1 * 1.09 = 29.98V. The peak will be 42.4V. This means that you need capacitors rated at at least 47V (always give the electrolytic cap voltage rating at least a 10% margin above the basic worst case !!) The rectifiers will only drop 0.7V at low load so yo can expect to see 41V on the input of the regulator. At full load the ripple will depend on your capacitors and transformer impedance but we can make a stab at it. Working on a maximum 3A DC load and using a reasonable 4700uF capacitor we can make an approximation of the ripple by assuming the capacitor charges to the peak voltage and discharges for 1/4 cycle (this isn't accurate - to be accurate you need to model the transformer. diode and capacitor impedance). AT 50Hz 1/4 cycle = 5ms so the ripple will be 0.005 * 3 /4700e-6 = 3.2V. (3A is pushing the VA rating of the transformer a bit - at nominal mains you are taking about 101W out at DC - and it will be higher at +10% mains. At low mains and full load the minimum voltage will be (25 * 0.94 * 1.414) - 2 -3.2 = 28.02 - so not enough for a 30V rail. You can take these calculations round the loop again - you'll end up with a 30V winding to get a reliable 30V output with a little headroom for the regulator at worst case low mains - and you'll need 63V capacitors and dump a lot of heat with high mains. MK

three-phase · Answer

Hello Andrew,

There are a couple of ways to approach this dependent upon what the load / isolation requirements are for the supply.

You can use a transformer with a 30V secondary and use a DC converter / regulator to then drop down to the 12V, but this will dissipate a fair bit of waste heat. There may even be options out there to use a 12V secondary and boost up to the 30V with a DC booster circuit. There are plenty of voltage regulator modules available on eBay, jus search for the desired voltage output and select one with an input range to suit.

Example 12V regulator / buck booster

You could use a transformer with a 15 - 0 -15 output and leave the two secondaries in series. The 30V is then achieved across both secondaries and then tap off the 15V from one of the secondaries. Both voltages would require regulating and you will need to ensure that the transformer VA rating is sufficient for both supplies.

Example 15 - 0 -15 transformer

With regard to the voltage rating and achieving the voltage output, that will depend upon the transformer rating relative to the load you put on it. At full load, the transformer will be designed to give out the rated voltage. At low loads this will rise and may cause damage to the circuit if not regulated against. if you use a transformer that has a higher VA rating than required, then the voltage drop will be less and you will be achieve a higher voltage output.

Whilst with bridge rectification and smoothing capacitors you can achieve a low load DC voltage greater than the nominal rating, up to the peak voltage level, as the transformer reaches full load, you will find that this will start to drop off.

Full wave rectifier tutorial

I am sure a few on here, with more experience than I, could provide some alternative solutions / products.

Kind regards

Andrew J · Answer

Thanks Donald,

I think in respect to providing the two rails, I had thought about tapping off one of the secondary windings but I may have been overthinking it (as per usual for me!) I can use a transformer, say 15-0-15 in series to provide the main power out, and hook up something like a PCB mount AC/DC PSU (example: https://uk.farnell.com/mean-well/irm-03-12/power-supply-ac-dc-12v-0-25a/dp/2815483). That way I can provide the AC to both but keep them isolated; only the transformer needs rectifying plus the AC/DC PSU is regulated. They're a reasonable price but I thought it should be possible to get a transformer that would have a high and low secondary windings, however I think keeping it simple has a lot going for it!

Power management was/is one of the concerns, and dropping from a high voltage down to 12v, even at low milliamps, could still be 4-5 watts, so this would seem to win on that score too.

What do you think?

That tutorial is very interesting and useful. I do have some understanding of rectification but that really cements it more.

Thanks,

Andrew

jw0752 · Answer

Hi Andrew, Here are a few observations. If you are using a full wave bridge you will likely loose at least 2 volts between the 2 diode drops. If you are using a linear voltage regulator to control the voltage down from there you will also have some drop out. If you want to minimize this diode junction loss check out the blog I did on the LT4320 https://www.element14.com/community/people/jw0752/blog/2017/03/03/incorporating-the-lt-4320-ideal-bridge-in-the-linear-bench-supply which uses MOSFETs to rectify the AC instead of doides. If you have a 26 volt secondary you will have approximately 33 volts DC after rectification and filtering but this is a no load voltage and will be drawn down by what ever load you apply to it. Once this loaded AC voltage get to the dropout voltage of your regulator you will also loose regulation. On my power supply builds I find that I will not be able to provide, for example 28 volts at 1500 mA, due to the load pulling the AC of the transformer below 30 volts. There are many DC-DC converters that will provide a second rail for power supplies like this but very few provide isolation. In most instances the ground will remain common to your two rails. I found that for isolation it was necessary to have 2 secondaries or two transformers. I found some toroidal transformers that were close to the specifications that I needed but not quite there. By adding additional windings to the transformers I was able to bring their voltages up to the 24 volts AC that I was looking for. Toroidal transformers are great for this as they are open in the middle and provide room for the extra windings. I believe that the power supply manufacturers must order in bulk their transformers according to their own specifications as I have not been able to find a good source for the transformers that I needed for my bench power supplies either. I have also experimented with using a Switch Mode 35 Volt power supply as the base source of power and then used a linear regulator circuit to allow me to dial down the voltage that I need. This has worked well as the internal regulation of the SMPS helps minimized the effect of loading on the base 35 volt supply. John

three-phase · Answer

Hello Andrew, As far as I can see, your proposed solution would work. You should only then need to design the regulation for the 30V rail as per Michael's and John's responses. That little PSU you have found seems to be quite good value for money. Kind regards.

Andrew J · Answer

Thanks John. Interesting post - you say that there wasn’t enough advantage to using the mosfets over the diodes in most cases - is that something you still think? I’ll do some more planning/designing on this I think and work out the best, and cost effective, way of providing the control circuitry rail. Some very helpful information provided by you and Donald. It’s certainly an interesting project. Regards, Andrew