Pros and cons of different rectifier designs

Status
Not open for further replies.
P

PhoenixOSU

Member
I am building a power supply for personal use (and for my own education), and am faced with the question of which rectifier and transformer combination to use. I will need at least 620W available out of the rectifier/filter. There will be a great deal more circuitry after the main rectifier, an H-bridge, some isolation transformers, and buck-boost converters.

I know that I should only use 90% of the rated VA, and am going to assume a 1W loss in the transformer and a diode drop of 0.8V. These are pessimistic assumptions, and I have looked at some diodes that I believe will make excellent choices. The MBR60100 will suffice.

I would like to get around 22-28Vdc out of the rectifier/filter to work with.

My design options are to use the two outputs of the transformer in parallel and build a bridge rectifier, the down side is the two diode drops, but the upside is the transformer can more easily handle the current this way. The other option is to put the two outputs in series and use the middle connection as a center tap and use a full wave rectifier.

The questions I am having with this are that it will mean that at any time only one of the two coils is conducting, will this cause a loss in efficiency? And with the use of the center tap, I am drawing twice the current through the coils as before, requiring me to get a bigger transformer, however, since the current is only present half the time, can I say the current is effectively the same as for the standard method of putting the coils in parallel?

I would like to use a full wave rectifier to reduce loss due to voltage drop, but the fact that it seems it would double the current through the coil means I may need a bigger transformer, which gives me pause in this decision. What are your thoughts/experiences on this?
 
H Bridge = Halfwave.
Clueless why anyone would use a half wave with a center tapped transformer. I have built a lot of DC power supplies for Ham and can tell you are spinnigng your wheels stuck in in the mud with your approach.
 
I concur with dereckbc. Center tapped transformers use a full-wave bridge arrangement. There is an absurd amount of information on how to design a TR (Transformer-Rectifier) circuit on the internet. I will guide you to http://www.belfuse.com/SignalTransformer/ and suggest you download the Product Catalog and start reading at page 35.
 
I have been considering thier MPI-650 for some time now. And as for an H-bridge, that is refering to the switching mechanisim for the high frequency isolation transformers, H-bridge is an arangement of 4 MOSFETs which reverse the polarity of the output at a given frequency to drive a transformer. I've decided to just go with a full wave rectifier from all my transformers, since it makes the most sense.
 
PhoenixOSU, what does the OSU stand for? OK State? If so hello fellow alumni.

What is the application of the DC supply? The easiest to design and build is a standard linear center-tapped transformer, bridge, brute force filter, and a regulator circuit using an LM-318 precision voltage regulator followed by Darlington transistors. They are not efficient but work very well and easy to work with. For a home-brewer they are the least expensive.

On the flip side are the switch-mode rectifiers being built today. They are lighter, more efficient, and less expensive to manufacture because the transformers being used at high frequencies. However very difficult to design and expensive for just one prototype model to fool around with.

However if you really want to do it, ARRL has an excellent design with photo-etched boards. Granted the design is already done, but you will get a lot out of it winding the transformers and building it.

I have to DC supplies. One is a traditional linear 13.6 VDC rated at 40-amps I use to power and repair Ham radio mobile equipment, and the other is a 27 VDC switch mode rated at 50-amps for commercial radio application. But to be honest you can buy one at less expense than you can build one.
 
The OSU stands for oregon state, sorry.

I am building the supply for the purpose of building one, because i've never done it before and building one is a good way to learn how they work.

I have speced out what I want at 4 isolated channels rated to 120W each using buck-boost converters. And probably a maximum of 10A out at 12V.

My original design idea was to take my input power into a 650W transformer rectify and filter then go through an h-bridge dc-dc converter to 4 1:1 isolation transformers, from there rectify again and filter again then off to the 4 buck-boost converters. the buck boost will be controlled with a PIC microcontroller, that way I can program it to behave however I want, potentially even serial control, though I don't think I will go that far.

I have decided that it may be a good idea to forgo the first stage and go directly to the 4 isolation transformers from the line power, then rectify, filter, and buck-boost regulate. this means I don't have to have an EMI filter before that h-bridge, finding an EMI filter that can handle that kind of current may be challenging, but there is one used by my employer that is very good at works at 15A.

My big concern is that a full wave rectifier using the center tap is only using half the TX at any given time, whereas a bridge uses all of it at any given time, this means that to attain the same power on any given TX the current in the active coil is twice that as it would be in a bridge setup, this worries me.

As for cost, I am trying to keep it reasonable, one thing I noticed is that high frequency transformers are dirt cheep, and low frequency transformers are very expensive at 650W they can get up to the $150-$200 range. but a good 150W transformer is only $35.

My other option is to build my supply off-line, which has some inherent risks, though the output would still be isolated by the 4 isolation transformers.
 
Do you really need a buck boost configuration? I doubt it since you are starting with a line voltage, a buck should work fine. And the buck topology is easier to stabilize. A 650W 60Hz step down transformer will be expenisive and totally unnecessary since there are so many available switcher IC's available today which let you use the smaller high frequency transformer.

You didn't mention if efficency is a design concern. Even at 95% efficency you could dissipate 32.5W in the power supply and that will require some type of air cooling and heat sinks for the MOSFETs. A full bridge configuration with H-bridge is almost a must to take full advantage of the core's B-H curve, which allows a smaller transformer.

Is this your first power supply design? It is a bit ambitious.
 
Yes, it is my first.

I had originaly wanted to go direct from line AC into a bridge, then into some filters, to get about 115-110Vdc, into a buck, but that means my buck will be at less than 50% duty cycle a lot of the time, loosing efficiency as you go down in duty ratio... and I want isolated outputs so I can put them in series if I like without shorting out one of them. so the idea of isolation transformers came in. my co-worker every time i ask, insists I don't try off-line stuff. I have considered taking the off line power of 110vdc into an h-bridge then into 4 isolation step down transformers, after which i can do a buck-boost. the reason I like buck-boost is we have some good designs here already made for those.

I have noticed that 650W 60Hz transformers are spendy, compared to high frequency transformers. Just what are the risks and extra safety steps that go into an off-line h-bridge powering isolation transformers?
 
By selecting the winding ratio to step down the secondary voltage you have more freedom to choose an appropriate duty cycle, so you don't need some ridiculously small duty cycle to buck 120V to 5V for example.

The biggest issue is the need for rectifier diodes and MOSFETs that can withstand ~170V, more expensive but they are available. Also the power supply must self-generate its own low voltage supply since that doesn't exist at startup, however there are IC datasheets that come with helpful example circuits. http://www.linear.com is a good start.

The biggest risk is that you are working with line voltage on the switching side whereas a stepdown transformer lets you more safely poke around in the circuit. See http://focus.ti.com/lit/an/slyt126/slyt126.pdf for pros and cons. There are issues with UL certification when line voltage is inside a product, which wall bricks bypass. This may be what concerns your co-worker.

It sounds like you have some proven circuits to start from instead of using a clean slate approach, in which case it may be simpler and cheaper (time is money) to do what you describe.
 
I?m not too concerned about UL requirements since I?m not selling this thing, it?ll just be for myself and won?t be on 24/7. But I will include EMI filtering etc to protect my other appliances. I may even take it down to OSUs motors lab and put it on a power analyzer to determine what kind of PFC I need if any, as a later option. The good thing about building it myself is I can upgrade stuff like that without too much hassle. I will also have fuses on the input to the PSU and the input to each of the 4 channels after the isolation transformers, all of which will be mounted on a fixture on the back of the PUS, covered by plastic of some kind, but easy access for replacement, all the metal work should be no problem for me.

I guess there are a few approaches I can take to get power to the ?channels?, I am going to be using 5 boards, 1 main board that takes the input power and splits it off into 5 isolated channels, and it?ll have some rectifiers and filters before sending power to each of the 4 channel boards. They will connect with heavy duty molex connectors; I can put mounting holes in the PCBs so they all fit nicely.

The first is I can take line power rectify it and then go thru an H-bridge into high frequency isolation transformers.

The second is I can take line power, not rectify it and take it into 4 60hz step down transformers (about 150VA rated each)

The third is using a single isolation transformer about 650-700VA then going into an h-bridge then into 4 high frequency isolation transformers. This was the idea I have been working on for a few weeks, but I?m finding that the alternatives are looking more attractive since this one involves a lot of power losses.

High frequency transformers are cheep, but the 150VA step down transformers aren?t too bad either, about $36 each.

I don?t want to design for cost, but for elegance, and to meet my objectives. I?m obviously still on the main board design, which seems to be more complex then expected since I?m doing it from scratch.

My problem with wall brick is that it can be more expensive then just getting a transformer and putting it in the box, they are big and heavy, but I don?t mind that, frankly weight is good, since it wont slide around when I plug stuff in.
 
Since you need four independent supplies, the second option seems good, you can use lower power components and the 150VA transformers are inexpensive (and a lot less heavy). Off-line design may be more trouble that you want at this point. The stepdown/off-line decision is mainly based on cost. With semiconductor costs dropping, off-line switchers become attractive from a weight and efficiency perspective, and PFC is easier to implement.
 
I agree about option 2, because it greatly simplifies the board, and I have a 144W (12V 12A) optimized design for a buck-boost converter from linear tech that I can use, from there there will be more investigation as to what I can add to it in terms of functionality. It also seems best to just use full wave rectifiers and a center tap on the transformers, then I greatly reduce the size of the mainboard and can at last start working on the buck-boost boards.
 
Status
Not open for further replies.
Top