Negative (low power) switching regulator operations? (For dual supply)

Thread Starter

TheLaw

Joined Sep 2, 2010
228
Hi everyone,

I've been trying to work on a design for a decent lab low power lab power supply, since I do not currently own one. Which method I take has jumped from linear to SMPS, current limited or not current limited, high power or low power....However, I've pretty much determined that I all I really want is a decent low-ish power supply capable of 1A/rail and with an adjustable output.

So I'm back on the SMPS track, but the biggest thing I have an issue with is that I do not know what a good negative switching part would be. Sure there is inverting stuff, but I don't need that since my bridge rectifier would be able to supply both positive and negative DC. I'd like to use matching parts for the + and - rail.

I've been looking at a wide variety of parts ranging from LM2575 to MC34063 and I just do not know how to go about. Granted, I could probably get away with a linear regulator on the negative rail because I'd probably hardly use it anyway, but for the sake of being professional and me being picky, I'd like to try switchmode there as well.

I don't know a whole lot about switchmode design internally, so I don't really know the dos and don'ts.

But, I would appreciate any input.

Thanks.
 

crutschow

Joined Mar 14, 2008
34,289
If I understand you correctly you have a conceptual problem.

One bridge rectifier will not generate a separate positive and negative output since that leaves no common connection for the positive and negative. A bridge rectifier will generate a positive voltage with the negative connection used as common or a negative voltage with the positive connection as common, but not both, unless you have a center-tapped transformer, and that can be used for the common connection.

If you don't have a center-tapped transformer, then you will need to use an inverting switching regulator to get a negative voltage.
 

CraigHB

Joined Aug 12, 2011
127
At one point, I wanted to build a battery driven switch mode regulator that uses a simple inverting buck-boost topology. That's the negative output one that uses only an inductor, rectifier, and switch. I wanted to use it to drive a purely resistive load.

The inverting buck-boost topology is great because it's very simple (component wise) and can put out whatever negative voltage you want. I never could find a controller that can accomodate the current output I was looking for (several amps). It's already hard to design the feedback compensation that many controllers require let alone come up with my own PWM control scheme. I finally gave up on the idea.

If you come across a controller that can do an inverting buck-boost with higher output, I'd definitely like to see it.
 

Thread Starter

TheLaw

Joined Sep 2, 2010
228
If I understand you correctly you have a conceptual problem.

One bridge rectifier will not generate a separate positive and negative output since that leaves no common connection for the positive and negative. A bridge rectifier will generate a positive voltage with the negative connection used as common or a negative voltage with the positive connection as common, but not both, unless you have a center-tapped transformer, and that can be used for the common connection.

If you don't have a center-tapped transformer, then you will need to use an inverting switching regulator to get a negative voltage.
Oi! Sorry there. Yes I will be using a CT transformer. Center tap would be shared ground and a + and a - coming from the diode bridge.
 

crutschow

Joined Mar 14, 2008
34,289
Oi! Sorry there. Yes I will be using a CT transformer. Center tap would be shared ground and a + and a - coming from the diode bridge.
OK. This is a nit, but technically that's not a full-wave bridge configuration, but two full wave rectifier (dual diode) configurations with both of the dual diodes in the one (bridge) package.

Have you checked, for example, National Semiconductor, Linear Technology, and Texas Instruments for a suitable negative switching regulator?
 

Thread Starter

TheLaw

Joined Sep 2, 2010
228
OK. This is a nit, but technically that's not a full-wave bridge configuration, but two full wave rectifier (dual diode) configurations with both of the dual diodes in the one (bridge) package.

Have you checked, for example, National Semiconductor, Linear Technology, and Texas Instruments for a suitable negative switching regulator?
I'm aware it's not full wave bridge. I wrote the original thread in a hurry...so it might be a bit blurred.

Yes I did. There are some that might fit the bill in one department, but can't do other things, like a wide range of voltage selection.

Grr....Am I leaning towards linear again? Hello abysmal efficiency!
 

SgtWookie

Joined Jul 17, 2007
22,230
I put together a couple of simulations using LTSpice for +1.2v to +30v and -1.2v to -30v switching supplies, 2.5A that were really quite efficient; ~80% worst case. They both operate from a single +34v supply.

Have a look at the attached.
The left image is the positive supply, the right is the negative. The positive buck regulator is showing ~11mV p-p max ripple with a 2A load.
 

Attachments

Thread Starter

TheLaw

Joined Sep 2, 2010
228
I put together a couple of simulations using LTSpice for +1.2v to +30v and -1.2v to -30v switching supplies, 2.5A that were really quite efficient; ~80% worst case. They both operate from a single +34v supply.

Have a look at the attached.
The left image is the positive supply, the right is the negative. The positive buck regulator is showing ~11mV p-p max ripple with a 2A load.
Hi thanks.

Can a lower input voltage/current be used? Like perhaps 15V @ 3A or something?
 

SgtWookie

Joined Jul 17, 2007
22,230
Sure, but the output voltage of the positive buck won't go above around 13.3v under load. Change the pot from 20k to 10k if you're going to run it under lower voltage.

The negative supply will still reach lower voltages under light load, so you can leave that one with a 20k pot.
 

Thread Starter

TheLaw

Joined Sep 2, 2010
228
Alright thanks for the input guys. On the LT1171 datasheet it has a schematic called "negative buck". There is also a regular "positive buck". If I have two windings on my transformer, can I just use negative buck to get the negative output and the positive buck for the positive? Did I interpret that wrong?
 

SgtWookie

Joined Jul 17, 2007
22,230
Alright thanks for the input guys. On the LT1171 datasheet it has a schematic called "negative buck". There is also a regular "positive buck". If I have two windings on my transformer, can I just use negative buck to get the negative output and the positive buck for the positive? Did I interpret that wrong?
Here is a link to the datasheet:
http://www.linear.com/docs/Datasheet/117012fg.pdf
The positive supply is basically the "Positive Buck Converter" on the top of page 15.
The negative supply is basically the "Positive-to-Negative Buck-Boost Converter" on the top of page 14.

The negative supply requires a positive input voltage. However, it can go much more negative than the positive supply.

The positive supply is a buck converter. It won't reach the input voltage.

It would be easier to use a single positive supply to power both the negative and positive outputs.
 

SgtWookie

Joined Jul 17, 2007
22,230
Why don't you start off by modifying an ATX or ATSplus12 form factor PC power supply into a bench supply, and then do some blue-sky dreaming about what other kind of supply you'll need?

An ATX form factor bench supply (google that term) is a really cheap way to get a very powerful supply that will actually be somewhat efficient. I built one a number of years back from a used Compaq 250W ATX form factor supply, and it's been quite handy. I also have a variable 0.25v-54v 2.5A Lambda linear supply that's built like a tank; between the two of them I can get by quite nicely. I actually threw out a bunch of other supplies I had because I simply wasn't using them.
 

Thread Starter

TheLaw

Joined Sep 2, 2010
228
Why don't you start off by modifying an ATX or ATSplus12 form factor PC power supply into a bench supply, and then do some blue-sky dreaming about what other kind of supply you'll need?

An ATX form factor bench supply (google that term) is a really cheap way to get a very powerful supply that will actually be somewhat efficient. I built one a number of years back from a used Compaq 250W ATX form factor supply, and it's been quite handy. I also have a variable 0.25v-54v 2.5A Lambda linear supply that's built like a tank; between the two of them I can get by quite nicely. I actually threw out a bunch of other supplies I had because I simply wasn't using them.
Yeah I've seens tons of them. Not a bad idea really, but I'd rather have something for the (semi)long term. Might not last me until the day I die, but something useful if you you know what I mean.

Though 12/5/3.3 are pretty useful voltages...

Hmmph. Let's do some self-examining.
 
Top