1. Home
  2. Electronics Design
  3. silly power supply

silly power supply

I'm doing an arbitrary waveform generator that uses a bunch of THS6022 monster opamps for the outputs. I have +-12 volt VME supples that could droop as low as 11.5 maybe. I need +-10 volts of output swing into 100 ohms, and the supplies are literally a couple of tenths of a volt away from working. How annoying.

Various boost regs look like overkill to do this, to boost the supplies a half volt or so. This dumb open-loop thing might work. I can get the 5v p-p square wave off the +5 supply, chop it up with a couple of mosfets.

This is another use for those cute Coilcraft dual-winding inductors. I can get the schottky bridge as a single surface-mount part.

The negative side is the same idea, using the same 5v p-p source.

Version 4 SHEET 1 1924 740 WIRE 80 -688 -16 -688 WIRE 256 -688 144 -688 WIRE 384 -688 256 -688 WIRE 544 -688 448 -688 WIRE 256 -656 256 -688 WIRE -16 -560 -16 -688 WIRE 80 -560 -16 -560 WIRE 256 -560 256 -576 WIRE 256 -560 144 -560 WIRE 384 -560 256 -560 WIRE 544 -560 544 -688 WIRE 544 -560 448 -560 WIRE 784 -560 544 -560 WIRE 880 -560 784 -560 WIRE 960 -560 880 -560 WIRE 544 -528 544 -560 WIRE -16 -496 -16 -560 WIRE 784 -496 784 -560 WIRE -16 -368 -16 -416 WIRE 544 -368 544 -464 WIRE 784 -368 784 -416 WIRE 192 -192 -16 -192 WIRE 368 -192 256 -192 WIRE 544 -192 448 -192 WIRE -16 -144 -16 -192 WIRE 544 -128 544 -192 WIRE -16 -16 -16 -64 FLAG 784 -368 0 FLAG 880 -560 POS FLAG -16 -368 0 FLAG -16 -16 0 FLAG 544 -128 0 FLAG 544 -368 0 SYMBOL ind2 352 -176 R270 WINDOW 0 -47 57 VTop 2 WINDOW 3 -57 56 VBottom 2 SYMATTR InstName L1

SYMATTR Type ind SYMATTR SpiceLine Rser=0.33 SYMBOL schottky 384 -544 R270 WINDOW 0 32 32 VTop 2 WINDOW 3 0 32 VBottom 2 SYMATTR InstName D1 SYMATTR Value RSX205L-30 SYMATTR Description Diode SYMATTR Type diode SYMBOL cap 528 -528 R0 WINDOW 0 65 34 Left 2 WINDOW 3 61 72 Left 2 SYMATTR InstName C2

SYMBOL current 784 -496 R0 WINDOW 0 78 6 Left 2 WINDOW 3 -168 174 Left 2 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName I1 SYMATTR Value PULSE(64m 800m 5m 0 0 5m) SYMBOL ind2 240 -560 M180 WINDOW 0 -44 75 Left 2 WINDOW 3 -50 40 Left 2 SYMATTR InstName L2

SYMATTR Type ind SYMATTR SpiceLine Rser=0.33 SYMBOL voltage -16 -512 R0 WINDOW 0 57 77 Left 2 WINDOW 3 56 115 Left 2 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName V2 SYMATTR Value 11.5 SYMBOL voltage -16 -160 R0 WINDOW 0 84 74 Left 2 WINDOW 3 50 129 Left 2 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName V3 SYMATTR Value PULSE(0 5 0 50n 50n 950n 2u) SYMBOL schottky 384 -672 R270 WINDOW 0 32 32 VTop 2 WINDOW 3 0 32 VBottom 2 SYMATTR InstName D2 SYMATTR Value RSX205L-30 SYMATTR Description Diode SYMATTR Type diode SYMBOL schottky 80 -544 R270 WINDOW 0 32 32 VTop 2 WINDOW 3 0 32 VBottom 2 SYMATTR InstName D3 SYMATTR Value RSX205L-30 SYMATTR Description Diode SYMATTR Type diode SYMBOL schottky 80 -672 R270 WINDOW 0 32 32 VTop 2 WINDOW 3 0 32 VBottom 2 SYMATTR InstName D4 SYMATTR Value RSX205L-30 SYMATTR Description Diode SYMATTR Type diode SYMBOL cap 256 -208 R90 WINDOW 0 -14 31 VBottom 2 WINDOW 3 46 29 VTop 2 SYMATTR InstName C1

TEXT 824 -64 Left 2 !.tran 15m TEXT 760 -160 Left 2 ;V377 Pos Rail Booster TEXT 792 -112 Left 2 ;JL Jan 10, 2017 TEXT 208 -384 Left 2 !K1 L1 L2 0.995 TEXT 48 -224 Left 2 ;5V p-p TEXT 840 -448 Left 2 ;LOAD TEXT 240 -344 Left 2 ;DRQ127

I guess I could also buy some standard 12-to-3.3 volt dc/dc converters and stack the 3.3's on top of the 12's.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics
Reply to
John Larkin
Loading thread data ...

There are a ton of really cheap boost converters out of the far east on ebay. Cheaper than the parts shipping cost if you try to build it in house.

formatting link

If they fail to work, you get your money back.

Have fun.

Reply to
tom

What's a VME supply?

--

Rick C
Reply to
rickman

For

formatting link
?

--
Adrian C
Reply to
Adrian Caspersz

A freaking 30 year old bus??? My first thought was why didn't he just get a supply that would provide the current needed without the half volt of droop? I suppose he is retrofitting some antique gear.

--

Rick C
Reply to
rickman

Here's my solution to this common problem. It relies on coupled inductors with their normal very tight coupling. I originallly needed to convert common 12V silicon MOSFET driver levels to +16 and -4V, to drive SiC MOSFETs. I had very little space available and wanted a simple fix. I used an LT3467 in a sot-23 package and a small srf0703 coupled inductor (it's running at 1.3MHz). Use a larger low-value inductor if you need more power. A simple re- arrangement makes this work for +/- 15V supplies from 12V.

formatting link 

--
 Thanks, 
    - Win
Reply to
Winfield Hill

I tried a few similar circuits,

formatting link

using parts that we have in stock, but they are relatively complex and the transient response of the non-regulated side is sloppy. My load can jump from around 60 mA (opamp idle current) to close to 1 amp. That argues for the + and - supplies to be independent.

The open-loop thing is sure simple, but I'm still open to suggestions.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics
Reply to
John Larkin

Yes, I'm redesigning an old (originally 1993) VME module. My old design had 7 FPGAs and a 68K CPU, and a mountain of analog stuff. The new one will use one big FPGA and a ARM chip and do most of the processing (DDS, summing, filtering, modulations) digitally.

This is the old one:

formatting link

The new board will be practically empty. The FPGA, CPU, four DACs, output amps, a bit of power supply. I'd use a ZYNQ SOC, but they don't have much on-chip SRAM for some reason. The power supply is a little interesting; the rest is mainly grunt work.

The history of VME is interesting; some other time.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics
Reply to
John Larkin

How does that help if you are running out of power on the 12V rail... (or is it only the drooping to 11.5 that is hurting you?) Maybe DC-DC converter from the 5V rail? (So far I'm liking the CUI three terminal 500mA DC-DC converters. they make 1.8 V ones $5 no muss, no fuss)

George H.

Or tweak the specs 9Vp-p into 100 ohm, or 10 Vp-p into 150 ohms.

Reply to
George Herold

There are plenty of amps available, but not quite enough voltage for the power opamps to reliably swing +-10 into the loads.

One idea is to get a couple of isolated dc/dc converters, 5 or 12 or

24 volts input, a couple of volts out, and stack their outputs in series with the existing +12 and -12 rails. But I need about 1 amp, and the only converter choices are expensive and sole-sourced.

What I need is a 1-volt output, 1 or 2 watt isolated dc/dc.

CUI does make good stuff.

Too late! I need to be compatible with the old product, +-10 behind 50 ohms. It used four paralleled RRO opamps per channel, but they were slow, and some customers want higher speed outputs. THS6022 is ideal, screaming fast and tons of current, and we have them, but they are not RRO.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics
Reply to
John Larkin

I have taken measurements on real circuits that show amazing regulation of the non-regulated side. It's a matter of the quality of your coupled inductors. Also, assuming you have +/-12V, keep in mind you're only making a pair of 2-volt supplies.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

Hmm I see more than one on DK. Some by Recom power. And then these

formatting link
(but not a lot of stock at DK, which always worries me.) Oh Cui has some too...
formatting link

That should do it!

George H.

Reply to
George Herold

Your output capacitors are too small.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

The -12 supply can actually be -11.6, maybe even a bit lower with connector and trace losses, so margin is even worse than it looks. I already have 330uF.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com
Reply to
John Larkin

Why not just change a resistor in the power supply to get a half volt more out of it? They might even have an adjuting pot with that much range.

Jon

Reply to
Jon Elson
100 ohm load? That sounds like a ~zero ohm buffer, followed by a 50 ohm source-termination resistor, followed by a 50 ohm load. Why not cut out the middle man, and reclaim nearly 50% of your supply voltage at the same time?

Tim

-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website:

formatting link

Reply to
Tim Williams

You mean not be a 50 ohm source? Too late, I'm updating an existing product design.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com
Reply to
John Larkin

OK, I had a big lunch and a drink, but I don't see how to use one of those to make +12 into something a couple of volts higher.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com
Reply to
John Larkin

Well, you can also do an unbalanced Howland to make any Zout you like while wasting less supply headroom, but resistor errors get multiplied if you try to be too fancy that way. I nearly always do the 50 ohm resistor thing myself.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net
Reply to
Phil Hobbs

It helps the current limit situation, too. Two 2010 100r resistors in parallel will dissipate 1 watt each into a short, which is tolerable with a bit of copper pour. My rise time will be around 140 ns, so real

50 ohm source termination is appealing.

THS6022 is a fabulous amp; I just need a little more supply voltage.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com
Reply to
John Larkin

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.