You can look up the diagrams of some commercial units. Or design your own.
Start with a power supply that will deliver at least 55 V (at ripple troughs) at full load current of 5 A. Put in a series regulator stage, probably several transistors in parallel on a heat sink (with emitter equalizing resistors). Drive with a reverse polarity transistor.
Drive said reverse transistor with another transistor, either a differential pair or an opamp or just a single transistor. Use zener as reference, pot to control other input. Or the classic uA723.
The only problem with this approach is that it won't take the output to zero. You can devise various methods to do that if it's needed, like tying the reference to a negative supply.
Start with a power supply that will deliver at least 55 V (at ripple troughs) at full load current of 5 A. Put in a series regulator stage, probably several transistors in parallel on a heat sink (with emitter equalizing resistors). Drive with a reverse polarity transistor.
Drive said reverse transistor with another transistor, either a differential pair or an opamp or just a single transistor. Use zener as reference, pot to control other input. Or the classic uA723.
The only problem with this approach is that it won't take the output to zero. You can devise various methods to do that if it's needed, like tying the reference to a negative supply.
This one is a single polarity 0-50V. The example shown is configured as a 0-2A, but it can be adapted to another current value by modifying R9 on the schematic.
Further down the thread, some accessories are described, such as automatic winding switching, and also the way to couple two units to make a +/-50V supply.
This circuit has the advantage of requiring no auxiliary supply.
http://forums.futura-sciences.com/p...mentation-de-labo-un-design-de-reference.html
Further down the thread, some accessories are described, such as automatic winding switching, and also the way to couple two units to make a +/-50V supply.
This circuit has the advantage of requiring no auxiliary supply.
http://forums.futura-sciences.com/p...mentation-de-labo-un-design-de-reference.html
Last edited:
Very innovating "Lab psu", how ever there's a much easier way, use a trafo. with a secondary of 2x36 - 40VAC, a rectifier bridge, 10mF 100V cap's and a VARIAC, on the mains..... They can be hard to get, but look on e-bay.... 3A 0 - 250VAC(or 115VAC US), and the "unregulated" traf of 2x36V~, which becomes about 50VDC rectified and noise free(bigger cap = less 'hum'), and 3A times 250VAC, comes to 750VA, which ought to be enough(but you'll need a pass transistor if you wanna ctrl the current, and the circuit looked kind of tempting...)
The LM10CH application notes include a floating lab power supply regulator. Motorola used to make a chip that was specifically designed for lab power supplies, but it's been discontinued for some time. Still, the application note may be useful since I'm pretty sure it talks about how to choose the power transistors.
Here's a power supply that recreates the Motorola MC1466 chip using discrete devices:
Super Power Supply, Current and Voltage Regulated
Here's a power supply that recreates the Motorola MC1466 chip using discrete devices:
Super Power Supply, Current and Voltage Regulated
Last edited:
For testing amplifiers I have used a Variac on a dedicated amplifier power supply -- 600VA transformer from Antek and 25,000uF on each rail. I took the amplifier parts out!
Sometimes HP 6271, 6274 rack mount supplies can be found at hamfests.
LM12CLK is still available... it can handle 50V@2A... and has a nice simple LAB-psu application on the datasheet.
This will do just fine -- HP6129C 0-50V, 5A Continuous:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Nice!!!
Very very very VERY nice! HP's test equipment has always been my kind of flavour, and B&K...
I'm in the "start blocks", to begin a new project, a dual Lab-PSU. I have spent many hours with Pspice(OrCAD v.16) to come up with a high V/A dual linear power supply, with all the necessary "trimmings".
Like(except the obvious feature, V-adj.), ampere adjustment, over voltage - i.e, if you want to drive an old HAM Radio equipment you don't want to get a spike - Digital volt/ampere display, a AC-millivolt meter to monitor noise, short circuit prof, and more...
I'm going to write a "diary" 'bout this project on my blog, the "heart" of the bench psu, the trafo, came with the post last week. It's ring trafo, 2x42VAC@650VA, it has 10-0-10 and 12V~, as well and is necessary for the project, LAB-PSU.
It's an "all discrete" design, I use a couple of OP-Amps for monitoring, and ICL7107 for voltage and ampere...
It(or I'm aim at...) will be capable to deliver +/- 50VDC, @ 7 - 10A per rail.... 6xMJ15003/15004 parallel coupled, output BJT's, MJ1500x is capable of 250W, 120V, 20A each, so about 1.35A@50V each, comes to approximate 70W, so cooling shouldn't be a problem...
But one might find a used unit, slightly defect on E-bay, that's worth while to repair... but it is more fun, building your own, or?
Very very very VERY nice! HP's test equipment has always been my kind of flavour, and B&K...
I'm in the "start blocks", to begin a new project, a dual Lab-PSU. I have spent many hours with Pspice(OrCAD v.16) to come up with a high V/A dual linear power supply, with all the necessary "trimmings".
Like(except the obvious feature, V-adj.), ampere adjustment, over voltage - i.e, if you want to drive an old HAM Radio equipment you don't want to get a spike - Digital volt/ampere display, a AC-millivolt meter to monitor noise, short circuit prof, and more...
I'm going to write a "diary" 'bout this project on my blog, the "heart" of the bench psu, the trafo, came with the post last week. It's ring trafo, 2x42VAC@650VA, it has 10-0-10 and 12V~, as well and is necessary for the project, LAB-PSU.
It's an "all discrete" design, I use a couple of OP-Amps for monitoring, and ICL7107 for voltage and ampere...
It(or I'm aim at...) will be capable to deliver +/- 50VDC, @ 7 - 10A per rail.... 6xMJ15003/15004 parallel coupled, output BJT's, MJ1500x is capable of 250W, 120V, 20A each, so about 1.35A@50V each, comes to approximate 70W, so cooling shouldn't be a problem...
But one might find a used unit, slightly defect on E-bay, that's worth while to repair... but it is more fun, building your own, or?
Last edited:
Joaquim:
I think that a simpler design strategy is to use a 500 watt Class-D amplifier and drive a power transformer, rectify and filter -- you can use an analog opto-coupler to provide feedback -- this way you burn many fewer watts with minimal energy dissipated in pass devices. I'd spend the time programming a DAC with a modified sine wave to boost the efficiency but spare the design the switching transients of a DC-DC converter.
Two of the HP supplies were purchased surplus from an aircraft/defense contractor in NY for $1, but I had to pick them up!
I think that a simpler design strategy is to use a 500 watt Class-D amplifier and drive a power transformer, rectify and filter -- you can use an analog opto-coupler to provide feedback -- this way you burn many fewer watts with minimal energy dissipated in pass devices. I'd spend the time programming a DAC with a modified sine wave to boost the efficiency but spare the design the switching transients of a DC-DC converter.
Two of the HP supplies were purchased surplus from an aircraft/defense contractor in NY for $1, but I had to pick them up!
Those HP6129's are just amplifiers -- uses a DAC made of discrete components -- each bit is trimmable.
I am just pulling a leaf from the book I worked in decades ago -- we used Mac's with 6L6's to power the ultra-low noise stuff in the lab. No reason it couldn't be implemented with some DSP at a higher switching frequency.
Well I've paid a smaller fortune for the trafo - 650VA... But I was going to build a SMPS as a LAB PSU at first... I have a 0-250V~ 3A Variac, to which I can hookup any trafo and adjust the output, I have a circuit board with a rectifier B1000C25000 and 100V, 10 000µF cap's on it, and old-school stabilization, two 40W light bulbs
And it's been used to death, or almost anyway... the panel instruments, V/A, has give in, and sure I can use a DMM to monitor the voltage/current...
But I'm a bit "goofy" that way, I want to build everything analogue and with discrete components if possible...
I work with AVR/PIC C programming day out and day in... and it's cheap, easy(if you can C or Assambler), one MCU say ATxmega - Almost any model - can both switch, drive a LCD, "show" noise, draw up graphs and more, and all that for less then €10 + x-hour of programming, but most code you can copy of the net, and modify to fit your purpose. And templates and other pre-done code, comes with AVR Studio SP2. I've used ATmega32 16MHz clock to almost everything, the draw back is that it has no DAC's and UARTs, in all other way it's a great easy to use MCU. I recommend it as a "beginners" circuit. It has JTAGICE(mk I) interface and is fully supported by AVR Studio 4 SP2, for "xmega" you'll need the JTAG ICE mk II, ATMEL's own is recommended, it has all features, Programmer/Debug/IPS etc. Now I'm way of topic.. sorry 'bout that.
It will be a linear all discrete components design! Maybe l8 on I'll get "down and nasty", with an all "digital" SMPS Lab supply... I can always start with programming the MCU, I reckon it'll take many hours behind the keyboard before I get it right... when you think you've got just right there's some bug in it, "to err is human", or ;-)
And it's been used to death, or almost anyway... the panel instruments, V/A, has give in, and sure I can use a DMM to monitor the voltage/current...
But I'm a bit "goofy" that way, I want to build everything analogue and with discrete components if possible...
I work with AVR/PIC C programming day out and day in... and it's cheap, easy(if you can C or Assambler), one MCU say ATxmega - Almost any model - can both switch, drive a LCD, "show" noise, draw up graphs and more, and all that for less then €10 + x-hour of programming, but most code you can copy of the net, and modify to fit your purpose. And templates and other pre-done code, comes with AVR Studio SP2. I've used ATmega32 16MHz clock to almost everything, the draw back is that it has no DAC's and UARTs, in all other way it's a great easy to use MCU. I recommend it as a "beginners" circuit. It has JTAGICE(mk I) interface and is fully supported by AVR Studio 4 SP2, for "xmega" you'll need the JTAG ICE mk II, ATMEL's own is recommended, it has all features, Programmer/Debug/IPS etc. Now I'm way of topic.. sorry 'bout that
.It will be a linear all discrete components design! Maybe l8 on I'll get "down and nasty", with an all "digital" SMPS Lab supply... I can always start with programming the MCU, I reckon it'll take many hours behind the keyboard before I get it right... when you think you've got just right there's some bug in it, "to err is human", or ;-)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.