Author Topic: Power Brick for DIY Linear Power Supply  (Read 6130 times)

0 Members and 1 Guest are viewing this topic.

Offline rstor22Topic starter

  • Regular Contributor
  • *
  • Posts: 104
  • Country: ca
Power Brick for DIY Linear Power Supply
« on: December 05, 2014, 06:56:06 pm »
Is there any issue or problems one may encounter if they use a power brick (such as a SMPS that belongs to a printer/scanner/laptop) to power a DIY linear power supply instead of using a transformer/rectifier circuit ?

The power brick I have is branded HP which provides 32V (1.1A) and 16V(1.6A) with a common ground.

Intended usage:
The 32V rail of the power brick will be used to provide a variable output of 1.2V to about 28V (current limited to a max 1A), and the 16V rail of the power brick will be used to provide a fixed output of 5V (current limited to a max 500mA).

Thank you
« Last Edit: December 05, 2014, 06:58:15 pm by rstor22 »
 

Offline ajb

  • Super Contributor
  • ***
  • Posts: 2867
  • Country: us
Re: Power Brick for DIY Linear Power Supply
« Reply #1 on: December 05, 2014, 07:50:09 pm »
The switching supply will have ripple at tens to hundreds of kilohertz as opposed to the line frequency ripple of a boat anchor transformer, how significant that is depends on the intended usage of the power supply.  There's a lot more to potentially go wrong in a switching supply, so be sure to use a high quality part, and consider adding some extra input protection to your circuitry in case the supply fails in an ungraceful manner.

As for your linear supply, regulating a 32V supply down to 1.2V at 1A means burning 30.8W in your pass transistor/sense resistor.  Since you've got a 16V rail with adequate current, consider incorporating a relay that switches the main regulator over to the 16V rail at low voltage setpoints (just like switching transformer taps in a more conventional bench supply design).  That will roughly halve your worst case power dissipation and should make your thermal design way easier.  The 5V regulator will burn 5.5W at max current (resulting in an efficiency of 32%), so unless you need particularly low noise, consider using a switching DC-DC converter there.  Murata et al make 7805-equivalent drop-in modules for a couple bucks.  If you need lower noise, you could do a switching pre-regulator followed by a 5V LDO.
« Last Edit: December 05, 2014, 07:54:33 pm by ajb »
 

Offline Chris Jones

  • Regular Contributor
  • *
  • Posts: 105
  • Country: au
Re: Power Brick for DIY Linear Power Supply
« Reply #2 on: December 06, 2014, 11:44:39 am »
If the negative terminal of the SMPS is tied to mains earth inside the SMPS brick then you will probably also have the problem that the negative output terminal of your linear regulator will not be at mains earth potential, but rather several tens or even hundreds of millivolts above it. This is just plain Ohm's law, if the negative conductor of the output cable from ths SMPS brick has tens of milliohms of resistance, and if there is an amp flowing through it, then both ends won't be at the same potential, and if the end at the SMPS brick is at mains earth then the other end isn't. This could cause problems in any sensitive measurements that you make with an oscilloscope, as your circuit's ground will differ from that of the oscilloscope even with the ground clip in place. It's the same reason why you hear CPU-related buzzing noises when plugging the audio output of a laptop into the single-ended line inputs of a grounded domestic hifi amplifier.

That particular problem can be avoided if you choose a class II SMPS power brick that does not need any mains earth connection for electrical safety. In this case you get a different problem in that they usually connect a ~2nF capacitor from the mains wiring to the output wiring, so that your output will have potentially a hundred or more volts of mains voltage on it, if not grounded. There can be very large instantaneous currents if you make a connection from the circuit you are testing to some other grounded circuit at an inopportune time. For example when connecting the audio cable between a grounded hifi amplifier and a class II CD player, if the signal pin of the cable touches first, a spike of current due to the ~2nF capacitor being charged to a couple of hundred volts could damage the input devices of the hifi amplifier. Although the capacitor has a high impedance at 50/60Hz, it does not have a high impedance at the much higher frequencies present in a sudden step such as caused when making a connection.

As long as you are aware of the above limitations, a SMPS power brick might be a sensible thing to use.

There should be much less capacitance from the mains to the output if you use a mains transformer. In the olden days you could buy mains transformers with an earthed screen between the mains windings and the secondary windings, which reduced capacitive coupling from the mains. For test instruments, they used to make transformers with two or even three separate screens between the primary and secondary. E.g. the mains winding would have its screen connected to mains earth, the secondary winding would have its screen connected to the 0V terminal of the secondary winding, and there would be no capacitive leakage currents to disturb the circuit under test. I wonder how people do this these days - can you still get such transformers?
 

Offline rstor22Topic starter

  • Regular Contributor
  • *
  • Posts: 104
  • Country: ca
Re: Power Brick for DIY Linear Power Supply
« Reply #3 on: December 11, 2014, 04:43:55 pm »
Thanks you for this information!

ajb> As the power supply I am building is a project in itself to get into electronics, I am uncertain of the all the future intended usages. What types of circuits would such ripple be a cause of concern?

Chris> My SMPS has a direct path from the common ground of the 32V/16V output to the mains earth connector. In this case you mentioned that the ground output of the linear regulator would be at a higher potentional than that of the SMPS mains earth ground/scope ground. Would you be able to provide some examples of the kinds of sensitive measurements which could be effected when using a scope with the above issue?
 

Offline Paul Price

  • Super Contributor
  • ***
  • Posts: 1443
Re: Power Brick for DIY Linear Power Supply
« Reply #4 on: December 12, 2014, 10:59:26 am »
You would likely not experience any noise problem using your brick to power your linear power supply project you are working on, most certainly not going to be a problem if your are using 3-wire power cords plugged into adjacent sockets on a mains power strip and you are probably not even needing this hookup unless working with seriously small microvolt sensor circuits.

Get on with it and build your project!
« Last Edit: December 12, 2014, 11:03:24 am by Paul Price »
 

Offline dannyf

  • Super Contributor
  • ***
  • Posts: 8221
  • Country: 00
Re: Power Brick for DIY Linear Power Supply
« Reply #5 on: December 12, 2014, 12:14:01 pm »
I have used them a lot, to power linear power supplies, to make solder stations, and to power motorized stuff like helicopters too.

No problem whatsoever.

My favorite is HP 20v 4a/6a power supplies - very compact.
================================
https://dannyelectronics.wordpress.com/
 

Offline jay

  • Regular Contributor
  • *
  • Posts: 62
  • Country: ee
  • Coder
Re: Power Brick for DIY Linear Power Supply
« Reply #6 on: December 13, 2014, 04:12:59 pm »
Is there any issue or problems one may encounter if they use a power brick (such as a SMPS that belongs to a printer/scanner/laptop) to power a DIY linear power supply instead of using a transformer/rectifier circuit ?

That's exactly what I used for my bench PSU project. Lacie power bricks suck (at least the models that I have). I had one old HP 90W laptop brick, but it failed immediately so no experience of those. The old IBM laptop (16V 4.5A) bricks are reasonably compact and the voltage doesn't vary much with load. The ripple was pretty bad.. 200mVpp in older and 100mVpp in newer model power brick. The low ESR input caps, that I used, brought the ripple below 20mV and after the regulators it was already below what my equipment can measure. The negative terminal wasn't connected directly to ground: multimeter showed 1MOhm. I dont' know if it's a coincidence or what but contrary to what many say most of the old power bricks, that I have, are such that the negative terminal isn't connected to ground directly although most of the bricks have ground on the input side. Weird..
SW engineer trying to design HW because it's more fun.
 

Offline tom66

  • Super Contributor
  • ***
  • Posts: 8883
  • Country: gb
  • Professional HW / FPGA / Embedded Engr. & Hobbyist
Re: Power Brick for DIY Linear Power Supply
« Reply #7 on: December 13, 2014, 04:20:57 pm »
Probably only the 16V rail will be regulated. Or perhaps only the 32V. However it is, this means if you load one rail, it may drop as it is not regulating that rail.

At a guess, the 32V will probably be for motors (so variation is tolerable) and the 16V will be for logic, so it will probably be more closely regulated.

To resolve this issue you need to put a small load on both rails, to keep the controller inside the SMPS switching.
 

Offline Chris Jones

  • Regular Contributor
  • *
  • Posts: 105
  • Country: au
Re: Power Brick for DIY Linear Power Supply
« Reply #8 on: December 17, 2014, 04:30:24 am »
Chris> My SMPS has a direct path from the common ground of the 32V/16V output to the mains earth connector. In this case you mentioned that the ground output of the linear regulator would be at a higher potentional than that of the SMPS mains earth ground/scope ground. Would you be able to provide some examples of the kinds of sensitive measurements which could be effected when using a scope with the above issue?
I would only expect problems if looking at tiny signals (a few millivolts). If you are aware of the possibility, then most of the time you will be fine and for those rare occasions with millivolt signals, you can find a battery or something similar.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf

 

-->