Shop OBEX P1 Docs P2 Docs Learn Events
Switching Regulators — Parallax Forums

Switching Regulators

LightfootLightfoot Posts: 228
edited 2006-07-06 14:14 in General Discussion
What switching regulator can supply 12 volts at 5 amps? The LM2575 (1 amp current) is the same as the 7805 but its switching.

Thanks [noparse]:)[/noparse]

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Well well, I'm seeing things, three of them.

-Stanley Blystone

Comments

  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2006-07-02 10:31
    Generally, I think you will find it hard to find a 12volt 5amp switching regulator in one chip.
    There is simply too much heat involved for the semiconductor to survive.

    Why struggle with this? The power switching supplies have torroidial transformers to handle the load.

    If you look in the back pages of Nuts and Volts, you will see that these are sold as small chassis mounted devices [noparse][[/noparse]roughly the size of a 3.5" floppy drive].

    Everything is there and set to go. You can even buy units with two or more voltages.

    Another convienent alternative is to use a computer power supply that is adapted to bench use. You MUST have about 500ma of load on the +5 to regulate voltages properly. If you have one with 5amps of 12volt available for free, that is your easy solution. Just attach a 5 watt 10 ohm resistor to one of the +5volt leads and go from there.· Without the resistor, the unit may self destruct.

    Try a conversion for 12amps at 12 volts with a true switcher -- A 350 ATX supply.· SEE below.

    http://www.wikihow.com/Convert-a-Computer-ATX-Power-Supply-to-a-Lab-Power-Supply

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    "If you want more fiber, eat the package.· Not enough?· Eat the manual."········
    ···················· Tropical regards,····· G. Herzog [noparse][[/noparse]·黃鶴 ]·in Taiwan

    Post Edited (Kramer) : 7/3/2006 2:03:01 PM GMT
  • PJAllenPJAllen Banned Posts: 5,065
    edited 2006-07-02 12:59
    http://vicorwebapps.vicorpower.com/upc/FirstGenDCDCAction.do
    ·
    Following another subject -- ·
    http://forums.parallax.com/showthread.php?p=590816
    , I thought that you were going with the 7812 using a pass transistor (for increased output current.)
    ·
    I guess you want a bench supply ( Y / N ?)
    ·
    Here is probably the best solution for your needs, an open-frame linear (you provide the input fusing, line cord, etc.), at about $1 per WattMax·--
    http://www.mpja.com/productview.asp?product=6609+PS
    http://www.mpja.com/productview.asp?product=6610+PS



    Post Edited (PJ Allen) : 7/2/2006 3:48:16 PM GMT
  • LightfootLightfoot Posts: 228
    edited 2006-07-02 18:42
    I want to build a power switching supply, not buy one shakehead.gif .

    Would it work to use the regulator for conditioning a gate input for a power mosfet?

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Well well, I'm seeing things, three of them.

    -Stanley Blystone

    Post Edited (Three of Them) : 7/2/2006 6:54:12 PM GMT
  • PJAllenPJAllen Banned Posts: 5,065
    edited 2006-07-02 19:26
    I have an 800+ page textbook on the design of switching power supplies.·

    A 12V, 60W+ switching supply isn't going to happen for an experimenter.
  • kjennejohnkjennejohn Posts: 171
    edited 2006-07-02 21:10
    So you want to build your own switching supply. Oh, dude, you just can NOT believe the journey you have set for yourself here. Especially for AC/DC supplies...

    The front end of a simple AC/DC switching supply is a simple bridge rectifier/filtering cap...that puts out 365V of DC, connected straight to the mains. More than enough power to kill you all day long. The parts have to withstand ALL the power you put in, so they are beefy, expensive parts, in single quantity.
    The next part is the controller section. This is where the fancy PID controller IC (read expensive) from any number of manufacturers comes in. This is pulsing a really high speed FET (read expensive) to switch the 365Vdc on and off. This in turn goes to an inductor/capacitor to "smooth things out". Part of this switching arrangement involves an extremely fast Schottky diode (read expensive). The output from coil/cap is fed back to the controller IC, which uses this info to determine the width of the switching signal to the FET.
    The final section is the filter. These things put out an incredible amount of spiking noise. This is a combo of inductors and capacitors. If you don't have this, nothing remotely like an analog circuit will operate reliably. However, digital circuits are mostly immune to noise like this, but not entirely. It plays merry Hobb with clocked devices if the noise levels get high enough.

    Notice I keep using the term "expensive". I can find AC/DC supplies in complete UL-approved enclosures with proper connectors, in a wide assortment of voltages and current capacities, all day, on the web, for $20 to $50. You can't BUY THE PARTS for a safe DIY switching supply for this amount. Not after you've burned the first two or three prototypes to the ground. Not to mention the hundreds of hours spent finding and reading tutorials, guides, data sheets and app notes on the proper design techniques (you will not BELIEVE the conditions you have to meet to build this successfully) and then get and assemble the parts to exquisite detail. And what, exactly, is your time worth to you?

    If you want to do a DC/DC converter, that is easily within the realm of the hobbyist. Go to the Maxim site, for openers, and get the data sheets and app notes for their line of controller ICs. It's still not cheap, but is is possible to do a simple switcher for less than $50 if you shop for the parts. In fact, you can probably get the IC(s) for free as samples from Maxim. For low power applications, you'll end up with the IC, a Schottky diode, an inductor, and a handful of resistors and caps. Pay attention to the notes about proper placement of ground, and parts placement.

    If you want to continue to do an AC/DC supply, more power to yah. First make sure your life insurance policy is up to date.

    Later!
    kenjj
  • BongoBongo Posts: 65
    edited 2006-07-02 23:55
    I use switchers a lot.· It negates the need for massive heatsinks and they waste much less energy. (in heat)· Have a look at national semi's website, search for "simple switchers"· They·(national) provide heaps of tips and software to do your design for you.

    I have found these items to be fast design, easy to use and very forgiving of hand wound inductors.· The only thing I really stick to is the esr rating for the recommended caps.· And maybe board layout.· Fairly standard single point ground.

    They are a bit dearer than some of the common switcher IC's, but the design time and difficulty factor is related to the reactions you have recieved above, I do believe.

    I would not consider anything else for large currents.· Except maybe a SX powered unit, soon to be started.· But this will be for 60 amps.

    bongo
  • LightfootLightfoot Posts: 228
    edited 2006-07-03 00:24
    I have built 5 volt switchers for powering IC's etc.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Well well, I'm seeing things, three of them.

    -Stanley Blystone
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2006-07-03 14:00
    Why for heaven's sake do you need a switcher?

    The Art of Electronics specifically discusses that they are more prone to noise that will cause occasional trouble in a bench supply. This is why the design is so complex and why they need a minimum load to be stable.

    It seems to me that if you want the cleanest bench supply AND you want to build it yourself; you can afford to waste a bit of energy in heat in exchange for the better quality.

    The 7812 with a pass transistor is quite good. Alternatively you could use a LM117 or LM150 and a pass transistor to have an adjustible design. Your transformer would be a simple 50 or 60 cycle step down.

    If you needed a small low volatge switching regulator to squeeze a bit more power out of a battery I could understand, but that is not the case. They can be done and are pretty good. They really are a different kind of switch, no torroidial transformer within the design.

    How about 12amps at 12 volts, plus other outputs +5v, -12v, and +3.3v ?· This is about the most 'bang for your buck' that you can expect.

    http://www.wikihow.com/Convert-a-Computer-ATX-Power-Supply-to-a-Lab-Power-Supply

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    "If you want more fiber, eat the package.· Not enough?· Eat the manual."········
    ···················· Tropical regards,····· G. Herzog [noparse][[/noparse]·黃鶴 ]·in Taiwan

    Post Edited (Kramer) : 7/3/2006 2:06:32 PM GMT
  • Kaos KiddKaos Kidd Posts: 614
    edited 2006-07-03 14:40
    About the ATX-> Bench power supply, I have read up and actually created and am using one of these ATX-> WB powersupplies... It's ok.
    It works like it's inteded to... BUT... it still NEEDs and HAS TO HAVE a load. And, the min amount of load is different from PSU to PSU
    AND... some of the PSU's need sepereat loads, like my ACER unit.
    It requires a load of 10 ohms @ 10 watts on the 12 rail for the +-12 to work...
    It requires a load of 13 ohms @ 10 watts on the 5 volt rail for the +- 5 and 3.3 rails to work.
    Yes, it does work, and not bad for a workbench. I'v 235 watts to workwith (5 Amps on the 12 rail, 20 amps on the 5 and 3 rails)...
    For me, it's messy, I would love to find a kit(s) for CLEAN PSU(s) with +-12, +-5, +-3.3 @ about 1 amp each, to build and install on my prototype system.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Just tossing my two bits worth into the bit bucket


    KK
    ·
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2006-07-03 16:05
    The problem is that I CANNOT find anything else that directly provides the 3.3volts in·more than 1 amp.· The best you can do is to step down the +5 to +3.3 with an LM2940-3.3 whenever the application calls for it.

    This may be a generational thing and eventually there will be 3.3v supplies available, but I haven't seen anything on the shelves in Taiwan yet.

    How did you find out that the Acer needed two loads rather than one?

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    "If you want more fiber, eat the package.· Not enough?· Eat the manual."········
    ···················· Tropical regards,····· G. Herzog [noparse][[/noparse]·黃鶴 ]·in Taiwan
  • SteveWSteveW Posts: 246
    edited 2006-07-03 16:20
    Mini-ITX power supplies are cheap and cheerful. The ones I've used take 18V from a laptop power supply and deliver the full set of PC rails, without necessarily needing loads to make them run.

    Something like http://search.ebay.co.uk/300001422980 should do the trick. I can't believe you can't get similar things locally. (they're certainly made there...)

    Steve
  • Robert KubichekRobert Kubichek Posts: 343
    edited 2006-07-03 17:13
    SteveW said...
    Mini-ITX power supplies are cheap and cheerful. The ones I've used take 18V from a laptop power supply and deliver the full set of PC rails, without necessarily needing loads to make them run.

    Something like http://search.ebay.co.uk/300001422980 should do the trick. I can't believe you can't get similar things locally. (they're certainly made there...)

    Steve

    Try out this power supply for a mini-itx mob, it runs off of 12v and they are 60-200W switchers...

    60W
    www.mini-box.com/s.nl/it.A/id.449/.f?sc=8&category=13
    MANUAL
    resources.mini-box.com/online/picoPSU-60WI/picoPSU-60WI-manual.pdf

    120W
    www.mini-box.com/s.nl/it.A/id.417/.f?sc=8&category=13
    MANUAL
    resources.mini-box.com/online/picoPSU-120/picoPSU-120-manual-engl.pdf

    200W
    www.mini-box.com/s.nl/it.A/id.300/.f?sc=8&category=13
    MANUAL
    resources.mini-box.com/online/PW-200-M/PW-200-M-manual.pdf

    Should be more than enough power for you... jumpin.gif

    Bob scool.gif
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2006-07-03 18:36
    Mini-switchers! Excellent stuff.

    I guess these were developed for laptops and/or automotive.
    I like the ones that take 6V to 26V.

    Still, they are certainly not DIY power supplies.
    But the small package, big power and ability to run remotely with multiple voltages is great.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    "If you want more fiber, eat the package.· Not enough?· Eat the manual."········
    ···················· Tropical regards,····· G. Herzog [noparse][[/noparse]·黃鶴 ]·in Taiwan
  • LightfootLightfoot Posts: 228
    edited 2006-07-03 18:59
    What would happen if a PC power supply was non switching?

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Well well, I'm seeing things, three of them.

    -Stanley Blystone
  • Chris SavageChris Savage Parallax Engineering Posts: 14,406
    edited 2006-07-03 19:35
    If PC Supplies were non-switching they wouldn't be able to handle the current required without being extremely expensive and inefficient (having to dissapate a lot of heat).


    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Chris Savage
    Parallax Tech Support
    csavage@parallax.com
  • PJAllenPJAllen Banned Posts: 5,065
    edited 2006-07-03 20:57
    I say that·for equivalent power,·linears would·make a computer weigh a lot more with a big honking multi-tap transformer --·that and·take up a lot more room (or maybe a room, hahaha), too.
  • LightfootLightfoot Posts: 228
    edited 2006-07-04 06:06
    Are switching supply's good for powering large inductive loads, such as a treadmill?

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Well well, I'm seeing things, three of them.

    -Stanley Blystone
  • PJAllenPJAllen Banned Posts: 5,065
    edited 2006-07-04 14:59
    Here's from The Art of Electronics:
    1. For digital systems, you usually·need +5V, often at high current (>10A).· Advice: (a) Use a line-powered switcher.· (b) Buy it.
    2. For analog circuits with low-level signals.· Advice: Use a linear regulator...
    3. For high-power anything.· Advice: Use a line-powered switcher.· It's smaller, lighter, and cooler (running.)

    ·· Further -- "...high-power switchers (generally line-powered) are complex and tricky and extremely trouble-prone.· If you must design your own, be careful, and test your design very thoroughly.· Better yet, swallow your pride and buy the best switcher you can find."

    Now, from me:

    · I will emphasize that a switcher should only be used in an application where there is a lot of current draw all of the time.· If you don't/aren't meeting the supply's minimum load then you're going to have trouble.· Switchers don't regulate from 0% - 100%, but typically 10% - 100%.

    · Yes, you could run a motor from a switching supply, but I wouldn't recommend that anything else run from (or be dependent upon) that voltage (output) unless that anything else will itself draw the minimum load.


    Post Edited (PJ Allen) : 7/4/2006 3:07:14 PM GMT
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2006-07-04 17:15
    In other words, it is best to keep the power to motors on a supply that is separate from the digital [noparse][[/noparse]because of inductive spikes and lots of noise] and separte from the low-level analog [noparse][[/noparse]for similar reasons].

    Motors have inductive feedback and some have brushes. Both make for a lot of noise. In some cases, diodes and capacitors at the motor can remove a lot of their EMI. Some companies put ferrite beads on the motor leads too.

    PJ,
    Will a 7805 clean up a +12volt switcher output for low level analog signals? I would hate to have to use three power supplies to avoid noise problems.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    "If you want more fiber, eat the package.· Not enough?· Eat the manual."········
    ···················· Tropical regards,····· G. Herzog [noparse][[/noparse]·黃鶴 ]·in Taiwan
  • SteveWSteveW Posts: 246
    edited 2006-07-04 17:41
    >Will a 7805 clean up a +12volt switcher output for low level analog signals?

    Depends what you mean by low level analog...
    If you look at a typical 78L05 datasheet, you'll see that ripple rejection is 60dB up to 10KHz, then plummets. If your switcher noise is at higher frequencies than 10KHz (and that's very likely), then the 78L05 will do very little to remove it. If your analogue stage is susceptible to power ripple at frequencies greater than 10KHz, then you have a problem.
    Fortunately, it's easy enough to remove frequencies above 10KHz with an LC filter. Cheap, too. Whether you remove the noise before or after the 7805 is up to you, there's no clear winner...

    Steve
  • BongoBongo Posts: 65
    edited 2006-07-04 22:44
    I've got real confused.

    Why would you use a regulated supply to feed a motor?

    There's more but I think I will sit on it.

    bongo
  • LightfootLightfoot Posts: 228
    edited 2006-07-05 21:42
    What are the noise profiles of PC chassie and cooling fans? What in the power supply compensates for this? I know the fans have something in them (probably a diode) that prevents them from running backwards. Is this part of a noise supression system?

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Well well, I'm seeing things, three of them.

    -Stanley Blystone

    Post Edited (Three of Them) : 7/5/2006 9:46:44 PM GMT
  • SteveWSteveW Posts: 246
    edited 2006-07-06 14:01
    > What in the power supply compensates for this?

    What makes you think anything does? PC power suplpies are (electrically) noisy. There will be tens to hundreds of milivolts of noise on their rails. Since PCs are fundamentally digital, that's not really an issue. If you need electrically quiet supplies for things like the audio & video sections, it's up to the designer to generate a locally quiet supply from what he's given. All part of the fun of system design.
    If you have a real design problem to solve, tell us what it is...

    Steve

    Post Edited (SteveW) : 7/6/2006 2:09:08 PM GMT
  • Mike GreenMike Green Posts: 23,101
    edited 2006-07-06 14:14
    The chips themselves contribute to the noise as parts of them turn on and off and surges of electricity wash back and forth in them. Their current needs vary widely over nanoseconds. That's why you're advised to put small capacitors (0.1uF usually) across the power leads very close to each chip. These tend to absorb/supply these short power surges. Larger capacitors (1uF-10uF usually) are placed regionally on boards for the same reason on a bit larger and slower scale.
Sign In or Register to comment.