| Author |
Message |
LetterBeacon
Joined: Mar 18, 2008
Posts: 407
Location: London, UK
|
 Posted: Fri Nov 27, 2009 10:46 am Post subject:
DIY Power supply that supplies 3.4A
Subject description: Does it exist? |
  |
|
Are there any schematics or layouts of power supplies that can provide around 3A?
All the designs I've found use the 7815/ 7915 which only provides 1A.
Can anyone shed any light? |
|
|
Back to top
|
|
 |
Dave Kendall
Joined: May 26, 2007
Posts: 354
Location: England
Audio files: 3
|
| Posted: Fri Nov 27, 2009 11:17 am Post subject:
|
|
|
Hi.
Dunno if this helps, but both the LM317 (+V) and LM337 (-V) adjustable V regs have outputs "in excess of 1.5A" according to several datasheets I've got. Many quote a typical output current of 2.2A, and one of them states a max current output of 3.4A (national semi)
I'm *NO* PSU expert - perhaps someone else could chime in here, but it seems that with large enough caps (4700uF or probably more), and some serious heatsinking, you might be able to get away with a LM317/337 based PSU that gives quite a bit more than 1.5A. Based on comments on the CGS site, it would appear that the physical proximity of components is important too. Here's what is written about the CGS66.
| Quote: | When fed from a 30 volt center tapped transformer (15v-0v-15v) you should be able to pull around 200ma at +/-15 volts from the PSU without problems, and notably more at +/-12V output. If using for +/-15 volts there will be minimal heat generated by the regulators because they will be running close to their minimum voltage drop, though you do risk ripple getting through to the output as the load increases.
When fed from a 36 volt center tapped transformer (18v-0v-18v) you should be able to pull around 600ma, at +/-15 volts, from the PSU without problems. There will be a fair amount of heat generated by the regulators and diodes, so make sure there is adequate airflow around the board. 600ma is close to the maximum D.C. current you can draw from this PSU when powered from a 1 amp transformer. You may also wish to use 3300 mfd 25 volt capacitors instead of the specified 2200 mfd.
Drawing more than 600ma, while possible with a larger transformer, is not recommended, as there is inadequate heatsinking on a number of components, and the electrolytic capacitors are much too close to the heatsinks.
Another factor limiting the maximum current that can be drawn from the unit is the size of the heatsink. With 600ma being drawn, each regulator can easily be dissipating 6 watts, and will require a heatsink with a rating of 12°C/W or better to operate in an ambient of 25°C. As heat in an enclosed space may well be greater than this, I'd recommend going for the best heatsink of the correct profile you can get. Hobby shops in Australia have them available at 8.5°C/W (50 mm) and 11°C/W (40 mm), depending on the shop, though I have used 63 mm heatsinks I salvaged from old switch-mode power supplies. The profile used is simply listed as PCB mount TOP-3 in the local catalogs. |
I have read on several sites that a good design based on LM317/LM337 will produce a more accurate output voltage than fixed LM7815s, which is important for things like VCOs, and less so for digital circuits.....
PSU experts, - I'd like to know too about this too -I've got a great big N.O.S. transformer waiting for a home.....
cheers,
Dave
(edited to include source of quoted text)
_________________
"I might be synthetic, but I'm not stupid" - Bishop |
|
|
Back to top
|
|
|
Blue Hell
Site Admin
Joined: Apr 03, 2004
Posts: 18181
Location: Netherlands, Enschede
Audio files: 95
G2 patch files: 310
|
|
|
Back to top
|
|
|
Dave Kendall
Joined: May 26, 2007
Posts: 354
Location: England
Audio files: 3
|
| Posted: Fri Nov 27, 2009 12:46 pm Post subject:
|
|
|
Thanks Jan.
Useful links. 
Being curious, is there any particular reason(s) why running a linear PSU at that sort of current might be dangerous or undesirable? If the components are uprated to say, 1N4004s, higher power-rated resistors, large caps, and possibly a chassis-mounted heatsink is used, what other problems do you think we might hit?
I'm interested in all this, as I also have a bunch of 50VA and 75VA 18-0-18 transformers, and some EFM PSU/distro boards (3501)
page 2 here - http://www.modular.fonik.de/files/EFMforum/orgEFMfiles/3500_doc.pdf
The caps on these boards are very close to the multiturn trimmers, and I'm wondering if that might be a problem too....
Sadly, spare money is nearly non-existent (except for the occasional beer) so I need to use what I have....
cheers,
Dave
_________________
"I might be synthetic, but I'm not stupid" - Bishop |
|
|
Back to top
|
|
|
Blue Hell
Site Admin
Joined: Apr 03, 2004
Posts: 18181
Location: Netherlands, Enschede
Audio files: 95
G2 patch files: 310
|
| Posted: Fri Nov 27, 2009 1:50 pm Post subject:
|
|
|
Heh, not dangerous ... well ... a short a 4A may give some fireworks but not too bad really ... it's just such a waste of energy to make it analog ... like about half of the power used will go into heat .. a bit more even likely.
_________________
Jan |
|
|
Back to top
|
|
|
LetterBeacon
Joined: Mar 18, 2008
Posts: 407
Location: London, UK
|
| Posted: Sat Nov 28, 2009 7:04 am Post subject:
|
|
|
Interesting replies, thanks guys!
Re: the +/-20VDC power supply schematic - to change the output to +/-15VDC would it just be a case of changing the values of R1/VR1 and R2/VR2? |
|
|
Back to top
|
|
|
Blue Hell
Site Admin
Joined: Apr 03, 2004
Posts: 18181
Location: Netherlands, Enschede
Audio files: 95
G2 patch files: 310
|
| Posted: Sat Nov 28, 2009 9:05 am Post subject:
|
|
|
| LetterBeacon wrote: | | would it just be a case of changing the values of R1/VR1 and R2/VR2? |
Yes, or R3 / R4
http://www.national.com/mpf/LM/LM338.html has a link to a data sheet, and the data sheet states that :
VOUT = 1.25 * ( 1 + R2 / R1 ) + IADJ * R2
Where R1 maps onto R3 / R4 in the schematic and R2 maps to R1-VR1 / R2-VR2, and IADJ ranges from < 45 µA to about 100 µA.
So let's guess 270 Ω for R3 / R4 to see what happens :
VOUT = 1.25 * ( 1 + 2200 / 270) + 45e-6 * 2200 = 11.5 V
for the lowest VOUT and
VOUT = 1.25 * ( 1 + 4400 / 270) + 100e-6 * 4400 = 22.0 V
for the highest VOUT
Maybe 330 Ω would give a somewhat better range even ... but that's the idea basically.
_________________
Jan |
|
|
Back to top
|
|
|
LetterBeacon
Joined: Mar 18, 2008
Posts: 407
Location: London, UK
|
| Posted: Sat Nov 28, 2009 11:32 am Post subject:
|
|
|
Excellent thank you - that's really useful. It looks like a 270Ω resistor will be fine if I just want +/-15V.
This is probably a very basic question (and gives away the fact I haven't had any formal electronics training), but how do I work out how much current the PS can supply at +/-15V? |
|
|
Back to top
|
|
|
|
Forum index » DIY Hardware and Software
FAQ
Calendar
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
Chat Room
