Searching \ for '[EE] More capable 7805 regulator alternative' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: techref.massmind.org/techref/power/regulators.htm?key=7805
Search entire site for: 'More capable 7805 regulator alternative'.

Exact match. Not showing close matches.
PICList Thread
'[EE] More capable 7805 regulator alternative'
2006\05\06@043733 by Gökhan SEVER

picon face
Hi everybody,

I've designed a small led luminarie board which has 5 rgb leds and 2 ICs
(PIC12F675 and MAX-485) and bunch of smd resistors and caps on it. For the
24V to 5V conversation i've used simple 7805 regulator circuitry. (Two caps
take part at the two side of 7805 regulator.) Actually previously I had used
lm2575 (Simple switcher 1A step-down voltage regulator) plus two caps and
one inductor with it for the 24v -> 5v power conversation stage.

The problem is when i've used 7805 with 24V input voltage it over-heats. I
dont have enough space for a heat-sink. There werent any heat problems when
i've used lm2575.  About one thousand of led pcbs were manufactured for the
7805 part and we're in urgent decision stage for proper component selection.
Do you know 7805 alternative that will work well on my circuit?

All of the suggestions are very welcome.

Thanks in advance,
Gökhan SEVER.

2006\05\06@051035 by Hector Martin

flavicon
face
Gökhan SEVER wrote:
> Hi everybody,
>
> I've designed a small led luminarie board which has 5 rgb leds and 2 ICs
> (PIC12F675 and MAX-485) and bunch of smd resistors and caps on it. For the
> 24V to 5V conversation i've used simple 7805 regulator circuitry. (Two caps
> take part at the two side of 7805 regulator.) Actually previously I had used
> lm2575 (Simple switcher 1A step-down voltage regulator) plus two caps and
> one inductor with it for the 24v -> 5v power conversation stage.

If the power dissipated by a linear regulator is too much in your
application, you should probably use a switcher.

>
> The problem is when i've used 7805 with 24V input voltage it over-heats. I
> dont have enough space for a heat-sink. There werent any heat problems when
> i've used lm2575.  About one thousand of led pcbs were manufactured for the
> 7805 part and we're in urgent decision stage for proper component selection.
> Do you know 7805 alternative that will work well on my circuit?

Any linear regulator will have the heat problem. The ideal linear
regulator will still have to dissipate 19*(current) W. Switchers are not
limited by that, and do not waste as much energy as heat.

If lm2575 worked, why not stick with that? Unless you can find a
switcher that fits in the footprint of a 7805 (I don't think any exist,
given that they all need a coil and some extra stuff), I doubt you can
use those PCBs as is (without some workaround to replace the 7805 with a
switcher)

--
Hector Martin (spam_OUThectorTakeThisOuTspammarcansoft.com)
Public Key: http://www.marcansoft.com/hector.asc

2006\05\06@051759 by Jinx

face picon face
> Do you know 7805 alternative that will work well on my circuit?

If you're going to stick with a linear, maybe you'll need to add
a dropping resistor. That will heat up, rather than the regulator.
You'd be looking at dropping the 24V down to 8V-ish. What
do you need ? About 300mA ? A 51 ohm resistor would drop
15.3V @ 300mA -> 4.6W. Some of the finer resistance wires
are up to 1 ohm/mm, so if you wound your own to save space
it might not be as large as a bodied resistor, but you'd need to
choose a gauge that could take the heat/current

2006\05\06@051853 by Jinx

face picon face
BTW, can you put the LEDs on the 24V side as they'll
be using most of the power ?

2006\05\06@053126 by Bob Axtell

face picon face
Gökhan SEVER wrote:
> Hi everybody,
>
> I've designed a small led luminarie board which has 5 rgb leds and 2 ICs
> (PIC12F675 and MAX-485) and bunch of smd resistors and caps on it. For the
> 24V to 5V conversation i've used simple 7805 regulator circuitry. (Two caps
> take part at the two side of 7805 regulator.) Actually previously I had used
> lm2575 (Simple switcher 1A step-down voltage regulator) plus two caps and
> one inductor with it for the 24v -> 5v power conversation stage.
>
> The problem is when i've used 7805 with 24V input voltage it over-heats. I
> dont have enough space for a heat-sink. There werent any heat problems when
> i've used lm2575.  About one thousand of led pcbs were manufactured for the
> 7805 part and we're in urgent decision stage for proper component selection.
> Do you know 7805 alternative that will work well on my circuit?
>
> All of the suggestions are very welcome.
>
> Thanks in advance,
> Gökhan SEVER.
>  
You need to use a switching supply rather than a linear supply. Take a
look at LM2576-5.0 by National;
it only needs a very few components. And it generates NO heat at 2A with
a 18V input.

--Bob

2006\05\06@054239 by Gökhan SEVER

picon face
Five rgb leds are driven with PWM and the given max dc current for each one
is 30ma. In total 30mA x 5 = 150mA.
We use 150W / 24V SPMS supply thus we can join about 20 boards to one
supply. And all the leds are driven with 24V source..

2006/5/6, Jinx <.....joecolquittKILLspamspam@spam@clear.net.nz>:
>
> BTW, can you put the LEDs on the 24V side as they'll
> be using most of the power ?
>
> -

2006\05\06@061900 by Jinx

face picon face
> And all the leds are driven with 24V source..

So you need the 5V for just the chips on 20 boards ?

2006\05\06@062841 by Gökhan SEVER

picon face
Yes, exactly..

I have forgotten to add that 5 rgb leds are parelled and there exist 3 pnp
transistor to handle max 150mA current from 24V power supply.



2006/5/6, Jinx <joecolquittspamKILLspamclear.net.nz>:
>
> > And all the leds are driven with 24V source..
>
> So you need the 5V for just the chips on 20 boards ?
>

2006\05\06@065604 by Diego Sierra

picon face
Hi!

> If lm2575 worked, why not stick with that? Unless you can find a
> switcher that fits in the footprint of a 7805 (I don't think any exist,
> given that they all need a coil and some extra stuff),

Here is one, which do not need anything but just the caps, as the 7805 does.

   http://www.recom-international.com/

And a few more (the first one do not need the caps):

   www.dimensionengineering.com/DE-SW0XX.htm
   http://focus.ti.com/docs/prod/folders/print/pt5101.html

Cheers,
Diego.

2006\05\06@071955 by Jinx

face picon face

> Here is one, which do not need anything but just the caps, as the
> 7805 does.
>
>     http://www.recom-international.com/

Hey, they look useful. Might solve a problem or two of my own

2006\05\06@081406 by Gökhan SEVER

picon face
Sorry for my previous posting; all the leds supplied with 5V via
transistors. I've confused with my earlier circuitry.
There are only 4 leds not 5.

Apparently;

DE-SW0XX (http://www.dimensionengineering.com/DE-SW0XX.htm)
R-78xx-0.5 (
http://www.recom-international.com/switching_regulator_R-78xx.html)

parts are exactly what i've been looking for. But the prices are too high
for our budget. The price of LM7805 is just $0.5 for 1K+ purchasing. The
others are about $10 :(

TI's PT5101 part is $7.35 which is still too expensive.

I've seen an article titled "What is a switching regulator?" on DE's page.
Exactly ( A beginner's guide to switching regulators -
http://www.dimensionengineering.com/switchingregulators.htm) I've read it
and do the math again for the wasted power on 7805:

Power Wasted:  (24V - 5V) * (load current)

load current = 4x22mA (for blue leds) + 4x22mA (for green leds) +
4x21.5mA(for red leds)  + 1mA (for PIC12F675) +
0.5mA (for MAX485)

(I hope i've had interpreted the dc electrical characteristics of two ICs..)

load current is about 264mA.

PW = 19 x 264mA
PW = 5.016 W

ohh, too much wasted power for 7805 to handle without proper heatsink. But
there is not enough space for heatsink in the circuit board. The worse
situation is the board will be buried inside the polyester material which
there wont be air circulation or conduction environment for over-heat...

Maybe i can halve the supplied voltage. Thus pw becomes 2.5W..

2006\05\06@083946 by Bernd Rüter

flavicon
face
part 0 44 bytes
his is a multi-part message in MIME format.
part 1 999 bytes content-type:text/plain; charset=ISO-8859-1; format=flowed (decoded quoted-printable)



Gökhan SEVER schrieb:
{Quote hidden}

Hello,

you have a real problem !

My tip: throw away this 1000 boards.
Create a new one with an LM2574 or better MC33063.

You have too much heat !


greetings...


part 2 328 bytes content-type:text/x-vcard; charset=utf-8;
(decoded 7bit)

begin:vcard
fn;quoted-printable:Bernd R=C3=BCter
n;quoted-printable:R=C3=BCter;Bernd
adr;quoted-printable;dom:;;W=C3=BClferoder Stra=C3=9Fe 10;Hannover;;30539
email;internet:.....postKILLspamspam.....bernd-rueter.de
tel;fax:+49 511 7636992
tel;home:+49 511 7636990
tel;cell:+49 163 2355583
version:2.1
end:vcard



part 3 35 bytes content-type:text/plain; charset="us-ascii"
(decoded 7bit)

2006\05\06@084952 by Russell McMahon

face
flavicon
face
> Power Wasted:  (24V - 5V) * (load current)
>
> load current = 4x22mA (for blue leds) + 4x22mA (for green leds) +
> 4x21.5mA(for red leds)  + 1mA (for PIC12F675) +
> 0.5mA (for MAX485)
>
> (I hope i've had interpreted the dc electrical characteristics of
> two ICs..)
>
> load current is about 264mA.
>
> PW = 19 x 264mA
> PW = 5.016 W

Add a series resistor before the regulator dimensioned to drop most of
the voltage.
Let's allow for 300 mA.
Vin = 24V
Allow 7805 3 volts "headroom" = 8V input

Vresistor = 24-8 = 16 V.
Resistor = V/I = 16/0.3 = 53.3R.
Use 56R
At max 300 mA resistor will drop IR = 0.3 * 56 = 16.8V
V Regulator in = 7.2V = OK

Power in resistor = V^2/ = 16.8^2/56 = 5W
Power in regulator = (Vin-Vout)*Imax
= (7.2-5)*.3 = 0.66 Watt.

TO220 Regulator with a smidgeon of heat sinking is fine.
Even without in free air is OK.
In a sealed package you need to make sure you can handle the 0.7 odd
Watt but it won't be too hard.
The 5W in the resistor would ideally be outside the package if you can
manage this.

As actual max current is (by your calculation) < 300 mA worst case
dissipation will be even less.

Ensure that the worst case current does not put regulator into
"dropout". I've allowed 2.2 V min. AFAIR 7805 can exceed this in some
cases depending on temperature etc.


       RM


2006\05\06@090949 by Gökhan SEVER

picon face
But there is still minor problem with your suggestion Mr. McMahon.

The led lighted products will be connected serially.  If i assume that each
board sink 300mA from 150W/6A smps power supply there will 20 parts in one
serie. But they will be seperated 1 meters apart from each other. So
probably the input voltage will drop under 7805's rated min input voltage
7.2V near the end of each serie..

2006/5/6, Russell McMahon <EraseMEapptechspam_OUTspamTakeThisOuTparadise.net.nz>:
{Quote hidden}

> -

2006\05\06@092918 by Russell McMahon

face
flavicon
face
> My tip: throw away this 1000 boards.
> Create a new one with an LM2574 or better MC33063.
>
> You have too much heat !

Resistor.
Series resistor.
Add a 10 watt 56 ohm series resistor.
Do it now. You'll like it :-).

And, as he says, for the next batch look at the MC34063 / MC33063.
Olde.
Cheap. Far cheaper than any equivalent modern part.
Exceedingly flexible and capable. Can drive any topology desired.
Unlike many modern parts.
Efficiency is OK if not right up there.
BUT even if it was 50% efficient (and it will be far better than that)
you'd only dissipate 1.5 watt in the switcher.
At 80% efficient, which you would achieve with a MC34063, switcher
dissipation is 0.2/0.8 * 5V * 0.3A = 375 mW.

A PCB that fits a 7805 footprint could be made.

MC34063 is 0-70 C version
MC33063 is -40 to +85C version

MC34063 / MC33063 stuff here

       http://others.servebeer.com/temp/mc34063smps

and

       http://others.servebeer.com/temp/mc34063_smps.htm

or for those with real browsers [ :-) ]

       http://others.servebeer.com/temp/mc34063_smps.mht








The fantastically useful App note AN920 here





       Russell McMahon.


       RM

2006\05\06@094157 by Mark Jordan

flavicon
face

       Build ONE SMPS converting your 24V to 8V and feed all boards from there.

       Mark Jordan


On 6 May 2006 at 16:09, Gökhan SEVER wrote:

{Quote hidden}

> > --

2006\05\06@095234 by Russell McMahon

face
flavicon
face

> But there is still minor problem with your suggestion Mr. McMahon.

If any given notional solution ever has only one minor problem then
the engineer is as good as home free :-)

> The led lighted products will be connected serially.  If i assume
> that each
> board sink 300mA from 150W/6A smps power supply there will 20 parts
> in one
> serie. But they will be seperated 1 meters apart from each other. So
> probably the input voltage will drop under 7805's rated min input
> voltage
> 7.2V near the end of each serie..

Either

reduce the resistor value for the end units,

reduce the resistor value for all units (still vast power reduction in
7805s),

use two resistors in series with a shorting link when required,

increase feed cable size,

feed in middle of run,

...  ... ... .

You COULD add a switch which drops out one resistor when voltage is
too low (could be a relay or solid state) or numerous other solutions.

My  'GSR'*  buck converter would do very well here at very low cost.
Inductor would be the greatest cost item. Ask more if/when you want to
use a switcher again.



       Russell


* GSR = "God's Switching Regulator", as He designed it.
Craziest design I've ever seen - but happens to work superbly in the
application it was intended for (as you'd expect), and in quite a few
others. Installed volume is presently only about 10,000 units but it
may be used in a 100,000 volume product shortly.





{Quote hidden}

>> --

2006\05\06@131839 by Byron A Jeff

face picon face
On Sat, May 06, 2006 at 11:37:31AM +0300, G?khan SEVER wrote:
> Hi everybody,
>
> I've designed a small led luminarie board which has 5 rgb leds and 2 ICs
> (PIC12F675 and MAX-485) and bunch of smd resistors and caps on it. For the
> 24V to 5V conversation i've used simple 7805 regulator circuitry. (Two caps
> take part at the two side of 7805 regulator.) Actually previously I had used
> lm2575 (Simple switcher 1A step-down voltage regulator) plus two caps and
> one inductor with it for the 24v -> 5v power conversation stage.

Why did you switch from the lm2575?

> The problem is when i've used 7805 with 24V input voltage it over-heats.

How many amps is your board drawing? Even at 100 mA you'd have to dissapate 1.9W
of power.

> I
> dont have enough space for a heat-sink. There werent any heat problems when
> i've used lm2575.

It's a switcher. The main advantage of a switcher is that it generally doesn't
generate much heat.

>  About one thousand of led pcbs were manufactured for the
> 7805 part and we're in urgent decision stage for proper component selection.
> Do you know 7805 alternative that will work well on my circuit?

Um the lm2575? It worked before right?

If you use a linear regulator of any sort you'll have to get rid of the heat.
The best you can do is transfer the heat to another component. But it'll still
get hot. For example you could create a 24V to 8V voltage divider to feed the
7805 circuit. It'll be cool, but the resistors of the voltage divider now would
take on the heat of dropping 16V.

BAJ

2006\05\06@141217 by fred jones

picon face
Hi,
I've got your solution but it's going to cost you dearly.  Go check out this
URL:

http://www.robotparts.com.au/dimension.htm

This is a drop in replacement for a 78XX regulator that is a switcher in one
component.  It should just drop in and take care of your problem but they
are quite a bit costlier.  Maybe it's better than trashing 1000 boards
though.
Good luck,
FJ


2006\05\06@153040 by Wouter van Ooijen

face picon face
> Yes, exactly..
> I have forgotten to add that 5 rgb leds are parelled and
> there exist 3 pnp
> transistor to handle max 150mA current from 24V power supply.

I don't think I have got the full picture. Can you give the total
circuit, preferrably as picture (instead of words)?

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2006\05\07@012100 by Gökhan SEVER

picon face
>
> Why did you switch from the lm2575?


Mainly for the cost reasons. The lm2575 and one coil adds at least
1$. Unfortunately, we have forgotten to balance the cost-reliabilty
trade-off well. Notwithstanding this was bad circuit design experience i've
learnt when / where to use linear regs and switchers.

> I dont have enough space for a heat-sink. There werent any heat problems
> when  i've used lm2575.
>
> It's a switcher. The main advantage of a switcher is that it generally
> doesn't
> generate much heat.


I wish i have attended to the smps class semester. My current goal will be
definetely, reading a good smps book and app notes from the vendors to be
prepared for the next challange of smps supply design.

>
>

2006\05\07@012359 by Gökhan SEVER

picon face
2006/5/6, fred jones <KILLspamboattowKILLspamspamhotmail.com>:
>
> Hi,
> I've got your solution but it's going to cost you dearly.  Go check out
> this
> URL:
>
> http://www.robotparts.com.au/dimension.htm
>
> This is a drop in replacement for a 78XX regulator that is a switcher in
> one
> component.  It should just drop in and take care of your problem but they
> are quite a bit costlier.  Maybe it's better than trashing 1000 boards
> though.
> Good luck,
> FJ


Although the solution was proposed previous postings but thanks anyway. The
cost of pin-to-pin matched switcher alternative of 7805's cost is nearly the
price of one pcb board :(

2006\05\07@104304 by Byron A Jeff

face picon face
On Sun, May 07, 2006 at 08:21:00AM +0300, G?khan SEVER wrote:
> >
> > Why did you switch from the lm2575?
>
>
> Mainly for the cost reasons. The lm2575 and one coil adds at least
> 1$. Unfortunately, we have forgotten to balance the cost-reliabilty
> trade-off well. Notwithstanding this was bad circuit design experience i've
> learnt when / where to use linear regs and switchers.

It's not a total loss yet. You've already gotten the suggestion for the series
resistor and running the LEDs directly from the 24V source. Each of these
suggestions should reduce or transfer the heat load from the 7805.

{Quote hidden}

A lot of the design issues have been packaged into integrated parts.

However, it can be done very cheaply with discretes. As a sample take a read
of Roman Black's two transistor 5V regulator. You can find it here:

http://www.romanblack.com/smps.htm

70 cents in parts. However, it takes up a bit of real estate on your PCB to
implement.

BAJ

2006\05\07@164244 by Yigit Turgut

picon face
how many items you are going to produce ?
if the job is that big,design or get a heat-sink @ your custom size.

going to be the cheapest solution,unless the components are being soldered
to pcb via an automation system or etc.

2006\05\08@041442 by Wouter van Ooijen

face picon face
Thanks for the circuit diagram, that makes things a bit more clear :)

You are feeding individual LEDs current drawn from a 12V 'bus' rail.
Without using switching techniques this will cause you to burn the
excess heat (12V - Vled) * LEDcurrent. It does not matter whether you
feed the individual LEDs from the 12V or the 5V, the heat is burned
somewehere. When total heat is a problem, you will have to use a
different conecpt. When the locality of the heat is a problem feeding
the LEDs from the 12V might be a solution, but I assume you use local
7805's to compensate voltage drop on the bus, so that voltage drop could
cause variations in LED current.

But you have 12V and you want to feed 4 LEDs with the same PWM signal.
So why not put the LEDs in series? That would reduce the current by a
factor 4. Choose a bus voltage that is sufficiently high to compenstate
the drop and still feed 4 LEDs, and switch each 4 LEDs with a constant
current supply. The 7805 can be replaced by an 78L05 or even a R+zener.
I think this will burn much less heat.

If heat is still a problem you can either use a switcher like you did
(but that still burns the heat ( 5V - Vled ) * current. If that is still
too much I would consider a current loop, with all LEDs in the loop, and
FETs over the LEDs to bypass them. This is a bit more complex, but burns
only the Vled * Iled heat. I don't think you can get rid of that.

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu
 

More... (looser matching)
- Last day of these posts
- In 2006 , 2007 only
- Today
- New search...