7824 regulator design to have current more than 1A ,how?
praveen
than 1A . please help me!!!!!!!!!!!
thank you,
k.praveen.
Frank Bemelman
praveen <praveen...@lycos.com> schreef in berichtnieuws 9n5vgc$d10...@zeus.ligtel.com...
> i want a simple regulated power supply design with 7824 to give current more
> than 1A . please help me!!!!!!!!!!!
Download a pdf file for these kind of regulators. Good chance there are
some application notes to increase the current. The most simple one
is something like this:
PNP POWER
+ -------E C--------------------
| B |
---2E2---|-------[7824]----------------- +24V >1A
|
|
GND---------------------------------------- GND
Output no longer protected against short circuit and overload, so take care...
Add the usual caps.
--
Frank Bemelman
(to email, remove 'x')
Ken Smith
Frank Bemelman <fbe...@euronet.nl> wrote:
[...]
> PNP POWER
>+ -------E C--------------------
> | B |
> ---2E2---|-------[7824]----------------- +24V >1A
> |
> |
>GND---------------------------------------- GND
>
>Output no longer protected against short circuit and overload, so take care...
>Add the usual caps.
If you use a transistor like the TIP36 for the PNP and put a resistor in
the E. leg, the circuit can still be made to withstand a short circuit.
--
--
kens...@rahul.net forging knowledge
Martin Sikström
I've seen yet another variant where they put a diode in
series with the base resistor and an emitter resistor.
Since the diode forward drop and the base-emitter voltage
drop will roughly cancel, the current limit should be rather
predictable.
Regards,
Martin
Winfield Hill
>
> i want a simple regulated power supply design with 7824
> to give current more than 1A . please help me!!!!!!!!!!!
"Martin Sikström" wrote:
>
> Ken Smith <kens...@rahul.net> writes:
>
>> Frank Bemelman <fbe...@euronet.nl> wrote:
>>
>>> PNP POWER
>>> + -------E C--------------------
>>> | B |
>>> ---2E2---|-------[7824]----------------- +24V >1A
>>> |
>>> |
>>> GND---------------------------------------- GND
>>>
>>> Output no longer protected against short circuit and
>> overload, so take care... Add the usual caps.
>>
>> If you use a transistor like the TIP36 for the PNP and
>> put a resistor in the E. leg, the circuit can still be
>> made to withstand a short circuit.
>
> I've seen yet another variant where they put a diode in
> series with the base resistor and an emitter resistor.
> Since the diode forward drop and the base-emitter voltage
> drop will roughly cancel, the current limit should be
> rather predictable.
The O.P. wanted _more_ than 1A, let's call that 2A, so
the CL would have to be set at 2.5A. To reliably get
24V out, he needs say 30 to 35V in. The normal 2A and
10V difference would dissipate a reasonable 20 watts.
But we know that when shorted, 2.5A and 35V = 88 watts,
which is WAY more than the power transistor / heat sink
will be able to safely handle. So, lacking a thermal
shutoff in the PNP pass transistor, we need a FOLDBACK
current limit. Sort of like this, note R1 and R2...
+ PNP POWER
(O)-+-- 0.18 --+-- E C -----------+---(O) +24V
| | B TIP36 | up to 2.5A
| R1 10 | Q1 |
| | | R2 |
| ZTX1055A +-----|----- 620 ----+
| small | Q2 | ______ |
| 2A pnp b | | | |
+------- e c --+-+--| 7824 |--+
'----+---- 22 -----+ |______| |
| | | |
=== === | ===
return | | | |
(O)-----+-------------+-----+------+---(O) GND
Q2 is a Xetex super E-line TO-92 transistor, with high
beta (300) at currents of 1A, rated for 1W dissipation.
But Q2 needs a little more thought, it has to take all
of the 7824's current during a short. Maybe with the
help of a collector resistor to limit dissipation...
But you get the idea. :-)
Thanks,
- Win
Winfield Hill
Rowland Institute for Science
100 Edwin Land Blvd
Cambridge, MA 02142
Winfield Hill
>
> praveen wrote:
>>
>> i want a simple regulated power supply design with 7824
>> to give current more than 1A . please help me!!!!!!!!!!!
>
> the CL would have to be set at 2.5A. To reliably get
> 24V out, he needs say 30 to 35V in. The normal 2A and
> 10V difference would dissipate a reasonable 20 watts.
> But we know that when shorted, 2.5A and 35V = 88 watts,
> which is WAY more than the power transistor / heat sink
> will be able to safely handle. So, lacking a thermal
> shutoff in the PNP pass transistor, we need a FOLDBACK
Actually, the TIP36 is a VERY rugged power transistor,
So OK, let's just go ahead and add a thermal limit.
+ STO-140 PNP POWER
(O)---( TIP-36
)<--+--- E C -----------+---(O) +24V
| B ______ | up to 2.5A
| | | | |
'- 22 -+-+--| 7824 |--+
| |______| |
=== | ===
return | | |
(O)-----------------+------+-----+---(O) GND
Bolt a Stancor disc thermostat in place on the heatsink.
I often select a STO-140 from our drawers, it opens at
140F = 60C, which is a pretty safe case temperature for
the TIP-36. With Rth_J-C = 1.0 C/W, that corresponds to
a 148C junction temperature at 88 watts with a 60C case,
which it can handle without trouble for a limited time.
Ken Smith
Winfield Hill <hi...@rowland.org> wrote:
[....]
> 10V difference would dissipate a reasonable 20 watts.
> But we know that when shorted, 2.5A and 35V = 88 watts,
> which is WAY more than the power transistor / heat sink
> will be able to safely handle. So, lacking a thermal
> shutoff in the PNP pass transistor, we need a FOLDBACK
> current limit. Sort of like this, note R1 and R2...
What I did, in a design very like this was to bolt the TIP36 and the
LM78XX to the same heat sink. The TIP36 heats up the LM78XX and the
LM78XX goes into thermal shut down well short of the point where the
transistor gets smoked.
BTW: This was for something at home and not something that has to be
repeatable,
Tony Williams
Winfield Hill <hi...@rowland.org> wrote:
> Actually, the TIP36 is a VERY rugged power transistor,
> So OK, let's just go ahead and add a thermal limit.
> + STO-140 PNP POWER
> (O)---( TIP-36
> )<--+--- E C -----------+---(O) +24V
> | B ______ | up to 2.5A
> | | | | |
> '- 22 -+-+--| 7824 |--+
> | |______| |
> === | ===
> return | | |
> (O)-----------------+------+-----+---(O) GND
The reason I had TIP36C's in stock for that 2V/6A
zener expt was due to a failed attempt at a dc
supply similar to this. It was 36V/5A, and I had
40-odd volts Vraw, a discrete regulator, and TIP36
pass element. There was a series current-sense
resistor and the usual shutdown on overload.
Shutdown did work ok, provided the overload was
increased slowly.
But (and this was the killer), a short on the
output always resulted in a destroyed TIP36.
After some time (and a quantity of TIP36's) I
came to the conclusion that a short circuit took
a (hot?) TIP36 outside the Safe Operating Area.
A move to a Mosfet series element did send all
problems away.
--
Tony Williams.
Win Hill
>
> In article <3B97FF2A...@rowland.org>,
> Winfield Hill <hi...@rowland.org> wrote:
>
>> Actually, the TIP36 is a VERY rugged power transistor,
>> So OK, let's just go ahead and add a thermal limit.
>>
>> + STO-140 PNP POWER
>> (O)---( TIP-36
>> )<--+--- E C -----------+---(O) +24V
>> | B ______ | up to 2.5A
>> | | | | |
>> '- 22 -+-+--| 7824 |--+
>> | |______| |
>> === | ===
>> return | | |
>> (O)-----------------+------+-----+---(O) GND
>
> The reason I had TIP36C's in stock for that 2V/6A zener
> expt was due to a failed attempt at a dc supply similar
> to this. It was 36V/5A, and I had 40-odd volts Vraw, a
> discrete regulator, and TIP36 pass element. There was
> a series current-sense resistor and the usual shutdown
> on overload.
Tell us more about your "usual shutdown on overload," do
you mean a simple current limit?
> Shutdown did work ok, provided the overload was increased
> slowly. But (and this was the killer), a short on the
> output always resulted in a destroyed TIP36.
>
> After some time (and a quantity of TIP36's) I came to the
> conclusion that a short circuit took a (hot?) TIP36 outside
> the Safe Operating Area. A move to a Mosfet series element
> did send all problems away.
That's a very useful story Tony, I'd like to explore it a bit
further. You're absolutely right, that would be a big reason
a foldback limit would be preferred. In the circuit above the
7824 will supply the TIP36 with all the base drive it needs
to operate at very high output currents, way out of the SOA.
http://www.onsemi.com/pub/Collateral/TIP35A-D.PDF
For example, as long as the + supply source and its storage cap
holds up the 7824 will deliver say 1A to the TIP36's base, and
with a beta of 30 at 20A (even at 25C) the TIP36 can certainly
deliver 20A to the short. According to the ON semi data sheet,
it can deliver 20A for 300us if the voltage is below 13V, which
is 260 watts for 300us. Following the square-root thermal mass
rule, we know this means it can deliver 20A*40V = 800W for only
(260/800)^2 * 300us = 31.7us before the junction temp goes over
150C. Your 40V power supply no doubt had a big enough storage
cap to provide 20A for much more than 32us (e.g. 200uF drops
only 3.2V). This means that the TIP36's junction had meltdown!
Looking at the SOA we see that at 20V the TIP36 can sustain 6A,
until the heatsink overheats. So one easy fix to my thermostat
circuit would be to limit the maximum power to 120W with a big
3.3-ohm power resistor.
+ STO-140 PNP POWER
(O)---( TIP36
)<-- 5A --+--- E C --- 3.3 -----+---(O) +24V
fuse | B 5W | up to 2.0A
| | ______ | continuous
| | | | |
'- 22 -+-+--| 7824 |---+
| |______| |
=== | ===
return | | |
(O)-----------------------+------+------+---(O) GND
Hmm, pretty ugly. At the rated 2A it drops 6.6V (and we can
only excuse about 3V, needed for the 7824 and the TIP36 Vbe),
and it's a big monster. We can use a 5W resistor if we add
a 5A fuse, etc.
Does a straight current limit, as suggested by Ken Smith and
Martin Sikström help us? All mounted on the same heat sink,
as is natural and was suggested by Ken?
+ R1 PNP POWER
(O)-+-- 0.5 -- E C ------------+---(O) +24V
| 3W B TIP36 | up to 2.5A
| | ______ |
| | | | |
'--- 1.2 ----+--+--| 7824 |--+
R2 | |______| |
| | |
=== | ===
return | | |
(O)-----------------+-----+------+---(O) GND
Keeping in mind the imprecise nature of the current limit in
this scheme, we'll set the CL a bit high at 3A nominal, which
drops 1.5V across R1, the 0.5-ohm emitter resistor. We have
to estimate Vbe at 3A, but if it's 0.9V the voltage across R2
will be about 2.4V. According to the SGS Tompson data sheet
the 7824's short-circuit peak current is between 1.3 and 3.3A,
call it 2.2 +/-1A, and the TIP base only needs about 0.02A,
so we can select R2 = 2.4/2.2 = 1.1 ohms, call it 1.2 ohms.
This means the instantaneous TIP36 short-circuit current will
be between 1.3 and 5.9A according to my calculator (2.6 to 9A
to the load). For the low value, 2.6A to the load has little
margin for our spec'd operation. Under normal operation at 2A
R2 has about 1.9 volts which requires the LM7824 to deliver a
rather high 1.6A, and dissipate 17.5W at 35V supply.
Considering the high value, 5.9A at 35V on the TIP36 is 212W,
which can only be sustained for about 500us. The question is,
can the 7824's thermal limit be relied upon to reduce its
short-circuit current within 500us? Perhaps, but it would be
a close call... It's thermal mass is less than the TIP36,
but it sees a lower instantaneous current...
Given all this, my foldback scheme doesn't look so bad after
all. When operating at 2A to the load the 7824 is delivering
less than 0.1A and dissipating under 1W.
+ PNP POWER
(O)-+-- 0.18 --+-- E C -----------+------(O) +24V
| | B TIP36 | up to 2.5A
| R1 10 | Q1 |
| | | R2 |
| ZTX1055A +-----|----- 620 ----+
| small | Q2 | ______ |
| 2A pnp b | | | |
+------- e c --+-+--| 7824 |--+
'----+---- 22 -----+ |______| +---,
| | | | |
=== === | === 1k
return | | | | |
(O)-----+-------------+-----+------+---+--(O) GND
Comments?
Thanks,
- Win
Winfield Hill
Rowland Institute for Science
100 Edwin Land Blvd
Cambridge, MA 02142-1297
Ken Smith
Win Hill <wh...@mediaone.net> wrote:
[....]
Check National's data sheet for the LM78XX.
Note that above about 7V of drop the device's peak current decreases with
increasing forward voltage. This is an internal SOA protection circuit
that protects the output element of the chip.
I don't have the LM7824's actual data sheet but lets assume that the SOA
protection in it works about the same and that the current is about 1A.
R2 will have 1.2V across it.
If we take Vbe to be 0.7
R1 will have 1.2 - 0.7 = 0.5V
IR1 will be 1A. (this is below the SOA knee of the TIP36)
PQ1 = 1A * (35-0.5) = 34.5W
Tony Williams
> > The reason I had TIP36C's in stock for that 2V/6A zener
> > expt was due to a failed attempt at a dc supply similar
> > to this. It was 36V/5A, and I had 40-odd volts Vraw, a
> > discrete regulator, and TIP36 pass element. There was
> > a series current-sense resistor and the usual shutdown
> > on overload.
> Tell us more about your "usual shutdown on overload," do
> you mean a simple current limit?
Yes. Which probably meant about 1.5x to 2x
normal full load current under a dead-short.
Which probably also means that my "5A" above
is in error, because I do remember that the
transistor power dissipation was in limits
during a dead-short.
Indeed, bringing the overload on slowly (to a
dead-short) did not result in destruction of
the pass-transistor.
> That's a very useful story Tony, I'd like to explore it a bit
> further. You're absolutely right, that would be a big reason
> a foldback limit would be preferred. In the circuit above the
> 7824 will supply the TIP36 with all the base drive it needs
> to operate at very high output currents, way out of the SOA.
I surmised that switching a short on quickly
caused a too-rapid temperature rise in a part
of the die, and this resulted in second-breakdown.
This is the forward-biassed second-breakdown, that
does not need the transistor to go into avalanche
first. Mullard Technical Note No.15 "SOAR, the
basis for reliable circuit design" was the basis
of the supposition.
> http://www.onsemi.com/pub/Collateral/TIP35A-D.PDF
> For example, as long as the + supply source and its storage cap
> holds up the 7824 will deliver say 1A to the TIP36's base, and
> with a beta of 30 at 20A (even at 25C) the TIP36 can certainly
> deliver 20A to the short. According to the ON semi data sheet,
> it can deliver 20A for 300us if the voltage is below 13V, which
> is 260 watts for 300us. Following the square-root thermal mass
> rule, we know this means it can deliver 20A*40V = 800W for only
> (260/800)^2 * 300us = 31.7us before the junction temp goes over
> 150C. Your 40V power supply no doubt had a big enough storage
> cap to provide 20A for much more than 32us (e.g. 200uF drops
> only 3.2V). This means that the TIP36's junction had meltdown!
I didn't have the TI data for the TIP36 at that
time, but the curves in the Mullard Tech Note
indicated that at high Vceo (forward-biassed)
second breakdown occurred at collector currents
far below Icmax. Once started, (in an elemental
transistor on the die) second breakdown becomes
self-sustaining.
There is an interesting calculation in that App
Note that shows that if an elemental transistor
on a die rises greater than 16C above the mounting
base of the die, then current-crowding in that
elemental transistor tries to increase further.
Note that the shift to a power Mosfet pass-element
did send the problem away.
> + R1 PNP POWER
> (O)-+-- 0.5 -- E C ------------+---(O) +24V
> | 3W B TIP36 | up to 2.5A
> | | ______ |
> | | | | |
> '--- 1.2 ----+--+--| 7824 |--+
> R2 | |______| |
> | | |
> === | ===
> return | | |
> (O)-----------------+-----+------+---(O) GND
[huge snip]
I think that a regulator operating with high Vraw
does have the capability to get ugly, very quickly.
I have the (financial) scars to prove it. :-(
> Given all this, my foldback scheme doesn't look so bad after
> all. When operating at 2A to the load the 7824 is delivering
> less than 0.1A and dissipating under 1W.
> + PNP POWER
> (O)-+-- 0.18 --+-- E C -----------+------(O) +24V
> | | B TIP36 | up to 2.5A
> | R1 10 | Q1 |
> | | | R2 |
> | ZTX1055A +-----|----- 620 ----+
> | small | Q2 | ______ |
> | 2A pnp b | | | |
> +------- e c --+-+--| 7824 |--+
> '----+---- 22 -----+ |______| +---,
> | | | | |
> === === | === 1k
> return | | | | |
> (O)-----+-------------+-----+------+---+--(O) GND
Yes, I wish I had been able to put foldback in, but
the load tended to pull constant-current.
What's a ZTX1055A? It's not in my Zetex data book.
--
Tony Williams.
Win Hill
>
>> Given all this, my foldback scheme doesn't look so bad after
>> all. When operating at 2A to the load the 7824 is delivering
>> less than 0.1A and dissipating under 1W.
>>
>> + PNP POWER
>> (O)-+-- 0.18 --+-- E C -----------+------(O) +24V
>> | | B TIP36 | up to 2.5A
>> | R1 10 | Q1 |
>> | | | R2 |
>> | ZTX1055A +-----|----- 620 ----+
>> | small | Q2 | ______ |
>> | 2A pnp b | | | |
>> +------- e c --+-+--| 7824 |--+
>> '----+---- 22 -----+ |______| +---,
>> | | | | |
>> === === | === 1k
>> return | | | | |
>> (O)-----+-------------+-----+------+---+--(O) GND
>
> Yes, I wish I had been able to put foldback in, but
> the load tended to pull constant-current.
There's a simple scheme one can use to insure startup,
and restart, of ill-behaved loads on a foldback-limited
power supply. You add a second regulator in parallel,
set to a slightly lower voltage, so it's not normally
operating. This 2nd regulator uses 3-terminal ICs that
have built-in thermal limiting, like the ubiquitous LM317.
Basically the scheme is for this 2nd pathway to operate
long enough to start or restore operation of the primary
foldback-protected supply. If the load is truly shorted
or defective, the thermal limiting quickly sets in and
reduces the current being continuously delivered to the
defective load to a low level.
> What's a ZTX1055A? It's not in my Zetex data book.
http://www.zetex.com/3.0/pdf/ZTX1055A.pdf
I found it in the DigiKey catalog, and didn't check to
see if it was in the book, probably not the best part
for the job - Zetex makes lots of E-line transistors
with high gain at high currents.
The ZTX1055A has h_FE over 400 from 10mA to 1.5A, and
it maintains beta over 100 to nearly 3A (Vce = 10V).
I suspect they have other cool parts that can do that.
Where do you buy your Zetex parts?
Win Hill
> IR1 will be 1A. (this is below the SOA knee of the TIP36)
> PQ1 = 1A * (35-0.5) = 34.5W
ST specs the "foldback" style short-circuit current limit at
0.2A, but at the same time they spec the short-circuit peak
current at between 1.3 and 3.3A, as I used in my calculation.
I don't know how long this high short-circuit current level
is present, but we need it to be over in less than 500us to
protect the TIP36 in your scheme. Looking at the schematic,
http://us.st.com/stonline/books/pdf/docs/2143.pdf it seems
the s-c transistor Q15 should act quickly, bypassing effects
from the slow-changing voltage on the compensation cap. So
it may be that the short-circuit peak current can be safely
ignored. [We need Tony to make a measurement!]
However, this still leaves us with a very uncertain overall
current limit in practise. Figure 7 shows the IC's typical
operating CL increasing from 1 to 2.2A as the input/output
voltage difference is increased from 2 to 5V, and back to 1A
for a warmed up part at 17V. We can expect a further 40%
change for the typ-to-max variation, or 3A max. This means
that the TIP36 current can range from 1A to over 5A, and the
total current to the load can range from 2A to 8A, not a
very satisfactory level of current-limit control. If I try
decreasing R2 to help this situation, we fail to get our
desired 2A output capability over part of the input range.
Win Hill
>
> However, this still leaves us with a very uncertain overall
> current limit in practise. Figure 7 shows the IC's typical
> operating CL increasing from 1 to 2.2A as the input/output
> voltage difference is increased from 2 to 5V, and back to 1A
> for a warmed up part at 17V. We can expect a further 40%
> change for the typ-to-max variation, or 3A max. This means
I just realized that I've been ignoring Martin Sikström's
suggestion of an extra compensating diode. This should help
to improve the bad performance of the simplified circuit.
+ R1 PNP POWER
(O)-+-- 0.5 --- E C ------------+---(O) +24V
| 3W B TIP36 | up to 2.5A
| | ______ |
| R2 D1 | | | |
'-- 0.5 -|<|--+--+--| 7824 |--+
1n5402 | |______| |
| | |
=== | ===
return | | |
(O)------------------+-----+------+---(O) GND
Now the 7824 current-limit can range from 1 to 3A, and if
crudely analyzed, the TIP36 current should simply track it,
thereby doubling the output current limit, which will then
range from 2 to 6A. This may be more acceptable.
Martin Sikström
> I just realized that I've been ignoring Martin Sikström's
> suggestion of an extra compensating diode. This should help
> to improve the bad performance of the simplified circuit.
>
> + R1 PNP POWER
> (O)-+-- 0.5 --- E C ------------+---(O) +24V
> | 3W B TIP36 | up to 2.5A
> | | ______ |
> | R2 D1 | | | |
> '-- 0.5 -|<|--+--+--| 7824 |--+
> 1n5402 | |______| |
> | | |
> === | ===
> return | | |
> (O)------------------+-----+------+---(O) GND
>
> Now the 7824 current-limit can range from 1 to 3A, and if
> crudely analyzed, the TIP36 current should simply track it,
> thereby doubling the output current limit, which will then
> range from 2 to 6A. This may be more acceptable.
Thanks. I knew there must have been a better reason for the
extra diode than to just make calculating the current ratio
easier. Can we now rely on the 7824's overtemp protection
to rescue the TIP36 in case of an overload?
By the way, you've drawn the diode backwards.
Regards,
Martin
Mark Daniels
<wh...@mediaone.net> writes
That is a cool idea, Win. I must make a note of it for future reference.
Now, if only A of E was better indexed . . .
>
>> What's a ZTX1055A? It's not in my Zetex data book.
>
> http://www.zetex.com/3.0/pdf/ZTX1055A.pdf
>
> I found it in the DigiKey catalog, and didn't check to
> see if it was in the book, probably not the best part
> for the job - Zetex makes lots of E-line transistors
> with high gain at high currents.
They certainly do. The E-line package is nice for heatsinking - beats
the hell out of TO92!
>
> The ZTX1055A has h_FE over 400 from 10mA to 1.5A, and
> it maintains beta over 100 to nearly 3A (Vce = 10V).
> I suspect they have other cool parts that can do that.
I like Zetex transistors. Many of their parts have extra-ordinary
specifications and seem to be reliable, too.
>
> Where do you buy your Zetex parts?
I buy mine from RS Components and Farnell.
>
> Thanks,
> - Win
>
> Winfield Hill
> Rowland Institute for Science
> 100 Edwin Land Blvd
> Cambridge, MA 02142-1297
--
Mark Daniels
Winfield Hill
>
> By the way, you've drawn the diode backwards.
==> groan <==
Tony Williams
Win Hill <wh...@mediaone.net> wrote:
> > What's a ZTX1055A? It's not in my Zetex data book.
> http://www.zetex.com/3.0/pdf/ZTX1055A.pdf
> The ZTX1055A has h_FE over 400 from 10mA to 1.5A, and
> it maintains beta over 100 to nearly 3A (Vce = 10V).
> I suspect they have other cool parts that can do that.
Ah, thank you.
ZTX649/749 have their hfe spec'd out to 6A, but
the 25V Vceo is somewhat limiting. I keep the
ZTX651 and ZTX751 in stock as general purpose
do-anything small transistors. Mind you, that
quoted "1.5W" for E-line can be a bit of an
ambush, it only applies when pcb-mounted, with
short leads, to at least 1 sq-inch of copper.
The free-air dissipation is nearer 300mW.
> Where do you buy your Zetex parts?
Farnell, of course. Are they still Newark in the US?
--
Tony Williams.
Win Hill
>
> i want a simple regulated power supply design with 7824
>
> thank you,
> k.praveen.
Mr. Praveen has written me asking for a more simple design.
He points out that he's not familiar with PNP transistors.
I half-heartedly replied that he could use an NPN transistor
placed _after_ the 7824, making up for its Vbe drop by raising
the ground leg of the 7824 with a diode (but this scheme has
NO current-limit protection). Or he could experiment with a
few paralleled 7824s (but they have an output-voltage spec.
spread of 2V, too much to equalize the currents with series
resistors). But a better answer may be to just forget using
fixed-voltage regulators like the 78xx family, and instead
to base the design on adjustable LM317-style IC regulators.
http://www.national.com/ds/LM/LM117.pdf
There are several possible simple approaches using LM317
chips, one goes like this,
_______
| | 0.025 to
o--+--| LM317 |----- 0.25 ----------,
| | ___|___ |
| |___| | |
+----|-| LM317 |----- 0.25 ------+
| | | ___|___ |
| | |___| | | +1.2 to 28V
+----|---|-| LM317 |--+-- 0.25 --+--+---o 0 to 4A
| | | | | | |+
| | | |_______| | === 4.7uF
=== | | | R1 47 |
| | | | | gnd
| '---+---+--------+
gnd | R2 = (Vo-1.25)*46.5
1k pot, 2W
|
gnd
Current-equalizing resistors are a good idea because there
may be some hysteresis in thermally-limited current from
the LM317. Because the tolerance on Vref = 1.25V is low,
under 5mV, small resistor values can be used, which means
low voltage drops and low power dissipation.
Thanks,
- Win
Winfield Hill
Rowland Institute for Science
100 Edwin Land Blvd
Cambridge, MA 02142-1297
Win Hill
>
> ... But a better answer may be to just forget using
Here's another approach, which has the advantage of avoiding
significant power dissipation in the voltage-control element.
It must be loaded with a minimum of 10mA per LM317 used. I
simply made a guess for the compensation-component values,
indeed none may be necessary. Tony?
_____
| LM |
o--+-| 317 |--- 0.025 -----------,
| |_____| to 1.0, etc. |
| | _____ |
| | | | |
+----|-| 317 |--- 0.25 -------+ Vo = 1 + 2.5 R2/R1
| | |_____| |
| | | _____ | +2.5 to 27.5V
+----|----|-| |--+- 0.25 -+-+----+---o 0 to 4A
| | | | 317 | | | |+
=== | | |_____| 470 | === 10uF
| | | | | 10nF | |
| '----+----+---+-+--||-, | gnd
gnd | | R2
__|__ 2.2k | 10k pot
| TL | | |
| 431 |----+----+
|_____| |
| R1 1.00k
gnd |
gnd
Thanks, that's my ASCII drawing for the day.
Win Hill
I received an email asking, "For symmetry why not connect the
470R to junction of the .25's?"
Good question.
As it is now the TL431A current is fixed at 1.25/470 = 2.66mA,
plus about 0.15mA of 317 adj-pin current, but if it's placed
at the junction of the current-equalizing resistors, then the
current drops with increasing load. If a choice of 1.0 ohms
is made (many folks can't easily get small values) this
happens early and starves the TL431 long before the 4A current
max is reached (The TL431 needs 1mA worst case, 0.4mA typical).
Incidentally, by using the reader's suggested location, and by
chosing the current-sensing resistor value appropriately, an
effective current limit can be created much lower than sum of
all the LM317 regulators.
Among the appealing features of this circuit and its predecessor
are the low cost and easy availablility of the parts and the
increased reliability due to over-temperature fault protection
of all the power components.
Thanks,