Is there any efficient and not too expensive way to generate the ~5V for
the PIC and some other devices out of the 230V AC Voltage from the net? Using a
7805 and other linear regulators is IMHO much too inefficient ...
I need about 2 - 150mA at 5V to drive my PIC 16c84 based IR-remote receiver
that will run 24 hours a day ...
I thought of switching regulators e.g. TOP210 but I don`t know where to get the
transformer (TDR1) mentioned in the data sheet and the TOP210 itself, because I
need only a few peaces for my project ...
Does anyone know how efficient normal transformers (230V => e.g. 6V) with diode
based rectifiers are? If it is high enough so that a DC/DC switching regulator
like the LM2574-5.0 (only about 77% efficiency :-( ) could be used? This would
be pretty easy, because the mentioned inductors are very easy to get ...
PS. Does anyone know about audio preamplifiers with digital Volume control?
Or is it better to use e.g. the TDA 1524 with digital potentiometers?
PPS. Sorry for my bad english, but school is far away now ;-)
Bernd Grunwald wrote:
>
> Is there any efficient and not too expensive way to generate the ~5V for
> the PIC and some other devices out of the 230V AC Voltage from the net? Using
a
> 7805 and other linear regulators is IMHO much too inefficient ...
If you have more time than money to spare, then reverse engineering
cheap mass-produced consumer electronics like alarm-clock(radio's) might
be interesting. You get a reasonable power supply and a case - often
with buttons, sometimes even a display and maybe some other interesting
parts that spawn ideas for additional features to your design.
> I need about 2 - 150mA at 5V to drive my PIC 16c84 based IR-remote receiver
> that will run 24 hours a day ...
>
> I thought of switching regulators e.g. TOP210 but I don`t know where to get
the
> transformer (TDR1) mentioned in the data sheet and the TOP210 itself, because
I
> need only a few peaces for my project ...
>
> Does anyone know how efficient normal transformers (230V => e.g. 6V) with
diode
> based rectifiers are? If it is high enough so that a DC/DC switching regulator
> like the LM2574-5.0 (only about 77% efficiency :-( ) could be used? This would
> be pretty easy, because the mentioned inductors are very easy to get ...
Why do you care so much about efficiency if you are going to feed it
from mains voltage anyway?
> Bernd Grunwald wrote:
> >
> > Is there any efficient and not too expensive way to generate the ~5V for
> > the PIC and some other devices out of the 230V AC Voltage from the net?
Using
> a
> > 7805 and other linear regulators is IMHO much too inefficient ...
I know you can get some very nifty little ic's that do just that with
no transformer etc. I know Maplin used to stock something like
that...If you can wait till tomorrow, I'll see if I can get more
details for you.
Jacques Vrey
Iscor Profile Products Newcastle
Tel:27-(0)3431-48759 jvreyKILLspamit.new.iscorltd.co.za
One of the application notes describes a transformerless power supply based
on a capacitor and zener if you want more info I can look it up and some old
posts to the list which include the maths.
>
> Why do you care so much about efficiency if you are going to feed it
> from mains voltage anyway?
>
>
> Joost
>
Maybe he realizes that the mains lines are often supplied by
burning fossil fuels and by making this small additional
effort in his design he'll be able to reduce the amount of
fossil fuels burnt each day? :)
> On Wed, 3 Sep 1997, Joost Kooij wrote:
>
> >
> > Why do you care so much about efficiency if you are going to feed it
> > from mains voltage anyway?
> >
> >
> > Joost
> >
>
> Maybe he realizes that the mains lines are often supplied by
> burning fossil fuels and by making this small additional
> effort in his design he'll be able to reduce the amount of
> fossil fuels burnt each day? :)
Life doesn't work out like that - witness computing power. You'd
expect programs to get faster. No. They get more complex instead.
(hi Bill)
Cars are another instance.
MikeS
<.....mikesmith_ozKILLspam.....relaymail.net>
>> Bernd Grunwald wrote:
>> >
>> > Is there any efficient and not too expensive way to generate the ~5V for
>> > the PIC and some other devices out of the 230V AC Voltage from the net?
> Using
>> a
>> > 7805 and other linear regulators is IMHO much too inefficient ...
>
>I know you can get some very nifty little ic's that do just that with
>no transformer etc. I know Maplin used to stock something like
>that...If you can wait till tomorrow, I'll see if I can get more
>details for you.
>
The series resistor/zener across the mains is hideously inefficient - and
if you want more than a few mA you need a very large heat-sinked resistor.
(I can provide circuit and design details if anyone wants.)
There is an IC HAR2405 from Harris Semiconductor that functions as a
direct-off-line 5V regulator. It uses a switching circuit to charge an
external reservoir capacitor to (Vout + a few volts) and an on-chip series
regulator to give a smooth output. You can add an external resistor to
adjust the output.
The Harris version supplies up to 50mA. There is also a Lucent (formerly
AT&T?) version that provides up to 100mA. I couldn't find the datasheet on
the web for the Lucent part, but I have the data-sheet for the Harris
(which I can eMail) so I can add:
It needs external components (especially at higher current draws)
to current-limit the reservoir charging circuit
It also needs a large series resistor (although not as big as the
zener/resistor method)
I think you should be able to build a direct off-mains supply using a
similar principle with an SCR as the switching element, and a
three-terminal micropower regulator (or even a 78L05). You could even use
the PIC to switch the gate of the SCR - if you can think of a way to start
the circuit up (and if the PIC locks up with the gate energised....BANG!!)
Oh yes... and with all these methods, remember that you have no mains
isolation, and one failed resistor results in 230v across the
pic/components/user/etc.
I think a "wall wart" is a very inexpensive way of getting power
for a PIC. I've now got several products that use a 12V, 500mA unit
driving a 7805 regulator. I think the wall warts are about $5 or less in
reasonable quantities.
If you can put up with a nonisolated supply, consider the Maxim
MAX600, 601, 602, 610, 611, 612. Info should be available at http://www.maxim-ic.com .
[..}
> > Does anyone know how efficient normal transformers (230V => e.g. 6V) with
> diode
> > based rectifiers are? If it is high enough so that a DC/DC switching
regulator
> > like the LM2574-5.0 (only about 77% efficiency :-( ) could be used? This
would
> > be pretty easy, because the mentioned inductors are very easy to get ...
>
> Why do you care so much about efficiency if you are going to feed it
> from mains voltage anyway?
I have already many devices at home that run 24hours a day (2 video recorder,
sat receiver, TV`s, ...) and I want to reduce costs ... So, why waste lots of
energy for *nothing* ????? (I have more efficient possibilities to heat my
rooms ;-) )
> > I need about 2 - 150mA at 5V to drive my PIC 16c84 based IR-remote
> > receiver that will run 24 hours a day ...
>
> Hmm, cannot you use the sleep mode to reduce power consumption ?
No, this is only the about 2mA of the whole thing ... :-(
> > > Is there any efficient and not too expensive way to generate the ~5V for
> > > the PIC and some other devices out of the 230V AC Voltage from the net?
> Using
> > a
> > > 7805 and other linear regulators is IMHO much too inefficient ...
>
> I know you can get some very nifty little ic's that do just that with
> no transformer etc. I know Maplin used to stock something like
No, sorry, I do need a transformer because I don`t want to get burned, if I
insert my Headphone in the included amplifier that will be connected somehow to
the digital part of the device (volume control ...) ;-)
Wall warts are indeed an economical supply for many projects, but I have so
many of them in behind my stereo stand (and of course, they are too wide
for one slot on the power bar), that they almost take up more space than
the equipment they power.
I realize they eliminate the designer's need to get safety approval for
their hardware, and sometimes the device being powered is just too small
for an internal P/S, but why doesn't someone sell reasonably priced
"enclosed" PSU's that can be designed into the case of the equipment being
powered. This strategy allows devices to get automatic approval for
telephone devices by using a pre-packaged and approved interface module.
Why not pre-approved simple PSU's. And I am not talking about those
commercial/industrial regulated supply modules that cost more than the
project. Something unregulated, maybe even unrectified or filtered. And
if the AC connection is the issue for safety, a power cord could be
'moulded' in, or the module could be designed to mount on the rear panel of
the project, with an AC receptacle (2 or 3 pin) and maybe a fuse socket
exposed on the back.
I think a "wall wart" is a very inexpensive way of getting power
for a PIC. I've now got several products that use a 12V, 500mA unit
driving a 7805 regulator. I think the wall warts are about $5 or less in
reasonable quantities.
If you can put up with a nonisolated supply, consider the Maxim
MAX600, 601, 602, 610, 611, 612. Info should be available at http://www.maxim-ic.com .
> Wall warts are indeed an economical supply for many projects, but I have so
> many of them in behind my stereo stand (and of course, they are too wide
> for one slot on the power bar), that they almost take up more space than
> the equipment they power.
Actually, what I'd like to see would be some smaller wall warts. Although
there have been a few that use HF transformers with optocoupled feedback
(an oscillator on the "hot" side generates a signal which is fed through the
transformer; the oscillator adjusts itself according to feedback about the
regulated voltage) they've tended to be expensive. What I'd like to see,
though, would be a wall wart that skipped the feedback step and simply gave
you pretty much unfiltered DC which was 18-24v peak. The user application
could then regulate the output in whatever manner was desired (7805 plus a
resistor and a couple caps being a nice choice). This should be cheaper
than an optocoupled feedback circuit, but would allow the wall wart to fit
in a much smaller plug than normal.
Alternatively... what if someone just produced a multi-output power supply
with output plugs suitable for use with all those appliances? Should be
much cheaper than a foot of wall warts; smaller too. Think that'd be at
all practical?
Martin R. Green wrote:
<snip>
> I realize they eliminate the designer's need to get safety approval for
> their hardware....
<snip>
I don't know if the rules have changed in the last few years but there
used to be a "Catch-22" in the CSA and UL (North America) rules. These
"Wall warts" are approved only as long as they are used "stand-alone".
i.e.: powering NOTHING (strange, but true). As soon as it's attached to
(or plugged into) something it was deemed a "new" and unapproved
device. You had to go through approval testing with both your product
and the wall-mounted power supply and both had to pass. If this has
changed I'm sure somebody will let me know.
--Matt
At 09:00 AM 9/4/97 -0600, you wrote:
>Martin R. Green wrote:
><snip>
>> I realize they eliminate the designer's need to get safety approval for
>> their hardware....
><snip>
>I don't know if the rules have changed in the last few years but there
>used to be a "Catch-22" in the CSA and UL (North America) rules. These
>"Wall warts" are approved only as long as they are used "stand-alone".
>i.e.: powering NOTHING (strange, but true). As soon as it's attached to
>(or plugged into) something it was deemed a "new" and unapproved
>device. You had to go through approval testing with both your product
>and the wall-mounted power supply and both had to pass. If this has
>changed I'm sure somebody will let me know.
>--Matt
In Canada, only the fire hasards and electrical shock is checked by the CSA.
Normally, all equipment below 40 volts doesn't need to be checked.
(The car equipment with 12 volts is an example.)
Jean-Pierre Bourdeau,
Courrier electronique (E-mail): TakeThisOuTbourdeauEraseMEspam_OUTdsuper.NOSPAM.net
Indicatif radio-amateur (Amateur radio callsign): VE2EXU
Page W3 (Web Page): http://oracle.dsuper.net/~bourdeau/
Courrier electronique via AMSAT (E-mail via AMSAT): RemoveMEVE2EXUTakeThisOuTAMSAT.nospam.ORG
Radio-amateur par paquets (Amateur radio packet):
VE2EXU@VE2CEV.#MTL.nospam.PQ.CAN.NOAM
Enlevez l'expression suivante lors de vos reponses par courrier electronique;
(Remove the following sentence in your reply with E-mail): .nospam .NOSPAM
On Thu, 4 Sep 1997 09:55:50 -0400 "Martin R. Green" <elimarEraseME.....BIGFOOT.COM>
writes:
>
>I realize they eliminate the designer's need to get safety approval
>for
>their hardware, and sometimes the device being powered is just too
>small
>for an internal P/S, but why doesn't someone sell reasonably priced
>"enclosed" PSU's that can be designed into the case of the equipment
>being
>powered. This strategy allows devices to get automatic approval for
>telephone devices by using a pre-packaged and approved interface
>module.
> Why not pre-approved simple PSU's.
Ya know, that's really an interesting idea. Actually, it'd be
just a wall wart with an IEC input connector (instead of the various
country mains connectors) and a couple tapped holes to bolt the thing to
the rear panel. Actually, I have seen little desk top enclosed supplies
that could probably be bolted into a cabinet in this fashion. It'd be
great, though, to have a 12 VDC 500mA unregulated supply that you just
bolt to the rear panel, and have it cost $5.
>>"Wall warts" are approved only as long as they are used
>"stand-alone".
>>i.e.: powering NOTHING (strange, but true). As soon as it's attached
>to
>>(or plugged into) something it was deemed a "new" and unapproved
>>device. You had to go through approval testing with both your
>product
>>and the wall-mounted power supply and both had to pass. If this has
>>changed I'm sure somebody will let me know.
>>--Matt
>
>In Canada, only the fire hasards and electrical shock is checked by
>the CSA.
>Normally, all equipment below 40 volts doesn't need to be checked.
>(The car equipment with 12 volts is an example.)
In the US I don't think that the government needs to approve anything
about a product before it can be legally sold other than it's potential
to cause RF interference. But in the litigious USA (of, by, and for
trial lawyers) it's pretty much essential to obtain product liability
insurance and the insurance company will likely require UL or similar
approval.
It isn't hard to conceive of a wallwart-powered product that is a
substantial fire and shock hazard, even if the wallwart isn't.
On Wed, 3 Sep 1997 22:34:19 +0200 Bernd Grunwald <EraseMEemboBSE.NETESTATE.DE>
writes:
>
>No, sorry, I do need a transformer because I don`t want to get burned,
>if I
>insert my Headphone in the included amplifier that will be connected
>somehow to
>the digital part of the device (volume control ...)
If your project has any hardwired connections to the outside they must be
isolated from the power line. Usually the best way to do this is to
isolate the entire project circuit from the power line by using a
transformer-type power supply, especially if it only needs a few watts of
power. RF connections can be inexpensively isolated by coupling the
signal and ground leads through safety-agency approved capacitors. Many
TV sets use this method on the antenna lead. For other types of signals
the cost of isolating them is usually higher than that of isolating the
power supply.
The efficiency of small transformers is low. They will use a watt or so
of magnetizing power without any load at all on the secondary. Using a
transformer at less than its full voltage rating should help this a lot,
since the iron will not approach saturation then. After the transformer,
loss in the regulator can be significant. Reducing the power demand of
the project is the best way around this. If most of the current goes to
a relay, try and run it off of the unregulated supply. The higher
voltage relay coil will use less current.
Well, I've thought about that one before, but the problem I've always had
with the idea is that it would be too easy for even experienced users to
plug a 9V device into a 15V socket (you get the idea). Of course, this is
possible with warts too if they use the same plug and you get them mixed
up. Most of the devices I'm currently running from warts are not
internally regulated. But the current situation, 1 wart per device, is
ridiculous.
> Wall warts are indeed an economical supply for many projects, but I have
so
> many of them in behind my stereo stand (and of course, they are too wide
> for one slot on the power bar), that they almost take up more space than
> the equipment they power.
Actually, what I'd like to see would be some smaller wall warts. Although
there have been a few that use HF transformers with optocoupled feedback
(an oscillator on the "hot" side generates a signal which is fed through
the
transformer; the oscillator adjusts itself according to feedback about the
regulated voltage) they've tended to be expensive. What I'd like to see,
though, would be a wall wart that skipped the feedback step and simply gave
you pretty much unfiltered DC which was 18-24v peak. The user application
could then regulate the output in whatever manner was desired (7805 plus a
resistor and a couple caps being a nice choice). This should be cheaper
than an optocoupled feedback circuit, but would allow the wall wart to fit
in a much smaller plug than normal.
Alternatively... what if someone just produced a multi-output power supply
with output plugs suitable for use with all those appliances? Should be
much cheaper than a foot of wall warts; smaller too. Think that'd be at
all practical?
> Is there any efficient and not too expensive way to generate the ~5V for
> the PIC and some other devices out of the 230V AC Voltage from the net? Using
a
> 7805 and other linear regulators is IMHO much too inefficient ...
> I need about 2 - 150mA at 5V to drive my PIC 16c84 based IR-remote receiver
> that will run 24 hours a day ...
> I thought of switching regulators e.g. TOP210 but I don`t know where to get
the
> transformer (TDR1) mentioned in the data sheet and the TOP210 itself, because
I
> need only a few peaces for my project ...
Bernd, using an off-line switcher for your application sounds like
crazy to me. On the other hand, the capacitor-zener-resistor
scheme is cheap but won't save power.
> Does anyone know how efficient normal transformers (230V => e.g. 6V) with
diode
> based rectifiers are? If it is high enough so that a DC/DC switching regulator
> like the LM2574-5.0 (only about 77% efficiency :-( ) could be used? This would
> be pretty easy, because the mentioned inductors are very easy to get ...
I don't think you can gain much efficency with a switcher if your
power requirements are so low. What I would do is use an ordinary
full-bridge-center-tapped supply with a reasonably sized transformer,
and if I was that concerned about power I'd use schottkys and
perhaps a low-dropout regulator like the LT1086. But have in mind
that we're talking of about 600mW of losses for the ordinary supply
when the output is at 150mA, and with the changes you would reduce it
to not less than 400mW by spending about $3 or $4 more, besides
bothering to get less-common components. I think that would take
years to be profitable.
> Well, I've thought about that one before, but the problem I've always had
> with the idea is that it would be too easy for even experienced users to
> plug a 9V device into a 15V socket (you get the idea). Of course, this is
> possible with warts too if they use the same plug and you get them mixed
> up. Most of the devices I'm currently running from warts are not
> internally regulated. But the current situation, 1 wart per device, is
> ridiculous.
While there are a few devices I've seen that rely on regulated wall-wart
supplies, the vast majority do the regulation internal. I think it's 'cos
the price difference between unregulated and regulated warts is higher than
the price of a 7805 and two caps (plus, a device with a regulator on-board
is less likely to go "boom" if connected to the wrong wart). A number of
them also have a protection diode between the input and the regulator (cheap,
but it means they'll run off just about anything).
What I'd like to see would be a standard for plugs/sockets such that a plug
would fit in any compatible socket. Probably a two-pin molex-style connector,
polarized, and with notches or projections on the top and bottom to code for
voltage, current, and AC/DC. A device which could accept anything from 9 to
15 volts AC or DC could have lots of notches in it, to fit a variety of plugs,
while a device that needs +5 DC would just have one notch for +5 and one for
DC (and one for its required current dose).
One extra thing that having separate warts for each devices gives us is
isolation _between_ devices. If you connect 2 or more devices to a common
universal P/S, you would run the risk of ground loops, and since many wart
powered devices handle audio and/or video signals, this could be a problem.
Any comments?
I like the idea of voltage compatibility coded plugs, but it would be nice
to come up with a solution that works for the thousands (millions?,
billions?) of wart powered devices already in people's homes. Maybe too
tall an order.
> Well, I've thought about that one before, but the problem I've always had
> with the idea is that it would be too easy for even experienced users to
> plug a 9V device into a 15V socket (you get the idea). Of course, this
is
> possible with warts too if they use the same plug and you get them mixed
> up. Most of the devices I'm currently running from warts are not
> internally regulated. But the current situation, 1 wart per device, is
> ridiculous.
While there are a few devices I've seen that rely on regulated wall-wart
supplies, the vast majority do the regulation internal. I think it's 'cos
the price difference between unregulated and regulated warts is higher than
the price of a 7805 and two caps (plus, a device with a regulator on-board
is less likely to go "boom" if connected to the wrong wart). A number of
them also have a protection diode between the input and the regulator
(cheap,
but it means they'll run off just about anything).
What I'd like to see would be a standard for plugs/sockets such that a plug
would fit in any compatible socket. Probably a two-pin molex-style
connector,
polarized, and with notches or projections on the top and bottom to code
for
voltage, current, and AC/DC. A device which could accept anything from 9
to
15 volts AC or DC could have lots of notches in it, to fit a variety of
plugs,
while a device that needs +5 DC would just have one notch for +5 and one
for
DC (and one for its required current dose).
>
>Actually, what I'd like to see would be some smaller wall warts. Although
>there have been a few that use HF transformers with optocoupled feedback
>(an oscillator on the "hot" side generates a signal which is fed through the
>transformer; the oscillator adjusts itself according to feedback about the
>regulated voltage) they've tended to be expensive. What I'd like to see,
>though, would be a wall wart that skipped the feedback step and simply gave
>you pretty much unfiltered DC which was 18-24v peak. The user application
>could then regulate the output in whatever manner was desired (7805 plus a
>resistor and a couple caps being a nice choice). This should be cheaper
>than an optocoupled feedback circuit, but would allow the wall wart to fit
>in a much smaller plug than normal.
>
I reverse engineered the power supply that Igloo uses to power their
thermoelectric cooled ice chests and it is semi-regulated without
opto-couplers, etc. It is a self-oscillating switch-mode converter, similar
to that used in low voltage florescent lamps but is powered off the mains.
It has three secondary windings. One to provide about 13.5 volt, 5 amp
output another returns surplus power back to the filter capacitor accross
the rectified mains input, and another to provide the oscillation feedback.
The system has fairly good load regulation, but its output is affected by
the mains voltage, just like a wall wart. A small version of this would be
just what you need. It uses 2 transistors, a MOSFET for the
oscillator/driver and I think the small bipolar one is used to limit the
duty cycle by cutting off the feedback/drive signal when the current
exceeds a preset limit. This keeps it running in the continuous mode (core
not allowed to saturate) for better efficiency. The circuit is extremely
simple and could easily be made into a small, cheap wall "pimple". I figure
Igloo would not appreciate me publicizing their design on the piclist.
Why not use a transformer and pre-regulate the voltage by clipping it
off when it exceeds a value 1.5V higher than the regulated output
voltage? A standard transistor based regulator with low voltage drop
regulate down to output voltage. Only 1.5V voltage drop makes it
possible to do it up to 500mA without using heat-sink. It is the same
methode as used in direct-main power supply's and they work without
heat-sink's, even if the input voltage is connected direct to main.
> Ya know, that's really an interesting idea. Actually, it'd be
> just a wall wart with an IEC input connector (instead of the various
> country mains connectors) and a couple tapped holes to bolt the thing
> to the rear panel.
Martin R. Green wrote:
>
> Well, I've thought about that one before, but the problem I've always had
> with the idea is that it would be too easy for even experienced users to
> plug a 9V device into a 15V socket (you get the idea). Of course, this is
> possible with warts too if they use the same plug and you get them >mixed up.
Did that with a AIWA walkman. Gave 12V instead of 3V. The wart was still
set to 12V and difficult to see... Walkman played at fastforward for
about 3sec and never again....
> Most of the devices I'm currently running from warts are not
> internally regulated. But the current situation, 1 wart per device, is
> ridiculous.
>
> JMNSHO - Martin R. Green
> KILLspamelimarspamBeGonebigfoot.com
>
bse.netestate.de
