simple buck switcher for LED drive from 9v?

[Walter Harley]

If it's just to indicate power is on, why not use a flasher of some
kind? The average power can be made quite low. Numerous designs have
been posted to alt.binaries.schematics.electronic

You ever try to play music at one tempo, while something is flashing at a
different tempo?

I did play around with just pulsing the LED at a rate too fast to see the
flicker, but perceived brightness seems pretty closely related to average
power, so that didn't buy me anything.

I have actually had suggestions to use the power LED as a metronome. But a
metronome that can't be turned off is painful to work with; and actually,
it's hard to synchronize well to a blinking light - audible metronomes work
better than visual ones. The visual ones distract without aiding.

One thing you sometimes see is LED's that turn on if the battery is getting
low; or, that turn on briefly when the device is turned on, and then fade a
second later (but don't turn on at all if the batteries are too low).
Informal feedback I got from musicians was that they'd forget to turn it off
if I did it that way - too bad, it would be very easy to implement, and
power-efficient!

The device at present is turned on by plugging an instrument into it. This
is actually a bad scheme because it means I can't use an ordinary
inexpensive input jack; but it corresponds to how most guitar stompbox
effects work, so my clientele are very comfortable with it.

Speaking as an amateur musician myself, I hate to denigrate the intelligence
of my customers - but let's just say that musical brilliance is often
accompanied by technological clumsiness.

The altogether best solution would be to automatically turn the power on
when they're playing, and off otherwise, and not worry about a power LED.
But I worry about RFI and whatnot causing it to inadvertently turn on. I
suppose I could work that out - maybe that's where I should focus my time.
(But I do need to learn about switchers somehow!)

 

[Winfield Hill]

Walter Harley wrote...

The real application is a battery-powered headphone amplifier
(used by electric bassists when practicing) that I make and
sell. I make 50-100 of them a year - not exactly the big time
but enough that manufacturing efficiency and parts cost matter.
Right now, the power LED is a big factor in the battery life.

I think a better approach is to lose the LED and instead
adopt a power-on pushbutton with a say six-hour auto-off
active time, etc. If your circuit is properly efficient,
its quiet-audio quiescent drain will be a small fraction
of the loud audio current, so the extra few hours sitting
around unused won't matter much.

It's the general wisdom that one to six hours is far too
long for a simple RC timer, although perhaps Tom Bruhns
might disagree, and the usual solution is along the lines
of "The Hour of Power" in AoE, page 972. Our concept was
solid: run a timer at a reasonably-short period and count
it down to the desired long time intervals. (Actually, I
almost throw up whenever I study the details of that old
circuit - hopefully it was one of Paul's contributions,
but probably not: I was a big fan of the '4536 with its
4-bit binary programmability.) As decades went by the
'4536 faded and a sister design from the same era (early
70's) became the dominant timer chip. I'm referring to
the mc14541 / cd4541 types. My replacement designs using
the '4541 have a simple elegance, and one will go into the
next edition. I suggest that you check it out. These days
they're manufactured by Fairchild, TI, ON Semi and ST, and
are sold for as little as 11.2 cents, qty 25. Very Cool.

Thanks,
- Win

whill_at_picovolt-dot-com

 

[Bill Bowden]

[email protected] wrote...

Not until it's exhausted, the inductor current drops a little,
given by the amount of hysteresis, see my answer to Michael.

Thanks,
- Win

whill_at_picovolt-dot-com

The circuit simulates ok with Circuitmaker using these values,
but I don't see the use for R6? Seems to work the same with
or without it. Maybe it improves operation at higher frequencies?
Switching points at junction of R1/R2 are 180mV and 480mV with
a 9 volt supply and frequency is about 66Khz. LED current is about
15mA with 9 volt supply and about 12mA with 5 volt supply.

L1 - 2mH
D1 - 1N914
Q1 and Q3 - 2N3904
Q2 and Q4 - 2N3906
R1- 22
R2- 10K
R3- 51K
R4- 510K
R5 - 100K
R6 - 4.7K or none?
R7 - 100K
R8 - 10K

-Bill

 

[John Popelish]

The circuit simulates ok with Circuitmaker using these values,
but I don't see the use for R6? Seems to work the same with
or without it. Maybe it improves operation at higher frequencies?
Switching points at junction of R1/R2 are 180mV and 480mV with
a 9 volt supply and frequency is about 66Khz. LED current is about
15mA with 9 volt supply and about 12mA with 5 volt supply.

L1 - 2mH
D1 - 1N914
Q1 and Q3 - 2N3904
Q2 and Q4 - 2N3906
R1- 22
R2- 10K
R3- 51K
R4- 510K
R5 - 100K
R6 - 4.7K or none?
R7 - 100K
R8 - 10K

-Bill

I think R6 has to be in the range of 270 to 470 ohms in order to
significantly speed up the turn off of Q4. Of course, this implies
some other adjustments to increase the total base drive a bit.

 

[John Woodgate]

I read in sci.electronics.design that N. Thornton <[email protected]>

Well, Win's circuit is pretty, tho I would also say there are easier
ways. An ultrabright LED and a resistor would I think be the easiest way
to get sub-miliiamp drain.

I agree. A **low-current** LED is likely to be adequately bright at not
much over 1 mA. The *total* power is thus 9 mW. Using a normal LED at
1.5 V/10 mA, the LED power alone is 15 mW, so your switcher would need
to be 150% efficient. No doubt you can find information on the design of
such circuits on the web. (;-)

 

[Winfield Hill]

Bill Bowden wrote...

The circuit simulates ok with Circuitmaker using these values...
LED current is about 15mA with 9 volt supply and about 12mA with
5 volt supply.

And the battery current was...?

Thanks,
- Win

whill_at_picovolt-dot-com

 

[Winfield Hill]

Walter Harley wrote...

"Winfield Hill" wrote...

At about US$0.50 in unit quantities from Mouser, that's a nice component,
handy for switching the sort of midrange currents we've been talking about;
I remember you mentioning it in AoE. When hunting for MOSFETs I figured
that one of the few things about electronics that might have changed
substantially since AoE 2 was published was the state of the art in
affordable/available MOSFETs, so I wasn't sure it was still a good reference
for that. Has so little changed? Or is it just that the changes have
focused on very-high-current components?

If you had wanted a surface-mount part, you'd have needed a new
part number (we'll have some of those in the next edition).

But most of the jelly-bean and "classic" stuff that we suggested
is still valid. For example, several times recently I conducted
intensive searches for an updated semi-precise low-cost jelly-bean
JFET opamp, with +/-15V supply capability in miniDIP packages.

Result: sticking with our book's recommendation, LF411 and LF412.
Seems most of the newer parts are tiny surface-mount low-voltage
thingamajigs aimed at the cell-phone market, etc.

Thanks,
- Win

whill_at_picovolt-dot-com

 

[N. Thornton]

Hi


There are panel mounting bulb holders that you can put the LED in,
where the LED will be completely out of reach of the bolted in place
cover.

FWIW I have used ultrabrights at below 1mA.

But the first thing I try to do with LEDs is to get them in the main
circuit somewhere, some place where they take the place of a bit of
resistance, where the average power flow is sufficient for them, and
where the ac component is symmetrical, maintaining constant
brightness. That method can cut your LED draw down to zero. Just look
for some where where you have a pretty much consistent Vdrop at 1mA or
so. Bias chains are probably the first place to look.


Regards, NT

 

[N. Thornton]

Hi


Once upon a time in Television land, power supplies used dropping
resistors, and viewers made toast over them. Then along came the type
of PSU in this thread, a dropping R plus switching tr or triac.
Because the switcher introduced some regulation, no loading R was
needed to stabilise the V, so these sets would not produce toast. Then
later along came the switchers we know today, using inductors for high
efficiency.

Very very occasionally a switched dropping R is still a reasonable
choice, but only very very occasionally.


BTW in the dropper supply era, there were some not very reputable
appliances using 'curtain burners'. These were mains leads with the
dropping R wound into the lead, and used the length of lead to
dissipate the heat. These were primarily used on small radios where
the heat could not be dissipated in such a small case. Unfortunately
the heat had a tendency to be dissipated either in the listener's
curtains, or the table or shelf on which the lead was unrealisingly
coiled.


Regards, NT

 

[N. Thornton]

Hi


Once upon a time in Television land, power supplies used dropping
resistors, and viewers made toast over them. Then along came the type
of PSU in this thread, a dropping R plus switching tr or triac.
Because the switcher introduced some regulation, no loading R was
needed to stabilise the V, so these sets would not produce toast. Then
later along came the switchers we know today, using inductors for high
efficiency.

Very very occasionally a switched dropping R is still a reasonable
choice, but only very very occasionally.


BTW in the dropper supply era, there were some not very reputable
appliances using 'curtain burners'. These were mains leads with the
dropping R wound into the lead, and used the length of lead to
dissipate the heat. These were primarily used on small radios where
the heat could not be dissipated in such a small case. Unfortunately
the heat had a tendency to be dissipated either in the listener's
curtains, or the table or shelf on which the lead was unrealisingly
coiled.


Regards, NT

 

[Walter Harley]

Thanks to all of you who've been offering suggestions!

Win wrote...

Linear Technology offers their LTC1779, a low-power
buck regulator with an 800mV Vref

I did look at the LTC1779 datasheet, and it looks like it would do the
trick, with efficiency probably around 70% at my supply voltage and current
requirements. Currently I'm getting around 22%, with the dropping resistor.

and also...

I think a better approach is to lose the LED and instead
adopt a power-on pushbutton

I'm dubious about a pushbutton, because inside a gig bag it would tend to
get accidentally pushed all the time; but I like the idea of detecting the
signal and turning on a power timer. It would be fine if it just stayed on
for 10 minutes after signal, assuming it turned on within half a second,
which should be easy.

The LED also serves as a low-battery indicator, in my circuit at present
(it's got a threshold; it turns off when the batteries are at 6v, when they
still have an hour or so of life left). But it doesn't need to be on
fulltime for that purpose; or maybe I don't need it at all.

And NT wrote...

An ultrabright LED and a resistor would I think be the easiest
way to get sub-milliamp drain

and later...

There are panel mounting bulb holders that you can put the LED
in [...] The first thing I try to do with LEDs is to get them
in the main circuit somewhere

Got a source for those panel-mount bulb holders? I haven't been able to
find any, at least not any made of metal.


-walter

 

[Michael]

Hi


Once upon a time in Television land, power supplies used dropping
resistors, and viewers made toast over them. Then along came the type
of PSU in this thread, a dropping R plus switching tr or triac.
Because the switcher introduced some regulation, no loading R was
needed to stabilise the V, so these sets would not produce toast. Then
later along came the switchers we know today, using inductors for high
efficiency.

Very very occasionally a switched dropping R is still a reasonable
choice, but only very very occasionally.


BTW in the dropper supply era, there were some not very reputable
appliances using 'curtain burners'. These were mains leads with the
dropping R wound into the lead, and used the length of lead to
dissipate the heat. These were primarily used on small radios where
the heat could not be dissipated in such a small case. Unfortunately
the heat had a tendency to be dissipated either in the listener's
curtains, or the table or shelf on which the lead was unrealisingly
coiled.


Regards, NT

LOL, It's funny the extreems dodgy people will go to to save a buck.

 

[Bill Bowden]

Bill Bowden wrote...

And the battery current was...?

Thanks,
- Win

whill_at_picovolt-dot-com

The duty cycle is close to 30% so the average battery current
at 9 volts should be about 4.5mA plus another 1/4 mA for R8.
The LED plus 22 ohm resistor voltage is 2.2 for an output power
of 33mW. Input power (.00475*9 = 43mW) is 10mW greater for an
efficiency of 77%. Some power is lost in the switching transistor
and diode, the inductor I believe is ideal.

-Bill

 

[Walter Harley]

[...] your switcher would need
to be 150% efficient. No doubt you can find information on the design of
such circuits on the web. (;-)

I think if you use just the right length of coax between the LED and driver
it works ;-)

 

[Robert Monsen]

The duty cycle is close to 30% so the average battery current
at 9 volts should be about 4.5mA plus another 1/4 mA for R8.
The LED plus 22 ohm resistor voltage is 2.2 for an output power
of 33mW. Input power (.00475*9 = 43mW) is 10mW greater for an
efficiency of 77%. Some power is lost in the switching transistor
and diode, the inductor I believe is ideal.

-Bill

I simulated it with "CircuitMaker", using your values. By using the
multipler, integrator, and divider, I built in power measurement. At 9V, the
entire circuit simulation measures at a little more than 50mW, and the
simulation across the LED (with a 1 ohm sense resistor in series) measures a
bit more than 28 mW. Thus, the efficiency is around 56%.

Why is the inductor optimal? I still don't have a good feel for choosing
inductors for these switchers, other than 'bigger is better'. What limits
the top end, other than price/size?

I tried out a few different sizes using circuitmaker, with Bill's part sizes
(other than the inductor value, which was varied)

(table in monospace font)

33uH 19.5%
390uH 40.1%
1mH 51.6%
2mH 56.0%
4mH 57.5%
10mH 58.8%
....

Regards,
Bob Monsen

 

[Tony Williams]

The LED also serves as a low-battery indicator, in my circuit at
present (it's got a threshold; it turns off when the batteries
are at 6v, when they still have an hour or so of life left). But
it doesn't need to be on fulltime for that purpose; or maybe I
don't need it at all.

Presumably then the LED is there as a reminder to switch
off the amp when not in use. Couldn't this be done with a
short 'blip' every now and again, say about every 10S?

Simple RC charging, voltage detector on the C, discharge the
energy on the C through the LED. Just wondering if this could
just be done with a CMOS 555 timer..... something like a 100k
and 100uF RC, connect the LED in series with the capacitor
discharge output of the 555. Maybe a 1k or so across the LED
to ensure discharge at lower supply volts.

 

[N. Thornton]

[email protected] (N. Thornton) wrote in message

LOL, It's funny the extreems dodgy people will go to to save a buck.

True, tho I think there was more than that to it with the curtain
burners. The desire was for smaller radios, and it constituted an
entire market sector. There was way more at stake than a buck.

Now there is only so much heat a small case can dissipate. Switchers
weren't a serious option back then: say youre the engineer, what are
you going to do to capture that market? It was a tough choice.

Time proved the curtain burner to be a Bad Thing. The question is,
when these things were designed, were they known to be a real fier
risk, or was it thought that a heat dissipating lead would be OK? I
dont know the answer.


So, this is a well skilled tron eng forum: Lets say youre the
designer back then, how could you tackle this heat in a small box
problem?


My first thought is to use relatively low voltage circuitry, divide
the circuits in two sections, and power them in series. This in
principle could halve the total current consumption, thus greatly
improving the heat problem.

Since the output in small rads was generally one class A valve/tube,
current draw would be acceptably constant, and quite possibly one
could have the audio ouput stage as one half, and the rest of the
stages as the other power comsuming block.

Any mileage?? BTW that would be for a 240v set, one would be pretty
pressed to do that on 110v.

Thought 2 would be to replace the hardboard back with a big flat
insulated resistor - a sandwich type construction of flat R and back
in one. Then half of the heat starts life outside the box, not inside.

I actually do have better things to do, but why not


Regards, NT

 

[Spehro Pefhany]

Thought 2 would be to replace the hardboard back with a big flat
insulated resistor - a sandwich type construction of flat R and back
in one. Then half of the heat starts life outside the box, not inside.

How about making the world's first combination night-light and radio?
(Using the bulb as a dropping resistor) That would have been within
the technology then.

Best regards,
Spehro Pefhany

 

[John Woodgate]

I read in sci.electronics.design that N. Thornton <[email protected]>

Time proved the curtain burner to be a Bad Thing. The question is, when
these things were designed, were they known to be a real fier risk, or
was it thought that a heat dissipating lead would be OK? I dont know
the answer.

Tens of millions of radios with line-cord heater droppers were made. Not
may cause fires. On US mains, the voltage across the line cord was only
a few tens of volts, and the current was 150 mA, so there was not so
much power dissipated. For 240 V mains, valves/tubes with 100 mA or even
50 mA heaters were used.

So, this is a well skilled tron eng forum: Lets say youre the designer
back then, how could you tackle this heat in a small box problem?


My first thought is to use relatively low voltage circuitry, divide the
circuits in two sections, and power them in series. This in principle
could halve the total current consumption, thus greatly improving the
heat problem.

The valve/tube heaters WERE connected in series. On 100 V mains, the B+
voltage needed to be as high as could be obtained, with a valve/tube
rectifier - about 120 V.

Since the output in small rads was generally one class A valve/tube,
current draw would be acceptably constant, and quite possibly one could
have the audio ouput stage as one half, and the rest of the stages as
the other power comsuming block.

I'm not sure what you have in mind. Kolster-Brandes did make a radio
with the heaters in series with the B+ supply, but it was
unconventional. (And it was a marketing boo-boo: it sold for less than
£10 and because of that you couldn't buy it on hire-purchase. So the
impecunious people it was aimed at couldn't buy it!)

Any mileage?? BTW that would be for a 240v set, one would be pretty
pressed to do that on 110v.

Thought 2 would be to replace the hardboard back with a big flat
insulated resistor - a sandwich type construction of flat R and back in
one. Then half of the heat starts life outside the box, not inside.

I actually do have better things to do, but why not

That's quite a good idea. Go back to 1960 and propose it to me. (;-)

 

[John Woodgate]

I read in sci.electronics.design that Spehro Pefhany <speffSNIP@interlog

How about making the world's first combination night-light and radio?
(Using the bulb as a dropping resistor) That would have been within the
technology then.

You'd need a special lamp: say 30 V 150 mA for 110 V or 120 V 50 mA for
220-240 V. It may have been done: it's the sort of thing Philips would
have looked seriously at. But I don't recall one in Britain.