555 timer alternative for short duty cycle?

John Nagle

unread,

Sep 16, 2016, 9:46:12 PM9/16/16

to

555 timers are useful, but they have a limitation
that they can't have a duty cycle of < 50% in astable
mode. I need 100KHz, about 25% duty cycle for a
switching power supply.

I know I can use an inverter and a power transistor
on the output of a 555, but that adds parts. In
the last 30 years, somebody must have developed an
improved 555 without going digital with counters.

John Nagle

Jim Thompson

unread,

Sep 16, 2016, 10:38:09 PM9/16/16

to

On Fri, 16 Sep 2016 18:46:02 -0700, John Nagle <na...@animats.com>
wrote:

alpibucky

unread,

Sep 17, 2016, 3:24:33 AM9/17/16

to

John Nagle avait soumis l'idée :

Hello,

You can use a 556 or 1/2 558 to build a circuit which offers a pulse
rate and width totally independent. The first timer gives the clock,
the second one is configurated as a one-shot.

There are alternatives for example with cmos/ttl digital circuits

Best regards,

alpibucky

alpibucky

unread,

Sep 17, 2016, 3:43:47 AM9/17/16

to

Après mûre réflexion, John Nagle a écrit :

Hello,

http://www.electroschematics.com/5834/pulse-generator-with-555/

Above is a possible solution with one 555 I think.

Best regards,

alpibucky

John Larkin

unread,

Sep 17, 2016, 10:44:20 AM9/17/16

to

On Fri, 16 Sep 2016 18:46:02 -0700, John Nagle <na...@animats.com>
wrote:

You might be able to use a schmitt trigger oscillator with an HC14
sort of gate.

There are tons of switcher controller chips around that do all sorts
of extra tricks.

--

John Larkin Highland Technology, Inc

lunatic fringe electronics

Jim Thompson

unread,

Sep 17, 2016, 11:21:38 AM9/17/16

to

On Fri, 16 Sep 2016 18:46:02 -0700, John Nagle <na...@animats.com>
wrote:

<http://www.analog-innovations.com/SED/555_Timer_25_Percent_Up_2016-09-17.png>

Phil Hobbs

unread,

Sep 17, 2016, 11:30:10 AM9/17/16

to

Or use a CMOS 555, and feed back the charge and discharge current
independently from the output pin. You just add one extra resistor and
a BAV99 dual diode, and you can have any duty cycle you want. It does
make the frequency depend on input voltage and temperature a bit.

That's only one extra part compared with the two-resistor astable
connection.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510

hobbs at electrooptical dot net
http://electrooptical.net

John Larkin

unread,

Sep 17, 2016, 11:41:00 AM9/17/16

to

A 555 works like a precise schmitt, if you ignore all that
trigger/discharge stuff. One RC from output to input makes it a square
wave oscillator. One added resistor to Vcc or to ground could make it
25% duty cycle. The OP didn't say if the 25% was constant or might
need to vary according to some feedback or something.

Phil Hobbs

unread,

Sep 17, 2016, 11:42:43 AM9/17/16

to

Good point, but that requires algebra. ;)

bloggs.fred...@gmail.com

unread,

Sep 17, 2016, 12:10:52 PM9/17/16

to

Just add a diode between the DIS terminal and TRIG/THR pair with anode to DIS. That's not so many parts.

Jim Thompson

unread,

Sep 17, 2016, 1:49:05 PM9/17/16

to

On Sat, 17 Sep 2016 11:42:34 -0400, Phil Hobbs
<pcdhSpamM...@electrooptical.net> wrote:

>On 09/17/2016 11:40 AM, John Larkin wrote:

[snip]

>>
>> A 555 works like a precise schmitt, if you ignore all that
>> trigger/discharge stuff. One RC from output to input makes it a square
>> wave oscillator. One added resistor to Vcc or to ground could make it
>> 25% duty cycle. The OP didn't say if the 25% was constant or might
>> need to vary according to some feedback or something.
>>
>>
>Good point, but that requires algebra. ;)
>
>Cheers
>
>Phil Hobbs

More than Algebra... gets _very_ unreliable (in the real, noisy,
world) for duty-cycle < 20%

John Larkin

unread,

Sep 17, 2016, 3:22:21 PM9/17/16

to

On Sat, 17 Sep 2016 11:42:34 -0400, Phil Hobbs

Or Spice.

Jasen Betts

unread,

Sep 17, 2016, 5:01:24 PM9/17/16

to

you're going to have to add parts.

eg:

Put a diode parallel with the discharge resistor, cathode to the
capacitor.

Alternativley pull the control-voltage pin up with a resistor (perhaps 2.2K - a
guess: I have not done any math)

--
This email has not been checked by half-arsed antivirus software

rickman

unread,

Sep 17, 2016, 5:07:38 PM9/17/16

to

Why can't the duty cycle be less than 50%? There are plenty of options
for how to do that, but to me the simplest is to invert the output.

--

Rick C

John Nagle

unread,

Sep 17, 2016, 6:03:30 PM9/17/16

to

On 9/17/2016 12:43 AM, alpibucky wrote:
> Après mûre réflexion, John Nagle a écrit :
>> 555 timers are useful, but they have a limitation that they can't
>> have a duty cycle of < 50% in astable mode. I need 100KHz, about

>> 25% duty cycle for a switching power supply. John Nagle

>
> Hello,
> http://www.electroschematics.com/5834/pulse-generator-with-555/ Above
> is a possible solution with one 555 I think.

That looks promising, so I modeled it in LTSpice IV.
Here's the file:

http://aetherltd.com/public/timerdutya04.asc

But the math in the article makes no sense. It says

"The duty cycle (n) is dependent on P1 and P2 in the following manner:

n = 1 + P2/P1

If P2 = 0 (n = 100%) then the frequency can be approximately calculated
with the following formula:

f = 0.69/((2*P1 + P2 + 4.7kO)*C1)"

So, for a 25% duty cycle,

0.25 = 1 + P2/P1
-0.75 = P2/P1
No way.

Actually, it seems that the setting of pot P2, from 0 to 1, directly
changes the duty cycle without affecting the frequency. In our
circuit, pot P2 is represented by R3 and R4, and the duty cycle seems to
be simply

n = R3 / (R3 +R4)

Changing R6 and C1 changes the frequency without affecting the duty
cycle, so this is convenient.

Here it is, set to 100KHz, 25% duty cycle.

http://aetherltd.com/public/timerdutya04.asc

Thanks.

I'm working on a switching power supply design for old Teletype
machines. These need a data signal at 120VDC (!) current limited
to 60mA to operate the selector magnet coil. The traditional solution
is to use a 120VDC supply with a 2K ballast resistor to drive the
55 ohm 4H (yes, 4 henries) selector coil. 95% of the energy goes
into heating the ballast resistor, and you need too much power
supply. The selector magnet only needs 120VDC for about the
first 2ms after turn-on; the high voltage is just to overcome
the huge inductance quickly. So an alternative is to charge
up a 1uf capacitor to 120V, and on turn-on, dump that into the
selector magnet. Then follow up with a low sustain voltage,
about 4V, to maintain 60mA.

Here's the current circuit:

http://aetherltd.com/public/ttydriver16.asc

V3 is simulating an input data stream at 45.45 baud, turning on and
off ever 22ms. L3 is the antique Teletype selector magnet.
(See www.aetherltd.com for what I do with antique Teletypes.)
Switch S2 is a stand-in for an optoisolator. When V3 is off,
the charging circuit pumps up C1 to about 120V in about 22ms.
If the off time is longer than that, 120V zener Z1 prevents
overcharging. When V3 turns on and S2 closes, C1 is dumped
into L3. The charging circuit continues to run, but with
the 55 ohm resistive load on it, it can only put out about
4V continuous, which provides the needed 60mA holding current
for the selector magnet.

This seems to work in LTSpice, but I want to see if I can do it without
surface mount parts. Those are easier for hobbyists to build.
With this 555 circuit, I can get rid of the inverting LTC1693-5
MOSFET driver. The 555's power transistor has enough drive to trigger
the MOSFET in the switcher. Now to find a suitable through-hole
transformer and MOSFET.

This is my first switcher. Any comments?

John Nagle

John Nagle

unread,

Sep 17, 2016, 7:08:13 PM9/17/16

to

On 9/17/2016 3:03 PM, John Nagle wrote:
Correction:

>
> Here it is, set to 100KHz, 25% duty cycle.
>

> http://aetherltd.com/public/timerdutya05.asc
>
> Thanks.
> John Nagle
>
>

P E Schoen

unread,

Sep 17, 2016, 9:52:45 PM9/17/16

to

"John Nagle" wrote in message news:nrkej8$oui$1...@dont-email.me...

> That looks promising, so I modeled it in LTSpice IV.
> Here's the file:

> http://aetherltd.com/public/timerdutya04.asc

[snip]

Here is the simulation for your ttydriver16 circuit.

http://enginuitysystems.com/pix/electronics/ttydriver16.png

As you can see, there is a lot of high voltage ringing (3.5 kV) when the
switch turns off. You really need a flyback diode across the coil to
eliminate this, but it will tend to keep current flowing and thus delay the
deactivation of the coil.

You could produce the 120 VDC by using the transformer of a 12V PSU, running
it backwards so 12V will produce at least 120 VDC and likely more like 350
VDC.

If you already have 120 VAC available, you could just rectify that to about
180 VDC and then use a 3k 15W resistor for the 60 mA hold current. Put a
10-20 uF capacitor across the resistor to get an initial high voltage pulse
for about 30-60 mSec to pull in the solenoid, and then it will settle down
to 60 mA. Of course, when the switch is opened, there will be a similar TC
to discharge the capacitor so it's ready for the next pulse.

It may be better to use a PIC and an optically isolated driver to a MOSFET
in the secondary circuit (or the 180 VDC rectified line voltage). You can
apply the full voltage for about 30-60 mSec and then set up a PWM of about
10-20% to get the holding current you need.

Paul

P E Schoen

unread,

Sep 17, 2016, 10:58:34 PM9/17/16

to

Adding the commutating or "flyback" diode gets rid of the transients, but
delays the current drop-off. I don't think this will work for what you want:

http://enginuitysystems.com/pix/electronics/ttydriver16-diode.png

BTW, don't worry about using SMT components for prototypes. If you use
somewhat larger packages such as SOIC and 0805 they can be soldered using
ordinary tools, and you can get adapters pretty cheap:

http://www.banggood.com/search/pcb-adapter-board.html

Paul

John Nagle

unread,

Sep 18, 2016, 2:02:17 AM9/18/16

to

On 9/17/2016 7:58 PM, P E Schoen wrote:
> Adding the commutating or "flyback" diode gets rid of the
> transients, but delays the current drop-off.

Right. Dumping the energy from a Teletype selector magnet fast is
difficult. The RC snubber shown is known to work on real-world
machines, despite the ringing. A diode makes it worse. The selector
magnet needs to dump in about 2-3ms or the mechanism won't work.

We need enough snubbing to keep from blowing out the drive
components. I use 400V opto-isolators. Actual ring voltage measured with
a similar circuit using a more traditional power supply is about 200V
peak. The SPICE parameters for L3 may need work. 4H and 55 ohm series
resistance are valid, but the parallel resistance is just a guess.

Pre-WWII Teletype machines often had visible arcs at the
keyboard contacts.

John Nagle

P E Schoen

unread,

Sep 18, 2016, 6:28:23 AM9/18/16

to

"John Nagle" wrote in message news:nrlakv$iju$1...@dont-email.me...

I put a 200 ohm resistor in series with the diode, and it dumps most of the
energy before the next pulse. Here is the simulation:

http://enginuitysystems.com/pix/electronics/ttydriver16-diode_200R.png

A higher value resistor might help. You might be able to use another
opto-isolator that turns on when the series switch turns off, and dump the
energy that way. The visible arcs probably helped to dissipate the energy.

Here is the simulation with 500 ohms:

http://enginuitysystems.com/pix/electronics/ttydriver16-diode_500R.png

There might be a way to store the energy from the coil into a capacitor
rather than wasting it as heat. It would be basically a tuned LC circuit
where you could "break" it at the current zero crossing where the capacitor
would be fully charged.

Here it is with 1 uF across the 500 ohms:

http://enginuitysystems.com/pix/electronics/ttydriver16-diode_500R_1uF.png

Changed the 500 ohms to 5k, and it looks just about perfect:

http://enginuitysystems.com/pix/electronics/ttydriver16-diode_5k_1uF.png

Paul

Lasse Langwadt Christensen

unread,

Sep 18, 2016, 7:34:56 AM9/18/16

to

dump it back into the high voltage supply?

Version 4
SHEET 1 880 680
WIRE 336 -336 -352 -336
WIRE -352 -320 -352 -336
WIRE -352 -224 -352 -240
WIRE -336 -160 -352 -160
WIRE 48 -160 -256 -160
WIRE 144 -160 48 -160
WIRE 240 -160 144 -160
WIRE -352 -128 -352 -160
WIRE 48 -128 48 -160
WIRE 240 -80 240 -160
WIRE 336 -16 336 -336
WIRE 192 48 192 0
WIRE 192 48 -112 48
WIRE -112 64 -112 48
WIRE 240 80 240 16
WIRE 240 80 -16 80
WIRE 336 80 336 48
WIRE 336 80 240 80
WIRE 240 160 240 80
WIRE 512 160 240 160
WIRE 576 160 512 160
WIRE 704 160 656 160
WIRE -112 176 -112 144
WIRE -16 176 -16 80
WIRE -16 176 -112 176
WIRE 144 208 144 -160
WIRE 512 240 240 240
WIRE 704 240 512 240
WIRE 144 288 144 272
WIRE 240 288 240 240
WIRE 240 288 144 288
WIRE 240 304 240 288
WIRE 192 384 -48 384
WIRE -48 400 -48 384
WIRE 240 464 240 400
WIRE -48 512 -48 480
FLAG -48 512 0
FLAG -352 -48 0
FLAG -352 -224 0
FLAG 48 -64 0
FLAG 240 464 0
SYMBOL ind 688 144 R0
SYMATTR InstName L1
SYMATTR Value 4
SYMBOL nmos 192 304 R0
SYMATTR InstName M1
SYMATTR Value FDS2734
SYMBOL nmos 192 -80 R0
SYMATTR InstName M2
SYMATTR Value FDS2734
SYMBOL diode 320 -16 R0
SYMATTR InstName D1
SYMATTR Value MURS120
SYMBOL voltage -352 -336 R0
WINDOW 123 0 0 Left 2
WINDOW 39 0 0 Left 2
SYMATTR InstName V1
SYMATTR Value 4.5V
SYMBOL voltage -352 -144 R0
WINDOW 123 0 0 Left 2
WINDOW 39 0 0 Left 2
SYMATTR InstName V2
SYMATTR Value 120V
SYMBOL voltage -112 48 R0
WINDOW 3 -387 50 Left 2
WINDOW 123 0 0 Left 2
WINDOW 39 0 0 Left 2
SYMATTR InstName V3
SYMATTR Value PULSE(0 10 10m 1u 1u 2m 50m)
SYMBOL cap 32 -128 R0
SYMATTR InstName C1
SYMATTR Value 20µ
SYMBOL voltage -48 384 R0
WINDOW 3 -387 50 Left 2
WINDOW 123 0 0 Left 2
WINDOW 39 0 0 Left 2
SYMATTR InstName V4
SYMATTR Value PULSE(0 10 10m 1u 1u 12m 50m)
SYMBOL diode 160 272 R180
WINDOW 0 24 64 Left 2
WINDOW 3 24 0 Left 2
SYMATTR InstName D2
SYMATTR Value MURS120
SYMBOL res -240 -176 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R1
SYMATTR Value 1k
SYMBOL res 672 144 R90
WINDOW 0 0 56 VBottom 2
WINDOW 3 32 56 VTop 2
SYMATTR InstName R2
SYMATTR Value 55
SYMBOL res 528 256 R180
WINDOW 0 36 76 Left 2
WINDOW 3 36 40 Left 2
SYMATTR InstName R3
SYMATTR Value 1Meg
TEXT -508 312 Left 2 !.tran .1

John Nagle

unread,

Sep 18, 2016, 5:05:41 PM9/18/16

to

On 9/18/2016 3:28 AM, P E Schoen wrote:
> "John Nagle" wrote in message news:nrlakv$iju$1...@dont-email.me...
>
>> On 9/17/2016 7:58 PM, P E Schoen wrote:
>>> Adding the commutating or "flyback" diode gets rid of the
>>> transients, but delays the current drop-off.
>
>> Right. Dumping the energy from a Teletype selector magnet fast is
>> difficult. The RC snubber shown is known to work on real-world
>> machines, despite the ringing. A diode makes it worse. The selector
>> magnet needs to dump in about 2-3ms or the mechanism won't work.
>

> Here is the simulation with 500 ohms:
>
> http://enginuitysystems.com/pix/electronics/ttydriver16-diode_500R.png
>
> There might be a way to store the energy from the coil into a capacitor
> rather than wasting it as heat. It would be basically a tuned LC circuit
> where you could "break" it at the current zero crossing where the
> capacitor would be fully charged.

Here's a real-world scope trace from a similar circuit:

The circuit (traditional drive, a 120VDC power supply and
a 2K resistor)

http://www.aetherltd.com/connecting.html

An actual selector magnet. 2 coils, in parallel, each about
1 inch in diameter and 1 inch long.

http://www.aetherltd.com/images/tty15-2/selectoraftercleaning.jpg

Multiple Teletype cycles on a scope:

http://www.aetherltd.com/public/selectorspike100vperdiv220ohm.jpg

One selector magnet decay time:

http://www.aetherltd.com/public/selectorspike100vdiv100usdiv220ohms.jpg

That's 100v/div (there's a 100x resistor voltage divider before the
scope, and a 10x probe), the spike is about 220V. It's decaying in
about 300us. The SPICE model for the selector has to be way off.
The current SPICE params are

L3 Vselector 0 4H Ipk=100mA Rser=55 Rpar=1000000

Rser is the measured resistance. Ipk is the safe maximum
continuous current for the coil. 4H is what some Teletype
literature says is the inductances. Rpar is completely bogus
chosen to keep the spike amplitude from going insane.

In practice, ringing isn't the problem. I just need to worry
about the power supply end, not the load end.

What all this Teletype stuff is for:

https://vimeo.com/124065314

I have more Teletype machines now and need to build about three more
power supply/controller units. I'd like to power them entirely from
a USB port. That's that this design is about.

John Nagle

bill....@ieee.org

unread,

Sep 20, 2016, 4:17:06 AM9/20/16

to

On Sunday, September 18, 2016 at 8:28:23 PM UTC+10, P E Schoen wrote:
> "John Nagle" wrote in message news:nrlakv$iju$1...@dont-email.me...
>
> > On 9/17/2016 7:58 PM, P E Schoen wrote:
> >> Adding the commutating or "flyback" diode gets rid of the
> >> transients, but delays the current drop-off.
>
> > Right. Dumping the energy from a Teletype selector magnet fast is
> > difficult. The RC snubber shown is known to work on real-world machines,
> > despite the ringing. A diode makes it worse. The selector magnet needs to
> > dump in about 2-3ms or the mechanism won't work.
>
> > We need enough snubbing to keep from blowing out the drive components.
> > I use 400V opto-isolators. Actual ring voltage measured with a similar
> > circuit using a more traditional power supply is about 200V peak. The
> > SPICE parameters for L3 may need work. 4H and 55 ohm series resistance are
> > valid, but the parallel resistance is just a guess.
>
> > Pre-WWII Teletype machines often had visible arcs at the keyboard
> > contacts.
>
> I put a 200 ohm resistor in series with the diode, and it dumps most of the
> energy before the next pulse. Here is the simulation:
>
> http://enginuitysystems.com/pix/electronics/ttydriver16-diode_200R.png
>
> A higher value resistor might help. You might be able to use another
> opto-isolator that turns on when the series switch turns off, and dump the
> energy that way. The visible arcs probably helped to dissipate the energy.

Visible short arcs typically drop between 20V and 30V, so they don't dissipate much energy. It typically takes a few kV to initiate them, but the glow-to-arc transition only takes microseconds, as you are only heating up a very thin surface layer to it's melting point. It doesn't do the surface any good - which is why spark plugs don't last forever - but an arc discharge can carry a lot of current with a relatively low voltage drop.

--
Bill Sloman, Sydney

Michael A. Terrell

unread,

Sep 23, 2016, 2:31:08 AM9/23/16

to

Jim Thompson wrote:
> On Sat, 17 Sep 2016 11:42:34 -0400, Phil Hobbs
> <pcdhSpamM...@electrooptical.net> wrote:
>
>> On 09/17/2016 11:40 AM, John Larkin wrote:
> [snip]
>>>
>>> A 555 works like a precise schmitt, if you ignore all that
>>> trigger/discharge stuff. One RC from output to input makes it a square
>>> wave oscillator. One added resistor to Vcc or to ground could make it
>>> 25% duty cycle. The OP didn't say if the 25% was constant or might
>>> need to vary according to some feedback or something.
>>>
>>>
>> Good point, but that requires algebra. ;)
>>
>> Cheers
>>
>> Phil Hobbs
>
> More than Algebra... gets _very_ unreliable (in the real, noisy,
> world) for duty-cycle < 20%

Do you remember this?

<https://ia802706.us.archive.org/35/items/Signetics555556Timers/Signetics555556Timers.pdf>

Jim Thompson

unread,

Sep 23, 2016, 10:34:29 AM9/23/16

to

Thanks, Michael!

Michael A. Terrell

unread,

Sep 23, 2016, 10:14:06 PM9/23/16

to

Jim Thompson wrote:
> On Fri, 23 Sep 2016 02:31:02 -0400, "Michael A. Terrell"
> <mike.t...@earthlink.net> wrote:
>
>> Jim Thompson wrote:
>>> On Sat, 17 Sep 2016 11:42:34 -0400, Phil Hobbs
>>> <pcdhSpamM...@electrooptical.net> wrote:
>>>
>>>> On 09/17/2016 11:40 AM, John Larkin wrote:
>>> [snip]
>>>>>
>>>>> A 555 works like a precise schmitt, if you ignore all that
>>>>> trigger/discharge stuff. One RC from output to input makes it a square
>>>>> wave oscillator. One added resistor to Vcc or to ground could make it
>>>>> 25% duty cycle. The OP didn't say if the 25% was constant or might
>>>>> need to vary according to some feedback or something.
>>>>>
>>>>>
>>>> Good point, but that requires algebra. ;)
>>>>
>>>> Cheers
>>>>
>>>> Phil Hobbs
>>>
>>> More than Algebra... gets _very_ unreliable (in the real, noisy,
>>> world) for duty-cycle < 20%
>>
>>
>> Do you remember this?
>>
>> <https://ia802706.us.archive.org/35/items/Signetics555556Timers/Signetics555556Timers.pdf>
>
> Thanks, Michael!

You're welcome. They are adding a lot of old databooks.

--
Never piss off an Engineer!

They don't get mad.

They don't get even.

They go for over unity! ;-)

tabb...@gmail.com

unread,

Sep 25, 2016, 5:50:59 AM9/25/16

to

The higher the voltage you can maintain across the coil during de-energising the quicker i drops. You could just pick a resistor that does 4mA at 120v = 30k 0.5W.

NT

Winfield Hill

unread,

Sep 25, 2016, 11:33:54 AM9/25/16

to

John Nagle wrote...

>
> 555 timers are useful, but they have a limitation
> that they can't have a duty cycle of < 50% in
> astable mode. I need 100KHz, about 25% duty
> cycle for a switching power supply.

In AoE III, Figure 7.10.B, page 429, we show
a simple configuration for a CMOS 555, that
gives a nearly 0 to 100% duty cycle adjustment
range with one pot, with constant frequency.

--
Thanks,
- Win

Jim Thompson

unread,

Sep 25, 2016, 5:26:24 PM9/25/16

to

On 25 Sep 2016 08:33:34 -0700, Winfield Hill

Voltage-controlled duty-cycle at constant frequency would be nice ;-)

P E Schoen

unread,

Sep 25, 2016, 5:53:14 PM9/25/16

to

NT wrote in message
news:88ab69dd-397f-459f...@googlegroups.com...

> The higher the voltage you can maintain across the coil during
> de-energising the quicker i drops. You could just pick a resistor that
> does 4mA at 120v = 30k 0.5W.

30k and 6H is a time constant of 180mSec. 60 mA into 30k is 108 watts and
1800V peak. The coil has stored energy of 3 * .06^2 = 0.0108 Joules, or
Watt-Sec. If the energy is dumped in 20 mSec the average power is 0.54
watts.

It might be possible to hold the voltage at 120V by using a zener of that
value. The peak power would be 7.2 watts, and would dump the energy in about
15 mSec.

Paul

tabb...@gmail.com

unread,

Sep 25, 2016, 6:04:59 PM9/25/16

to

On Sunday, 25 September 2016 22:53:14 UTC+1, P E Schoen wrote:
> NT wrote in message
> news:88ab69dd-397f-459f...@googlegroups.com...
>
> > The higher the voltage you can maintain across the coil during
> > de-energising the quicker i drops. You could just pick a resistor that
> > does 4mA at 120v = 30k 0.5W.
>
> 30k and 6H is a time constant of 180mSec. 60 mA into 30k is 108 watts and
> 1800V peak. The coil has stored energy of 3 * .06^2 = 0.0108 Joules, or
> Watt-Sec. If the energy is dumped in 20 mSec the average power is 0.54
> watts.

I thought the coil took 4mA

> It might be possible to hold the voltage at 120V by using a zener of that
> value. The peak power would be 7.2 watts, and would dump the energy in about
> 15 mSec.
>
> Paul

certainly a faster option.

NT

Lasse Langwadt Christensen

unread,

Sep 25, 2016, 6:14:56 PM9/25/16

to

use a diode and dump it back into the 120V supply

-Lasse

piglet

unread,

Sep 26, 2016, 2:52:23 AM9/26/16

to

On 25/09/2016 22:53, P E Schoen wrote:
> 30k and 6H is a time constant of 180mSec. 60 mA into 30k is 108 watts

How do you get 180msec? I make it 200 micro-sec. T=L/R

piglet

P E Schoen

unread,

Sep 26, 2016, 4:35:05 AM9/26/16

to

"piglet" wrote in message news:nsagip$85d$1...@dont-email.me...

> On 25/09/2016 22:53, P E Schoen wrote:
>> 30k and 6H is a time constant of 180mSec. 60 mA into 30k is 108 watts

> How do you get 180msec? I make it 200 micro-sec. T=L/R

You are correct. But I think my other calculations show that a 30k resistor
is impractical, because of the 1800 volts. And the 108 watts would dissipate
the 0.0108 J in 0.0108/108 = 100 uSec if the voltage were maintained.

Thanks for catching the error.

Paul

piglet

unread,

Sep 26, 2016, 5:46:28 AM9/26/16

to

On 17/09/2016 23:03, John Nagle wrote:
>
> This is my first switcher. Any comments?
>
> John Nagle
>

For a first switcher a 555 may not be the best option. Have a look at
the UC3842-45 family they are nearly as cheap and plentiful as 555s and
will do a beautiful boost converter with a lot less agony and heartache
than a roll-you-own controller. Under-voltage lockout, cycle by cycle
current limit, tight regulation and a 5V reference you can use for a few
milliamperes of aux power are features hard to bolt on to 555 but come
free with a UC3842!

piglet