clock, data, and power on same line?

[Jon Slaughter]

Is there any theorical reason why this can't be done? and if not then why is
it not practical?

I'm just curious... but I know the egomanics like Eyeore are going to chime
in with "Your a moron" bs to feed there ego's but luckily I have learned how
to use the ignore feature of my usenet app.

Thanks,
Jon

 

[Jon Slaughter]

Is there any theorical reason why this can't be done? and if not then why
is
it not practical?

I'm just curious... but I know the egomanics like Eyeore are going to
chime
in with "Your a moron" bs to feed there ego's but luckily I have learned
how
to use the ignore feature of my usenet app.

Thanks,
Jon

What I mean by this is that I have never seen any mention of something like
this(although I think I do recall a 1-wire protocol somewhere but don't
remember much). I do know that you can encode a clock with data(I think its
manchester encoding?) but wonder why not move to the next "logical" step and
add power. (I know its possible but it seems rarely used)

I imagine, say, if I have some speakers that need power for something(say
some lights or maybe an active filter) but instead of running an extra wire
for power(and I'll probably need one for ground) why I couldn't just run the
power over the same analog audio lines.

Since the power is DC I could just use a cap to drain the power into and
another to block DC for the audio part? (of course this is for analog stuff
but I figure the same could be applied to digital)

I do realize that noise would probably become a much larger issue but I'm
sure there are other things and such... which is why I ask.

 

[DJ Delorie]

Is there any theorical reason why this can't be done?

It has been done. The Dallas/Maxim 1-wire protocol has power, clock,
and data all on one wire.

How?

The line is normally at +5v (providing power, devices have a diode and
cap to maintain it during signalling).

Each falling edge is the "clock" signal.

Each rising edge is the "data" signal - short duration low pulse =
binary 1, long duration = binary 0.

 

[Jon Slaughter]

Is there any theorical reason why this can't be done? and if not then why
is
it not practical?

I'm just curious... but I know the egomanics like Eyeore are going to
chime
in with "Your a moron" bs to feed there ego's but luckily I have learned
how
to use the ignore feature of my usenet app.

Thanks,
Jon

Thanks guys. I'll look into it some more later now that I know it works and
its actually used.

Jon

 

[Jamie]

Is there any theorical reason why this can't be done? and if not then why is
it not practical?

I'm just curious... but I know the egomanics like Eyeore are going to chime
in with "Your a moron" bs to feed there ego's but luckily I have learned how
to use the ignore feature of my usenet app.

Thanks,
Jon

yes. just modulate the line and keep the level up
to where the device will not suffer for it's rail
supply. isolation with something like a diode and
cap will help maintain the rail line to the device
while the device can modulate it.

 

[Eeyore]

Is there any theorical reason why this can't be done? and if not then why is
it not practical?

I'm just curious... but I know the egomanics like Eyeore are going to chime
in with "Your a moron" bs to feed there ego's but luckily I have learned how
to use the ignore feature of my usenet app.

There's no shortage of 'self-clocking' interfaces.

Look at the AES/EBU interface for example.

Power is low Z though. Do you really want to drive a low impedance line with
high speed data ?

Graham

 

[Spehro Pefhany]

Is there any theorical reason why this can't be done? and if not then why is
it not practical?

I'm just curious... but I know the egomanics like Eyeore are going to chime
in with "Your a moron" bs to feed there ego's but luckily I have learned how
to use the ignore feature of my usenet app.

Thanks,
Jon

http://pdfserv.maxim-ic.com/en/an/AN3989.pdf


Of cours just because it *can* be done doesn't mean it *should* be
done. A few more connections can make things a lot better.


Best regards,
Spehro Pefhany

 

[David L. Jones]

http://pdfserv.maxim-ic.com/en/an/AN3989.pdf

Of cours just because it *can* be done doesn't mean it *should* be
done. A few more connections can make things a lot better.

Not so in the case of the Dallas/Maxim 1-Wire iButton Products.
Perfect example of how a 1-Wire interface can be used to make a unique
and very effective product.

I just put a 1-Wire ID chip onto a whole bunch of boards that go into
making up a product, and the ability to program the chip with info
though a single wire interface while the board is un-powered is very
flexible indeed.
Then when the product powers up it can interrogate each board and know
what boards are plugged in, what revision they are etc, and can then
auto-configure itself accordingly. And it only takes a single wire
which is great for retrofitting into existing designs.

Dave.

 

[Terry Given]

Not so in the case of the Dallas/Maxim 1-Wire iButton Products.
Perfect example of how a 1-Wire interface can be used to make a unique
and very effective product.

I just put a 1-Wire ID chip onto a whole bunch of boards that go into
making up a product, and the ability to program the chip with info
though a single wire interface while the board is un-powered is very
flexible indeed.
Then when the product powers up it can interrogate each board and know
what boards are plugged in, what revision they are etc, and can then
auto-configure itself accordingly. And it only takes a single wire
which is great for retrofitting into existing designs.

Dave.

And I just yanked a DS1821 digital thermometer from a design cos it
costs $12! Ive replaced it with a smt thermistor, a resistor and a
capacitor, total cost $0.20. Funnily enough the IO pin it was connected
to was also an ADC input.....

Cheers
Terry

 

[David L. Jones]

And I just yanked a DS1821 digital thermometer from a design cos it
costs $12! Ive replaced it with a smt thermistor, a resistor and a
capacitor, total cost $0.20. Funnily enough the IO pin it was connected
to was also an ADC input.....

Welcome to the principle of horses for courses.

My $30 ThermoChron iButton is worth every cent when I can attach it to
my shoelaces or harness to measure water and air temp when I go
canyoning, or put them throughout my house or inside a product cabinet
to log thermal profiles.

Can't use your 20 cent solution on a sole remaining I/O pin that
doesn't happen to have an ADC input.
And of course the DS1821 can also be used standalone when your product
is too busy doing other things or went to sleep or something. Want to
add some more sensors to the your only remaining I/O pin? - oops.
You could of course add a $.50 PIC in that case if you were counting
your pennies.

That's why there is a product for every purpose.

BTW, you were getting ripped off, the DS1821 is only $6.31 at Digikey.

Dave.

 

[Terry Given]

Welcome to the principle of horses for courses.

indeed. the original decision to use this part (not, of course, mine)
was downright daft.

My $30 ThermoChron iButton is worth every cent when I can attach it to
my shoelaces or harness to measure water and air temp when I go
canyoning, or put them throughout my house or inside a product cabinet
to log thermal profiles.
nice!


Can't use your 20 cent solution on a sole remaining I/O pin that
doesn't happen to have an ADC input.

yes, you just need to think harder (and maybe spend another $0.10)

And of course the DS1821 can also be used standalone when your product
is too busy doing other things or went to sleep or something. Want to
add some more sensors to the your only remaining I/O pin? - oops.
You could of course add a $.50 PIC in that case if you were counting
your pennies.

That's why there is a product for every purpose.

I suspect the "purpose" of the DS1821 is to allow softweare people to
hook temperature sensors to micros without having to do any electrical
engineering. Hell, I could use a 100nF Z5U cap to sense temperature
here, I have a micro......

BTW, you were getting ripped off, the DS1821 is only $6.31 at Digikey.

I built 5 prototypes of the rev 3 circuit, so didnt really care too much
about the pricing; my BOM is being quoted as we speak (?!) but US$6.31
is about NZ$10, so its still more expensive than my large SCRs.

Dave.

Cheers
Terry

 

[DJ Delorie]

Of cours just because it *can* be done doesn't mean it *should* be
done. A few more connections can make things a lot better.

AH, but if you don't *have* those connections? My home furnace
control system uses 1wire because I have four conductors to each
thermostat, and can't easily run more. Power, ground, LCD serial, and
1wire. The 1wire does temperature, humidity, and four pushbuttons.
Outdoor weather stations use 1wire to report temp, humidity,
precipitation, air pressure, wind speed, and wind direction - all on a
single twisted pair.

 

[Paul Burke]

Is there any theorical reason why this can't be done? and if not then why is
it not practical?

It is done. Model train buffs have a system called DCC, which uses a
switched clock, the edge separation defining 1s and 0s. Communication in
the reverse direction is by current loading. It's a bit slow and clunky,
but it works.

I'm just curious... but I know the egomanics like Eyeore are going to chime
in with "Your a moron" bs to feed there ego's but luckily I have learned how
to use the ignore feature of my usenet app.

Why the gratuitous razz?

Paul Burke

 

[Guy Macon]

I imagine, say, if I have some speakers that need power for something(say
some lights or maybe an active filter) but instead of running an extra wire
for power(and I'll probably need one for ground) why I couldn't just run the
power over the same analog audio lines.

Look here:
[ http://en.wikipedia.org/wiki/Phantom_power ]

(of course this is for analog stuff but I figure the same could be
applied to digital)

A variation of the basic idea is here:
[ http://en.wikipedia.org/wiki/Power_over_Ethernet ]

 

[Jan Panteltje]

Is there any theorical reason why this can't be done? and if not then why is
it not practical?

I'm just curious... but I know the egomanics like Eyeore are going to chime
in with "Your a moron" bs to feed there ego's but luckily I have learned how
to use the ignore feature of my usenet app.

Thanks,
Jon

Sure it can be done, using different frequencies is one way.
But the extra circuits needed are more complex then 2 extra copper wires.

 

[Winfield]

http://pdfserv.maxim-ic.com/en/an/AN3989.pdf

Of course just because it *can* be done doesn't mean it *should*
be done. A few more connections can make things a lot better.

Best regards,
Spehro Pefhany

The new PoE ethernet-cable-powered network
devices are going to be pretty convenient.

 

[Winfield]

http://pdfserv.maxim-ic.com/en/an/AN3989.pdf

Of course just because it *can* be done doesn't mean it *should*
be done. A few more connections can make things a lot better.

Best regards,
Spehro Pefhany

The new PoE ethernet-cable-powered network
devices are going to be pretty convenient.

 

[Winfield]

1-wire = 2 conductors minimum, 3 used at times

There are ocean-going systems that only use
one wire. Plus the ocean, of course.

 

[Tim Shoppa]

Is there any theorical reason why this can't be done? and if not then why is
it not practical?

Devil is always in the details. There are many examples of it being
done, and being done for at least 70 or 80 years.

I actually like the old Westinghouse SCADA equipment that used 807's
to put the carrier onto the power lines. I saw it being deinstalled as
recently as ten years ago.

Tim.

 

[Spehro Pefhany]

The new PoE ethernet-cable-powered network
devices are going to be pretty convenient.

PoE = Two pairs, minimum, right? = 4 conductors.
USB = 4 conductors
IEE1394 = 6 conductors (if power provided)
POTS = 2 conductors minimum
1-wire = 2 conductors minimum, 3 used at times



Best regards,
Spehro Pefhany