1-32 Khz in 1khz divisions

Jim: Very Nice.. I'm actually just across town from you here in Phoenix.. (Over near 101 and 51)

Well I'd like to thank the group for all the helpful suggestions.. My tunnel vision was clearly getting in the way..

__________________ Note: To reply, replace the word 'spam' embedded in return address with 'mail'.

In your case, the frequencies would be audibly separable if they were exponentially spaced. That is each one would be some factor multiplied by the next lower one. This is the way notes are separated on a musical scale. Which brings up a simple possibility. There are so called top octave generators integrated circuits that produce the

12 notes in one octave of a musical scale from a single high frequency input. Adding 12 flip flops produces the next lower octave, etc.

If you don't have perfect pitch, you might want to use only every other note, or pick the notes of a chord, and use more flip flops to cover more octaves.

'mail'.

Decoding can be done accurately with a DVM, IF you restore logic levels on the receiving end before measuring.

I might also do it by counting time at the receiving end... use watch crystal (or similar) clocking on the transmitter.

How about that! We lived up that way at Desert Cove and 66th St. for

25 years, then moved down here to Ahwatukee Foothills after we no longer had kids in college... 4 all together.

Youngest son and his wife had a cracker-box house at 106th Way and Thompson Peak. Sold it for 4X (Californicated) what they paid for it and bought a house in far east Mesa... big house on an acre.... with money left over.

Youngest son just bought in Queen Creek... so far out there in the boonies I need my GPS to find him ;-)

Naaah! That's what the group is actually for... throw out a problem and let everyone take a swing at it.

...Jim Thompson

--
|  James E.Thompson, P.E.                           |    mens     |
|  Analog Innovations, Inc.                         |     et      |
|  Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
|  Phoenix, Arizona            Voice:(480)460-2350  |             |
|  E-mail Address at Website     Fax:(480)460-2142  |  Brass Rat  |
|       http://www.analog-innovations.com           |    1962     |
             
"Winners never quit, quitters never win", Jack Bradley Budnik ~1956

I'd suggest the voltage divider is probably the most practical solution on perfboard.

You could use 10mV..320mV divided down from a bandgap reference. You could monitor the voltage at the top of the divider as you're attaching the wires-- any major deviation would indicate a short. Just a string of 1% resistors (eg. 10R or 100R), a voltage reference and a pot+ resistor to trim in the 320mV.

What's the resistance of 700' of coax?

That is best and most simply done by something entirely from your original "abstraction." People like you need to just come out and say what you're trying to do rather than pass on your confusion.

The OP pointed out that with the 32 different frequencies he could use an existing off-the-shelf receiver - a handheld frequency meter - instead of making yet another custom device.

That makes me think, why not just use a resistive divider to generate:

0.1v 0.2v 0.3v . . . 3.2v

and then use a handheld voltmeter to see what is what?

-Robert Scott Ypsilanti, Michigan

Oops. Not quite that bad.

I'm assuming that you'll only want some 13 outputs or so - the rest comes from a couple of chips of external dividers - or another chip.

You only need 4MHz/16KHz = 250, so it's only 8 bits.

Dec counter Jz set Nop Jmp next set: xor porta, bitmask sto counter, divisor

So, maybe 4*16, but that still adds up to 288Mhz.

Anyway, if the OP wanted to do the cable-sorting problem using frequencies, an optimisation would be to use maybe a couple of extra instructions on each, and run a 5 bit counter from each of these initial counters, and write the

5 bits as one to the output.

So you'd get

16 8 4 2 1 18 9 4.5 2.25 1.125 24 12 6 3 1.5 ...

Which'd reduce the problem to something that's doable on a 100Mhz microcontrollercontroller

Jim Thompson wrote:

Digital-izing your idea, I come up with this simple thing. I don't think measuring duty-cycle is too swift. But he has 50% slop on the one-shot timing which is a no-brainer. There will be cross-coupling so this has to be squashed: View in a fixed-width font such as Courier.

. . XMTR . . 32-bit SR . LINE # . +----------+ __ . .... __|__ . .--------|RST Q0|------| \\---[1k]---> 32 . | | | .----|__/ . | | | | __ . | '1'-|D Q1|------| \\---[1k]---> 31 . | | | +----|__/ . _ _ | | | | . ... _| |_| |_-------+--->CLK | | . | | | | | . 1KHz | | | | | . | | | | | . | | | | | . | | | | | __ . | | | Q31|------| \\-+-[1k]---> 1 . | | +----------+ +----|__/ | . | | | | . | | | | . | '----------------' | . ------------------- | . | Q | | . | B |------------------------' . | CD4538 | . | | . ------------------- . 3mS ONE-SHOT . . . -------------- . | __ __ | . RCVR | | | | | | . | -- -- | . | | | | | | . | -- -- | . ------------- . | . / 14 . CD4093 | . __ -------------------- . | \\ __ | | . >---[RF]-+--| o-+----| \\ | CD40110 | . | |__/ | | o-->CU | . === | .-|__/ | | . |CF | | | LATCH RST | . | | | -------------------- . --- | | | _ _ | . | | | | |_|_ . | | | | . | | '--------. +---. . | | | | | . | +---[100K]---+----+---|-||-' [10K] . | | | | | | . | === | | | --- . | | | | | com . | --- ------------------- . | com | Qa Bb /Qb | . '------------|Ba | . | CD4538 | . | | . ------------------- . 1,5mS retriggerable . . .

This should work significantly better: View in a fixed-width font such as Courier.

. . XMTR . . 32-bit SR . LINE # . +----------+ __ . .... __|__ . .--------|RST Q0|------| \\---[1k]---> 32 . | | | .----|__/ . | | | | __ . | '1'-|D Q1|------| \\---[1k]---> 31 . | | | +----|__/ . _ _ | | | | . ... _| |_| |_-------+--->CLK | | . | | | | | . 1KHz | | | | | . | | | | | . | | | | | . | | | | | __ . | | | Q31|------| \\-+-[1k]---> 1 . | | +----------+ +----|__/ | . | | | | . | | | | . | '----------------' | . ------------------- | . | Q | | . | B |------------------------' . | CD4538 | . | | . ------------------- . 3mS ONE-SHOT . . . -------------- . | __ __ | . RCVR | | | | | | . | -- -- | . | | | | | | . | -- -- | . ------------- . | . / 14 . CD4093 | . __ __ -------------------- . >---[RF]-+---+-------| \\ | \\ | | . | | | o--| o-->CU CD40110 | . === | .-|__/ |__/ | . |CF | | | | . | | | | LATCH RST | . --- | | -------------------- . | | | _ _ | . | '------. | | |_|_ . | | | | . | __ | '--------. +---. . | | \\ | | | | . '--| o-. +---[100K]---+----+---|-||-' [10K] . |__/ | | | | | | . | === | | | --- . | | | | | com . | --- ------------------- . | com | Qa Bb /Qb | . '-------------|Ba | . | CD4538 | . | | . ------------------- . 1,5mS retriggerable . . .

That makes life a lot easier...

Just use a micro with 32 or more i/o ports, and serially output a unique number on each.

Use a micro on the remote side to decode the serial number & display on a LED/LCD. A fairly trivial task using "bit banging" techniques.

An analogue approach as has already been suggested.. use a resistor divider chain to give 32 outputs in say, 1V increments. You can use a simple DVM at the remote end, even over 700ft of co-ax, as the DVM won't draw enough current to cause a noticable voltage drop.

TCM

What is that block labeled "sync"- that was what I was trying to avoid- you can make the test pulses self-synchronizing by chopping them up.

How would you connect the coax cables (RG-6 / RG-11) conductor and shield to the resistances? I'm thinking that would be a lot of alligator clips. (64!)

These are unterminated cables.

__________________ Note: To reply, replace the word 'spam' embedded in return address with 'mail'.

Hello Robert, I have used that method myself decades ago. For each wire, 1 volt steps were used to suit. an old analog panal meter with 0 to 30V scale with 1 volt minor divisions. Very easy to read.

---------------------

Hello Barry, Have a bit of a think here. Lets say the voltage divider used a bunch of 1K resistors to get 1 Volt steps.

Your digital voltmeter has 10 Megohms input resistance.

Wire number seven has 7 Volts applied to it. At the far end of wire or coax number 7 you will read 7Volts. Even if the DMM does read 6.99 Volts at the far end. Surely that will still be good enough to identify it?

I only suggest one volt steps to make it very easy to identify at the far end. Use whatever voltage steps you like.

Regards, John Crighton Sydney

...

OK. :-)

The first computer I ever owned was a "Scelbi 8-H", which had a(an?)

8008. A popular thing to do in those days was to write a little loop that would play music through your transistor radio, more or less.

Naturally, I had to go one better, since those programs were based on making intentional RFI. I used an output port and a speaker. :-)

Of course, it was based on timing loops, and the thing I remember concentrating a lot of effort on was making the duration of the note come out the same for each pitch.

I wrote interleaved timing loops. Check the master counter against each of your event timer registers, and if one of them has been triggered, toggle that output bit, otherwise do nothing for the same amount of time. Or if it's the "tempo" register, then advance the note counter.

Geez, with today's processors, you should have practically all day to toggle 32 KHz! (so to speak. :-) )

Have Fun! Rich

Quick! What line are you reading when the voltage is 2.273V?

--
It\'s a "D" type flip-flop used to disable the counter and freeze the
rest of the circuitry when the pulse on the conductor being probed
isn\'t present.

The test pulses can\'t be chopped up because they\'re being used as
gates for the receiver\'s display counter.  Basically what happens is
that the transmitter sends a one millisecond pulse down conductor
number 1, a 2 ms pulse down #2... 

When the receiver gets the test pulse on conductor #1 the counter is
gated on for 1ms and accumulates one clock which is displayed as "1"
and updated every time a new 1ms TX pulse is received.  The pulse on
conductor #2 will be 2ms wide, on #3, 3ms wide and so on, with the
result being that the counter will be allowed to count to 2 when
conductor #2 is being probed, to 3 when #3 is being probed...

The initial two clocks during reset DOES cause some heartburn.

Don't know whether it'd be worth the logic to fix the reset, or just devise a nice receiver that lights an LED.

...Jim Thompson

--
|  James E.Thompson, P.E.                           |    mens     |
|  Analog Innovations, Inc.                         |     et      |
|  Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
|  Phoenix, Arizona            Voice:(480)460-2350  |             |
|  E-mail Address at Website     Fax:(480)460-2142  |  Brass Rat  |
|       http://www.analog-innovations.com           |    1962     |
             
"Winners never quit, quitters never win", Jack Bradley Budnik ~1956