Panning over 4 audio-channels using only 1 pot ?

[watchjohn]

.

Hi,

It`s a little difficult to explain:


I have 4 audiochannels (mono)

I have 1 amplifier to make them audible

I want to use only 1 potmeter to mix/pan them
to the amplifier.

Lets say these are the 4 channels

1...2...3...4

Ideally i like to be able to "pan" over the channels
to hear mostly one channel, either 1...2...3...or 4
(while the other channels should be much weaker)

So if i adjust for max volume of channel 1
then channel 2,3,4 should be muchweaker
slightly moving the pot towards another channel
should change the output in the same way.

When to pot is at channel 2...then the other channels
are much weaker...etc

pot between 2 channels gives both of them at half strength
...etc


Same as having a pan control for 2 channels (stereo)
where you can "pan" from the left to the right channel
Instead of 2 channels i now have 4 channels.

I like to be able to do it with only 1 (single channel)
digital potmeter (under uP control)

Yes i know i could do it with 4 digital pots
but i dont want that.

How to do it with only 1 pot ?
...maybe i could have somekind of resistor-bridge ?

Thanks in advance !


WatchJohn

 

[itsmoked]

There may be some analog scheme that would make this work but I can't think of a tenable one - off hand.

I would run one pot, or better, one cheap "encoder knob",(you know, a dial that you can turn forever in any direction), into a microprocessor, then use the micro to run four digital pots, or a standard quad digital pot, or A/Ds, or variable gain amps, or ??, to get the result you're looking for.

You could also look at digital pots that use up/down pulses. Then possibly use four that all look at your up and down buttons.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[watchjohn]

.

Thanks Keith,

I understand your suggestions
and i´m shure it will work.
...but i am looking for a solution
where i would only have to use one single digital pot

I see you perfectly understand what i want to do.

For others i just thought of another way to better
explain what i am looking for.

Think of a "slider-pot" you know those bars used in
audio-mixer-panels.

Such pots normally have 3 connections
1 at the lower end
1 at the upper end
1 wich is the slider

What i´m trying to have is such a bar-pot
with 4 connections (spread over the range)
and 1 slider.

Instead of a mechanical bar-pot i would like to use
1 digital pot to create the same effect.

Thanks again !


WatchJohn

 

[MacGyverS2000]

I could imagine a needlessly complex analog solution with analog comparators switching in/out resistors networks as the pot passed over the various ranges, but why? It would be physically smaller, less costly, and easier to implement with 4 separate digipots and a uC. I see no valid reason to do otherwise.

This doesn't have anything to do with your wristwatch bug detector, does it?


Dan - Owner

http://www.Hi-TecDesigns.com

 

[VEBill]

You might look into multitap potentiometers. They're rare and normally custom. Watch out for log taper pots with center tap because that isn't quite what you're looking for.

The basic idea is that the pot would have the various signals tapped in at various positions along the resistance track. You'd have to load the output so as to provide a voltage divider. The least attenuation would occur when the slider was next to that particular tap.

You'd be limited in range by the likely-linear track. It would take some modeling to make it all work.

 

[IRstuff]

If it's a digital pot, then connect it to a microcontroller and program in the behavior you want.

TTFN

Eng-Tips Policies FAQ731-376

 

[nbucska]

Keith has designed the hardware -- now you just have to determine the algorithm you want to use.

I suggest temporarily use your PC -- build a temp.interface
using the parallel-- printer -- port to the digital pots
and the input encoder ( not POT, why complicate things
with the A/D ? )

You can write the program in GWBASIC, it is fast enough
for what you need - so you can experiment with different
control rules -- when done you can decide to transfer it
to a cheaper uC or use a PC compatible for $50 to $150.




----------------------------
Please read FAQ240-1032
My WEB: <

http://geocities.com/nbucska/>

 

[itsmoked]

nbucska's modelling scheme is a good idea.

I've done that many times to good effect! If you can't have a PC in the model for some reason, you can always use a PICSTAMP running basic to model it too.

We all like the uC angle because what happens is you do a ton of work to get some analog solution to almost work the way you want it but not quite. There, you crash into an insurmountable wall. Stalled. With a micro you do things like create any taper you need with the pots or provide complete blanking of the other end channel or ??

Anyway you can then handle the initial conditions and funny business that shows up at the end points, etc.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[VEBill]

I missed 'digital' - sorry.

If it is going into a suitable micro for subsequent processing, then you could use almost any input device. Is a digital pot particularly easy to interface to a uC?

 

[nbucska]

The digital pot is output device ( basically a D/A converter where the output
is proportional with the ref. voltage AND the loaded number). The input device is an encoder with two phase output, so the direction of the rotation can be determined.


Both are easy to interface



----------------------------
Please read FAQ240-1032
My WEB: <

http://geocities.com/nbucska/>

 

[MacGyverS2000]

Actually, digital pots can be both input and output, so the OP needs to clarify... I was assuming their use in the feedback path of an op-amp (input from uC), whereas it seems some were considering their use as a user input (output to uC).

Dan - Owner

http://www.Hi-TecDesigns.com

 

[watchjohn]

.

Ok...i understand all arguments...thanks !
digital-pots seem to be the best solution


But let me explain a little more:
---------------------------------


Indeed it is for my "Wristwatch-Sized-Bugdetector"

That thing is going to have a very wide frequency-range
from at least 1 Mhz to 10 Ghz.

As i explained in another thread:

It will receive RF signals
...show the signal-strengths on a led-bar.
...make the signals audible on a speaker.

Since it is so small...i have very little space
but i don´t want to compromise on quality/sensitivity.

Normally the RF-part of such detector consists of:

Antenna > RF-Amplifier > Diode(detector)

The rectified (detected) signal (AM/DC-level)
is then fed to a led-bar and a audioamp with speaker

Here´s an example of such very basic bugdetector-circuit
it only shows the signal-strength...it´s not sensitive and there´s no audio.

http://geocities.com/ajpotts19/wavemeter.html?20073

I have come up with a better idea that i have never seen
before in such detectors...allthough i have seen
it in some fieldstrength-meters but always without audio.

Instead of a power-consuming RF-amplifier
i will use a so called LOG-detector (chip)
such Log-detector lets you direcly put-in an RF-signal
and at the output you will have a DC-signal
representing the signal-strength.

A (temperature compensated) and fast diode
is not needed either !

Best of all is that since the DC-output-signal
represents the RF-input...it´s also an (AM)demodulator
at the same time...so the audio is there also !

Check out this 3 U$, 8 pin, 3x2 mm chip
it goes up to 10 GHZ !

http://www.analog.com/UploadedFiles/Data_Sheets/AD8317.pdf

There´s hundreds of Logdetectors all with different specs.

So here you have a 3x2 mm chip...connect a antenna
and at the output you will have:
the signal-strength...and...the (AM) audio !

Here´s a radio-amateur implementation for a 10 ghz
portable fieldstength-meter (no audio).

http://www.w1ghz.org/new/portable_powermeter.pdf

Here´s another one (no audio again):

http://www.qsl.net/n9zia/wireless/pics/LED_sig_meter.png

If you search around you might find some more examples,
amazing is that no-one never includes a simple
audio-amplifier to also make the received audio available
instead of only the fieldstrength indication.

It would only cost 2 U$ extra for an LM386 and a speaker,
that little extra lets you hear/verify what you are actually receiving
You could then hear:
AM,Digital,FSK,Pulsed and even FM(edge-detected).

An advantage is also that you can have such Logdetector-chip
directly at the antenna...and from there you only have to transport the DC-signal (LF) wich is a lot easier then RF.

Advantage is also that the Logdetector provides
a lot more gain while drawing much less power
compared to RF-amplifiers (MMIC´s or Gain-blocks)
I even found one that needs only 500uA !

Also it is very hard to find an (Ghz)RF-amplifier that will
work at 3 Volts...most Logdetectors work from 2.7 Volts !

Since i have only 1 single LI-ON battery (3.7 to 3 volts)
i need everything to work at that low voltage.

So i decided to use a Logdetector(s)



Now to the 4-channel-panning-pot-question
-----------------------------------------

Getting good reception over a range from 1 mhz to 10 Ghz
is almost impossible with just 1 small antenna.

So for my detector where the antennas MUST be built-in
i have decided to use 4 small separate antennas.
(lets not discuss the antennas..that´s another thread)

each antenna for a part of the very wide frequency-range,
each with its own simple RF-filter,
each with it´s own logdetector(mounted right at the antenna)

As result i will now have 4 DC-levels representing
the signal-levels in each band...including the demodulated
audio from each band.

Now finally to the point...

It would be very helpful if i could select a certain
frequency band...that will give me lots of additional
features since i could then:

1) Listen and measure on a certain frequency-band only.
2) Determine roughly the freq-band of an unknown signal.
3) Lock out certain annoying (unwanted) signals.
...etc

Offcourse i could use 4 switches...that´s simple.

BUT

i would prefer to "pan" over the range so i could
fine-adjust the received frequency range...or even
roughly "pullout" a certain signal or attenuate others.

Infact i would then have somekind of simple
antenna-tuner/filter/selector !
wich is normally done at the RF-side and more complicated
and very difficult for such a wide range up to 10 Ghz !

Using such Logdetector it is now much much simpler since
the mixing/filtering can be done at the DC-side !

Yes having 4 single or 2 stereo digital-pots is
very doable...i have a PIC-processor in there anyway
and all fuctions are executed using 3 pushbuttons.

BUT...

i want to save on space...so less components as possible.
The processor is already very busy so i like to save
on processor-load and in/ouput-lines also.

So that is why i am looking for a way to do it
with as less as much dig-pots as possible.

I think my Logdetector-idea as Wideband-Detector with
audio is a very good and new idea...the mixer/panner
acting as a tuner/filter would make it even better.

I have also thought of 4 LDR´s or other opto-elctronics
but...probably draws to much power.

A custom mechanical 4 tap pot will work
...but too difficult to get that fit in my
wristwatch-sized-bugdetector
i like to control it using the already available
uP and pushbuttons.

I suppose you get the point...i´m still hoping that
there might be a way to use some form of resistor-network
and only 1 digital pot...wich would allow me to pan over the 4 channels.


Thanks again !


WatchJohn

 

[VEBill]

The world's first 10 kg (22 lb) 'wristwatch'. ;-)

One of the primary distinctions between antennas is size. So what's the difference between one tiny antenna and the next tiny antenna?

 

[MacGyverS2000]

"The world's first 10 kg (22 lb) 'wristwatch'."

I'm with Bill on this one... I see no physical way possible you're going to fit this into anything remotely resembling a watch (and I'm including those huge honkin' calculators they once had). Between the LED display, processor, LOG detector, battery, etc. you're already way beyond the size of a wristwatch. Add in impedance-controlled PCB with antenna traces and connectors for larger antennas and you're way beyond an uphill battle.

Have you considered drawing some of this out with proper dimensions to get a feel for what you're up against? I think you're going to be unpleasantly surprised at the size...

Dan - Owner

http://www.Hi-TecDesigns.com

 

[watchjohn]

.

Hi VE1BLL,

It´s really another story and needs a thread on its own.

But a short answer to your comment:

I can´t have 1 single small antenna for 1 Mhz to 10 Ghz
so i devide that range into 4 practical pieces
for each i can make an antenna that will cover such piece.
bands are devide roughly like this:

a) 1 to 100 Mhz
b) 100 to 500 Mhz
c) 500 to 3000 Mhz
d) 3000 to 10000 Mhz

Each of those ranges can be covered with 1 antenna(RX-only)
I choose these ranges since in practice
they also represent certain bands that are used for
certain purposes.

For example:


1-100 Mhz
---------

For 1 to 100 mhz i will use a multiloop
wich is contained inside the wristband.


3000 - 10000 Mhz
----------------

For 3000 to 10000 Mhz i will use an antenna
wich is a mix between a "upside-down discone" and
a wideband "stubby monopole" (both UWB-antennas)
the upper circuitboard will also serve as a groundplane.

Here´s an example of such "stubby monopole"

http://tinyurl.com/2aq5wy

The actual antenna is only the round disc (ground)
and the little cylinder in the middle is the radiator,
mine will be much smaller since it will do
a higher freq-range.

To make it even more wideband that middle radiator
will have the shape of an "upside-down discone"

So in the end it will look something like this:

http://uei.ensta.fr/roblin/papers/Paper1049.pdf

For the freq-range of 3000-10000 mhz that antenna
will have exactly the size wich fits in my case
(the disc has the size/diameter of a wristwatch)
and the upper circuitboard will act as the disc

You understand that such antenna can not be effective
for the lower frequencies...wich i also want to receive,
so i need other (small) antennas for those ranges.


I´m still working on the other 2 antennas
probably they will be a dipole or a folded dipole
with wideband charactristics...i´m still looking a that.



Resuming
--------

They are all small but tuned to a different freq-band
and they are constructed differently.



Antennas act as filters also
----------------------------

Because they are tuned or best for only one freq-range
they also act as filters...meaning that they will
give highest levels at their own bands and a lot less
on other bands.
Wich is exactly what i want...since i want to be able
to switch or fine-adjust for certain bands.

Also because the antennas already act as filters
i can have simpler RF-filters (LC´s) wich lowers
the component count...and lowers problems inherent
to the use of LC filters.


Greetings


WatchJohn

 

[nbucska]

John:
Have you tried to estimate your market? Only very few of
us average Joe Does are worried about being bugged.
I think your market is 10% of the millionaires (CEO etc)

The big problem of invertors -- and I am one -- is marketing. 7 patents and hundreds of ideas earned for
me less than $100.-.

If you are businesman, you don't need to be an inventor;
you buy the inventors cheap. And if you are not,
you are out of luck...






----------------------------
Please read FAQ240-1032
My WEB: <

http://geocities.com/nbucska/>

 

[richs]

Hiya-

One thing that I think was not mentioned on the digital pots is that some of them (including microchip) have a SPI interface which, I believe, is not a multi adderess interface or an I2C interface which has multiaddresses. A "bussed"
configuration could have all the digital pots together saving the "I/O pins" of the pic. Once the value is output, the
value remains set and the pic's efforts are no longer needed until the value is changed of course.

I will not go into the various "schemes" for asserting the chip select pins on the desired digital pots as there are many that can work and still not use an excess of pins and/or extra ICs.

BTW, large wristwatches are not unheard of. I did a little job on an atomic clock product recently. One guy turned the clock into a wristwatch......

The link is here:

http://www.leapsecond.com/pages/atomic-bill/

Cheers,

Rich S.

 

[RajeevCell]

Dear WatchJohn,

Your idea is indeed interesting. Two points jumped to my mind.

1. If you make the bands of your antenna/filter/logdetector groups to overlap somewhat on the sides, then you won't need the "panning" feature. A simple selector would do. Indeed some overlap is already there. You cannot design that sharp cut-off.

2. Bugs transmit data in encrypted form, so the audible signal won't be legible, in case you want to decode the bug data.

Cheers
Rahul

Yeasir Rahul

http://www.voltsmith.com

 

[geekEE]

If the digital pots are going to be used for audio and they are going to be changing all the time, you should probably look at chips that have zero-crossing detectors. Otherwise you will end up with audible artifacts (zipper sound) as the pot changes settings.

I think that a quad digital pot will probably take up less space than any external circuitry needed to do the four-channel panning.

 

[itsmoked]

Very funny richs.......

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com