Audio Switcher Help

On a sunny day (Mon, 19 Nov 2007 07:31:45 GMT) it happened "Jon Slaughter" wrote in :

It has only ground and +? Then you must likely bias your audio at 1/2Ub, else bottom get clipped.

analog signal range is from 0 to +V!! +V | R |---------- chip input audio in --C---| | R | ///

That chip sucks in my view, their 'no click' switching seems to be a slow 8 ms bump.

The 74HC4053 rocks, has negative supply.

Anyways, Maxim is hard to get sometimes.

Some of there chips allow -V on the signal. I guess they use a charge pump to generate -V. Kinda sucsk that this one doesn't.

I need a very large number of switches(about 80) that uses a small space as possible.

Having to bias it really sucks but I guess I don't have much of a choice ;/ For my application I guess the clickless switching is not all that important. (I suppose I could do the switching myself by monitoring the signal with the Pic and switch when needed using the high speed mode).

On a sunny day (Mon, 19 Nov 2007 18:27:09 GMT) it happened "Jon Slaughter" wrote in :

It is an interesting subject. For some time I wanted to make an audio switch for on my desk, that connects all possible DVD outputs, recorder inputs, PC in and outputs, headphones, speakers, other stuff, in a crossbar way, that can be controlled by a simple small keyboard. So, autonomous, not PC controlled, as PCs may be off.

I found one chip that did it all and it costs 135$. Doing it with 74HC4053 and the like, in DIL, takes 2 eurocards (100x160mm),

Many years ago I did it for somebody else, controlled by a Z80 with CD22101

4x4 crossbar switches, and it took half a 19inch rack... (needed logic level changes). Of course it can be done mechanically too.

Oh, I am only talking about a say 16 * 16 crossbar, with headphone or line drive capability, in stereo. And perhaps one for video too.... Maybe that is why I still switch connectors manually.

An other thing with crossbars is that a not-connected bar will pick up noise, so you need to switch an unused bar to ground, requiring an extra bar. Using + and - supplies greatly simplifies things.

crossbar, crosspoint, what is the difference?

Price about 45 USD

Maybe one day :-)

The devices don't need any level shifting or anything. I'm going to have to troubleshoot and see whats wrong. As far as that app note is says, it should work the way I original thought. Might have screwed up chip(not sure how though but I'll try another one soon).

On a sunny day (Tue, 20 Nov 2007 06:28:15 GMT) it happened "Jon Slaughter" wrote in :

YOU NEED 1/2 V BIAS FOR AUDIO. + | R |--------- chip audio -- C --| | R | ///

lol, you didn't take a loot at that pdf did you?

On a sunny day (Tue, 20 Nov 2007 12:16:55 -0600) it happened "Jon Slaughter" wrote in :

Actually I looked twice. Have you tried this? May solve your problem :-)

It won't solve my problem really. I guess it probably will work but ultimately its something I need to avoid(not only more components but more noise).

Page 1,

"Connecting an n-channel MOSFET in parallel with a p-channel MOSFET allows signals to pass in either direction

with equal ease. Whether the n- or the p-channel device carries more signal current depends on the ratio of

input to output voltage. Because the switch has no preferred direction for current flow, it has no preferred input

or output.

"

Now thats telling me that it will take AC just fine such as an audio signal? I could be wrong in interpreting it though?

On the data sheet for the 4610 it says

"Each switch handles V+ to GND analog

signal levels"

Which says it doesn't take AC ;/ So I guess I'm wrong but I can't see how the first quote(from page 1) says AC won't work.

Doesn't Figure 1 show a circuit that lets AC go through? Note it has -V so maybe -V = GND? But then that doesn't explain the quote that switch has no preferred directionf or current flow. This is telling me that I could treat the output as the input and vice versa but then that means that AC should flow?

On a sunny day (Tue, 20 Nov 2007 19:23:22 GMT) it happened "Jon Slaughter" wrote in :

Look at it this way: It is a piece of silicon. If you have Vss grounded, then driving below Vss will see diodes. And on top of that, one of them chips has comparators that actually switch the thing off, if the voltage is close to Vdd or Vss.

lmfao...somebody comes in here for help, somebody else gives him the right answer, and he comes back with a snappy retort like this.

The fellow is correct; it needs vcc/2 bias for capacitively coupled audio. Swings above and below ground, don'cha know.

Jim

Um if you cared to read the entire thread or even the first you would have noticed that I never asked for a solution so the problem but asked why the device was behaving the way it was. Obviously you don't care about that though... I suppose your on PMS and have to find something to whine about.

Also, the app note specifically say that current can go in both directions and my simulation of the circuit in the app note shows that it works that way too. Either the device I'm using doesn't work that way or there -V is ground. (there is no -V on the device though)

In any case there are other methods that are potentially better than passing biasing... such as running a +- power supply so that I can avoid all all the resistors an caps(which introduce noise).

Unfortunately your just a whinner that I guess has nothing better to do with your time?

Again though, as I'm sure you completely miss what I'm saying(chances are you won't even read it but reply with the same BS). My first post was about the device itself and how it was described in the app notes and datasheet(i.e., no mention of biasing, example circuits not showing any biasing, app note specifically saying "both directions", etc...)

Surely the biasing method would work but it would work even if AC could be passed... in that case it is far from the optimal solution. If the device I'm using will not pass AC(which it seems that way) then I need to know for sure.

Theres a huge difference between using a method that works and a method that works optimally. I guess your one of those people that don't give a shit about that though?

Um, that would be look

Ooookay, if we're going to play the horse's ass game, I can play as well as the next feller ...

That would be Um, (comma).

if you cared to read the entire thread or even the first you would have

I've been reading the entire thread. I wanted to see how long you would rave on about your superior knowledge of data sheets until somebody showed you how the real world works.

Obviously you don't care about that

That would be "you're", and I'm way the hell too old for PMS.

Yep, positive voltages at either end will be transmitted quite nicely. Negative voltages (below the Vss rail) will turn the parasitic diodes on and you will wind up with nothing.

Oh, Christ, another "my simulator works just fine." Great. Pack your simulator into your product.

Either the device I'm using doesn't work that way or there -V is

That would be "their". You have no clue as to how CMOS devices work if you say there is no V-.

You'd rather go to the expense of a +/- supply than use a voltage divider? Haven't marketed any products, have you? (And that would be "and".) Never seen a capacitor yet introduce noise. That's why we use 'em, don'cha know.

That would be "whiner". I bided my time until somebody that knows what he is talking about told you what the problem was. No sense trying to teach my grandmother how to suck eggs.

So do the Motorola data books. So do the National data books. Nobody bothers showing input biasing for an AC signal because that would defeat the specmanship of the part. You only have to mess with an experiment for about

15 minutes and then look at the internal design of the part to figure out in a couple of seconds why you can't force AC through this sort of device without bias.

That sentence fragment makes absolutely no English sense...

in that case it is far from the optimal solution.

You wouldn't know an optimal solution if it bit you in the butt.

If the device

Trust me, sonny, I've been using these things in real product design for almost thirty years. Jan told you an inconvenient truth and you seem unwilling to accept it.

By the way, there are several ways of low noise biasing when you've got literally infinite input impedance but you apparently won't listen to folks that know how to do it, so go figure it for yourself.

I've never known an optimal design to make it in the marketplace yet. A product is a bunch of engineering compromises flying in loose formation.

I guess your one of those people that don't give a shit

That would be "you're". I give a shit about a lot of things. One thing I do NOT do well is suffer fools graciously.

Jim

Plonk.

Jim