Maker Pro
Maker Pro

frequency feed through MUX

ami85t

Feb 19, 2014
71
Joined
Feb 19, 2014
Messages
71
Hi to all,
I'm trying to build an oscillator (frequency of ~400hz) using a LC tank and I have a few questions connected:

1. Is it possible to do that when one leg of the LC tank (a parallel one) is wired through a MUX (a TC74HC4051 analog 8 to 1 bidirectional )?
The frequency is supposed to be oscillated using a 555 timer.

2. will the same be for a shift register? Is the SR suitable only for digital bus or that it can
be fed with an analog signals?

Thanks, Amitai
 

Harald Kapp

Moderator
Moderator
Nov 17, 2011
13,700
Joined
Nov 17, 2011
Messages
13,700
Amital:

1) A 555 oscillates without an LC tank circuit. Also what's the use of the multiplexer? I'm not sure what your circuit looks like and why so. Can you elaborate in more detail, showing a schematic and givingbackground information what you want to achieve and why your circuit is designed that way?

2) A shift register is a purely digital circuit. It's been quite some time that analog shift registers have been used (bucket brigade device, see here under "history"). Today's shift registers cannot handle analog signals.

Harald
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
As Harald says, a 555 is not designed to work with an LC circuit. You might be able to make it work (I've never heard of it being done) but you should use a circuit that's designed from the start to work with an LC tank.

You can switch different inductors using a 74HC4051 but (a) the analogue switches in the 74HC4051 have significant resistance which will reduce the Q of the tuned circuit, and (b) because of the voltage magnifying effect of the Q of the tuned circuit, it's possible for the voltage at one or both end nodes to go outside the power supply rails; in this case, protection diodes in the 74HC4051 will conduct and clip the voltage. In other words, the 74HC4051 may prevent the oscillator from working properly if the oscillator needs the voltages on the tank circuit to go outside the power supply rails. Those are just potential problems that I should warn you about. They may not turn out to be problems in practice.
 

ami85t

Feb 19, 2014
71
Joined
Feb 19, 2014
Messages
71
Thanks for the quick answers.
I tried to draw an idea of a schematic for the circuit:
Capture.PNG
Can the effect of the resistance of the MUX be reduced somehow?
I didn't understand what's the meaning of the voltage magnifying effect -
You mean that trying to integrate the oscillated frequency in a circuit might cause disturbance to
the oscillators frequency?

Amitai
 

Harald Kapp

Moderator
Moderator
Nov 17, 2011
13,700
Joined
Nov 17, 2011
Messages
13,700
Where does the 555 come into play? It is not in your schematic and it does not operate from an LC tank circuit.

The resistance of the MUX can be reduced by selecting a MUX with selected low ON resistance. It cannot be eliminated completely.
By voltage magnifying effect I think Kris refers to the resonance point of the LC tank where the voltage across the LC circuit can easily reach levels high above the operating voltage of the circuit.

You mean that trying to integrate the oscillated frequency in a circuit might cause disturbance to
the oscillators frequency?
I'm note shure to which part of Kris' or my answer you refer. Of course, loading an oscillator (drawing curent from it) will at least reduce the amplitude of the oscillator - with the possibility of stopping the oscillation if the load is too high (too much current drawn). Also the frequency of the oscillator may be influenced, how much depends on the particular construction of the oscillator.
I think you may refer to Kris mentioning the protection diodes in the MUX, do you? In that case what happens is that if the voltage on the LC tank is higher than Vcc or lower than GND (which can happen, see above) the protection diodes from the pins of teh MUX to Vcc or GND (those are integrated into the chip, no way of removing them) wil become conductive and draw current. This current will damp the oscillation, see above.
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
Right, you can use a multiplexer with low ON-resistance and/or parallel two or more of them, to get a lower resistance. But if you're multiplexing between separate LC tank circuits, instead of switching just the inductor, the resistance is not so critical because it's not inside the tank circuit, so it won't affect the Q (much).

As Harald says, an LC tank circuit can generate voltages much higher than the circuit can apply to it. It's like when you push a child in a swing. You don't need to push the swing all the way to its furthest point. Once it gets swinging (oscillating), it only takes a little push each time to make it swing higher and higher. The protection diodes in the multiplexer are like two walls; the swing can only swing between those walls. So yes, they would put a big damper on the swing, and can greatly affect the frequency and the waveform; you would have to make sure that your circuit only pushes the swing softly, so it never swings far enough to hit the walls. Back in the realm of electronics, that equates to limiting the drive power, and possibly limiting the Q of the tuned circuit as well.

We can't say more than that, until you show the rest of the circuit.
 

ami85t

Feb 19, 2014
71
Joined
Feb 19, 2014
Messages
71
Thanks Kris and Harald,
Do I understand correctly and the diodes you speak about are integrated in the MUX and
therefor they will see to it that the output will stay in the rail of the voltage supply?

I now that the Q factor is the ratio of resonant frequency / Δf, but what is the importance of
it? the amount of the stability and accuracy of the oscillator?

I have been told that for a cmos mux:
" Since your inductors connect to ground then the outputs that drive the inductors must be able to swing positive and negative which is impossible with a single positive supply voltage. "
Is that correct? can anything be done about that if so?

For now I can't be more specific and give more details about
the project, I hope i will be able soon enough.
Thanks again,
Amitai
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
Thanks Kris and Harald,
Do I understand correctly and the diodes you speak about are integrated in the MUX and
therefor they will see to it that the output will stay in the rail of the voltage supply?
Yes, exactly.
I now that the Q factor is the ratio of resonant frequency / Δf, but what is the importance of it? the amount of the stability and accuracy of the oscillator?
Yes. But the diodes could prevent the oscillator from working altogether.
I have been told that for a cmos mux: "Since your inductors connect to ground then the outputs that drive the inductors must be able to swing positive and negative which is impossible with a single positive supply voltage." Is that correct? can anything be done about that if so?
Yes, and yes. But you haven't shown us the complete circuit yet. Once you've done that, we can suggest ways to avoid that problem, if it's possible.

While you're at it, explain what you want to ACHIEVE with this circuit. There may be better ways of doing what you want to do. Explain the whole project and why you want to switch different tank circuits.
For now I can't be more specific and give more details about the project, I hope i will be able soon enough.
For now, I can't give you any more specific advice. I hope I will be able to soon enough.
 
Top