Electrical Switch Providing Capacitance

[mario]

Hello guys,

I would like to know if the above circuit is viable. Basically, I want to exert varying capacitance at the 74HC6044 output pins 1,11,4, and 8. In other words, when I flip Switch A, will the circuit deliver 640pF capacitance of +5 VCC to Pin 1? I know the voltage characteristics of VCC will change when the capacitor is used, and I would like to know if 6044 is the correct component to use as an electrical switch. In addition, I would like to use switch B,C, and D cumulative or exclusively: will flipping switch B and C provide 9nF (6.8nF + 2.2nF=9nF) of capacitance?

Pins 1,11,4, and 8 will then drive another component with different voltage characteristics from the original +5 VCC. They will have continuity (The schematic doesn't show that they are connected, and I'm sorry for not showing their connection).


The datasheets for 74HC6044 can be found here:

If the the above is not viable, can some one suggest an alternative solution? I researched using veractors, but those only go up to 100pF. They're not practical for my application since I want to exert over 25nF of capacitance.


Thanks in advance

[Krzysztof]

Hi Mario.
You can use as a supplement to the generator. Frequency = 20kHz max! (See Fig.12 PDF).
Expect more, and by this wave of low distortion dynamic variable resistance 40-50ohm (see Fig.10 PDF).
Use the generator control - small power.
You can block capacitors connected in parallel.
Expected voltage capacitors within (0,5-4,5 V) when powered by 5V.
I think that is not suitable for balancing coil TX! (High voltage).
Using the RX input (tuning) expect a large increase in noise! (When it is before the IC).
Best regards Chris.

[pebe]

The CD4066 is a CMOS device and needs its inputs to be connected at all times. You should not let them 'float'. So you should fit a pullldown resistor of about 47K between each input and -Vcc.

When a CTRL pin 13 is switched to Vdd the switch element is turned on and pin1 is connected to pin2. You can use the switches for several purposes, but the voltage on either of those pins must not fall outside the range Vss to Vdd. So you cannot use it to tune a Tx coil because of the high voltages. The other 3 switches in the package can be used in a similar manner.

Your +5V supply rail should be marked Vdd, not Vcc. If you use the 74HC4066 you will be limited to a maximum supply voltage (Vcc to Vdd) of 11v, but if you use the CD version it will be OK up to 18V.

[mario]

The CD4066 is a CMOS device and needs its inputs to be connected at all times. You should not let them 'float'. So you should fit a pullldown resistor of about 47K between each input and -Vcc.

When a CTRL pin 13 is switched to Vdd the switch element is turned on and pin1 is connected to pin2. You can use the switches for several purposes, but the voltage on either of those pins must not fall outside the range Vss to Vdd. So you cannot use it to tune a Tx coil because of the high voltages. The other 3 switches in the package can be used in a similar manner.

Your +5V supply rail should be marked Vdd, not Vcc. If you use the 74HC4066 you will be limited to a maximum supply voltage (Vcc to Vdd) of 11v, but if you use the CD version it will be OK up to 18V.

Hi pebe,

Thank you for your input!

I made several revisions. Can you tell me if the switch is now viable? I would like to create my test board after some one gives me the ok.

Also, the miniboard will be part of an oscillator on the trace found HERE

[Mechanic]

G'day Mario,

In theory that should work!

I have tried making a small board with an adg453 to switch in different value caps into a low pass filter at a later stage, but it just did not want to work. It may have had something to do with the gain in that stage or something, but it caused the opamp to go to the rail.

My advice would be try it and see what happens, looks like it should work the way you have it set up.

Cheers Mick

[pebe]

It should be OK provided that X, Y or Z do not go outside the range of your supply rails, 0V to +5V. Remember that if the cap is tuning an oscillator coil, then the volts across the coil may well exceed the rail voltage.

I should point out that each switch has an inherent resistance of about 470ohms with a 5V supply, or 180ohms @ 10V, or 125ohms @15V - and that would be in series with the switch.

Your 25nF cap at 950kHz would have a reactance of 6.7ohms and 470ohms in series with that would be the worst case. I haven't got the circuit of your oscillator so don't know what effect that would have, but as Mechanic said, try it and see.

[DavidB]

I worked on a downhole Nuclear Magnetic Resonance Porosity tool once for Schlumberger and we used the micro canned relays to switch different capacitor values across the antena to tune it to the formation. It operated at about 2mhz. We had to test all the relays first and pick out the ones with the lowest noise through the contacts.

[Mechanic]

Hi DaveB,

That is a very good way to do it for sure! Good thinkin '99!

Cheers Mick

[mario]

Wish me luck

I had some time to finish the PCB Layout in Sprint...I also ordered the presensitized boards to create my pcbs....going to be a few days until I get the parts


Thanks guys, I'll let you know how this turns out

[Noemio]

floating inputs

Hi Mario,
I apologize, but… if you really have to use a mechanical switch why don’t you connect directly the capacitors through the switch?
If, instead, your switch is just an example and you will drive pins 5, 6, 12 and 13 with a logic port or a microcontroller, the inputs are never floating so you don’t need to use pull up or down resistors.

[mario]

Hi Mario,
I apologize, but… if you really have to use a mechanical switch why don’t you connect directly the capacitors through the switch?
If, instead, your switch is just an example and you will drive pins 5, 6, 12 and 13 with a logic port or a microcontroller, the inputs are never floating so you don’t need to use pull up or down resistors.

You basically answered the question!

A micro can be a master switch charging the caps directly, thus eliminating the 4066 units in addition to the mechanical switch.

However, why not have a programmable option to disable/enable the mechanical switch? With the above design, people have the option of charging the caps with a micro if so they choose. In other words, the design is future proof! The design also saves resources if used via the micro route as the numbers of i/o pins required to charge the caps is 8 vs 4 if you use two 4066s.

Don't forget, using a micro also complicates the project. Many members just want something that works without concerning themselves with the additional monetary/time investment required to drive a micro.

I haven't yet written the program required to boot the detector; but with this design, I now have the means of controlling the frequency with the micro when I eventually do.

I just hope it works, this was a **** to do in Sprint.

[pebe]

Mario,
You did not give a full circuit so I didn't know how you were going to use the caps. I assumed points X, Y, and Z were not necessarily going to ground.

But remember, with a micro you can only switch a port to +V or -V, you cannot quickly make the port an open circuit. So if a cap has to be quickly switched in or out of circuit, a micro will not do.

[Mechanic]

Hmmmmm

K.I.S.S

Ok now if you were to go to the trouble of using a micro to something, I would not bother using it to control some 4066's to switch in different caps to an oscillator. You can only run the oscillator in 1 setting(frequency) at any one time. When you switch from one frequency to another you need to allow time for the circuits to settle and then re ground balance the detector. The switched caps will work fine as you have them set out currently, no micro needed.

What I would do would be use the micro controller to create a square wave and drive pin 12 of 74hc00 with the desired frequency. A high speed micro would be best for this, that way you can easily fine tune out emi. You would need to use a rotary encoder to increase or decrease the frequency and perhaps a button to jump between pre-set frequency's. But then again if you were going to go to all of this bother, you might as well just set up the micro to control the tx and sample timings and be done with it and free from the pre-set minelab timings and cycle times. However to do this you would need to really understand how it all works, as it would be easy to end up with a detector that you just couldn't figure out!

Best of luck with the little board, show us some piccys when your done.

I use diptrace to do all of my cct layouts. I found it really easy to learn how to use. It just seemed to make sense.

Cheers Mick

[Mechanic]

Hi Mario,

If you do go down the micro path, there is code for the sd2000 timing for the atmega644. http://www.geotech1.com/forums/showp...0&postcount=79
If you do use this code, just keep in mind that you may have to invert the operation of the pin that controls tx. Also leave the front end fets to work on their own, don't worry about hooking them up to the micro as this can be problematic for a number of reasons(I'm not saying it can't be done!)

Cheers Mick

[mario]

Hmmmmm

K.I.S.S

Ok now if you were to go to the trouble of using a micro to something, I would not bother using it to control some 4066's to switch in different caps to an oscillator. You can only run the oscillator in 1 setting(frequency) at any one time. When you switch from one frequency to another you need to allow time for the circuits to settle and then re ground balance the detector. The switched caps will work fine as you have them set out currently, no micro needed.

What I would do would be use the micro controller to create a square wave and drive pin 12 of 74hc00 with the desired frequency. A high speed micro would be best for this, that way you can easily fine tune out emi. You would need to use a rotary encoder to increase or decrease the frequency and perhaps a button to jump between pre-set frequency's. But then again if you were going to go to all of this bother, you might as well just set up the micro to control the tx and sample timings and be done with it and free from the pre-set minelab timings and cycle times. However to do this you would need to really understand how it all works, as it would be easy to end up with a detector that you just couldn't figure out!

Best of luck with the little board, show us some piccys when your done.

I use diptrace to do all of my cct layouts. I found it really easy to learn how to use. It just seemed to make sense.

Cheers Mick

Once we get a micro, the approaches to tackling some of our issues are endless.

I really don't have much experience in engineering, so my design approaches may be cumbersome (ex, this switch). Sorry if my solutions are not very elegant. However, with the help of you guys, I'm looking forward to implementing what I hope best works for my limited understanding.

I'll post pics once I get this little board done and let you know if it works.