How to implement an "auto-tune" circuit?

[janetsmith2000@yahoo.com]

Hi,

How to implement an "auto-tune" circuit, which could be found in digital radio/TV set. when setting tuning a channel, I could choose Auto-tune or Manual tune.

Actually i want to make a tunable Analogue Filter. But i think how good it is if i can make this has the capability to auto tune itself.

What's the difference between Analogue Filter and Digital Filter?

 

[Nigel Goodwin]

Hi,

How to implement an "auto-tune" circuit, which could be found in digital radio/TV set. when setting tuning a channel, I could choose Auto-tune or Manual tune.

Actually i want to make a tunable Analogue Filter. But i think how good it is if i can make this has the capability to auto tune itself.

What's the difference between Analogue Filter and Digital Filter?

Auto-tuning is fairly simple to do, but you need some way of detecting it's tuned in - on a TV they use the AFC output, this is normally sat at a specific middle level, as you tune through the signal it will go either high or low (depending on the design), move smoothly through the middle (accurately tuned) point, and then go low or high again at the other side. Outside any station tuning it will simply stay at the middle, accurately tuned, level - so you need to detect and recognise the changes positive and negative either side of a signal, then tune it to the middle point. The AFC circuit is basically an FM demodulator.

What kind of analogue filter are you talking about?, and what can you use to lock to? - any auto-anything has to have someway of recognising when it's found something.

Assuming you are talking audio filters, there's not much practical difference between analogue and digital (switched capacitor presumably?) ones - the main advantage is the ease with which you can tune digital ones.

 

[janetsmith2000@yahoo.com]

I was designing a portable EKG system. normally there got 3 electrod (ground, +ve, -ve) to measure voltage (in microvolt range) from a patient. Since it is portable, it doesn't have any Ground lead, so the reading is very susceptible to outside interference / noise.

I suppose that a filter system can be used to filter out the noise.

 

[Russlk]

I expect that the signal does not have a discrete frequency, so the better approach is to prevent the interference rather than try to filter it out. Shielding and common mode rejection should be all that is needed.

 

[Nigel Goodwin]

I expect that the signal does not have a discrete frequency, so the better approach is to prevent the interference rather than try to filter it out. Shielding and common mode rejection should be all that is needed.

Yes, I believe such devices commonly use fixed low-pass filtering, the signals you are looking for are all very low in frequency so you can reject any above your highest frequency. But you want to start with the best possible signal you can - don't start with a rubbish signal and then try and improve it, make it as good as you can in the first place.

 

[janetsmith2000@yahoo.com]

So,

FFT is not required as well?

Shielding, is it kind of special cable?

 

[Russlk]

As I understandit, FFT is a transform and does not do anything by itself, but does allow different forms of signal processing. Coaxial cable is one form of shielding.

 

[janetsmith2000@yahoo.com]

**broken link removed**

noise in a image file can be reduced using fft, so i think it might help in my design project.

 

[janetsmith2000@yahoo.com]

I expect that the signal does not have a discrete frequency, so the better approach is to prevent the interference rather than try to filter it out. Shielding and common mode rejection should be all that is needed.

Yes, I believe such devices commonly use fixed low-pass filtering, the signals you are looking for are all very low in frequency so you can reject any above your highest frequency. But you want to start with the best possible signal you can - don't start with a rubbish signal and then try and improve it, make it as good as you can in the first place.

I enclosed a picture taken from **broken link removed**

It stated it was 100Hz, .... was it high enough to be considered 'high frequency'? was it high enough to implement various advance technology to get a better waveform?

 

[bmcculla]

Hello,

When I suggested a tunable filter i was thinking of something along the lines of the TRAC020LHQ36 (available on digikey.com). This IC uses switched capacitors to allow you to program in various analog filters. I probably should have said "programable filter" instead of tunable filter. A programable filter will allow the user of your EKG to specify a high pass filter and low pass filter with specific frequency cutoffs to get a better view of the signal they want.

A programmable filter can also be implemented with a digital filter. A digital filter uses mathematical formulas on digital data to filter it. Look up FIR filters and IIR filters for more information. Using these types of filters is called "Digital Signal Processing" or DSP.

A FFT is another type of digital signal processing. The FFT is a Fourier Transform optimised for digital systems. The Fourier transform breaks a signal down into its frequency components letting you see what frequencies are present in the waveform. In your project the FFT would be only for viewing - using a FFT to filter data takes a lot of processing time ( can be slow even on a pentium 4).

Shielded cables are very important for a good EKG reading. Shielded cables have a metal foil wrapped around their internal conductors. The metal foil is connected to ground and "shields" the internal wires from noise. Just cut open a standard TV coaxial cable to see the foil and internal conductor. The standard BNC cables in your electronics lab should be fine. The sheild works as a faraday cage preventing electromagnetic intrferance from reaching the wires carrying the signal. For extra sheilding use aluminum foil connected to ground to wrap problem areas (careful not to short anything out). Nigel is right about noise - reducing it with properly sheilded cables and electronics is much better than filtering.

Hope this helps

Brent

 

[Nigel Goodwin]

It stated it was 100Hz, .... was it high enough to be considered 'high frequency'? was it high enough to implement various advance technology to get a better waveform?

It states 25mm/Sec - which presumably is the horizontal axis, so the signals are no where near 100Hz. I presume the 100Hz probably refers to the low pass filter used?.