how to design an antenna ?

[schneiderj]

Hello,

I have build my first own coil. And now I have to mesure them. As first approch I did as follow. I build the following circuit :

**broken link removed**

• 
  And
• mesure U1, U2 adn U3 for various frequence (sinusoidal)
• calcul the intensity with the value of U2 and R (50 Ω)
• Calcul Z from the calculated I and U3
• calcul the value of the inductance

I obtain these results :
**broken link removed**

With the help of my son we do the calculation with the imagenary approch :

RL = E1*R/E2 * cos(φi) - R
L = -E1*R/(E2*ω) * sin(i)

at a frequancy of 5500 Hz, I have :

• E1 = 134 mV
• E2 = 18.4 mV
• φ=i 4.4 radians

That give RL = 11.6 Ω (found with ohmeter : 3.8 Ω) and for L = 10.3 mH.

This valie for L is not far from one found with the first method, but why I have so large difference for RL (between mesure and calculation) ?
Did that means that my approch is not good ?

Thanks you for your help !
Jean-Marie

 

[schneiderj]

Hello,

I have confirmed the inductance of the coil by searching the frequence of the resonnance of a LC circuit. I use for that a capacitor of 97.9 nF and found a maximum at 4970 Hz (see the pdf document, x axis in Hz, answer without unity), which give a value of 10.5 mH.
Is that correcte ?

I take a picture of the signal obtained with that coil and my cardio belt. It is in the attached gif document. Before performing amplification of the signal did I have to reduce the noise or is it better to perform that operation after amplification ?

Thanks you for your help,
Jean-Marie

 

[RadioRon]

Hello,

I have confirmed the inductance of the coil by searching the frequence of the resonnance of a LC circuit. I use for that a capacitor of 97.9 nF and found a maximum at 4970 Hz (see the pdf document, x axis in Hz, answer without unity), which give a value of 10.5 mH.
Is that correcte ?

I take a picture of the signal obtained with that coil and my cardio belt. It is in the attached gif document. Before performing amplification of the signal did I have to reduce the noise or is it better to perform that operation after amplification ?

Thanks you for your help,
Jean-Marie

To be honest Jean-Marie, I was not able to double check your calculated values for L and Rl by reviewing your calculations. I am just a bit confused about your terms. However, I can confirm that your experimental method by finding the resonant voltage frequency of 4970 Hz is correct, and if your measurement of the capacitor value is correct then 10.5 mH is indeed correct. Happily, this agrees with the previously determined values but it does show that your experimental error in the first method is significant, since you calculated value of L ranging from 10 to 21 mH (I ignore the value 73mH as the value of U1 in that calculation was very small and probably included a very large error).

I worked backwards from your graph of E vs frequency to calculate the Q and hence the total circuit resistance. I find that the Q is 43 and with an inductive reactance of 322 ohms, this means that the series R in the circuit is about 7.5 ohms. I am wondering what your experimental setup was for this test since the total circuit R would include the inductor R as well as the source resistance too.

The amount of signal that you see from the cardio device on your scope is very encouraging. I think it is a great success so far. The noise is a function of the measurement bandwidth which, when using an oscilloscope, is extremely large. The best way to avoid noise and to increase your receiver sensitivity is to restrict the receiver bandwidth to the minimum necessary to pass the carrier and all of the useful sidebands needed to show the modulation correctly. The total width of this signal depends on how the 5KHz is modulated, that is, is the modulation sinusoidal in nature, or square wave or something else? Anyways, to answer your question, we need to include a band pass filter. We can put this before or after the amplifier. If we put it before the amplifier, we will have to make the filter have very low loss so that we do not excessively degrade the noise figure of the system. If we put it after the amplifier we don't have to worry about the loss nearly as much. Normally, putting the filter after the amplifier means that the amplifier will be amplifying a broad bandwidth of signals and sometimes this means that stronger interference signals will also be amplified. These are not a problem as long as the amplifier is operating in its linear region, but as soon as the amplifier begins to limit its gain will drop and your sensitivity will drop too. For these reasons, more advanced receivers usually put a band pass filter before the amplifier and another band pass filter after the amplifier.

I suggest that since this is a simple system, a filter after the amplifier will be adequate. I'm not sure what the lower and upper frequencies should be for this band pass filter. Certainely, we can probably put the lower frequency at about 4500 Hz, but the upper one depends on several factors, mainly the shape of the modulation in the time domain.

 

[schneiderj]

Thanks for these great explations !!

OK, I understand that my previous explanation are not very clear, I made confusion between several abreviations !! I use H instead of L.
U1, U2 and U3 are the value obtained with the help of a multimeter, and E1, E2 are the value found with the oscillo (mVolt pic to pic).

I think that the bad reproductibily of the mesure are coming from the generator I am using : it is a very basic sound card with a simple sofware ! I will made one in the next weeks with XR-2206 element (https://www.elektronline.hu/konyvtar/tan005/0003.pdf.

Here are other picture of the signal obtained with a simple coil (2 different in fact). First have been post here :https://www.electro-tech-online.com/threads/how-to-design-an-antenna.84693/#post658644. It seams that there is no modulation. Did I am correcte ?

I will search for band pass filter. What did you suggest as first approch : an active or a passive one ?

Jean-Marie

 

[RadioRon]

Thanks for these great explations !!

OK, I understand that my previous explanation are not very clear, I made confusion between several abreviations !! I use H instead of L.
U1, U2 and U3 are the value obtained with the help of a multimeter, and E1, E2 are the value found with the oscillo (mVolt pic to pic).

I think that the bad reproductibily of the mesure are coming from the generator I am using : it is a very basic sound card with a simple sofware ! I will made one in the next weeks with XR-2206 element (https://www.electro-tech-online.com/custompdfs/2008/11/0003.pdf.

Here are other picture of the signal obtained with a simple coil (2 different in fact). First have been post here :https://www.electro-tech-online.com/threads/how-to-design-an-antenna.84693/#post658644. It seams that there is no modulation. Did I am correcte ?

I will search for band pass filter. What did you suggest as first approch : an active or a passive one ?

Jean-Marie

Well, we did not expect any modulation other than on-off keying, which is also a type of modulation. When the transmitter is keyed, it is only sending a 5KHz tone with no other modulation other than the obvious ramping up and down at the beginning and end of the burst. As we recall, the manufacturer described the operation as one or three bursts of 5KHz tone for each heartbeat, one or three depends on which model of transmitter. So it appears to be operating according to their description.

Since there is nothing more than a 5KHz tone bursting on and off fairly slowly, we can make the bandpass filter quite narrow. For example, 4700 Hz to 5300 Hz is not unreasonable. I would make it narrower, but I'm not sure how stable the 5KHz frequency is, nor am I sure how stable our filter might be, so we need to be conservative. The easiest way to make a narrow filter is to use parallel resonant tank circuits. The schematic diagram attached shows a typical configuration of such a circuit.

C2 resonates with L1 at 5KHz. L2 resonates at 5KHz with the net capacitance of C4 and C5 in series. C3 is kept quite small, with Z perhaps about 1000 ohms or a bit less. In this way you have two resonant tank circuits which are loosely coupled through C3. The coupling is kept small so as to avoid overloading the tank. L1 and L2 are kept away from each other so that they do not couple. In fact, if you choose to make L1 and L2 close to each other to intentionally couple, you might be able to leave out C3.

I have chosen to show two different methods of coupling energy into and out of the circuit. The input is using a tap on the inductor to match the input impedance to the tank. The lower the source Z, the "colder" the tap should be on L1 (ie. the fewer turns from ground where the tap connects). The output is using two capacitors to create a voltage divider, again to match the load impedance to the tank. If your external load is very high impedance, you can short out and remove C4 and just resonate the second tank with L2 and C5.

Obviously this is not the only way to make a narrow bandpass filter. You can use op amps to make an active filter if you like.

 

[schneiderj]

Thanks for this explanation. I wurf on the net and many pages on active filters, but only few on passive one. The best I have found is this one : . By the way a discover 5Spice sofware. It is very easy to use at least for the smal simulation I have done...

But I would like to have a better understanding on filter designing, not only with simple one with a RC or LC... did you have some recommendations (book, web site) ?

I perform the first amplification of the signal of my coil... it is very encouraging : I have a good signal at 1 meter. And this with a common AOP (LM324) and without filtration !!

Jean-Marie

 

[audioguru]

The lousy old LM324 quad opamp and its sister the LM358 dual opamp have very poor high frequency performance. Their max gain at 5kHz when the supply is 10V to 15V is only 100. The max gain at 5kHz is less when the supply voltage is less than 10V.

An MC34071 single opamp, MC34072 dual opamp and MC34074 quad opamp has a max gain at 5kHz of 800 at a supply voltage as low as 3V.

 

[schneiderj]

The lousy old LM324 quad opamp and its sister the LM358 dual opamp have very poor high frequency performance. Their max gain at 5kHz when the supply is 10V to 15V is only 100. The max gain at 5kHz is less when the supply voltage is less than 10V.

An MC34071 single opamp, MC34072 dual opamp and MC34074 quad opamp has a max gain at 5kHz of 800 at a supply voltage as low as 3V.

Hello audioguru !!

That was for testing purpose only !!

I have found an interresting document from TI (https://focus.ti.com/lit/ml/sloa088/sloa088.pdf. Quite simple to understand an very informative !

From this and with the help of 5Spice I have done various simulation to see what is happening with the various solutions.
Then with the help of FilterPro I have try to find a good solution. The following schem was issued from my last experiment (see pdf document). As aop I want to use is TLC27L4CP.

But before going head I would like to know if that way is correct.
So any comments are wellcome

Jean-Marie

Jean-Marie

 

[audioguru]

The LM324 has poor high frequency response because it is low-power.
The TLC2714 is ultra-low-power so guess what? It has extremely extremely poor high frequency response. it is also extremely noisy.
At 5khz its open-loop voltage gain is only about 20 and it cannot be used in a 5kHz filter.

Use a half-decent normal audio opamp like a TLC071 single, TLC072 dual or TLC074 quad opamp that work well with low noise up to 600kHz.
Make a bandpass filter tuned to 5kHz.

 

[RadioRon]

Hello audioguru !!

That was for testing purpose only !!

I have found an interresting document from TI (https://www.electro-tech-online.com/custompdfs/2008/11/sloa088.pdf. Quite simple to understand an very informative !

From this and with the help of 5Spice I have done various simulation to see what is happening with the various solutions.
Then with the help of FilterPro I have try to find a good solution. The following schem was issued from my last experiment (see pdf document). As aop I want to use is TLC27L4CP.

But before going head I would like to know if that way is correct.
So any comments are wellcome

Jean-Marie

Jean-Marie

Your filter seems overly complicated and the performance could be better. Try the attached circuit.

The graph shows the relative performance. To see the true performance, add 20 dB of gain to the graph result.

edit: the second op amp on the right functions as a high impedance buffer, so its input resistance should be kept higher than 50K ohms. It's voltage gain should actually be set for 0.1 not 10 as shown. You can adjust the gain as you see fit.

The circuit may suffer from ringing. If this happens, it can be reduced by adding some resistance at various points

 

[schneiderj]

Hello !!

I till working arround that project and it is almost completed. I till have probleme to find the good way to store it in a friendly box...

Several weeks ago I found that componant : **broken link removed**. Very small if I compare to my antenna (4*1 cm), but a little bit costly for my application.

I try to find the technologie used to reach this solution, but till now I am unsecussfull.
What is your felling ?

Thanks for your comments,
Jean-Marie