RESONANCE EFFECTS FOR EVERYONE TO SHARE

[duff]

Hi Gyula,

I am also puzzled by the half wave rectifier's performance here,  maybe the oscilloscope waveform across your L2 would reveal something  but I would expect a normal sinusoidal 305 kHz waveform with symmetrical zero crossings...

Unfortunately my camera has lost the ability to communicate with the computer. I have yet to determine whether its the camera or the software.

I did look at the waveform across the secondary and its more like a square wave with rounded corners with jagged sides. Yesterday I looked at the output of the FWB and is seemed ok.

If I can get the camera talking again I'll upload the pics.

You surely know this method when the series inserted resistance equals the unknown input impedance (at resonance of course) then the input voltage from the SG gets halved across the unknown, (a normal voltage divison happens).

Yes I know the method and I'm embarrassed to say it did not occur to me. <cowering in shame> 


@Luc

Using the above principle I was able to determine the impedance at resonance. I started at 302KHz where the impedance was 741 ohms (741 +j0) and by adjusting the capacitance and varying the frequency the impedance decreased and the voltage increased. Finally I reached the point where the voltage started droping again. At the sweet point I had the following output:


V_out = 4.521VDC
F = 121.9 KHz
C1 = .0012uF
L1 = 1.38mH
R_L1 = 8.26Ω
R_L = 63.6Ω
Zin = 139Ω @ 0 degrees.

SG = square wave 23Vpp / 8.13Vrms

The above voltage was with the full-wave bridge in the circuit.
Note: I determined yesterday that a half-wave bridge, in this circuit,  yields more volts out.

Impedance measurements were taken using a sine wave.


Gyula suggested a matching network and I look at that some today. That would provide a little more output . Also wondering how harmonic attenuation (due to the matching network) would influence the output???


-Duff

[Edit]
Change component names to correspond with schematic
Corrected value of L1
Now LC works in resonance formula

[gyulasun]

Hi Duff,

Would you mind updating your schematics with the new component values because at the moment I am puzzled a little by your new symbols like R_L C_series etc. ?    (I mean the circuit you measured Z=139 Ohm real input impedance.)

I can assure you:  you DO NOT have to feel yourself embarrased at all for anything , you do an excellent and outstanding job here!

Thanks,  Gyula

EDIT: 

I did look at the waveform across the secondary and its more like a square wave with rounded corners with jagged sides. Yesterday I looked at the output of the FWB and is seemed ok.

Would you recall that you used a square wave input in this case?  It is odd you found waveform like that, though the loaded Q of L1C1 might justifies this?  It would be also good you could check the 3dB bandwidth in your latest circuit when you have time for that.

[duff]

Would you mind updating your schematics with the new component values because at the moment I am puzzled a little by your new symbols like R_L C_series etc. ?    (I mean the circuit you measured Z=139 Ohm real input impedance.)

Sorry for the confusion.  I renamed the components to be inline with the previous posts...

Schematic added above...

EDIT:       Would you recall that you used a square wave input in this case?  It is odd you found waveform like that, though the loaded Q of L1C1 might justifies this?  It would be also good you could check the 3dB bandwidth in your latest circuit when you have time for that.

I'll get the 3db bandwidth for you - no problem

Yes - I used square wave input for voltages across RL
I used sine when taking impedance readings.

As always I value your input.!


-Duff

[duff]

Gyula,

Vpeak = 5.6V
3db voltage = 5.6 * .707 = 3.96V

Fr = 121.9 KHz
3db points: 92.7, 140.6 KHz

3db BW = 140.6 - 92.7 = 47.9KHz

Q = 121.9/47.9 = 2.545

I think I did that correctly - it's been a long time...


-Duff

[gyulasun]

Hi Duff,

Yes you did it correctly, and the loaded Q of  2.5 means something ruins too much the unloaded (high) Q of L1... and this low value also explains the waveform you described across L2.

(The unloaded Q of L1 in itself  is  (2pi*f*L1)/R_L1=1056.43/8.26 gives about  Q=128.)

Because basically the SG output of 50 Ohm is loaded by a series resonant circuit that is further loaded by the transformed bulb or your transformed 66.6 Ohm resistance, the big mismatch seems to come from the 50 Ohm that is in series with the coil:  if you consider the loaded Q like (2pi*f*L1)/(R_L1+R1+50)  then you get Q=1056.43/59.1=17.8 and this is loaded further by the transformed load of 66.6 to your value of 2.5, agree?
So improvement can come by matching somehow the primary tank circuit to 50 Ohm.  One possibility is to make some taps on L1 by unwinding about 1/6 or 1/5 part of it, make a few cm long tap and continue rewinding again what was unwound to receive back the 145 turns;  another possibility is indeed use a matching network (can be a pi filter or L+C low pass filter etc).

However we should not forget what Luc was asking: verifying his anomaly finding.  So I do not suggest any further alteration in his original circuit for now, it is up to you how to proceed. At this stage you are now, we know pretty much correct data about this circuit. I will be available further on. 

Regards,  Gyula