Re: Voltage balancing resistors for series capacitors

Robert, your first sketch, the inductor and 16NF capacitor are a parallel tuned circuit, so when its at resonance, its high impedance, so when you feed it you need to feed it from a high impedance circuit and have lots of volts to drive current into it.
If you break the earthy end of the 16 NF capacitor, and insert a 160 NF capacitor and feed your 50 ohm source across that capacitor, the tuned circuits impedance appears to have fallen to 1/100 of what it was before. Unfortunately if you needed 16 NF to resonate the inductor, then you now have only 16 NF - 10% (capacitors in series) resonating the inductor. So the transformer action is slightly different.
Frank

Output impedance of audio amplifier is near 0ohm. To get max power from it load resistance must be close to specified. Othervice amplifier will be overloaded or not at all. Rare audio amp chips if any can give max power at 58kHz. Coil as antenna with parallel tuning capacitor has high impedance and need proper tap from cold side (maybe 1 or two turns) to lower it. How coil is "matched" to amplifier can be tested by measuring current it draws from amplifier output.

The discussion has been started about a series resonant circuit with reasonably selected capacitors.

There's no point of achieving good results with a parallel resonance circuit. I'll consider the circuits in post #11 as bad idea and irrelevant for the time being.

In case that the real impedance part of the coil (in series equivalent circuit) can't be brought into the range of nominal amplifier output impedance, an impedance matching circuit may required, involving additional LC elements.

I will continue with the series capacitance circuit and try a coil constructed with litz wire and see how that performs.

The circuit below is the one I am testing at the moment. This gives 1400Vac Peak to peak.
I could lower the resistance by using a larger diameter wire or litz wire as FvM suggested, but lowering the resistance below 8 ohms is not good for the amplifier. Am at 7 ohms now and going lower could subject it to overloading as Borber mentioned. Lowering the resistance also reduces the bandwidth whilst I am looking to increase the bandwidth.

Attachment 121022


Frequency response for TDA 7297

Attachment 121023

The voltage out of the amplifier is very low (~6 volts from memory)
Maybe this type of amplifier is not suited to the application.

You didn't yet mention any bandwidth requirements. But it's true that for higher bandwidth it may be necessary to create intentional losses and thus to provide more amplifier power.

Frank mentioned about having a lot of volts to drive the circuit. Would using a transformer between the amplifier and the series capacitors help to drive the antenna?
Or maybe a different type of amplifier that could feed into 75 Ohms rather than 8.
If a transformer would work, I presume the winding of the first coil in the transformer would still need to be seen by the amplifier as 8 ohms.

I have not thought about bandwidth until you mentioned it FvM. Then I looked at the variation of the peak frequencies of a set of RF tags and realized that this need to be considered.
With the existing circuit configuration and amplifier the bandwidth is ~ 150Hz. Ideally it should be about 600 Hz.

Bandwitdh of 600 Hz means that the resonator Q must be lowered to 100, involves larger series R and more amplifier output power for the same coil ampere turns, presumed it's driven by a linear amplifier. A transformer can't change the situation.

The TDA 7297 is a class AB amplifier so is linear. If I increase series resistance, I loose voltage across the coil.
I can only conclude then that I need another type of amplifier as the 7297 will only produce 1400V with a narrow bandwidth.

Do you have any suggestions where I might start looking as amplifiers or what type of amplifier might be best suited.
I don't have a lot of knowledge in this area.

Thanks.

If 1400V stands for generated field strength, respectively ampere-turns with a specific coil size, yes. Your requirements aren't yet clear, "digital" modulation with a certain rise/fall time or fully analog modulation? Is 1400 - 1800 Vpp (635 Vrms) across a 4.8 mH coil yout final specification? Theoretically, it involves only 2.5 W real power with Q = 100.

1800V PtP and bandwidth is what is required. Can adjust number of turns, wire thickness, amplifier type and power, circuit design, etc to suit.
Modulation is analog.
Only restriction is the size of the coil which is 30cm * 25cm.

Thanks

I woule assume that the significant parameter is generated field strength (or may be voltage/turn) , not total coil voltage. Adjusting the number of turns would generally change the total coil voltage.

If you are using one section of the TDA 7297, then if you return the tuned circuit to the unused output and feed this amplifier with an out of phase signal, you will have doubled your output voltage.
Frank

I constructed two type of coils and tried 2 different bridge amplifiers. Each coil was tuned to 58kHz.

TX coils
0.5mm single strand - 71 turns - 6.8 ohms resistance - 4.8mH - 1450V PtP
Litz - 93 turns - 2.6 ohms resistance - 5.68mH - 1440V PtP


I set up a tuned RX coil 150mm from the edge of the transmitting coil and recorded the voltage for both TX coils

RX readings for each TX coil[/U]
0.5mm = 10.67Vrms
Litz = 10.7Vrms


The amplifier for these measurements was the TDA7297 which is a 15W bridge amplifier designed for 8 ohm speakers
The other amplifier is a TDA7375 being a 35W bridge amplifier designed for 4 ohm speakers
The TDA7375 gave slightly lower readings than the TDA7297

If the coil only requires 2.5W to produce 1800V PtP with a Q of 100, I can only conclude there is something about audio amplifiers that limits the current in this application.
Any thoughts appreciated,
Robert

I reckon, that your circuit impedance is in the order of 20 ohms ( Xl = 2K @ 60 KHZ/Q), so your amps are not delivering enough current. Put a capacitor from each amp O/P to earth, say 20% of Ctuning. Swing the frequency or adjust Ctuning , you should see more voltage. Value of Cadded?, try 20 % of Ctuning. The amps should draw more current. Carry on to maximise the Vout. Watch out for the ICs overheating and closing down on thermal protection, as they are running out of gain at 60KHZ.
Frank

Thanks Frank. Will give that a try first thing tomorrow morning (Aus time) and post the results.

Added the a capacitor from each amp output to earth. The tune capacitance for the coil was ~ 1.35nF so added a 220pF to each output. Result was no change to the signal.
Added a capacitor between the series tuning capacitors and the coil and the other side of the coil (parallel). The antenna went out of tune and when I re-tuned it, the signal was lower than the original signal.
Added 2 / 220 pf capacitors, one each side of the coil and returned them to ground. Signal virtually disappeared.

Took the signal from one amplifier and fed it back into the bridge amplifier on the same chip and there was no signal on the output of the bridge amplifier. (TDA 7375)

Checked the amplifier during these tests and there seemed no change in the temperature.
Seems logical from all that you have suggested that the amplifier should have managed to provide the necessary current to achieve the 1800V with a reasonable bandwidth.

Will have to leave this where it is for now until I can find a suitable amplifier circuit.
Thanks everyone for your suggestions and assistance. Much appreciated.

Robert

If the input one amp goes +, its output goes +, if you feed this back as an input to the second amp (as a +) then the output of the second amp goes + as well. So you should get little or no output across the tuned circuit, rather both end of it are going up and down at the same rate.
Frank

How about putting another litz conductor in parallel with your existing coil (wind it on the same frame), should not change the inductance too much but should reduce the Rloss - the DC resistance will be halved for a start! The extremely high Q coils I once got involved with were 119/42 SWG !
Frank

balancing HiV caps in series with R's is a bad idea, maybe MOVs

Best is to get a 2kVac Film cap often rated for 5~10x for DC
See RF ceramic Catalog here
http://www.vishay.com/capacitors/list/product-22146/

Pay attention to kVAr and D dissipation factor. YOu may need film instead of ceramic.. not cheap. but best solution.