Question about the half bridge in this circuit

Can any one explain to me how Steve might have arrived at the value of the 0.68uF capacitors?
And would I need to change their value if I was to run this off say 12V instead of 120V?

Capacitor value is given by the switching frequency and coil inductance (coupled to the long output coil). None indicated, so I cannot say. Why don't you ask Steve, the author?
Reducing DC voltage from 120 to 12 will reduce the output power, even 100 times. You should try it, also pulse width contributes , again not indicated.

jiripolivka said:

Capacitor value is given by the switching frequency and coil inductance (coupled to the long output coil). None indicated, so I cannot say. Why don't you ask Steve, the author?
Reducing DC voltage from 120 to 12 will reduce the output power, even 100 times. You should try it, also pulse width contributes , again not indicated.

Click to expand...

There is no direct path to ground, other than through the capacitors. So can I assume that the larger the capacitance of C7 and C8, the larger the peak current through the primary coil?

Steve Ward contributes to this forum: https://4hv.org/signup.php

But unfortunately I can't figure out how to get past their registration page - there appears to be no way of getting their forum system to acknowledge that I have read their rules. So it just keeps throwing me back to the same start page.

boylesg said:

There is no direct path to ground, other than through the capacitors. So can I assume that the larger the capacitance of C7 and C8, the larger the peak current through the primary coil?

Steve Ward contributes to this forum: https://4hv.org/signup.php

But unfortunately I can't figure out how to get past their registration page - there appears to be no way of getting their forum system to acknowledge that I have read their rules. So it just keeps throwing me back to the same start page.

Click to expand...

If the switching frequency is variable, then yes, larger capacitors are better. It seems to me that the frequency may be fixed, so he capacitors should resonate with the output transformer inductance.

You did not indicate the purpose of the circuit. Looks like a HV generator, or a driver for Tesla transformer. Quite complex where a mechanical buzzer can work.

When I have simulated these kinds of layouts, I watch the capacitors gain a DC charge, and then it varies up and down with each cycle.

I would say the capacitor values are large enough when the charge varies about 5 or 10 percent per cycle. You can make them larger, and it gains you a little, but not that much more.

BradtheRad said:

When I have simulated these kinds of layouts, I watch the capacitors gain a DC charge, and then it varies up and down with each cycle.

I would say the capacitor values are large enough when the charge varies about 5 or 10 percent per cycle. You can make them larger, and it gains you a little, but not that much more.

Click to expand...

Right. They're functionally equivalent to a DC blocking cap. You'd want to size them so the voltage change per cycle (impedance) is low at the anticipated current/frequency.

Changing to 12V may have have a big effect on currents (depending on your load) so they may want to be re-sized.

jiripolivka said:

If the switching frequency is variable, then yes, larger capacitors are better. It seems to me that the frequency may be fixed, so he capacitors should resonate with the output transformer inductance.

You did not indicate the purpose of the circuit. Looks like a HV generator, or a driver for Tesla transformer. Quite complex where a mechanical buzzer can work.

Click to expand...

Sorry....here's the whole circuit: https://www.stevehv.4hv.org/SSTC5/miniSSTCfnlsch.JPG

It's a driver for a solid state tesla coil. The frequency can be varied via R6, r% and a capacitor sub-point.

How would you go about calculating the peak current based on the square wave frequency, duty cycle, the primary inductance and the capacitor values?

I have seen the formula involving Pi but clearly that formula is not applicable here given that a square wave is involved rather than a sine wave.

I have seen the formula involving Pi but clearly that formula is not applicable here given that a square wave is involved rather than a sine wave.

Click to expand...

If the transformer is operated in series resonance at the fundamental of the square wave you get an roughly sinoidal current. So you can calculate the resonant circuit as if it would be driven by a sine waveform.

FvM said:

If the transformer is operated in series resonance at the fundamental of the square wave you get an roughly sinoidal current. So you can calculate the resonant circuit as if it would be driven by a sine waveform.

Click to expand...

What is the name of that equation? So I don't have to search for it again blindly.

F = 1/ 2 X PI X (LC)^.5, where C = C7 + C8
Frank

chuckey said:

F = 1/ 2 X PI X (LC)^.5, where C = C7 + C8
Frank

Click to expand...

What is the name of this law Frank?

Resonant frequency of LC circuits? https://en.wikipedia.org/wiki/LC_circuit
Frank

FvM said:

If the transformer is operated in series resonance at the fundamental of the square wave you get an roughly sinoidal current. So you can calculate the resonant circuit as if it would be driven by a sine waveform.

Click to expand...

If its resonant it would be sinusoidal but I don't see any reason to think it's resonant.

Note that myself and a couple others suggested earlier that the C should be sized for a low voltage change. Doing that would preclude resonance (which requires voltage contribution from the cap). So using a resonant current formula to size the cap to prevent resonance doesn't make any sense.

Not operating in resonance, with negligible voltage change on the C during a cycle would nominally produce a triangle. The size and shape of the triangle would be defined by I'=V/L and the frequency.

Note that the triangle wave is the worst case compared to the resonant calculation. Though neither factors in the secondary load.

asdf44 said:

If its resonant it would be sinusoidal but I don't see any reason to think it's resonant.

Note that myself and a couple others suggested earlier that the C should be sized for a low voltage change. Doing that would preclude resonance (which requires voltage contribution from the cap). So using a resonant current formula to size the cap to prevent resonance doesn't make any sense.

Not operating in resonance, with negligible voltage change on the C during a cycle would nominally produce a triangle. The size and shape of the triangle would be defined by I'=V/L and the frequency.

Note that the triangle wave is the worst case compared to the resonant calculation. Though neither factors in the secondary load.

Click to expand...

From this and the reading I have done it looks as though I will need to determine the resonant frequency of the secondary coil plus top load, and then match the square wave frequency and the resonant frequency of the primary circuit. That will involve calculating the appropriate capacitor values given a fixed inductance value for the primary coil.

I have previously set up a little device will tell me, roughly, the resonant frequency of my secondary coil.

I was fiddling with the circuit in that java circuit simulator and it seems that at resonance, the peak current in the primary is between 100mA and 200mA. Which would fit in with the FETs Steve has specified that have a maximum continuous drain current of 3A or so.

I am going to use some 62A max FETs and interchangeable caps so that this part of the circuit is more versatile.

boylesg said:

From this and the reading I have done it looks as though I will need to determine the resonant frequency of the secondary coil plus top load, and then match the square wave frequency and the resonant frequency of the primary circuit. That will involve calculating the appropriate capacitor values given a fixed inductance value for the primary coil.

I have previously set up a little device will tell me, roughly, the resonant frequency of my secondary coil.

I was fiddling with the circuit in that java circuit simulator and it seems that at resonance, the peak current in the primary is between 100mA and 200mA. Which would fit in with the FETs Steve has specified that have a maximum continuous drain current of 3A or so.

I am going to use some 62A max FETs and interchangeable caps so that this part of the circuit is more versatile.

Click to expand...

I'm missing the explanation for why you want to do any of those things. Do you have additional sources talking about resonance in this context?

Lots of power supplies operate completely independent of resonance and/or deliberately avoid it. We have what looks like a straightforward topology here where any AC signal on the primary (including square) will produce an amplified signal on the secondary...and I think that's the goal here.

asdf44 said:

I'm missing the explanation for why you want to do any of those things. Do you have additional sources talking about resonance in this context?

Lots of power supplies operate completely independent of resonance and/or deliberately avoid it. We have what looks like a straightforward topology here where any AC signal on the primary (including square) will produce an amplified signal on the secondary...and I think that's the goal here.

Click to expand...

I am going by what I saw when I set that part of the circuit up in that java circuit simulator (https://www.falstad.com/circuit/).

The 680nF caps seem to indicate a resonant frequency of about 30kHz according to that LC equation. Assuming an inductance value of around 20uH for the primary coil - I have not set that up yet so I don't know exactly what it is and Steve has not specified what the inductance of his primary coil is. But based on his form diameter and number of primary turns he specifies it should be in the vicinity of 20uH. I wrapped a few turns of wire around a coffee mug and measured the inductance.

When I set the square wave input to the FET gates at this frequency I observed a maximum current in the primary coil of between 100mA and 200mA. It was hard to see an exact figure.

If I set the square wave at 1Hz or changed the value of the caps significantly then that peak current dropped to uA's

So I am assuming that to get peak power transfer from the primary to secondary coil you need them operating at the same frequency and you 'tune' the primary circuit accordingly.

That is what you do with a spark gap tesla coil, i.e. you tune the primary circuit by means of the spark gap and capacitor bank.

I also assume that Steve Ward's primary circuit has some inherent current limiting in it that would prevent it burning out his rectifier. Perhaps by slightly de-tuning the primary circuit to prevent perfect resonance? I have not quite 'nutted' this bit out yet but I am trying to learn.

On my first attempt to run the thing I might put a 15A or so fuse on the wall socket side to make sure I don't fry any wiring in the wall etc and perhaps a 10A or so fuse on the transformed/rectified side to make sure I don't fry my transformer and rectifier.

Oh I just remembered - it is the duty cycle of of the square wave, that turns the FETs on and off, that determines the peak current in the primary. The larger the duty cycle and the longer the FETs remain on each cycle, the greater will be the peak current in the primary given that it is also resonant.

boylesg said:

I was fiddling with the circuit in that java circuit simulator and it seems that at resonance...

Click to expand...

I guess you mean the tesla coil model in Falstad's java simulator. I'm running that now. The resonant action is not in the first transformer primary, but in other areas of the circuit.

A spark event occurs during the rising waveform of the 60 Hz power supply. The transformer primary is 10H. Step up ratio is 100 x.

The resonant frequencies which I observe are much faster:

(a) the righthand loop at 200 kHz. Once it gets oscillating it continues to some extent for a long time.

(b) The middle loop oscillates during a spark event at 200 kHz, modulated at 2 or 3 kHz.

On old TV sets that used a line output transformer to step the 200v up to 16 KV, the transformer was actually tuned to three times the operating frequency. This gave a "squarer" current pulse then tuning to the fundamental, so more power was transferred from primary to secondary.
Frank

BradtheRad said:

I guess you mean the tesla coil model in Falstad's java simulator. I'm running that now. The resonant action is not in the first transformer primary, but in other areas of the circuit.

A spark event occurs during the rising waveform of the 60 Hz power supply. The transformer primary is 10H. Step up ratio is 100 x.

The resonant frequencies which I observe are much faster:

(a) the righthand loop at 200 kHz. Once it gets oscillating it continues to some extent for a long time.

(b) The middle loop oscillates during a spark event at 200 kHz, modulated at 2 or 3 kHz.

Click to expand...

No, no, no.

I actually re-created the FET portion of Steve Ward's circuit in the circuit simulator.

The FET gates were being triggered by the square wave generator provided in the circuit simulator, one of the gates through an inverter so the were turned on alternately.

boylesg said:

No, no, no.

I actually re-created the FET portion of Steve Ward's circuit in the circuit simulator.

The FET gates were being triggered by the square wave generator provided in the circuit simulator, one of the gates through an inverter so the were turned on alternately.

Click to expand...

Sorry I misunderstood.

I had a look at your post #7 link. I believe the half-bridge and capacitors is a substitute for a full H-bridge. It is simpler to design and operate a power supply this way.

Switching frequency depends on how fast the mosfets are driven. They are switched according to what arrives at the antenna (at extreme left in the schematic).

My guess is that the antenna picks up EM radiated from the tesla coil, as the secondary resonates (or perhaps self-resonates). In this way the system automatically fixates on the resonant frequency.

I don't believe you need to adjust capacitor values to achieve resonance.
The capacitor values only need to be large enough to store juice during one half of the power cycle and then release it during the other half, without dropping too much in voltage.