Flyback Transformer Driver

[boylesg]

In flyback transformer / ignition coil drivers of this form, would folks recommend any additional protection meaures for the linear voltage regulator other than the snubber I have implemented. I have played around with various snubbers in the javascript circuit simulator and, as far as I can see from that, all of the kick back current is diverted into the snubber network. But just in case......

 

[BradtheRad]

Yes, the customary plain diode will gobble up all the energy from the collapsing flux field. Leaving nothing in the secondary.

Try different arrangements and combinations. I think you want a resistor in series with the diode or zener or capacitor.

I'm not sure there's a science to it.

The snubber network only needs to absorb the strongest part of the spike, so that it doesn't get at fragile components.

 

[boylesg]

Yes, the customary plain diode will gobble up all the energy from the collapsing flux field. Leaving nothing in the secondary.

Try different arrangements and combinations. I think you want a resistor in series with the diode or zener or capacitor.

I'm not sure there's a science to it.

The snubber network only needs to absorb the strongest part of the spike, so that it doesn't get at fragile components.

In the jaycar electronics kit version of this circuit, that simply run off 1 car battery, they just had a resistor in place of my voltage regulator.

From what I have learned so far it seems snubber design is as much trial and error as calculation.

Then inside that they had a reverse biased 16V / 1W zener and a 470uF capacitor. I am assuming that was another snubber network to absorb any remaining over voltage that would otherwise have damaged the 555.

Perhaps I should re-instate a 25V or so zener + 47u capacitor snubber between the 1N4007 and the voltage regulator to make sure that the input to the voltage regulator never exceeds its rated maximum.

 

[jiripolivka]

The circuit looks correct. I would advise to find an old CRT monitor or a TV set if you can find its schematic. Their flyback circuits also utilized a "booster" diode in the primary circuit, for better efficiency.
I would add one more diode, 1N 4007, across the IRF MOSFET, cathode to drain, anode to GND, to protect the IRF.

 

[FvM]

I would add one more diode, 1N 4007, across the IRF MOSFET, cathode to drain, anode to GND, to protect the IRF.

Protect against what? The diode is already built-in to the MOSFET. In addition, the diode won't be ever forward biased in a flyback switch.

 

[boylesg]

Protect against what? The diode is already built-in to the MOSFET. In addition, the diode won't be ever forward biased in a flyback switch.

Can I take it that it would do no harm to calculate the the RCD snubber to handle a slightly larger kickback voltage than what you actually expect?

 

[jiripolivka]

Had you ever tried to build a flyback circuit? If you do, be cautious about HV spikes, not only in the secondary circuit.
Even if the IRF has a protection diode, I would still add one 1N4007, to protect its channel from overvoltage.
Use a scope with a caution- spikes on the IRF may grow higher than its channel can survive.

 

[boylesg]

Had you ever tried to build a flyback circuit? If you do, be cautious about HV spikes, not only in the secondary circuit.
Even if the IRF has a protection diode, I would still add one 1N4007, to protect its channel from overvoltage.
Use a scope with a caution- spikes on the IRF may grow higher than its channel can survive.

This is not the first one I have built. I have built a few versions now and fried a few mosfets and a zener diodes. The mosfets came from discarded tvs so that's OK. Have been learning as I go along.

I have found that 4 x 75V zener diode snubber is insufficient to dissipate the kick back when driving a flyback from a tv with 24V at about 20kHz - I smoked the zener diodes. Hence I calculated and added an RCD snubber. I have tested this briefly but yet to give it a longer test.

I also killed a 1.5A 12V voltage regulator possibly because the mosfet gate driver is drawing more current than it is rated for. This happened as soon as I switched flyback transformers to one that ran at 20KHz rather than perhaps 7kHz. I have since modified my circuit to use an LM317 supplying just over 13V and with a pass transistor to supply up to 10A - the mosfet gate driver can supply up to 9A.

The mosfet I have been using are rated at 500-600V and around 15A.

 

[jiripolivka]

This is not the first one I have built. I have built a few versions now and fried a few mosfets and a zener diodes. The mosfets came from discarded tvs so that's OK. Have been learning as I go along.

I have found that 4 x 75V zener diode snubber is insufficient to dissipate the kick back when driving a flyback from a tv with 24V at about 20kHz - I smoked the zener diodes. Hence I calculated and added an RCD snubber. I have tested this briefly but yet to give it a longer test.

I also killed a 1.5A 12V voltage regulator possibly because the mosfet gate driver is drawing more current than it is rated for. This happened as soon as I switched flyback transformers to one that ran at 20KHz rather than perhaps 7kHz. I have since modified my circuit to use an LM317 supplying just over 13V and with a pass transistor to supply up to 10A - the mosfet gate driver can supply up to 9A.

The mosfet I have been using are rated at 500-600V and around 15A.

The Zeners , four in series, would require parallel resistors to divide the voltage evenly. Their response may also be slow for this application. In switching DC/DC converters, varistors are used to prevent spikes. For safety, I would also add a permanent load resistor to the transformer secondary.

Years ago I have repaired flyback DC/DC converters which needed to replace blown switching transistors (MOSFETs were then not available). I had to select one from four-five pieces. Though specified to 800 V reverse, not all did survive. You can see that mostly those power switches were dying in CRT monitors. It is a risky business.

 

[FvM]

The Zeners , four in series, would require parallel resistors to divide the voltage evenly.

Capacitors would, Z-diodes can do this on their own, fortunately.

 

[boylesg]

For safety, I would also add a permanent load resistor to the transformer secondary.

The RCD snubber seemed to work OK for the brief time I tested it. Although the small 3W resistor got rather hot. I have since replaced the single 3W resistor with about 8 x 5W resistors in series that add up to the same value.

For safety, I would also add a permanent load resistor to the transformer secondary.

What is your resoning for this? And I have doubts that I could easily or inexpensively obtain a resistor that could withstand 20kV or so. Perhaps a small permanent spark gap along the lines of a tesla coil?

 

[BradtheRad]

Originally Posted by jiripolivka
For safety, I would also add a permanent load resistor to the transformer secondary.

What is your resoning for this? And I have doubts that I could easily or inexpensively obtain a resistor that could withstand 20kV or so. Perhaps a small permanent spark gap along the lines of a tesla coil?

A load on the secondary will absorb energy created by the collapsing flux field. The high-V spikes in the primary will have less amplitude.

It will also greatly reduce emf in the secondary, so the resistor need not be a high-V type. (If you do need them, there are individuals who can fill small orders and have their own websites.)

Because of reduced emf in the secondary you may not get any sparks. Try different ohm values.

I have played around with various snubbers in the javascript circuit simulator

If this is the Falstad Javascript simulator, then it has a Tesla coil in the menu of circuits. (Misc Devices >> Spark Gap >> Tesla Coil)

I realize it's not the same as a jacob's ladder but it may provide some insights into how both operate. (I recommend reducing the breakdown voltage so you don't have to wait so long between arcs.)