Rebuilt Mazilli ZVS flyback driver
Mads Barnkob, Sat May 16 2009, 04:01PMWhile cleaning out various boxes in the lab, I came across some of my first high voltage projects, among them, a mazilli driver built on a veroboard, which resulted in quite a few blown traces due to the high currents :)
I decided to rebuild it in the same style as Myke do most of his small projects, but since I had to reuse the old components it did not turn out as good :)
I tried with 4 different flybacks, a standard TV flyback, portable tv flyback, photocopier flyback and a very old disc flyback with rectifier tube, the photocopier and rectifier tube did not surviwe.
Read the rest and see the HD film of it all here:
Re: Rebuilt Mazilli ZVS flyback driver
Arcstarter, Sat May 16 2009, 08:03PM
Looks very nice :D.
I cannot tell, but about how many turns is your inductor? It should be around 50-100 or something like that, probably much more. Another member suggested around 100-200 is correct (it was '...' iirc). He actually put an inductor in the migratory junk box he said had about half enough turns i think, and i took it and that is what i used in the ZVS. The first mosfets on my ZVS where from the junk box, the UF4007's where from there, and i used Ultra7's large green 100 watt 1kohm resistor, tapped in the middle for 2 500ohm 'ressitors'. It heats quite a bit despite the very large wattage rating. The energy you do not lose in heat on the mosfets is lost in the resistor XD. At least at 700 watts -.-
Arcstarter, Sat May 16 2009, 08:03PM
Looks very nice :D.
I cannot tell, but about how many turns is your inductor? It should be around 50-100 or something like that, probably much more. Another member suggested around 100-200 is correct (it was '...' iirc). He actually put an inductor in the migratory junk box he said had about half enough turns i think, and i took it and that is what i used in the ZVS. The first mosfets on my ZVS where from the junk box, the UF4007's where from there, and i used Ultra7's large green 100 watt 1kohm resistor, tapped in the middle for 2 500ohm 'ressitors'. It heats quite a bit despite the very large wattage rating. The energy you do not lose in heat on the mosfets is lost in the resistor XD. At least at 700 watts -.-
Re: Rebuilt Mazilli ZVS flyback driver
Mads Barnkob, Sat May 16 2009, 08:15PM
I got no heat losses in resistors or any other part than only a little on the mosfets.
the inductor is 12 windings of 12 or 14AWG, measured 170uH
Mads Barnkob, Sat May 16 2009, 08:15PM
Arcstarter wrote ...
Looks very nice :D.
I cannot tell, but about how many turns is your inductor? It should be around 50-100 or something like that, probably much more. Another member suggested around 100-200 is correct (it was '...' iirc). He actually put an inductor in the migratory junk box he said had about half enough turns i think, and i took it and that is what i used in the ZVS. The first mosfets on my ZVS where from the junk box, the UF4007's where from there, and i used Ultra7's large green 100 watt 1kohm resistor, tapped in the middle for 2 500ohm 'ressitors'. It heats quite a bit despite the very large wattage rating. The energy you do not lose in heat on the mosfets is lost in the resistor XD. At least at 700 watts -.-
Looks very nice :D.
I cannot tell, but about how many turns is your inductor? It should be around 50-100 or something like that, probably much more. Another member suggested around 100-200 is correct (it was '...' iirc). He actually put an inductor in the migratory junk box he said had about half enough turns i think, and i took it and that is what i used in the ZVS. The first mosfets on my ZVS where from the junk box, the UF4007's where from there, and i used Ultra7's large green 100 watt 1kohm resistor, tapped in the middle for 2 500ohm 'ressitors'. It heats quite a bit despite the very large wattage rating. The energy you do not lose in heat on the mosfets is lost in the resistor XD. At least at 700 watts -.-
I got no heat losses in resistors or any other part than only a little on the mosfets.
the inductor is 12 windings of 12 or 14AWG, measured 170uH
Re: Rebuilt Mazilli ZVS flyback driver
..., Sat May 16 2009, 09:30PM
It depends on the core, for the yellow/white powdered iron cores in an ATX psu you need 100-200 turns (50 would be pushing it depending on the core), but it looks like he is using a ferrite core in this project which needs only 10-20 turns.
It should be noted that the circuit will operate differently with the ferrite core than a powdered iron one, because the ferrite is going to saturate quite quickly--although it has been shown time and time again that this does not seem to hurt the output very much. If anything the only result is more heating in the fets (I generally see the fets mounted in gigantic heatsinks with using a ferrite core, when using a powdered iron core with enough turns I run 500w for hours using heatsinks about 1/10 the size of the one shown in the first post)
Looks great overall! I might suggest doing something to stabilize the output connector a little, it sucks to dig a project out of a box and realize that you snapped off one of the pins on your fets
..., Sat May 16 2009, 09:30PM
It depends on the core, for the yellow/white powdered iron cores in an ATX psu you need 100-200 turns (50 would be pushing it depending on the core), but it looks like he is using a ferrite core in this project which needs only 10-20 turns.
It should be noted that the circuit will operate differently with the ferrite core than a powdered iron one, because the ferrite is going to saturate quite quickly--although it has been shown time and time again that this does not seem to hurt the output very much. If anything the only result is more heating in the fets (I generally see the fets mounted in gigantic heatsinks with using a ferrite core, when using a powdered iron core with enough turns I run 500w for hours using heatsinks about 1/10 the size of the one shown in the first post)
Looks great overall! I might suggest doing something to stabilize the output connector a little, it sucks to dig a project out of a box and realize that you snapped off one of the pins on your fets
Re: Rebuilt Mazilli ZVS flyback driver
Myke, Sat May 16 2009, 10:45PM
It looks really nice.
I like how you ran the wires that go to the flyback between the two heatsinks.
With my design, it can be kinda annoying to remove the MOSFETs because they are soldered in. I made the diode/resistor thing have little feet at the end of the leads so that it can just sit on top of the MOSFET's leads.
Myke, Sat May 16 2009, 10:45PM
It looks really nice.
I like how you ran the wires that go to the flyback between the two heatsinks.With my design, it can be kinda annoying to remove the MOSFETs because they are soldered in. I made the diode/resistor thing have little feet at the end of the leads so that it can just sit on top of the MOSFET's leads.
Re: Rebuilt Mazilli ZVS flyback driver
Arcstarter, Sat May 16 2009, 11:12PM
All my components laid down, to give just enough clearance for the lid on the box i made for it. I used terminal blocks, which i would highly recommend. If you lived in USA i would ship you a few no charge -.-.
On the two drain wires, i cut some insulation off a few cm's down, and soldered the capacitor to it right there. This means short leads to the capacitor.
Arcstarter, Sat May 16 2009, 11:12PM
All my components laid down, to give just enough clearance for the lid on the box i made for it. I used terminal blocks, which i would highly recommend. If you lived in USA i would ship you a few no charge -.-.
On the two drain wires, i cut some insulation off a few cm's down, and soldered the capacitor to it right there. This means short leads to the capacitor.
Re: Rebuilt Mazilli ZVS flyback driver
Dr. Dark Current, Sun May 17 2009, 07:13AM
Very nice.
The reasons I don't like this circuit much are the arc screaming before going out and the flyback core getting smoking hot. But it has its place, mainly because of its simplicity-to-power ratio
Anyway, I should mention this
P.S. I have some experiments with my super simple self-oscillating bridge going on, so be prepared for another simple and powerful driver unfortunately I don't have much time now, but when I get to it, I'll surely post the results.
Dr. Dark Current, Sun May 17 2009, 07:13AM
Very nice.
The reasons I don't like this circuit much are the arc screaming before going out and the flyback core getting smoking hot. But it has its place, mainly because of its simplicity-to-power ratio
Anyway, I should mention this
P.S. I have some experiments with my super simple self-oscillating bridge going on, so be prepared for another simple and powerful driver
unfortunately I don't have much time now, but when I get to it, I'll surely post the results.Re: Rebuilt Mazilli ZVS flyback driver
..., Sun May 17 2009, 08:03AM
If your core gets hot it means your are saturating, and you need to use more turns on the core (or possibly add more of an airgap). The screaming can be solved by using a smaller tank capacitor.
The mazzilli driver does have its drawbacks, but the two you mention are pretty easily solvable.
..., Sun May 17 2009, 08:03AM
If your core gets hot it means your are saturating, and you need to use more turns on the core (or possibly add more of an airgap). The screaming can be solved by using a smaller tank capacitor.
The mazzilli driver does have its drawbacks, but the two you mention are pretty easily solvable.
Re: Rebuilt Mazilli ZVS flyback driver
Dr. Dark Current, Sun May 17 2009, 08:47AM
Well, this circuit from design needs huge amount of reactive power circulating between the tank cap and the core, I think someone (Sulaiman?) said the Q must be at least 5, so if you're outputting 500W, the core sees 2.5kVA (AT LEAST!)
Airgap seems to solve the screaming but makes the core even hotter, because it oscillates on a higher frequency so the hysteresis losses are higher.
Dr. Dark Current, Sun May 17 2009, 08:47AM
... wrote ...
If your core gets hot it means your are saturating, and you need to use more turns on the core (or possibly add more of an airgap). The screaming can be solved by using a smaller tank capacitor.
If your core gets hot it means your are saturating, and you need to use more turns on the core (or possibly add more of an airgap). The screaming can be solved by using a smaller tank capacitor.
Well, this circuit from design needs huge amount of reactive power circulating between the tank cap and the core, I think someone (Sulaiman?) said the Q must be at least 5, so if you're outputting 500W, the core sees 2.5kVA (AT LEAST!)
Airgap seems to solve the screaming but makes the core even hotter, because it oscillates on a higher frequency so the hysteresis losses are higher.
Re: Rebuilt Mazilli ZVS flyback driver
Arcstarter, Sun May 17 2009, 07:16PM
Well, with none of the 15 or so flybacks i have killed have i had screaming arcs, or core heating. I did with one absolutely amazing flyback, with about 50 volts idle and 45 volts under load. It pulled almost 25 amps, was about 1000watts and VERY high voltage. It was DC.
My homemade HV transformer never had this problem either, but the screaming and heating has to do with the characteristics of the core and primary inductance and cap, primarily. (Unless the secondary also plays a major role?)
By the way, i am looking forward to your driver! For like a year now i have considered you the king of flybacks. I did make a multipurpose inverter that i drove a flyback with, using a stupid 555, thanks to the idea i got from you on youtube
.
Arcstarter, Sun May 17 2009, 07:16PM
Dr. Kilovolt wrote ...
Very nice.
The reasons I don't like this circuit much are the arc screaming before going out and the flyback core getting smoking hot. But it has its place, mainly because of its simplicity-to-power ratio
Anyway, I should mention this
P.S. I have some experiments with my super simple self-oscillating bridge going on, so be prepared for another simple and powerful driver unfortunately I don't have much time now, but when I get to it, I'll surely post the results.
Hmm, to me it seems this thread might result in good info :P.Very nice.
The reasons I don't like this circuit much are the arc screaming before going out and the flyback core getting smoking hot. But it has its place, mainly because of its simplicity-to-power ratio
Anyway, I should mention this
P.S. I have some experiments with my super simple self-oscillating bridge going on, so be prepared for another simple and powerful driver
unfortunately I don't have much time now, but when I get to it, I'll surely post the results.Well, with none of the 15 or so flybacks i have killed have i had screaming arcs, or core heating. I did with one absolutely amazing flyback, with about 50 volts idle and 45 volts under load. It pulled almost 25 amps, was about 1000watts and VERY high voltage. It was DC.
My homemade HV transformer never had this problem either, but the screaming and heating has to do with the characteristics of the core and primary inductance and cap, primarily. (Unless the secondary also plays a major role?)
By the way, i am looking forward to your driver! For like a year now i have considered you the king of flybacks. I did make a multipurpose inverter that i drove a flyback with, using a stupid 555, thanks to the idea i got from you on youtube
.Print this page