Sorry C,
I was talking about urb nerd's circuit.
I can't see any reason why it shouldn't work unless it's the grounds or layout.
But I like your floating supply.

 

At 20kHz PWM frequency, the gate current requirement is around 4.2mA, which is nothing for LM2596. The peak loads can also be supplied by caps. LM2596 also has switch current protection in-built, apart from thermal protection. That is why I was looking at a different mechanism that made it fail.
With zero input to the opto, output bottom mosfet is on, but not the top one. So, Pins 5, 6 & 7 are shorted, but not pin no.8. So, there is literally no load on the LM2596. LM2596 output should have the +10V, irrespective of the duty cycle.
I was thinking on the lines of the input votlage being higher to the converter (Absolute max is 45V) as the transformer output voltage could have been higher with no load (load regulation) and the converter was in the process of destruction right on power on.

@ronv:
The minus pins of the LM2596 boards appear to be tied together, but part of the track is used as a shunt resistance (reduce cost) in CC-CV regulators.

 

I would recommend checking out the input voltage to the LM2596, even before connecting it. If the voltage is found within limits, then connect the converter and check its output voltage. With zero duty, FOD3180 pin 8 should be checked to be 10V (LM2596 output) before proceeding.

 

Would it be correct to assume both minus pins on the regulator are tied together?
How big is the cap on the input to the regulator.
High frequency switchers like this can be fussy. It may not like a breadboard.

For my circuit, and shematic above. The negative terminals are indeed connected directly.

@ cmartinez, i misread the post above - didn't actually realise the shematic you linked was a more expansive one that i had seen.
Thanks for showing that.

 

To try it you could use just a 12 volt wall wort. That would tell you if something was wrong with the 2596 circuit.
What FET are you using?
Still 150 watt bulb?

 

To try it you could use just a 12 volt wall wort. That would tell you if something was wrong with the 2596 circuit.
What FET are you using?
Still 150 watt bulb?

still the same bulb, the FET is an IRp4110pbf.
A battery array allows me to control the get at high speeds, but i am not really looking to use batteries.
Got to be the way i used the converter circuit, i got an alternative in the post today.
It even came with a novelty heatsink that looks so insufficient it might even intrinsically radiate heat haha.

 

At 20kHz PWM frequency, the gate current requirement is around 4.2mA, which is nothing for LM2596. The peak loads can also be supplied by caps. LM2596 also has switch current protection in-built, apart from thermal protection. That is why I was looking at a different mechanism that made it fail.
With zero input to the opto, output bottom mosfet is on, but not the top one. So, Pins 5, 6 & 7 are shorted, but not pin no.8. So, there is literally no load on the LM2596. LM2596 output should have the +10V, irrespective of the duty cycle.
I was thinking on the lines of the input votlage being higher to the converter (Absolute max is 45V) as the transformer output voltage could have been higher with no load (load regulation) and the converter was in the process of destruction right on power on.

@ronv:
The minus pins of the LM2596 boards appear to be tied together, but part of the track is used as a shunt resistance (reduce cost) in CC-CV regulators.

i dun goofed describing the gate drive opto operation, it indeed makes more sense that pin 8 is not pulled low - not sure how i got that impression now i think about it.

Its looking like im chalking this failure up to the observer effect.
The wave function collapsed and the converter failed, as it is expected to given my track record. Only when i don't observe do things go to plan

 

i dun goofed describing the gate drive opto operation, it indeed makes more sense that pin 8 is not pulled low - not sure how i got that impression now i think about it.

Its looking like im chalking this failure up to the observer effect.
The wave function collapsed and the converter failed, as it is expected to given my track record. Only when i don't observe do things go to plan

I believe the technical term for what you've described is "you're jinxed"...

 

[QUOTE="urb-nurd, post: 814952, member: 243253]

A battery array allows me to control the get at high speeds, but i am not really looking to use batteries.

So just replace the batteries with a $5 12 volt wall wort.

 

[QUOTE="urb-nurd, post: 814952, member: 243253]



So just replace the batteries with a $5 12 volt wall wort.

i should of phrased that better - i am trying to power this all from one supply.
If all else fails and my deadline looms, i will definitely be using a separate dc source

 

i dun goofed describing the gate drive opto operation, it indeed makes more sense that pin 8 is not pulled low - not sure how i got that impression now i think about it.

Its looking like im chalking this failure up to the observer effect.
The wave function collapsed and the converter failed, as it is expected to given my track record. Only when i don't observe do things go to plan

Well, it happens to everyone!
The reason for the bulb turning on with no voltage on pin 8 of the opto, cannot be explained still.
The MOSFET switching could be slowed down with a gate resistance (say 100 ohms) which would make it less susceptible to layout issues. This resistance will increase the power dissipiation which can be looked at later. Another resistor of about 10k between gate source, close to the MOSFET. Opto to MOSFET to be as close as possible, preferably with the gate, source wires to opto directly with a twisted pair.

 

Well, it happens to everyone!
The reason for the bulb turning on with no voltage on pin 8 of the opto, cannot be explained still.
The MOSFET switching could be slowed down with a gate resistance (say 100 ohms) which would make it less susceptible to layout issues. This resistance will increase the power dissipiation which can be looked at later. Another resistor of about 10k between gate source, close to the MOSFET. Opto to MOSFET to be as close as possible, preferably with the gate, source wires to opto directly with a twisted pair.

I just got another 5 converters in the post - ill get back with my findings this evening!

 

Hi,

Please show the entire schematic of what you are doing.
Thanks

 

Hi,

Please show the entire schematic of what you are doing.
Thanks

View attachment 79862

The only thing missing from that is my the resistor on the PWM output

 

If your new parts don't fix it, here is a budget power supply for it.

 

If your new parts don't fix it, here is a budget power supply for it.

Thats greatly appreciated ron!
Thanks alot!

 

Turns out, the gate drive optocoupler failed also!
Meaning the output current surpassed its 2.5A peak rating (which is likely given i forgot the gate series resistance!), either that or the output of the DC-DC converter (the supply for the opto) spiked above the 25V limit of the gate drive opto.

Whatever caused the issues resulted in both failing.

 

I pulled out a gate driver IC i had laying around but the datasheet has me a little confused.
I am not quite sure how to hook it up from the descriptions in the datasheet.
The terminals: VC , OUT and PWR GND seem like collector voltage, output and Power ground to me.

But i am not sure the function of the VS Pin, could it be for connection to the source of the fet?

Seems like VC will be the high potential for the output with PWR GND being the low output potential.

My pwm output looks to connect to the non inverting input, and i guess the return path for this PWM will be the logic ground.

I am not too sure if i need an opto to isolate the PWM also.
as always, assistance is greatly appreciated.
Thanks AAC!

 

Vs and logic ground are meant to go together as like the micro supply, while Vcc and power ground are the high voltage supply.
But having said that, this is not a good chip for you as it does not have under voltage lockout. What that means is the output will follow the pwm no matter what Vcc is. Since in your case Vcc is a nice slow power supply, that mean the gate voltage to your FET might be 4, 5, or six volts while the current in the lamp will still be quite high. Since the FET needs 7 or 8 volts to turn on "hard" you may have trouble with it smoking on power on.
As to the opto. No it is not necessary since you have a transformer for isolation. But at least until you get it working it's not a bad idea since it sounds like it was the only thing that stood between your micro and something bad.

 

Vs and logic ground are meant to go together as like the micro supply, while Vcc and power ground are the high voltage supply.
But having said that, this is not a good chip for you as it does not have under voltage lockout. What that means is the output will follow the pwm no matter what Vcc is. Since in your case Vcc is a nice slow power supply, that mean the gate voltage to your FET might be 4, 5, or six volts while the current in the lamp will still be quite high. Since the FET needs 7 or 8 volts to turn on "hard" you may have trouble with it smoking on power on.
As to the opto. No it is not necessary since you have a transformer for isolation. But at least until you get it working it's not a bad idea since it sounds like it was the only thing that stood between your micro and something bad.

Thanks ron, i think ill wait till another fod3180 comes then, i must have fried £20 worth of Fets already!
If i do make an attempt at using the gate driver ic (to give me the illusion of progress in my project, brains are strange like that), i will be using 3 18350's for the dc power, hopefully that would be more reliable than my non isolated converter.