why 555 based flyback drivers explode 555 ICs

[attaatta]

Hi all
i built this attaached circuit and it worked then i started losing 555 Ics.Can some one help me what to add or change?I also replaced MOSFET with D1047 transitor but the problem of heating 555 and ultimately malfunctioning is still on.Need help. I am submitting it with new thread anticipating many will benefit.Thanks in advance

 

[panic mode]

you don't have any suppression so the transients are bound to kill something.
try putting flyback diode across primary for example

 

[Jony130]

By adding flyback diode to the circuit you're going to kill all the high voltage on the secondary side .

 

[crutschow]

You should add a zener diode from the transistor drain to source to clamp the transient voltage from the transformer inductance. As Jony130 noted, adding a normal diode will kill the output voltage of the flyback. The zener voltage should be no more than about 75% of the maximum transistor Vds rating.

Also add a 0.1μF ceramic cap directly across the 555 power to ground pins.

 

[WTP Pepper]

As the inductor switches off, the voltage on the FET drain will rise (that's physics and how a coil works). As you have no transient suppression in the form of a catch diode, it will also lift the supply of the 555 to what could be several hundred volts depending on the primary inductance and blow it.
Much as already said by others.

 

[johansen]

As the inductor switches off, the voltage on the FET drain will rise (that's physics and how a coil works). As you have no transient suppression in the form of a catch diode, it will also lift the supply of the 555 to what could be several hundred volts depending on the primary inductance and blow it.
Much as already said by others.

well, not quite. the mosfet will go into avalanche and clamp the transformer waveform to whatever voltage or whatever, depending on the wattage this is actually acceptable, and its how most car fuel injection solenoids work.

however, it is the supply side inductance that blows up the 555, not the coil inductance. you need a low inductance capacitor across the supply. 0.1uF ceramic, etc. but you may need more than just .1uF depending on the actual supply side inductance. 100 uF per average amp consumed by the flyback is a good start.

 

[crutschow]

As the inductor switches off, the voltage on the FET drain will rise (that's physics and how a coil works). As you have no transient suppression in the form of a catch diode, it will also lift the supply of the 555 to what could be several hundred volts depending on the primary inductance and blow it.

Neglecting any stray inductance or capacitance, the inductor will only generate the voltage spike at the FETs drain, it will not transfer that voltage to the supply side of the inductor and the 555. The source of spikes for that is likely stray circuit inductance as mentioned by johansen.

 

[bountyhunter]

You need some good quality (low ESR) switcher grade caps hear the FET to supply the current pulses for the converter. You also need an R-C snubber across the transformer to snub switching spikes.

What's happening is likely spikes are getting coupled to the gate and killing the 555 by banging the output below ground or above the maximum limit which I recall is only about 16V.

 

[attaatta]

Thanks all.
I will try folllowing today in order of pri and post results.
1.Capcitor accross supply
2..1uf cap accross 555 power to ground pins.
3.zener accross mosfet
4.Seperating power to oscilator and mosfet(feeding from different sources).
5.Put diodes accross primary.

 

[johansen]

you can feed the 555 from a different power supply but they both need to be on the same ground, tied together at the source of the mosfet.

 

[attaatta]

Thanks I will do that

 

[bountyhunter]

You are still going to need an R-C snubber across the transformer winding to reduce the voltage spike caused by transformer leakage inductance. FETs have internal capacitance from drain to gate which couples the spike and kills your 555. Put a scope on the circuit while it's running and see what I mean.

 

[WTP Pepper]

well, not quite. the mosfet will go into avalanche and clamp the transformer waveform to whatever voltage or whatever, depending on the wattage this is actually acceptable, and its how most car fuel injection solenoids work.

however, it is the supply side inductance that blows up the 555, not the coil inductance. you need a low inductance capacitor across the supply. 0.1uF ceramic, etc. but you may need more than just .1uF depending on the actual supply side inductance. 100 uF per average amp consumed by the flyback is a good start.

Well quite. The MOSFET will not clamp the waveform to "whatever voltage or what ever". The voltage across the coil will eventually collapse having lifted the supply to what could be several hundred volts.

I think you are confused to what an avalanched breakdown is and more confused about low inductance capacitors.

 

[johansen]

lol, i am quite confident that an irf510 breaks down at about 100-110 volts and that the supply rails are not affected at all, except by the supply side inductance destroying the 555

here's an example of what the supply side voltage would look like At The 555. View attachment 65113
the lower trace is dc coupled and it was 10 volts per div. the mosfet was a irfz48.
the reason it is clamped is because the mosfet and the flyback diode conduct in the reverse direction, the mosfet breaks down and conducts in the other, but only for the first microsecond or however long it takes. The remaining decaying oscillation is the supply side inductance resonating with the local capacitance on the board.
Now i should make clear that the second portion of the decaying sine wave is what the voltage looks like across the 555, when there is a local capacitor connected across it.
the entirely of the waveform is actually the source to drain voltage.

top waveform is the gate source voltage.

since he originally had no cap on the 555, it would have been destroyed instantly upon turn off. the voltage on the dc bus at the 555 would rise to whatever the mosfet avalanches at plus the dc voltage drop through the transformer coils. (assuming the capacitance of the mosfet is negligible) but after the energy was dumped into the mosfet the supply would ring down just as it does in the oscope shot.

had a low inductance 1uF cap and a few hundred uF electrolytic to back it up, only the mosfet would be affected.


View attachment 65114

here is what his actual schematic looks like.
with two modifications, the transformer replaces R
and a virtual 555 would be connected across the capacitor. (or more accurately, lack of a capacitor)

 

[Jaguarjoe]

In lieu of a conventional diode across the primary, you can use a zener across the coil rated at some voltage lower than the transistor breakdown voltage. As was said, the lower the zener voltage the lesser the spark. You can avoid that by using high voltage coil driver transistors which can eat 450 volts or maybe more.

Another thing you can do is add a small sacrificial' driver transistor between the 555 and the power transistor. Many auto makers did this to protect the costly ECU from thunderbolts. A 2n3904/06 is cheaper than a 555.

 

[attaatta]

Thank u all for nice suggestions.
I changed circuit as per your suggestions .I have successfully operated the circuit without changing any 555 today.I went like this:-
1. Used 4700 uf capacitor accorss 555 power to ground---------555 stabilized.
2.Adjusted my biasing resistors as they were malfunctioning due to some previous overheating.
3.The circuit went well in oscilator section but output voltage accross flyback was too low.
4.I reduced 470 ohm resistor to give mosfet more drive.But I lost my mosfet
5.I noticed that if 470 ohm resistor is reduced 555 also started overheating.
6.I have replaced mosfet but the output voltage is far from desired.
7.I now have used zener accross MOSFET for safety of transistor and 555.

I think my problem was biasing 555 wrongly as well as transient spikes.

I am getting very small arcs accross flyback secondry.

But now how to increase the output???

Thanks all

 

[rumpfy]

Hi atat,
Most of the answers seem to be saying the same thing that the circuit arrangement is not satisfactory.
Firstly, the drive for the mosfet is not well defined for negative voltage spikes at the gate. There is no impedance to ground except for what may be inside the 555.
Secondly, there is no apparent arrangement to detect saturation of the coil at the drain.
Thirdly, the 555 is not the preferred type for this kind of arrangement.
I havent got the data sheet for the 555 so I cant say what its output circuit is like but the lack of control of the gate drive is critical, to my mind.
There is a device UC3842. It is a current mode controller. Get the data sheet from ONSEMI and have a look at Fig 18 and see the drive arrangement for the gate drive. Bit different to the 555 I'll bet.
Have a look at the way the 3842 switches at a certain coil current This type of control will prevent the inductance in the drain circuit from saturating. Your arrangement with a couple of pots to set this and that, is NOT the way its done if you want a reliable system. If you do not prevent the coil from saturating, the mosfet current will suddenly increase to destructive levels. I note you want more output and are blowing up mosfets now. post #16. I'm not surprised.
The choice of mosfet is dependent on the peak voltage developed at the drain at switch off. The selected device will be OK if it has sufficient withstand voltage rating at the current level required at switch off.
You seem to be guessing this thing, and the need to understand you transformer characteristics is paramount for you. To do more development you will need an oscilloscope. By observing the source current at the current sensing resistor, you will be able to see aturation of the coil inductance as the current increases. A longer ON time will give a greater current in the current sensing resistor. When the slope of current increases from that at the start, you can be sure saturation of the coil is happening.
Hope this helps.

The suggestions to use zener diodes is not really useful because the speed of zeners is too slow. They will not protect mosfets because the mosfet swiching speeds are far too high for a zener to follow. Zeners will not protect the mosfet for coil saturation either.
If you want to vary the output voltage from your machine, then use the feedback voltage pin on the UC3824.
It's all there. The datasheet is 18 pages of the most useful stuff you'll ever want.

 

[attaatta]

Hi atat,
Most of the answers seem to be saying the same thing that the circuit arrangement is not satisfactory.
Firstly, the drive for the mosfet is not well defined for negative voltage spikes at the gate. There is no impedance to ground except for what may be inside the 555.
Secondly, there is no apparent arrangement to detect saturation of the coil at the drain.
Thirdly, the 555 is not the preferred type for this kind of arrangement.
I havent got the data sheet for the 555 so I cant say what its output circuit is like but the lack of control of the gate drive is critical, to my mind.
There is a device UC3842. It is a current mode controller. Get the data sheet from ONSEMI and have a look at Fig 18 and see the drive arrangement for the gate drive. Bit different to the 555 I'll bet.
Have a look at the way the 3842 switches at a certain coil current This type of control will prevent the inductance in the drain circuit from saturating. Your arrangement with a couple of pots to set this and that, is NOT the way its done if you want a reliable system. If you do not prevent the coil from saturating, the mosfet current will suddenly increase to destructive levels. I note you want more output and are blowing up mosfets now. post #16. I'm not surprised.
The choice of mosfet is dependent on the peak voltage developed at the drain at switch off. The selected device will be OK if it has sufficient withstand voltage rating at the current level required at switch off.
You seem to be guessing this thing, and the need to understand you transformer characteristics is paramount for you. To do more development you will need an oscilloscope. By observing the source current at the current sensing resistor, you will be able to see aturation of the coil inductance as the current increases. A longer ON time will give a greater current in the current sensing resistor. When the slope of current increases from that at the start, you can be sure saturation of the coil is happening.
Hope this helps.

The suggestions to use zener diodes is not really useful because the speed of zeners is too slow. They will not protect mosfets because the mosfet swiching speeds are far too high for a zener to follow. Zeners will not protect the mosfet for coil saturation either.
If you want to vary the output voltage from your machine, then use the feedback voltage pin on the UC3824.
It's all there. The datasheet is 18 pages of the most useful stuff you'll ever want.

Thank you rumpfy

You have perfectly analysed the situation as i am still a learner of the subject.
I will seek guidance after studying and analysing UC3824 ASAP.

 

[Jaguarjoe]

In the mid 70's through the 80's, GM had a device called an "HEI" (high energy ignition). It was based on a '4-pin' HEI module that was triggered from a reluctor/star wheel in the dizzy. This could charge and dump 5 amp ignition coils all day long- 700 to 7000RPM.

They were installed in about 3 million GM cars during their lifetime.

The neat thing about this is that it can be used as a switch for things like large lamps or other coils by controlling its pretty hi-z input.

So in a nutshell, you have a complete, cheap, robust, spark generator driver for about $.50 used from a JY, or $15 new from an auto parts store. Oh yeah, it's current limited too.

You can get matching ignition 'coils' very cheap too.

 

[dr pepper]

Taking the supply leads to the flyback primary and fet source back to the supply seperate from the supply to the 555 is a good idea, and put decoupling caps on both.

To really get good insulation you could use a opto isolator to trigger the fet, and use seperate supplys for both the fet and the 555, but I dont think thats really called for.