555 and flyback not the same frequency?

Thread Starter

magnet18

Joined Dec 22, 2010
1,227
Hi everyone,
I'm currently attempting to build a plasma speaker using a 555 timer, an IRFS654 MOSFET, a lead acid battery, and a flyback; from what I gather it's not an uncommon procedure.
My problem is that the frequency coming out of the flyback doesn't seem to match the 555's frequency. I scoped the 555 pin 3 as being 25 kHz, which is well beyond audible, but when I attempt to create an arc it makes a nasty tone approximately that of nails on a chalkboard, and the plasma arc is erratic and jumps about. Also, as I change the distance of the spark gap, the frequency changes audibly, growing lower as the distance is increased and higher as the distance is decreased. I have tried this with both 20 and 10 turns on the primary and got the same result, with the 10 turn's providing more voltage, but the same effect.
Does anyone know where I might be going wrong?

I can post a pic/video of my setup if anyone thinks it will help.

Thanks in advance :)
 

debe

Joined Sep 21, 2010
1,389
If the transformer has built in rectifiers for the EHT then the output wont be an AC arc it will be a DC with HF ripple. The transformers they use generaly dont have the rectifier components in the HV side. This maybe the cause of the problem. Perhaps a picture of the T/F you are using?
 

Thread Starter

magnet18

Joined Dec 22, 2010
1,227
@Bill, yes I do, I'll scope the voltage, control, and reset pins and see what the noise is.

@debe, I heard about that somewhere, I attached the primary in both directions, one direction gave me a very weak arc that was highly sporadic, the other direction gave me the much more powerful, less erratic arc I am currently dealing with. I'll post a pic also.

Also, if this is a noise issue, I happen to have three 10,000uF electrolytic caps laying around, would strategically placing these help with this?
 

SgtWookie

Joined Jul 17, 2007
22,230
It sounds to me like a single pulse from your 555 circuit is not enough to build enough voltage on the secondary side to jump the gap - and that there is an HV diode in the secondary like debe mentioned.

Once the voltage on the secondary does build high enough to ionize the air and jump the gap, the HV rapidly bleeds off until the plasma arc stops; whereupon the voltage has to build up again to bridge the gap.

You might be able to build the secondary voltage faster to jump the gap if you (not necessarily in order):
1) Decrease the size of the spark gap.
2) Increase the wire gauge of the primary side winding.
3) Decreased the number of primary side windings.
4) Use a better MOSFET (lower Rds(on))
5) Decrease the length of the wiring on the primary side.
6) Add capacitor(s) on the V+ supply side of the coil primary (assuming that you're switching the ground side).

It takes 3 million volts to jump a 1 meter air-filled gap at sea level.
Scaling that down, it takes 30kv to jump an 0.39" gap, or ~19.05kv to jump a 1/4" gap.
 

CDRIVE

Joined Jul 1, 2008
2,219
Did you do a good job of decoupling the 555 supply bus? If the HV spikes find their way back to the 555 it will surely time erratically. The CV pin shouldn't be left floating either.
 

SgtWookie

Joined Jul 17, 2007
22,230
Well, it also seems to me like the wires on the board were kind of long. Square waves imply lots of bandwidth needed. Long wires have lots of inductance. Caps with very short leads between supply and ground help a great deal to reduce those parasitic effects.
 

Thread Starter

magnet18

Joined Dec 22, 2010
1,227
It sounds to me like a single pulse from your 555 circuit is not enough to build enough voltage on the secondary side to jump the gap - and that there is an HV diode in the secondary like debe mentioned.

Once the voltage on the secondary does build high enough to ionize the air and jump the gap, the HV rapidly bleeds off until the plasma arc stops; whereupon the voltage has to build up again to bridge the gap.

You might be able to build the secondary voltage faster to jump the gap if you (not necessarily in order):
1) Decrease the size of the spark gap.
2) Increase the wire gauge of the primary side winding.
3) Decreased the number of primary side windings.
4) Use a better MOSFET (lower Rds(on))
5) Decrease the length of the wiring on the primary side.
6) Add capacitor(s) on the V+ supply side of the coil primary (assuming that you're switching the ground side).

It takes 3 million volts to jump a 1 meter air-filled gap at sea level.
Scaling that down, it takes 30kv to jump an 0.39" gap, or ~19.05kv to jump a 1/4" gap.
Thanks very much, in addition, would wiring 2 12V batteries in series or parallel and powering the arc as such help? I will try the things you suggested and see how they help.

@CDrive, I think that may be part of the problem, I'm adding a 1000uF decoupler, that should do the trick ;)

Is there any value for a decoupler that is too big and would have a negative effect?
 

SgtWookie

Joined Jul 17, 2007
22,230
The ESR (equivalent series resistance) of capacitors usually increases as the value of capacitance goes up. That's why caps are often paralleled. You might see two or three in parallel, like the 0.1uF and 10uF caps on the output of a voltage regulator, or 0.1uF and 1uF caps on a 555's power pins.
 

Thread Starter

magnet18

Joined Dec 22, 2010
1,227
Alright, I've attached some pics of my setup and some scope captures, the first one is pin 8, the second is pin 5 (that appears to be a problem), the third and fourth are pics of my physical setup. (the LARGE electrolytic you see isn't connected)

I'm about to switch from 20 to 18 ga wire on the primary and use 5 turns (same as these were taken with)

Wookie, you mentioned a cap to the primary high side, would that just be from the primary high to ground?
 

Attachments

Thread Starter

magnet18

Joined Dec 22, 2010
1,227
Alright, I applied 24V to the primary, and it "seemed" to work ultrasonically, its torture on the MOSFET though, I may need a bucket of icewater to cool the thing :(
 

debe

Joined Sep 21, 2010
1,389
Looking at the T/F in your pic i suspect it will be internaly wired with HV diodes to give 25KV DC which is very common in TVs per Circuit shown. It was only in very old TVs that the T/F put out AC & was rectified external to the T/F.
 

Attachments

CDRIVE

Joined Jul 1, 2008
2,219
@CDrive, I think that may be part of the problem, I'm adding a 1000uF decoupler, that should do the trick ;)
Sgt.Wookie pretty much answered this but even using .1uF decoupling caps aren't going to help much when using a Protoboard. The protoboards exhibit far to much inductive properties and capacitive coupling to parallel conductors for this type of circuit.
 

Thread Starter

magnet18

Joined Dec 22, 2010
1,227
@debe, Thanks, I expected that, but it's nice to have conformation.
Why do I get the feeling you're a TV repairman? :)
Fun fact, the electric field I'm making is strong enough to cause my radio to go all white noise, even though its 15' away and playing my computer's music.
 

SgtWookie

Joined Jul 17, 2007
22,230
Ahhh, wait a minute - in that pin 5 O-scope plot, I'm seeing 4 pulses per 2mS, which is 0.5mS per pulse, or 2kHz in frequency. I thought you said that the output was 25kHz?

What are you actually seeing on pin 3? Is that 25kHz?
 

Thread Starter

magnet18

Joined Dec 22, 2010
1,227
Sgt.Wookie pretty much answered this but even using .1uF decoupling caps aren't going to help much when using a Protoboard. The protoboards exhibit far to much inductive properties and capacitive coupling to parallel conductors for this type of circuit.
Thanks, I'll build it on a perfboard, hopefully that improves performance.
I think I'm reverting to my original idea for the project, and seeing the maximum distance at which it will arc at different resistor settings (different frequencies/duty cycles)
 

Thread Starter

magnet18

Joined Dec 22, 2010
1,227
Ahhh, wait a minute - in that pin 5 O-scope plot, I'm seeing 4 pulses per 2mS, which is 0.5mS per pulse, or 2kHz in frequency. I thought you said that the output was 25kHz?

What are you actually seeing on pin 3? Is that 25kHz?
it was 25 when unconnected, lemme check when its attached to the flyback real fast...
 

Thread Starter

magnet18

Joined Dec 22, 2010
1,227
OH, I was wrong, here is what the scope actually shows on pin 3
Also, when I removed the flyback and put hot glue over the other pins on the bottom that it was arcing to, it stopped being ultrasonic like it was in post 11 and resumed its old behavior. That is what the samples are showing.
 

Attachments

SgtWookie

Joined Jul 17, 2007
22,230
Looks like your 'scope probe doesn't have a good ground? The signal appears to be going negative, and it shouldn't do that.

[eta]
I count ~22.5 divisions out of 25; so that's 0.9*200uS = 0.00018 = 5.555kHz.

The frequency's being counted from the "T" marker to some unknown point - I guess you have x5 or x10 enabled?

It could be that your 555 ground terminal or MOSFET source terminal ground path isn't very good?
 
Last edited:
Top