Grens sg3525 half bridge plasma speaker / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Grens sg3525 half bridge plasma speaker

1 2 next

Move Thread

LAN_403

ZakWolf

Fri Mar 16 2012, 03:07AM

Registered Member #3114 Joined: Sat Aug 14 2010, 08:33AM
Location:
Posts: 608

Im am making a new plasma speaker and had a couple questions.

This is a great circuit and i have all ready made it, i just wanted to make a new one and fix any errors or problems that have risen since it was designed.

First, is there anything that needs to be improved?

I read something on instructables that said i could use an lm7818 for powering the IC side, will that kill the chip?

why does the electrolytic caps have to be so big? will 2200uf work ?

lastly for the "audio in" cap should it be electrolytic or nonpolerized?

sorry for all the questions

EDIT: forgot the link

Ben Solon

Fri Mar 16 2012, 03:33AM

Registered Member #3900 Joined: Thu May 19 2011, 08:28PM
Location:
Posts: 600

Well it would help to know what the circuit you are using looks like. But to answer your other questions:
18v does seem a little high for a typical ic chip. Look at the ic data sheet to find out the max vcc. If 18 is too high use a 12-15v regulator for the ic section, and a higher voltage for supplying the mosfets power.
The caps( I assume these are decoupling/bus caps for the switch/bridge section) are just an energy reserve. If you decrease the capacitance it just means you will run out of power sooner in a more demanding situation. 2200 is easily enough IMHO.
And last but not least, I find that at low audio levels(1-5v) a polarized lytic holds up well, but don't push your luck. I don't garentee that they will last long. But you don't need a huge capacitance just for dc blocking purposes. 100nF- 1uF should be fine, and ceramic caps are plentiful at this range.

ZakWolf

Fri Mar 16 2012, 03:44AM

Registered Member #3114 Joined: Sat Aug 14 2010, 08:33AM
Location:
Posts: 608

awesome, thank you Ben.

Im was going to use a 50k trim pot for a wider frequency range, that will be ok right

Ben Solon

Fri Mar 16 2012, 04:09AM

Registered Member #3900 Joined: Thu May 19 2011, 08:28PM
Location:
Posts: 600

No problem. But again, I don't know what the circuit looks like. But frequency range doesnt really matter. What you want to do is find the resonant frequency of the fly back for longer sparks. This is a fixed frequency so all the pot would do is give you a larger window for this frequency. If you had say, a 1k pot, you wouldn't have a big enough window to find this frequency. One other note though. The frequency must be over 20khz so that all you hear is the music. If you had <20khz you would get music overlaid with loud high frequency buzzing.

ZakWolf

Fri Mar 16 2012, 04:14AM

Registered Member #3114 Joined: Sat Aug 14 2010, 08:33AM
Location:
Posts: 608

ben123324 wrote ...

No problem. But again, I don't know what the circuit looks like. But frequency range doesnt really matter. What you want to do is find the resonant frequency of the fly back for longer sparks. This is a fixed frequency so all the pot would do is give you a larger window for this frequency. If you had say, a 1k pot, you wouldn't have a big enough window to find this frequency. One other note though. The frequency must be over 20khz so that all you hear is the music. If you had <20khz you would get music overlaid with loud high frequency buzzing.

I added it to the original post in the edit... sorry

So i am looking at the bus caps and i was wondering gren says 10,000uf! and you said in your honest opinion 2200uf was plenty, why such a big difference. Could i get away with 2000uf, or even 1000uf?

ZakWolf

Fri Mar 16 2012, 10:32AM

Registered Member #3114 Joined: Sat Aug 14 2010, 08:33AM
Location:
Posts: 608

Just finished drawing the pcb with a sharpie now all i have to do is etch and order the parts, i got the board size down to 4x 4 1/2 inches.

Alex M

Fri Mar 16 2012, 11:16AM

Registered Member #3943 Joined: Sun Jun 12 2011, 05:24PM
Location: The Shire, UK
Posts: 552

Here is what I found when I made this circuit.

Use separate supply's for the chip and flyback, I used 15v for the chip as it helps the MOSFET's saturate.

Making the GDT was a pain and many ferrite rings do not work very well, you need one that is of high permeability or else the chip will get very hot.

Use tri-filiar windings on the GDT! You get much better performance, before winding the GDT simply twist your coil wires together. Uzzors has a good guide to this

Make sure the GDT secondary's put out more than 12v at all frequency's in order to saturate the MOSFET's or else they will get very hot. Add more secondary turns to the GDT if needed.

Put a DC blocking capacitor in series with the primary coil, stops the core saturating.

Use 15v or 18v zeners, 12v ones will waste some energy (my 3w ones got warm) that should be going onto the gate making the MOSFET saturate.

Fast diodes in parallel (in reverse, cathode to the chip) with the 22 ohm gate resistors to make for faster switch off.

Use BJT totem pole pairs or between the chip and GDT, it really helps reduce chip heating and allows for faster MOSFET switching (less heat generated). Or use some gate drive chips, UCC37321 and UCC37322.

I used mostly film caps on mine, except for the electrolytic's on the bridge. I have heard that they only need to be 100uF max but since mine are working fine I have left them in there.

If your audio distorts at lower volumes then add more DC blocking capacitance, with 3uF I was able to go much louder without distortion occurring. I used a film cap here as its the first one I found in my capacitor box, seems to work well.

Use fuses to protect your supply's. Although I have not had any MOSFET's die with this circuit its always good practice just in-case.

Lastly a tip about primary coils, I have found adding or removing just one turn with this driver makes a BIG difference to arc length and current draw. Also arcs don't start "getting good" until you input 24v or more, on 12v they are quite small no matter what. Also on 36v with 5 primary coil turns I got some 3-4 inch arcs then the flyback arced over. so bear that in mind when you want "moar power".

Hope I haven't board you to death but hope you find these pointers useful. This circuit is awesome once you get it going and its quite loud too! If you have ever made a 555 based plasma speaker then you will find this one is MUCH louder and less distorted.

Dr. Dark Current

Fri Mar 16 2012, 11:23AM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

I think this is the FM plasma speaker circuit I designed a long time ago, maybe with some little flaws, eg. I can't see a bypass cap on the 12 volt supply rail. Here is the original design:

(it is by no means perfect, I have since designed a better sounding PWM plasma speaker, but this old one still seems to circulate on the net).

ZakWolf

Fri Mar 16 2012, 11:37AM

Registered Member #3114 Joined: Sat Aug 14 2010, 08:33AM
Location:
Posts: 608

thanks everyone

Sigurthr

Wed Aug 22 2012, 01:35AM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

Dr. Dark Current,
I'm currently using a variant (from Le Quack / Inducktion) of your old schematic to audio modulate one of my SSTCs by feeding the Out2 (pin14) to the UCC pair's enable pins. The circuit I have looks nearly identical except that there are no ultrafast diodes (or any diodes).

There seems to be very low variation in the output's modulation; I'm guestimating a change in duty cycle from 35% to 65% maximum, with only a -+10% (40-60%) swing most cases. The Pulse repitition frequency stays stable and the SG3525's output looks clean. I'm feeding the audio in from an ipod's headphone jack in to a 6:1 step-up audio transformer to increase the Vp-p as without it you could barely hear the audio.

Could you please post your "better sounding" PWM plasma speaker schematic?

Also, perhaps you are more familiar with this chip than I am, is there a way to configure the chip to FM modulate via PRF instead of PWM? Adjustable constant duty cycle with the PRF = to the audio input frequency would be awesome.

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob