Is it true that for a 30W output stage operating at constant 200KHz I would need a switching frequency of 10 times that at least, ie 2MHz or more?
Has anyone designed a 30W/2MHz stage they could share?
Has anyone designed a 30W/2MHz stage they could share?
30w
Why would you want a 2MHz switching frequency unless you are a bat and I guess even they cannot hear that high?
If you have a switching frequency of 200KHz where do you come up with 2MHz?
200KHz theoretically allows a 20KHz amplifier bandwidth and that is pushing it.
yes the theory is that the switching frequency should be 10 x the upper frequency you are trying to produce.
I guess you want a 30 watt amplifier to listen to music and typically with a 12dB/octave output filter in a switchmode amplifier = class D you need more than this so as not have too many switching artifacts superimposed on your audio waveform. A switching frequency of at least 350KHz is a always good.
Above 400KHz the losses in the MOSFETs becomes high.
Why would you want a 2MHz switching frequency unless you are a bat and I guess even they cannot hear that high?
If you have a switching frequency of 200KHz where do you come up with 2MHz?
200KHz theoretically allows a 20KHz amplifier bandwidth and that is pushing it.
yes the theory is that the switching frequency should be 10 x the upper frequency you are trying to produce.
I guess you want a 30 watt amplifier to listen to music and typically with a 12dB/octave output filter in a switchmode amplifier = class D you need more than this so as not have too many switching artifacts superimposed on your audio waveform. A switching frequency of at least 350KHz is a always good.
Above 400KHz the losses in the MOSFETs becomes high.
My "listening" frequency is 200KHz and I assumed the internal switching frequency ought to be 10 times over, or as you have indicated 17 times over (20KHz-->350KHz). That would make it around 3.5MHz switching frequency. Are there any class D designs at those frequencies?
So this is not for audio? Depending on your application, the "10 times" rule might or might not be applicable.
Not for audio, but the output must be as pure as possible. It currently works based on a simplistic Class AB design but it could also double as a portable room heater.
Now you really got me curious - if you need as pure as possible, it clearly isn't a rodent repeller 🙂
You could do worse than start with a square wave at 200kHz (switching frequency same as "listening" frequency) and filter it with a LC tank circuit resonant at 200kHz to remove the harmonics.
"As pure as possible" is not a specification. There must be some maximum percentage distortion that the application will tolerate. Good engineering doesn't waste effort improving things more than necessary.
"As pure as possible" is not a specification. There must be some maximum percentage distortion that the application will tolerate. Good engineering doesn't waste effort improving things more than necessary.
My current device in simulation at least is like 0.1% distortion, hoping to being it down to soon.
OK, so we have a square oscillator op-amp, say +/- 10V, plug this on a push-pull powerful FETs, follow by a LC filter that rolls off steeply both sides of 200KHz and will this give me the required watts (like 30W) ?
I could build it on simulation at least?
OK, so we have a square oscillator op-amp, say +/- 10V, plug this on a push-pull powerful FETs, follow by a LC filter that rolls off steeply both sides of 200KHz and will this give me the required watts (like 30W) ?
I could build it on simulation at least?
You can design it for any amount of watts you want. I built a solid state Tesla coil that puts out 2kW average and 50kW peak at 200kHz. I haven't figured out how to connect a THD analyser to it yet.
Would you have some basic design for me to start off with? All I need is between 180KHz-220KHz (pure) sine wave into around 30W.
Currently it works like this so you get an idea:
Supply +/- 10V to +/- 13V (Two 12V batteries)
Op-amp Wein bridge oscillator at 200KHz @ 7.5V peak
Current amp stage (3 pairs of output transistors)
RM12 transformer 1:10 (giving out around 70V peak)
Nominal load 100R, therefore Power on Load = 24.5W
Power expended as heat around 25 Watts
***************************
There is no reason I am not happy with the above except the heat generation and wasteage of the batteries.
Thanks for all the help
Currently it works like this so you get an idea:
Supply +/- 10V to +/- 13V (Two 12V batteries)
Op-amp Wein bridge oscillator at 200KHz @ 7.5V peak
Current amp stage (3 pairs of output transistors)
RM12 transformer 1:10 (giving out around 70V peak)
Nominal load 100R, therefore Power on Load = 24.5W
Power expended as heat around 25 Watts
***************************
There is no reason I am not happy with the above except the heat generation and wasteage of the batteries.
Thanks for all the help
I would start by just overdriving your current amp stage so the peaks of the sine wave are clipped off. This will reduce losses as the devices spend less time in the linear region.
Then redesign your transformer to be resonant, so it puts the peaks back.
Then redesign your transformer to be resonant, so it puts the peaks back.
Just some thoughts : if I overdrive the stage, or if I drive it with a square wave directly, isn't that when the transistors stay on and stay on and cross-conduct as it takes them time to rise and fall? Isn't that the reason why for powerful square waves people use FETs?
And for a square wave power output, would it not be better to have a single FET rather than push-pull, so that there is no possibility of cross-conduction? I have never designed any so am thinking aloud really.
And for a square wave power output, would it not be better to have a single FET rather than push-pull, so that there is no possibility of cross-conduction? I have never designed any so am thinking aloud really.
For efficient power output at 200kHz you should certainly be using MOSFETs. It doesn't much matter whether you use a single device or a push-pull pair.
Yes, my current working design Class AB achieves less than 0.1%. It also works well against a varying load.
Just to understand what was said above, instead of a class D design where the output is a PWM square wave, which would need to be in the MHz range, instead I can use a 200KHz square wave at 50% duty cycle and then somehow convert it to sine wave using filters or tuned circuits. Am I correct?
And if I am correct, then could you please describe the topology of this system?
So far I have simulated :
1) supply +/-10V
2) op-amp producing square wave at +/-7V and 200KHz
3) push-pull MOSFETS taking this to +/-10V (IRF120 and IRFD9120)
4) Feeding final POWER MOSFET (eg PSMN017-80PS) .
5) Next ?
Just to understand what was said above, instead of a class D design where the output is a PWM square wave, which would need to be in the MHz range, instead I can use a 200KHz square wave at 50% duty cycle and then somehow convert it to sine wave using filters or tuned circuits. Am I correct?
And if I am correct, then could you please describe the topology of this system?
So far I have simulated :
1) supply +/-10V
2) op-amp producing square wave at +/-7V and 200KHz
3) push-pull MOSFETS taking this to +/-10V (IRF120 and IRFD9120)
4) Feeding final POWER MOSFET (eg PSMN017-80PS) .
5) Next ?
What I'm saying is, do you need 0.1%? eg, do harmonics at 400/600/800/1M/etc cause any issues with the application?
How tuneable do you need it to be? (if it's 200KHz and nothing else, I'd build something resonant - if you need to sweep +-50KHz or something, that rules that out)
How accurate does the output voltage need to be?
How tuneable do you need it to be? (if it's 200KHz and nothing else, I'd build something resonant - if you need to sweep +-50KHz or something, that rules that out)
How accurate does the output voltage need to be?
I need it to be 0.1% even better if possible. Harmonics may cause problems as the energy transfer increases with frequency so ideally we want as little unwanted as possible else we will be applying unwanted energy on the load.
It needs to be able to produce 180KHz to 220KHz.
The output voltage will also vary a bit, depending on the load. The load is somewhat variable, it can go low or high, from -20% to +50%. When the load goes low the voltage must be decreased to maintain an average current through. When the load goes high the voltage must be increased again to maintain the current as much as possible.
It needs to be able to produce 180KHz to 220KHz.
The output voltage will also vary a bit, depending on the load. The load is somewhat variable, it can go low or high, from -20% to +50%. When the load goes low the voltage must be decreased to maintain an average current through. When the load goes high the voltage must be increased again to maintain the current as much as possible.
This paper gives a sort of overview of what is going on, try the references too.
http://www.academicjournals.org/article/article1380795826_Chang et al.pdf
http://www.academicjournals.org/article/article1380795826_Chang et al.pdf
- Status
- Not open for further replies.