Switch mode noise

[NickC]

I in the process of purchasing a off the shelf 150w 9 volts 15amp power supply. Its an offline switching ps.

Here the problem its switching freq is only 25khz. What's the highest frequency i can mod the power supply without effecting its stability. WOuld 50khz be to much to ask?. I plan on using this for an audio supply and 25khz is just to low, the noise will be somewhere along the audio part round 12.5khz. Noise is a big issues so i would like to keep it down. Also any pitfalls i should avoid when trying to mod the ps to switch at 100khz?

Also is there any filters i can use to cut down on the noise value? I suspect a LC after the power supply should help to cut down the noise. Perhaps a inductance made from the 3c90? Any suggestion on how to do this?

 

[Eva]

I'm just curious: What is such an audio application requiring 9V and 15A?

Also, I think that you are assuming that subharmonics of 25Khz will appear in your audio signal, while this will probably not happen at all, since radiation phenomena happen well above 25Khz.

 

[sagarverma]

I'm just curious: What is such an audio application requiring 9V and 15A?

me too.

tune the freq to atleast above 30khz,well above most of the audio amps' freq response range.

 

[NickC]

Actually the ps is going to heat a tube's cathode/fillament. The filament voltage is 10v 10a. The ps can be modified to 10volts. The noise on the filament/cathode will be directly seen on the output of the valve. So i need to keep this low.

Would the noise function from the voltage been seen on the output? Not the emitted energy but on the 10v?

 

[Eva]

The output ripple will probably be 100mV or 200mV peak to peak in worst case, and will probably be dominated by a triangle wave of double the switching frequency (the concept of 'noise' turns useless as waveforms are analyzed in detail, there is no noise but ripple in SMPS, like in any 50Hz PSU).

This may be easily attenuated to sub-milivolt levels by means of LC filters.

 

[jackinnj]

I in the process of purchasing a off the shelf 150w 9 volts 15amp power supply. Its an offline switching ps.

Here the problem its switching freq is only 25khz. What's the highest frequency i can mod the power supply without effecting its stability. WOuld 50khz be to much to ask?. I plan on using this for an audio supply and 25khz is just to low, the noise will be somewhere along the audio part round 12.5khz. Noise is a big issues so i would like to keep it down. Also any pitfalls i should avoid when trying to mod the ps to switch at 100khz?

Also is there any filters i can use to cut down on the noise value? I suspect a LC after the power supply should help to cut down the noise. Perhaps a inductance made from the 3c90? Any suggestion on how to do this?

there a couple of ways of attacking the noise issue:
1) make sure that the switching supply is well enclosed, subject to needs of ventillation. High frequency switching transients make their way out of openings in their cabinets!

2) if you have a scope, examine the a.c. waveform under load. you should be able to determine the frequency of the spikes (if any) -- given that you are powering a constant load your job is made a bit easier -- my guess is that you are going to need a few hundred micro-henries at most (remember, it isn't the switching frequency it is the time of the transients you want to mitigate).

3) you can consider a linear "post-regulator" but these burn energy -- an LM317 or an LM723 with a couple of beefy NPN transistors in parallel should work.

I wouldn't touch the timing of the controller or the compensation network -- for this you need a network analyzer to make sure that the control loop keeps the device stable.