Synchronizing SG3524

[darw82]

Hi,
I want to build power supply with 3 output voltage.
and I use SG3524 for each output.

Is it okay to sync. all of SG3524 together? They run at DIFFERENT DUTY CYCLE.

The osc & CT pin connected together, RT leave open.
Only the master have RT resistor and CT capacitor.

Thanks

 

[Eva]

If I remember properly, this is the right way to synchronize them (I've suffered a hard-drive crash in the computer where I stored all my datasheets, schematics and electronics stuff 🙁 ). However, in case not all the outputs require tight voltage regulation you may consider a single control circuit and a single switching stage instead, with multiple transformer secondaries and/or a coupled output inductor to achieve several output voltages.

 

[Onra]

Hi,
what is the advantage of synchronized regulators?
Is`nt it better to spread the distortions (by switching) on the supply rails?
Onra

 

[Danko]

For example, you build 2 regulators. They are unsynchronized. One SG3524 is switching at 199khz, and the other is switching at 200khz. Then the difference between theese frequencies is 1khz, which is audible. and this can disturb some things. For example, if you are running an amplifier from common rails of theese SG3524s, then the amplifier will "subtract" 200khz-199khz, and this will be on the speaker.

As far as I know... 🙂

If I'm wrong, please, somebody correct me! 🙂

 

[gmarsh]

Hi,
what is the advantage of synchronized regulators?
Is`nt it better to spread the distortions (by switching) on the supply rails?
Onra

I've worked on designs which have multiple switching regulators, synchronized at the same frequency but also running at different phase angles relative to each other.

Synchronization will reduce low frequency "beat" ripple on the input side, and phasing will also reduce the peak primary side current making life easier on your input capacitors, etc.

 

[Eva]

Danko is absolutely right concerning beat frequencies (intermodulation). Also, delaying some regulators half cycle respect to the others is useful for current ripple and EMI reduction.

 

[Onra]

Hi,
Thanks a lot to all.
My question was triggered by an appnote, i read a while ago. They added a bit of jitter to their SMPS- oscillator to reduce EMI by spreading the noise spectrum.
Onra

 

[Eva]

Adding jitter noise to the oscillator spreads EMI better across its frequency range, thus producing less EMI in average. However, note that jitter noise is not the same as intermodulation between two oscillators closely tuned. The same EMI-reduction principle could be employed in synchronized regulators by adding a random delay to the sync pulse or a random component to the master oscillator.

 

[Onra]

Hi,
That is understood.
I was fixed on the distortions produced by switching on several devices at the same time and did´nt realize the problem of intermodulation at all.
Onra