Choose for dc stabilizer for lab power supplay

[Emil88d]

Hello, i would like to build my own lab power sypply. It will be linear PS made out of toroidal transformer bridge rectifier and then stabilizer with purpura voltage of about 12-15 v and then to power LTC3780 dc-dc converter. I would like to have about 0-30V and 0-8A supply with with small disturbances and there is my question what stabilizer i must use to have stable voltage? I buy some LM338T but they are about 5A and i need connect them to have more power and what schematic should i use? I found them in datasheet of the LM.

[David Hess]

Either will work but the lower example requires one fewer power resistor and no separate values.

Why is a DC stabilizer even required in your application?  The switching regulator should handle the ripple from the rectifier just fine.

[Emil88d]

So i dont need this DC stabilizer before LTC3780 to work properly? Will it work fine with not so big distortion? If so i will skip this part.

[Kleinstein]

The SMPS chips are perfectly fine without a regulator in front. For those chips even residual ripple is just a "slow" changing supply they can handle.

If combined would be first the SMPS chip and behind that a linear regulator to get faster transient response.

[wraper]

It would be simply waste of power gone into heat. And you want to do it backwards. If you want cleaner output from DC/DC converter, you put linear regulator AFTER DC/DC converter, not before. Because DC/DC converter by itself is a major source of ripple.

[tunk]

I think you would want a capacitor as a "stabilizer" after your rectifier.
Also note that your transformer and LTC has to handle 22A, (30V+3V)/12V*8A.

[David Hess]

So i dont need this DC stabilizer before LTC3780 to work properly? Will it work fine with not so big distortion? If so i will skip this part.

That is correct.  You have to stay within the maximum input voltage of the LTC3780 but it will have no problem rejecting input ripple.

It would be simply waste of power gone into heat. And you want to do it backwards. If you want cleaner output from DC/DC converter, you put linear regulator AFTER DC/DC converter, not before. Because DC/DC converter by itself is a major source of ripple.

Usually it is better to use a passive LCR network to remove switching ripple because the ripple rejection of a linear regulator falls with increasing frequency.  Before I knew better, I tried that several times but later better measurement technique showed that the input and output decoupling networks combined with the series inductance of the power traces had more effect  than the the linear regulator itself.

If you go this route, careful probing technique is necessary to get consistent results which means high voltage differential probes and maybe coaxial connections.  Circuit layout is critical to prevent ground loops and that includes any probe grounds.