LM338 regulated snubberized PSU for audio amplifiers

[AndrewT]

Hi,
it does both.
the extra R creates an RC filter that removes some of the ripple on the first cap.
The rectifier sees one cap to charge and a resistor in series with another cap. The rectifier thus sees a smaller load and supplies less current.

But,
the first cap suffers more ripple current than when the R is omitted. It must be selected to be able to survive this demanding position.

The last set of caps (after the R) supplies almost all of the transient current demand of the following stages. These will determine the sound quality and bass/mid performance. Size them accordingly.

 

[samsagaz]

Hi,
it does both.
the extra R creates an RC filter that removes some of the ripple on the first cap.
The rectifier sees one cap to charge and a resistor in series with another cap. The rectifier thus sees a smaller load and supplies less current.

But,
the first cap suffers more ripple current than when the R is omitted. It must be selected to be able to survive this demanding position.

The last set of caps (after the R) supplies almost all of the transient current demand of the following stages. These will determine the sound quality and bass/mid performance. Size them accordingly.

Andrew if i remove the R, i dont will loose power that are disipated in this R. You think that is better to use that R? I was thinking about use an Soft Start or something like that instead that R.

 

[AndrewT]

soft start is a different animal.
Soft start is to limit the start up current of a large transformer.

 

[sasmit]

can we build a dual regulated supply using a center tapped transformer.. and a single bridge rectifier...using LM338 ? It seems possible ...can anyone give a valid reason or something I am missing out , both the secondaries are anyways going to be connected

 

[AndrewT]

If you use a +ve regulator along with a -ve regulator you can use either a a Dual secondary or a centre tapped secondary.

If you intend using two +ve regulators for a dual polarity supply then you must use a dual secondary along with dual bridge rectifiers.

 

[Ted205]

how about a tracking preregulator added into the mix:

An externally hosted image should be here but it was not working when we last tested it.

 

[AndrewT]

how about a tracking preregulator added into the mix:

if reg1 needs diode protection then why does reg2 not require it?
Remember to install reg2 the right way round.

Don't let the charging pulses through C1 contaminate the clean ground of reg1.

C2 makes a big difference to the performance of reg1. Is 10uF big enough?

 

[Ted205]

c2 - 10uf is plucked from the datasheet.

confused by
"Don't let the charging pulses through C1 contaminate the clean ground of reg1."

 

[AndrewT]

c2 can benefit from being ten to 20 times larger.

The pulses charging the smoothing cap are connected to the clean ground and will pulse this ground giving a pulsing output from the regulator which is referenced to that now contaminated clean ground.
The smoothing cap must be connected to the output not strung though the clean ground.

 

[paulb]

if reg1 needs diode protection then why does reg2 not require it?

The top protection diode is because of the output capacitor, guarding against current coming back through the reg when the input is shorted. Maybe it should go back to the input.
The other is for the capacitor on the adj terminal.

 

[Ted205]

corrected layout

An externally hosted image should be here but it was not working when we last tested it.

the preregulator in the above picture is the wrong way round . Also included a 1uf cap between the regs.

All the R and C values are just estimates really

 

[sasmit]

Can we have the schematic you're talking about...I can't see any schematics in the previous few posts

 

[Sadface]

I couldn't find much mention of this anywhere in the thread.


How much heatstink is required for each LM338 IC?


It seems to me like these would need a decent sized heatstink on them?

 

[asuslover]

Depends on the voltage drop and the current needed.

 

[Sadface]

Depends on the voltage drop and the current needed.

Lets say one wanted the PSU to be capable of 5A continuous.
28VAC secondaries, giving 39Vdc into the reg. The standard 30V out of the regs.

Would the dissipation be as simple as:
P=E x I
=9.5v*5A
=47.5W

 

[asuslover]

I think lm338t can dissipate max 15w , lm338k will dissipate 50w but is discontinued and very expensive .
But 47w is a lot of heat anyway....

 

[Sadface]

Is my working correct for how much the regs would have dissapate or have I missed something?

 

[traderbam]

Yes and no. At that current the RMS voltage at the regulator input would be less than 9.5V due to transformer losses. On the other hand, your mains voltage at its specified maximum could make it higher. You really need a power transistor to shunt most of the current.