Do I want fast settling or massive filtration in PS?

[Magz]

Hello all,

I'm working on an upgraded power supply for my tube preamp (line and phono stages combined). It draws a total of 146mA, and is fully class A with a GZ33 rectifier and 350B tube shunt regulators. The current power supply uses a CLCRC filter with 33uF/10H (165ohm)/100uF/344ohm/66.7uF and I'm moving to a CLCLCRC design with something like 6.0uF/10H (36ohm)/47uF/10H (36 ohm)/100uF/300ohm/66.7uF. Transformer has ~75 ohms impedance.

I have been modeling this with PSUDII and have been able to obtain various configurations which pass the current step test (fast settling, no ringing). My questions is this - am I better off optimizing for fast setting time after a current step, or going instead for the most filtration (and thus lowest ripple) I can obtain? I can get the settling time to ~1/4 second for a 20mA step and about 30uV ripple, or get even lower ripple but have the settling time increase to almost 1/2 second. Does the presence of the tube regulators decrease the importance of the settling time?

Any help in this area is greatly appreciated!

 

[Magz]

By the way, the shunt regulator is located between the R and the last C on each channel. As PSUD has no provision for simulating this I did the sim without it just to look for potential issues.

 

[jan.didden]

By the way, the shunt regulator is located between the R and the last C on each channel.

I assume that the shunt is in parallel with the load, at the very end of your CLCR etc, as is customary?

The shunt keeps the total current draw constant, which means that with an ideal shunt there is NO current step. So if you sim this without the shunt it gives a wrong picture.

jan didden

 

[Vinylsavor]

Hi,

Your regulater at the end of the filter chain should make any changes in front of it mostly inaudible, if it works correctly.

Best regards ... Thomas

 

[DF96]

Provided there is no ringing, and the settling time is well below the LF limit of the amp, I'm not sure it matters. Adding a regulator makes it irrelevant anyway. The real issue is the stability of the whole thing including regulator.

I would go for minimum hum, consistent with no ringing.

 

[Magz]

Ok, so it seems that the consensus is that with the shunt regulator in the curcuit, I should go for the max capacitance and thus filtration that will remain stable, and not worry about making the settling time as fast as possible.

Thanks to all!