Filter cap before the plate supply

[g-man]

Deleted

[sluckey]

It's not just the cap. It's the resistor also. That RC smooths the B+ a bit even before it reaches the plates. Not so important to a push-pull amp, but does a lot for reducing hum in a single ended amp.

[g-man]

Deleted

[Fresh_Start]

It's not just the cap. It's the resistor also. That RC smooths the B+ a bit even before it reaches the plates. Not so important to a push-pull amp, but does a lot for reducing hum in a single ended amp.

Duncan's PSU lets you model power supplies.  It's not the most intuitive software ever, but it does the job. You can see the impact on ripple of doubling a resistor, adding a PI filter, or using a choke.


http://www.duncanamps.com/psud2/


Merlin has a section on filtering:  http://www.valvewizard.co.uk/smoothing.html

As far as your sketch goes, B+ 5 is much better filtered than B+ 3 because of the larger resistor in that branch of the power rail.

IIRC the hum rejection of a push-pull power amp is improved when negative feedback is applied. So an extra "reservoir" cap might be more important in a push-pull amp without NFB. To reinforce SLuckey's statement, I can tell you from direct experience that adding a reservoir cap and resistor - a PI filter - before the plate node reduces noise a lot in a single ended power amp.


Last but not least, don't forget that the design of the power rail impact the touch response, sag, etc. of the whole amp. Your design is a relatively stiff power rail even without the extra PI filter up front. That's probably what you're going for but it took me a long time to grasp how everything ties together.  Solid state rectification with a 100uf reservoir cap would be a rigid power supply.  Putting a choke and extra filter caps in a 5E3 circuit may seem like a good idea, and you may like it better, but it won't feel like a Tweed Deluxe anymore.


Hope that helps,
Chip

[Willabe]

The 50R before the standby will give back some of the sag.

[jjasilli]

IMHO you need a reason to add a "pre-filter".  E.g., a certain tube rectifier is not rated to handle a filter cap in the 20+uF range. Or, it's an SE amp w/o the hum reduction of PP.

[g-man]

Deleted

[Fresh_Start]

IMHO you need a reason to add a "pre-filter".  E.g., a certain tube rectifier is not rated to handle a filter cap in the 20+uF range. Or, it's an SE amp w/o the hum reduction of PP.

For a push-pull amp, I'd focus first on getting the cleanest power to the first couple of gain stages.  Ideally one power rail node for each pair of out-of-phase triodes.  IOW one node per 12AX7.

Also, in the OP drawing, why not have the branch for B+5 come off of B+3?  Extra ripple reduction. If you're trying to keep the voltage for B+5 higher than B+4, just use a resistor somewhat lower than 47K.  The extra pi filter will more than make up for lowering that resistor - even down to 5K or 10K.

Cheers,
Chip

[Fresh_Start]

OH, G-man - is that supposed to be a fuse on the power transformer center tap?  I hope not.  If you lose the center tap connection to "ground", bad things happen very fast.

Respectfully,
Chip

P.S.  I stand corrected.  Years ago when I was working on my first build, my voltages were too high.  I came across the idea of putting a zener diode under the PT center tap.  When I asked about this approach on 18watt.com (which no longer exists), the flame throwers, blow torches and conspiracy theorists attacked me without mercy and insisted that anything between the PT center tap was strictly forbidden.  The zener works fine but I mistakenly thought that a fuse would be a problem.

Please refer to quotation from PRR below:

[mresistor]

..and the first reservoir capacitance in the power supply should be before any standby switch....   personally   I don't use standby switches much any more

[PRR]

> a fuse on the power transformer center tap?  I hope not.

That's fine. Ampeg even put the StandBy switch in that location.

[g-man]

Deleted

[jjasilli]

The Standby SW is in the wrong place.  It should go Before the plate supply.  (Or in the CT as PRR suggested above).  Probably a drawing mistake, but don't wire the amp that way (with plate voltage but no screen voltage). 

[g-man]

Deleted

[PRR]

> don't wire the amp that way (with plate voltage but no screen voltage). 

A couple amps did do it that way.

In theory, it does no harm. Plate current goes very-very low, but not zero. Output will be very very low, but maybe not zero (but maybe not enuff to hear). Tube "wear" is not a real concern, especially with such low current. If I were to leave a vital radar-station this way for months on end, I might want to run some tests. Standby for breaks should not be an issue.

I agree it is NOT usually done this way, and I can't see any reason to do it.