How Much is Too Much Filtering?

[Glowin Plates]

I have replaced the filter caps on one of my amps that has a tube rectifier
(6CA4) using a JJ 50uF multi-section.
I think the originals were 40uF so I felt pretty safe going a bit higher.
I have heard a lot about too much filtering putting stress on the rectifier tube
and I do notice this amp has a low frequency vibration coming from the power supply.
Should I be worried?

[Vinylsavor]

Hi!

check the datasheet of the tube and go with the max values there.
50uF is the max for this tube, but it requires a minimum transformer secondary resistance, which depends on the voltage.

It is a common mistake to load rectifier tubes with too much capacitance. Rectifier tubes don't like high caps, they cause high repetetive peak currents and a high inrush current.

Also soundwise smaller caps or choke input filters are the better choice

Best regards ... Thomas

[ryuji]

Use a pi filter and you can put as much capacitance after the inductor as you like, within reason. Best suggestion is to look up detailed data on tube and possibly transformer and simulate in psud2 to check peak rectifier current. In most cases a 10 uf input cap does the job from what I have personally simulated

[cbdb]

New to tubes, didnt know they didnt like current spikes. Can someone tell me why? So whats the advantage of using a tube rectifier over solid state for (I'm assuming) a power supply?

[ryuji]

Voltage drop, Lack of switching noise, soft power up.

Silicon has current limits too, recification only pushes current at ac peaks.

[Arty]

tubes are high voltage and low current stuff -usualy- , this is why they do not like high repetative peak currents. Same reason why most of them use an output tranformator. To match load impedance, as tubes usualy are unable to produce high current at output

As for using tube in a rectifier circuit, yet i have found no reason to do so.
Surely there has gotta be something, i just do not know why.
For some instruments -mainly guitar- a tube rectifier may be desired, it has different properitys then diodes. They can add a distorsion, and this can enhance a guitar sound.
But i personaly would not use a tube rectifier for anything else than instrument amplifier.

[Eli Duttman]

Vacuum rectifiers don't generate the switching noise PN junction diodes do. OTOH, the introduction of high PIV Schottky diodes made from silicon carbide deals with the noise issue. No PN junction, no noise.

Vacuum rectifiers constructed with cathode sleeves start up slowly. Good "high" current vacuum rectifiers with cathode sleeves are getting harder and harder to find. OTOH, directly heated vacuum rectifiers, like the 5U4, 5R4, and 5Y3 start almost as quickly as SS diodes do. Negative temperature coefficient (NTC) inrush current limiting thermistors do a respectable job of taming start up surges. So, I will not design a HIFI power amp with anything other than SS B+ rectification. Low current vacuum rectifiers remain available and their use in small signal applications remains a reasonable choice.

Silicon has current limits too, recification only pushes current at ac peaks.

Cap. I/P filters exhibit pulse conduction. The larger the cap., the shorter the pulse length and the higher the pulse current. Search out my posts on "hash" filters in combination with large value cap. I/P filters.

A properly executing choke I/P filter causes the rectifier winding to exhibit a 100% conduction duty cycle. The critical current is that value which causes that 100% behavior to occur.

[Miles Prower]

Originally Posted by cbdb

New to tubes, didnt know they didnt like current spikes. Can someone tell me why?

Hollow state is high voltage, low current. Consider the 5U4GB, one of the beefiest vacuum twin diodes. It has an Isurge rating of 1.0A per plate (2.0A if you use two and parallel). That really isn't all that much, considering that even small silicon diodes can have an Isurge= 10A. You get these high current surges when the voltage rises above the voltage on the reservoir capacitor, which has 1/120 secs to discharge, but only a few milli-secs to recharge.

Why you can't simply use a huge reservoir capacitor, as you would do in a solid state amp. (For one project using 5U4GBs for the main DC rail, 34uF meant an Isurge= 800mA, and 47uF was too much.) You have to keep that capacitor small enough, and use a subsequent LC or RC LPF to get rid of the AC ripple. The only other option is an L or R input ripple filter that holds down the recharge current at the expense of voltage output.

So whats the advantage of using a tube rectifier over solid state for (I'm assuming) a power supply?

Glowey bottle kewelness. The one and only reason I did a project using the 5U4GB was that a vintage power xfmr that was otherwise a good fit for the project overvolted badly with Si diodes. With a 5U4GB, the DC was spot on. Obviously, this xfmr had been designed with the 5U4GB in mind. So that's what I used as opposed to dropping voltages some other way.

Otherwise, SS gives higher output voltages, has better regulation under load, can use larger filter capacitors (if the xfmr can withstand the higher peak currents) doesn't need heater power or another hole in the chassis. You can always use some sort of soft start/delay to let other VTs warm up before hitting 'em with the high voltage.

[ryuji]

I know its nitpicking but I use a 5c8s full wave indirect heated cathode rectifier, ipeak is 1.3~1.4 amps and rms current rating is 420 mA max. Also has 1700v reverse voltage rating.

Already knew it was capacitor behavior that does the pulse behavior but his current circuit uses capacitor input

[ryuji]

For best ss power supply for regulation etc I think a custom built smps would be best? Filter stage before switching components filter out the ac line ripple and set switching rate at 30 to 40 khz and you'd have to be a dolphin our something to hear the hash noise? You'd also gain percise voltage control and regulation