Any advantage using DC for filament?

[uneumann]

"The usual 6.3v AC winding works fine with a bridge rectifier and a filter cap."

If you don't mind all the ripple that remains.........AND you don't mind using 20,000 uF capacitor to kill the ripple.

Well - nothing is free. You can have 6.3v of ripple or .5v ripple or .05v ripple... it's just a question of how much you need to reduce the ripple. That means you don't NEED 20,000uf caps. You can use any size you can fit or afford and get the benefit of the reduction it gives you.

Look... a bridge costs about a 1$. A 6,800uf/15v cap costs under $3 (Parts Express).
I use two of those caps to filter the filaments for two preamp tubes leaving a measured ripple of < 0.1vpp.
Why is this controversial? If you have any sort of hum problem due to filaments, this is a huge win and a really cheap fix - less than the cost of a new 12ax7 tube.

In production, these $5-$10 costs matter so most stock amps don't use DC. But, if you're retrofitting a boutique amp or building one yourself, the added cost and effort is miniscule compared to the benefit and extra noise margin you gain. It probably costs more to use colored cloth wire just to make the chassis look cool than to run DC filaments. Nothing against cool looks - I do it too - but I wouldn't cut $5 in cost and wind up with a humming amp.

[supawah]

trans secondary noise coupling? across diodes?

All the Ampegs I've seen from that era have the filament supply tied to DC ground (either via a centre tap or a hum balance pot).

The addition of a 0.1 uF (or higher) cap to ground from the filament supply is a good idea as it provides a low impedance path at HF and RF.

The main problem with heater hum is NOT coupling of 50 Hz or 60 Hz filament hum into the audio, but the capacitive coupling of rectifier switching artefacts from the HT.

Switching artefacts capacitively couple from the HT winding to the LT winding in the PT, and thence couple from the LT winding to the valve cathode.

Tying the LT winding to AC ground using cap helps to reduce these artefacts on the heater supply. Tight twisting of the LT wires also increases capacitance between the two windings, reducing noise.

I also like to deal with the problem at source by using snubbing caps on the HT rectifier.

Other ways of dealing with HT artefacts on the LT would be to employ an electrostatic screen between the HT and LT windings of the transformer, or (better) use a separate LT transformer.

I'd like to learn more about this. I just had to fix a sunn concert bass that was going ballistic at 124 volts ac line, which seems to be pretty regular here in Austin. turns out the three secondary windings were coupling all kinds of junk through the amp. now how this could happen across the full wave rectifiers is what's eluding me.

[kg]

A low noise tube is also a low gain tube.

technically, if we're speaking about "unavoidable" noise generated in the tube itself, mu has nothing to do with it: it's all about gm.

[Chuck H]

And I'll add that the noise relative to gain is the only consideration in SGM's statement. I only mention it because I've had MANY preamp tubes that were had unacceptable hiss (shot noise?), hum or were too microphonic right out of the box but were absolutely NOT higher gain tubes. Also, Sovtek made a 12ax7 that had higher capacitances but still had most of the gain. The result is a fairly dull sounding tube, but NOT lower gain in the LF and mids. And distinctly not noisy or microphonic. They're a real peach for some high gain amps in the first position.

[Alan0354]

Noise is referred to the input, on the first pass, even if the first stage has only gain of 10, all the noise from the plate and after is going to be divided by the gain ( 10) of the first stage when referring back to the input.

I am not familiar enough with how to determine low noise tube and the mechanism. But I can tell you that for BJT, noise of the transistor mainly due to the base spread resistance WHICH is like the resistance in series with the base at the input........WHICH is like the grid stop resistor ( 68K for Fender). That is what I am more concern when comes to noise. The thermal noise of that 68K is in series with the guitar signal and it is un-attenuated. You see the full thermal noise of the 68K resistor.

People kept saying lower the 1M load resistor at the input, that does not do anything as that 1M is in PARALLEL with the impedance of the guitar which is the impedance of the pickup. Lowering the 1M will muffle the sound of the guitar.

Input current that cause shot noise should be low as tube don't draw grid current when voltage is always negative. Shot noise from the plate is not as important as the noise has to divide by the gain of the first stage to refer back to the input. Unless you run ulber plate current, it should not be that bad.


Noise calculation is quite exact......if you take into consideration of all the noise contributors. It is not as obvious as people think, most of the time, you really have to work out the noise of every contributing source to see which one is the gating factor.

[bluto]

Interesting thread! FWIW running a separate filament xfmr (like a 12v for $10 from RS) can be a low buck way to build a PS, have done that myself combined with a 230V/$20 Triad xfmr connected to a diode bridge rectifier, plenty of filament and B+ supply for a total of $30. Some old hifi amps had DC supplies with the preamp filaments wired in series( of course they would all need to have the same current rating, like 150mA), filtering and regulation is much easier this way. Here is part of a schematic from an old Sherwood that shows the series filament DC supply, off the same PT tap as the bias supply...

[Tooboob]

I'm glad that folks finally pointed out that every time you have a quiet amp, some guy plugs in one of those 'lectric guitars and spoils everything.

[trobbins]

So you could cross-check that by making your own test 'mains antenna' with say an electric kettle load, to confirm that poorly run wiring that is many meters away is a contributor to hum. As compared to some conducted noise voltage issue.

[echoride]

Although it wasn't a high-gain amp, a while back I refurbished an old RCA power amp ( a console pull). All AC filaments; no hum balance pot; filaments weren't tied to cathode; filament wiring wasn't even twisted and just laid in the chassis willy-nilly; grounding was individually going to a big buss wire going the length of the chassis. Completely against what is preached. I completely recapped it and replaced all resistors, then fired 'er up to see how noisy it was. Except...it wasn't noisy. In fact I couldn't even tell it was on: no hum whatsoever, no hiss. Did I screw up? I connected an audio source and she sounded great!

Granted, it wasn't a high-gain amp, but I was impressed. Maybe RCA knew a thing or two about making a simple, perfectly functioning amp? (6BQ5's in PP, which incidentally were being run at plate voltages considerably higher than their very own RCA handbook ratings.)

People don't realize RCA and Zenith and Magnavox (but apparently not GE and definitely not Curtis Mathes!) had engineers just as good as Marantz and Mcintosh, and paid them more, but their assignments were different: they had to get the cost down WHILE RETAINING PERFORMANCE and minimizing repairs. The RF heads in the JFK/MM era Zenith consoles were fully the equal of most component FM tuners, and the power supply amp chassis usually works for 50 years if kept free of roach eggs and excessive smoke and dust, and run often enough to keep the caps formed.