Any advantage using DC for filament?

[Enzo]

2) If you tie the filament to the cathode resistor of the output tube, then the filament will be bouncing up and down with the signal, how does that help?

Only if that cathode is unbypassed. A cap will steady it up. But hum abatement, that wouldn;t matter anyway, as long as the heaters always remain more positive than your preamp cathodes.


There are many sources of hum, and each one has its own cure independent of others. Not all appear in all amps. Electron emission from heater to cathode is one possibility, and elevating heaters fights that. That is separate from radiated hum from heater wires and other causes.

[Leo_Gnardo]

There are many sources of hum, and each one has its own cure independent of others. Not all appear in all amps. Electron emission from heater to cathode is one possibility, and elevating heaters fights that. That is separate from radiated hum from heater wires and other causes.

"Floating" the filaments on a filtered positive voltage seems to work well. Looking at some mid-60's Ampegs, they floated the filaments on a 0.1 uF cap, typically 200V rated. And unbiased, the cap joins the filament winding to chassis ground. I find this technique works in some amps, and the advantage is, it's simple and cheap. HOW it works, I'd like to hear an explanation.

[eschertron]

"Floating" the filaments on a filtered positive voltage seems to work well. Looking at some mid-60's Ampegs, they floated the filaments on a 0.1 uF cap, typically 200V rated. And unbiased, the cap joins the filament winding to chassis ground. I find this technique works in some amps, and the advantage is, it's simple and cheap. HOW it works, I'd like to hear an explanation.

No pull-up resistor or divider? I'd like to hear about it too!

[Alan0354]

"Floating" the filaments on a filtered positive voltage seems to work well. Looking at some mid-60's Ampegs, they floated the filaments on a 0.1 uF cap, typically 200V rated. And unbiased, the cap joins the filament winding to chassis ground. I find this technique works in some amps, and the advantage is, it's simple and cheap. HOW it works, I'd like to hear an explanation.

That's a very bad idea, really bad idea. I worked with hv for years. any conductor floating will be charged up in the vicinity of high voltage and can arc to other circuits. That was one BIGGEST source of problem for us. We had to account for all the metal trace and loose ends on the pcb and this happens EVERYTIME.

These amps only have 500V, you might get away with this, but this is a bad practice and show the ignorance of those people.

[Leo_Gnardo]

That's a very bad idea, really bad idea. I worked with hv for years. any conductor floating will be charged up in the vicinity of high voltage and can arc to other circuits. That was one BIGGEST source of problem for us. We had to account for all the metal trace and loose ends on the pcb and this happens EVERYTIME.

These amps only have 500V, you might get away with this, but this is a bad practice and show the ignorance of those people.

I've put the meter on and seen no voltage pullup, in fact no DC reading at all on top of those Ampeg filament-floater caps. FWIW when I use this technique to de-hum an amp, 600V cap is what I use, keeping in mind the rare but possible plate or SG to filament short.

For such "ignorant people" Ampeg sure made great sounding and long lasting amps back in the 60's.

[Alan0354]

I've put the meter on and seen no voltage pullup, in fact no DC reading at all on top of those Ampeg filament-floater caps. FWIW when I use this technique to de-hum an amp, 600V cap is what I use, keeping in mind the rare but possible plate or SG to filament short.

For such "ignorant people" Ampeg sure made great sounding and long lasting amps back in the 60's.

This is because your meter has 20Mohm to ground, that grounded the filament right away. It's only two resistors extra, you don't want to mess around. We had to turn off all the lights, make it arc to find every single loose metal on the circuit board.

Just because it was done does not mean it's right. It might not be as bad as it's only 500V, but it's a really bad practice.

[NateS]

Lots of people with varying degrees of proficiency have made some great sounding amps. Some of which want to kill you. If you need further convincing, one only need to look at the number of 'classic vintage sounds' which are copies of one another, yet sound nothing like each other.

I'm guilty of it too. At some point we go from calculating to 'ehhh looks about right', or worse 'if we don't do xyz the electrical spirits will be angry with us. Ken Amdahl wrote a book called 'There are no electrons' where he explains electricity using little green people. It's fun reading, and actually is a really good explanation of basic electronics, even though it's completely fictional.

Point being, just because they made some good sounding amps don't mean they knew what they were doing.

[Mick Bailey]

"Floating" the filaments on a filtered positive voltage seems to work well. Looking at some mid-60's Ampegs, they floated the filaments on a 0.1 uF cap, typically 200V rated. And unbiased, the cap joins the filament winding to chassis ground. I find this technique works in some amps, and the advantage is, it's simple and cheap. HOW it works, I'd like to hear an explanation.

Some Ampegs reference the heaters to the bias supply, too.

Interestingly I found this extract from RDH4; "Emission from heater to cathode is caused by impurities in the heater insulation material or from the deposit, during valve manufacture, of cathode material on to the heater. To overcome this fault the heater should be biased positively with respect to the cathode. As the voltage required may be of the order of 50 volts, the minimum required to give satisfactory hum reduction should be used to avoid exceeding the maximum permissible heater-cathode voltage.

Cathode-to-heater emission can also take place and cause hum, although it is not common. The cure for this is the biasing of the heater negatively with respect to cathode."

The second paragraph is probably why they used the -50v reference in the B15 and others.

[jpfamps]

I use two resistors to get the correct voltage - just a simple voltage divider. I usually connect it to the 1st cap node. If you place it too far down the filtered B+ nodes it can introduce noise of its own into the preamp. A cap can be added to the lower resistor to stabilise the DC reference - 22uf is fine. When choosing resistor values, bear in mind the dissipation. I use the lowest values possible so they're not running hot but give the correct ratio.

So, you've then got the DC reference. Just connect it to a 100 Ohm resistor each side of the Heater AC supply the same as you would if you were using resistors to ground.

20v seems a sensible level to me for a regular amp. If you took the node up to (say) 80v than this would reduce the heater-cathode voltage difference on a CF by that amount.

EDIT: Aim for between 220k and 470k for the overall divider resistance. If you have a bleeder across the 1st cap this replaces it.

A cap should always be tied to the voltage divider so that they heaters are tied to AC ground.

Having a high impedance path from the heaters to ground makes HF/RF interference more likely.

[jpfamps]

"Floating" the filaments on a filtered positive voltage seems to work well. Looking at some mid-60's Ampegs, they floated the filaments on a 0.1 uF cap, typically 200V rated. And unbiased, the cap joins the filament winding to chassis ground. I find this technique works in some amps, and the advantage is, it's simple and cheap. HOW it works, I'd like to hear an explanation.

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.

[Alan0354]

So there is no advantage of DC filament supply?

I am designing an amp, I'm not going to even entertaining the idea of DC filament unless there's a compelling reason for it.

[trobbins]

For indirectly heated cathodes, AC signal on the heater can couple to cathode as one form of hum generation (note there are others, such as parasitic coupling from heater to grid). The coupling can be both from the resistance between heater and cathode, and the capacitance.

I've measured a batch of old 12AX7, and the resistance characteristic was very varied, but had the common (although not exclusive) form of an increasing resistance as DC voltage differential increased (both pos and neg polarities). As such, DC elevating the heater would move the resistance level to a higher region, and hence reduce hum from that mechanism. Some tubes had a quite low resistance, and so would be more prone to hum issues than others in the batch. This characteristic may be associated with some peoples experience that heater DC elevation improved their amp.

Note that the capacitance level is not affected by the elevation level, and that any signal on the elevation voltage is undesirable. Also note that HT rectifier noise can couple to the heater winding within the PT, and being higher in frequency can more easily couple through heater/cathode capacitance, or couple to grid via parasitic heater/grid capacitance.

[Alan0354]

Right, I am open to raising and adjusting the DC offset of the filament. But I just have to say I never notice problem with amp hum by itself except when the two power tubes are mismatched.

Biasing the filament is no big deal, making a regulated DC supply is a big deal as you don't have low voltage in an all tube amp and PT usually don't come with 18 or 24VAC . I feel people doing DC supply just for the sake of doing it.

[Enzo]

Alan, it is just another tool in the box, you can use it or not. Many of the things we do have no compelling reason to be done that way. I have my choice of several tone stack configurations, several phase inverter configurations, various grounding schemes, power supply filtration schemes, and so on. None of the choices I don't take are "wrong." Sometimes when you have three ways to do something, we pick one way over the others not for its superiority at the job, but because it fits the physical situation. For example I might select which triode in a tube to use for something based on which side is closer to the rest of the circuit in the chassis.

If you haven't encountered heater-cathode hum, then you have not felt the need for its abatement. That doesn;t make it invalid or a waste of time, it just makes it moot for YOUR situation.

[Leo_Gnardo]

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.

Coincidentally I was thinking along the same lines while I took a grocery-shopping break earlier this afternoon. Although it's hardly ever done, yes a separate transformer for filaments would likely be a solution. Just because it isn't done by the manufacturers, doesn't mean it can't be done by the determined hobbyist/home builder. Costs extra, weighs more, drill more mounting holes, but so what if you're aiming for perfection.

Alan, depending on how much gain you're going to design into your amp, DC filaments may or may not be necessary. You may have to experiment to find whether you need to do this or not. One thing that's for sure, some preamp tubes put more hum/buzz into the audio than others. Not just a brand to brand variation, but even within a batch that are otherwise presumably identical. If you're running DC filaments on hi gain pre tubes, it's insurance you won't be hearing hum/buzz that's due to filament powering. You'll find you're hearing what the tube does to your guitar signal clearly, instead of hearing it behind a fog of buzz. In addition more preamp tubes will be acceptable with the hum source eliminated, so you'll be wasting less time picking thru preamp tubes trying to find the minimum hum ones.

Nate, regardless of proficiency or not there wasn't a grounded AC cable on any of those old amps. Neither were there grounded outlets to plug them into - grounded power outlets were a rarity until @ late 60's. Hooray for grounding! I really don't like getting shocks, & nobody but Uncle Fester does as far as I've seen.