• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

when do filament supplies need to be floated

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I want to build franks 6sn7 preamp. I have read that the 6.3 filament supply should be floated. the circuit shows The filament winding has two resistors in series across the 6.3 winding with ground between the two resistors. I have built a Pas preamp with out floating the filaments. why is floading recommended. If the filament is DC or used with tube rectfiction have any thing to to with floating or not floating the supply? I believe the reason for the floating filament has to do with 60HZ hum? The pas preamp I built used a 12v regulator for the filaments without floating the filaments with no hum. When is floating needed or recommended? Thanks in advance, Roy
 
Last edited:
if you look at datasheets for tubes there is a heater cathode voltage rating for indirect heated cathodes...

cascodes, mu-followers, srpp tubes circuits or even dc coupling grids can have these ratings exceeded.....that is why filaments need to be floated....else to avoid hum dc biasing filaments are resorted to....
 
Thanks Tony, So if I use franks interstage capacitor coupled version which is one stage amplification and a A cathode follower stage. would I still need to float the heaters. I'me not sure how to float a filament supply. I can follow the circuit diagram, but not really understand totally. Was floating done on early designs? I thought I have seen the filament and B+ referenced from the same point. Which is circuit ground? Thanks all, Roy
 
yes......when one side or even the center tap of the filament winding is connected to ground then the chances of exceeding the heater cathode voltage rating is a real possibility....

biasing the filament to 1/4 of B+ is recommended for Aikido preamps....and many others...
 
What most people call "floating" is when there is no connection, directly or indirectly, to a fixed voltage reference.

"Filament" is the common name for a directly heated cathode.

If it's an indirectly heated cathode, we call the resistor thingy inside a "heater".

Your heaters are not floating, they have a fixed voltage reference.

Heaters should never be left floating. Always respect the heater-cathode voltage rating including the max signal swing and startup swings.
 
I have heard that, as a rule of thumb, you should never exceed 100V from heater to cathode.
When I originally built my 6SN7 mu-follower, I tried a single heater supply for both the "Upper" current-source tube and "Lower" signal tube referenced to a voltage divider on the B+. It was set at 1/3rd of a 220V Plate supply.
As my tubes got about a few years old they would start to intermittently pop and crackle. Nothing else seemed to be wrong. I could only figure that I was near the maximum heater cathode voltage for the "Lower" tube.
On my last redesign, I added an additional regulated heater supply referenced to ground for the "Lower" signal tube. and left the heater supply for the "Upper" CCS tube referenced to 1/3rd the B+ voltage. This allowed me to finally reduce the resistance in the CRC B+ supply and raise it over 300V DC, where the Octals are much more happy.
It has been several months now, and I have experienced no crackling or pops.
Octals Rule!
 
I want to build franks 6sn7 preamp. I have read that the 6.3 filament supply should be floated. the circuit shows The filament winding has two resistors in series across the 6.3 winding with ground between the two resistors. I have built a Pas preamp with out floating the filaments. why is floading recommended. If the filament is DC or used with tube rectfiction have any thing to to with floating or not floating the supply? I believe the reason for the floating filament has to do with 60HZ hum? The pas preamp I built used a 12v regulator for the filaments without floating the filaments with no hum. When is floating needed or recommended? Thanks in advance, Roy

"Floating" is a very ambiguous and context-sensitive word. I use it in the sense I think you mean it.

Resistor values are generally around 100 ohms, chosen to minimise "floating", and wasted current, whilst still providing an effective centre ground connection. The reason for a centre ground is so that the heaters mean potential remains static: one end of the heater rises as the other end falls. Sometimes a pot is used for the purpose, so it can be adjusted for minimum hum. It's then imaginatively called a "hum pot". A grounded centre-tap transformer is arguably better because no resistor is necessary, so there's no "floating" at all.

Doesn't apply to regulated DC, obviously.

Ground potential may not be the best choice, BTW. Apart from the maximum allowable value for heater-cathode voltage difference, there is arguably an ideal value in each case. I can't remember why, or what the value is...possibly such that the heater is always positive wrt the cathode, which would avoid thermionic emission from h to k...anybody?
 
"Floating" is almost universally used in electronics to define a high impedance or undefined state

It gets ambiguous when others misuse common terms. Which is my intention to avoid in correcting the initial poster.

As in "floating paraphase" perhaps?

Taking your more generally correct meaning, there are degrees of floating, depending on the meaning of "high impedance", from loosely tethered to cast adrift. A relative matter, I guess, and 50 ohms or so is quite high relative to the resistance of several heaters in parallel. I would say that's quite loose.

I assumed that's the kind of direction the OP was coming from. Better to be positive about understanding than precious about ignorance, I think.

Others appear to have been misled by the usage, so in this company you do have a point. "Misuse" is a bit strong though. Intelligible, certainly in common usage, and in the specialist sense too, I thought, with not much effort.
 
Yero,

Is your question about "lifting" the filaments, as in applying a DC to them to bring them closer to cathode voltage?

Or is it about seperating the filament circuit from a hard ground point, ( so that it can be "lifted") at the winding (centertapped or end grounded)?
 
PlasticIsGood said:
A relative matter, I guess, and 50 ohms or so is quite high relative to the resistance of several heaters in parallel. I would say that's quite loose.
The correct comparison is with typical valve impedances of k ohm. The aim is to stop the heater from picking up any voltage significantly different from valve electrodes, especially the cathode. A truly floating electrode could pick up low velocity electrons and gradually become negative, or it could suffer from secondary emission caused by high velocity electrons and become more and more positive. Many datasheets therefore specify that the DC resistance between heater and cathode should not exceed a few 10's of k, except for the special case of a CF, LTP or cathodyne.

So from this point of view, 50 ohms is almost the same as a short circuit.
 
I want to build franks 6sn7 preamp. I have read that the 6.3 filament supply should be floated.

You have the terms reversed. "Floating" means the heaters are not referenced to any DC voltage. They should NEVER be left floating.

The simplest thing to do is tie the heater circuit to ground by grounding the center tap of the 6.3V secondary. Lacking a CT you can connect a couple of 100R resisters and tie the mid point of those to ground. Next step up in sophistication is to tie them to an elevated voltage. Next step is to use a pot in place of the two resistors you you can adjust for best balance.

But never leave them floating.
 
The correct comparison is with typical valve impedances of k ohm. The aim is to stop the heater from picking up any voltage significantly different from valve electrodes, especially the cathode. A truly floating electrode could pick up low velocity electrons and gradually become negative, or it could suffer from secondary emission caused by high velocity electrons and become more and more positive. Many datasheets therefore specify that the DC resistance between heater and cathode should not exceed a few 10's of k, except for the special case of a CF, LTP or cathodyne.

So from this point of view, 50 ohms is almost the same as a short circuit.

I agree with this analysis. The cathode to heater current is _tiny_ and so by Ohm's law almost any resistance will work.

I have one amp where I did in fact use a dead short. I let the heater circuit float except I use a piece of wire on a power tube sock to short the cathode to one heater pin. This biased to the entire heater loop to the same voltage as the cathode. Worked great for killing the hum and the parts found was a small as it gets.

The key to minimizing 60Hz hum from heaters is to experiment. There are many known tricks and which one will work for you on your amp? No one can know in advance. Even DC heaters may not work some times
 
The correct comparison is with typical valve impedances of k ohm. The aim is to stop the heater from picking up any voltage significantly different from valve electrodes, especially the cathode. A truly floating electrode could pick up low velocity electrons and gradually become negative, or it could suffer from secondary emission caused by high velocity electrons and become more and more positive. Many datasheets therefore specify that the DC resistance between heater and cathode should not exceed a few 10's of k, except for the special case of a CF, LTP or cathodyne.

So from this point of view, 50 ohms is almost the same as a short circuit.

You're right to point out that common mode impedance is important, but you may be wrong to suggest that differential mode can be ignored: there can be more than one "correct comparison".

I'm sceptical about your hypothesis for several reasons:

The heater is different from the valve's electrodes in that it is insulated so the effective resistance is several orders of magnitude higher than "typical valve impedances of k ohm".

The heater is enclosed within the cathode which forms a Faraday cage, which isolates it from other electrodes, so only the cathode is significant.

The hypothesis doesn't explain why resistors of such low values are used to establish the heater centre ground. Why not use the max allowable value, which would presumably be several k ohm at least?

It ignores the contribution of heater/cathode capacitance, which is the principle conveyor of heater hum and other noise to the cathode.

It doesn't explain why any of the effects it claims are bad, except for the assertion that a totally floating heater circuit may continue to "pick up voltage" indefinitely...presumably until the maximum heater/cathode voltage is exceeded. This seems implausible because I believe an equilibrium voltage would soon be established.

It ignores the fact that there are usually several heaters in parallel, belonging to different valves.

It ignores other considerations originating in the heater circuit itself, such as common mode hum and other noise.

It abandons concern about differential mode impedance.

I would be pleased if you could demonstrate your case. Perhaps you could derive from your hypothesis a method for calculating the ideal value of resistors to use? Then the relative significance of all the parameters would be plain.

My view is that, for common-mode AC at least, it is sufficient simply to consider heater/cathode capacitance. Calculations using this simple expedient appear plausible.

Note that by this account the common mode impedance of the heaters depends on frequency. A good test of the efficacy of a method would be to consider a gain stage with unbypassed cathode resistor, followed by a cathode follower, with both heaters sharing the same heater circuit. How much hum would result if the heater ground were one-sided? How much of a 20kHz signal would be fed back to the gain stage cathode from the CF, via the heater circuit? How much do these answers depend on the resistance of the heater ground connection?

Can you use your hypothesis to calculate answers to these questions?
 
PlasticIsGood said:
I'm sceptical about your hypothesis for several reasons:
Your scepticism is noted. Read a few valve datasheets, or even textbooks. Isolated electrodes in valves can and do pick up voltages, as I described. The heater is not entirely enclosed in the cathode.

The reason people often use lowish value resistors across the heater circuit with a grounded junction is to help short out capacitive noise from the mains transformer, either incoming noise or rectifier spikes. A secondary CT is better, but can cost more. For the purposes of establishing a reference level anything below a few 10's of k is fine. For shorting noise the lower the better, but too low starts eating up heater power so a few hundred ohms or so is a compromise.

So there are two issues in play: establishing a reference, and grounding noise. In the context of the former, a few hundred ohms is a short circuit not a 'loose' connection as I believe you said. Forgive me if I have misunderstood you.
 
And the coupling of heater voltage to control grid can become noticeable, due to amp wiring, and internal tube and valve holder construction. And the impact can be exacerbated by the heater containing additional frequencies, such as from HT rectification or flat-topped mains in your house. A humdinger pot can certainly help to null out the contribution from each heater lead transfering to the control grid - Merlin has presented a dramatic example of this impact.

Note that dc elevation of heater would only alleviate any hum contribution from the dc resistance leakage between cathode and heater, and would not influence other coupling mechanisms such as capacitance between heater and cathode, or control grid coupling. DC elevation also requires the applied voltage not to have significant AC components on it (eg. some bypassing of a HT divider is needed).
 
Your scepticism is noted. Read a few valve datasheets, or even textbooks. Isolated electrodes in valves can and do pick up voltages, as I described. The heater is not entirely enclosed in the cathode.

The reason people often use lowish value resistors across the heater circuit with a grounded junction is to help short out capacitive noise from the mains transformer, either incoming noise or rectifier spikes. A secondary CT is better, but can cost more. For the purposes of establishing a reference level anything below a few 10's of k is fine. For shorting noise the lower the better, but too low starts eating up heater power so a few hundred ohms or so is a compromise.

So there are two issues in play: establishing a reference, and grounding noise. In the context of the former, a few hundred ohms is a short circuit not a 'loose' connection as I believe you said. Forgive me if I have misunderstood you.

I haven't argued that there is no heater cathode leakage, there is and it is specified in datasheets, and measurable. It is very small. There are more than two issues at play, but two's a worthwhile advance from one.

The issue for me here from the start has been the OP's use of the word "floating". I took this as an encouraging observation for a novice but warned against ambiguity. Others have taken a strict line, claiming that there is a true unambiguous meaning that should be used exclusively for fear of misunderstanding. I understand the importance of jargon, but feel that plain English should also be acceptable in a DIY forum as long as it is intelligible, albeit with a little effort, and that electronics engineers should ease up a tad, and not play dumb in the face of common language.

My point was that the circuit is, in some sense and to some degree, floating on the ground resistance. I had in mind wrt to differential mode, low-impedance signals, such that typical centre-grounding resistance is large compared to resistance around the heater circuit. Whether or not this is floating or not in the strict sense the strict people wish to apply is of no consequence: I understood what the OP meant. I think I can see why he used the word "floating". I introduced "loose" to distinguish between "cast adrift" and "loosely tethered". It is arguably a specious point, because differential mode signals are cancelled by the cathodes as long as the centre ground holds firm, maybe, ish. But a novice might not know that.

My only question in the thread remains unanswered, btw.

I've read some books, thanks, and checked with datasheets and a real valve before I wrote what I wrote. I don't believe the answer to my question is in any of them.

How can I calculate the best reference voltage for a heater?

Thanks for engaging with the technical issue: it's been useful to refresh my mind, and learn a little, too. Please reconsider, in your own mind, the importance of heater/cathode capacitance, and downgrade emissions to "rarely significant".

On the social issue that originally attracted me, I feel resigned to ploughing a lonely furrow, paradoxically. I'm quite good at helping novices, but my method stops working in the presence of sticklers.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.