That high value resistor was suggested as a kludge to help with a bad design demand, that the heater supply be compatible with a random external device. Since this is DIY, we don't have to accept that as a requirement.
In the Dyna case, "high value" means large enough to not significantly load whatever foolishness the attached device is doing, and small enough to pull the heater down near signal ground voltage. But that's spinach and I say heck with it.
All good fortune,
Chris
In the Dyna case, "high value" means large enough to not significantly load whatever foolishness the attached device is doing, and small enough to pull the heater down near signal ground voltage. But that's spinach and I say heck with it.
All good fortune,
Chris
Yes, so can I just flip a coin, or close my eyes and pick any K of the 3 for this resistor to heater? Or does every K have to have its own resistor?
Last edited:
Guess I'm not sure what you mean. The "high value resistor" is intended to connect between the heater winding center tap and signal ground, in place of or in parallel with the .02uF capacitor.
Hope that's a good enough description.
But I still say it's spinach. Best results are with a defined, optimum DC voltage, bypassed to signal ground with sufficiently small impedance at both DC and AC.
All good fortune,
Chris
Hope that's a good enough description.
But I still say it's spinach. Best results are with a defined, optimum DC voltage, bypassed to signal ground with sufficiently small impedance at both DC and AC.
All good fortune,
Chris
That's different from DF96's post #37 proposal.. " the correct solution is a high value resistor from the cathode to the heater supply. A floating heater supply will go to whatever potential is set by the leakage paths it sees and the charge it happens to pick up inside the valves."
You jumped in there in his stead... it seems. He suggested a H-K resistor not a cap bypass resistor. If we put in the H-K resistor, what do we do with the CT connection?
Last edited:
But some commercial amps still put that cap serial to the heater center tap in amps without external heater socket taps like the ST70 and claim it has it's advantages. I was met with hostility when asking one amp designer about the advantage and I still never got a response. 😱
My take on it is that it clearly works but it really isn't ideal. As an engineer (which I am far from) I would think one would want to eliminate any and if not all variables to the active devices operating points by defining all the electrodes within the envelopes DC conditions. One could theoretically define the DC conditions of the ST70 heater circuit but they are set by parameters which are sometimes hard to define and therefore I would think inconsistent. Not a great design but hey it works right? 😛
Either you or I missed his point completely. The reason for him proposing putting a high value R from H-K is because the CT isn't being used and the the heater supply is floating. If the CT was being used you wouldn't need the R from H-K then now would we? 🙄
DF96 is quite knowledgeable and certainly didn't intend any resistor _directly_ between heater and cathode. That would be silly.
Perhaps. It could be he didn't complete his thoughts.... if the cap is to remain, then a H-K resistor needs to be added.... words in his mouth? If so, then once a heater biasing voltage is applied to the winding will that voltage find its way to every cathode with the new resistor connection?
From my POV it looked like he did finish his thoughts.
The cap doesn't remain, it is replaced with a resistor was his exact thought, or at least my understanding of his thought.
The two 3.15vac "taps" are out of phase by 180 degrees, the current on each phase won't be exactly be equal so there will be a small amount of current flowing on the center tap (neutral), a high value resistor will develop a voltage across it and set a DC reference. The high value resistor can easily be dominated by another DC reference if a preamp was ever introduced via the external sockets.
The cap doesn't remain, it is replaced with a resistor was his exact thought, or at least my understanding of his thought.
The two 3.15vac "taps" are out of phase by 180 degrees, the current on each phase won't be exactly be equal so there will be a small amount of current flowing on the center tap (neutral), a high value resistor will develop a voltage across it and set a DC reference. The high value resistor can easily be dominated by another DC reference if a preamp was ever introduced via the external sockets.
Last edited:
It isn't a shorting resistor.20to20 said:
It dissipates charge, not creates it.
It connects to the heater winding CT.
If sufficiently low resistance, then yes it is a short. That is unwanted, because it is assumed that the heater wiring has a reference voltage and is not floating. If you are supplying just one valve heater and do not care about charge pickup inside the envelope then you could leave the heater supply floating; still not a good idea in my opinion, but you will probably get away with it.
You appear to be confusing poor h-k insulation with a resistor designed in. They are not the same thing. Think about the current loops and voltages.
As I said, the resistor goes to the CT.
If there is no CT, then the resistor could just go to one side of the heater supply. Note: we are preventing the floating of a heater supply so by definition the supply has no other connections apart from the heaters. Hence connecting it to one cathode does no harm and does some good.
You can choose. Or just put the resistor to chassis instead of a cathode.
In their ECC83 datasheet Mullard specify that the DC resistance in the circuit between heater and cathodes should not exceed 20k, except in the case of LTP.
I originally suggested a resistor to ground in parallel with the useless 0.02uF cap. Then someone said that the intention was to allow the heater supply to float up to the cathode voltage via h-k leakage - that is when I said that the correct way to do this is with a resistor from the cathode. I don't really care which is chosen. The important thing is that the heater supply has a DC reference, and good quality NOS valves have good h-k insulation so they don't care too much what that reference is.
Provided there was only one such resistor and the heater supply was otherwise floating then it would do no harm.Chris Hornbeck said:
The heater supply needs a bias (which may be zero). Therefore there must be one and only one DC connection to something with a defined voltage. "One" to establish the bias; "only one" to prevent hum currents.20to20 said:
Whether the cap remains or goes is irrelevant, as the cap does nothing; it is too small a value to have any useful effect on noise and too low leakage to have any useful effect on DC conditions. I can only assume it was included either because the designer didn't know what he was doing (unlikely) or he assumed that constructors would expect to see something 'grounding' the heater supply so included it even though he knew it did nothing.
Fair enough. We should include a good bypass at audio frequencies to eliminate any signal modulation of heaters to signal ground. Downside is the drain resistor appears as a load on signal.
I still say it's spinach.
Thanks, as always,
Chris
If we assume perhaps 200pf for power transformer interwinding capacitance coupling noise from the power line into the heater winding, then 0.02uf to ground forms a divider attenuating that noise by a factor of 100. Why would this not be useful?
A 20k resistor might be more useful - or even a short (if the heater supply doesn't need to be used elsewhere). I once found that several uF was insufficient, but a short stopped rectifier spikes from getting into the heater supply - so grounding the supply was better than elevating it.
- Status
- Not open for further replies.