Although, just to ensure I have some tunes over the festive season, I'm tempted by the short-term fix of putting 22R across the 4.1R on each of the LM317s, to increase the current through the EH valves to something closer to 3.65mA...
Where are you, Maplins, when I need you?
Alex
Yes, of course - a 100R pot across the resistors I have now would do this quite neatly. I am just reluctant to do too much tuning for individual valves.
For now, my plan is to use fixed resistors to increase the current to the EH 6922s so the voltages are closer to the data sheet value, and then in due course to change to voltage regulation.
Alex
Valves intended for a voltage heater supply should be fed voltage, especially when near-equivalents are likely to have different current draw. Current supply is OK for valves designed for that use, or valves which you know are all identical and you are not too bothered about correct heating.
My recent experiences with the filament undervoltages on my VSE preamp have reminded me of this bit of Allen’s philosophy:
Every point of the circuit that is required to “move” must be able to move, free and fast; and those points not meant to move must be tied down very tight! (from the Tube Preamp Cookbook)
The RTP3 is the ultimate embodiment of this approach: every stage has a current sink on its cathode and another triode on its anode, locking down the operating conditions of each active device. This means that circuit operation is less sensitive to valve parameters, so even a tired old set of valves (or, in this case, a set with emission well under spec) will give at least good sound. As has been pointed out in this thread, though, low emission will skew the transconductance curves, and so increase distortion, and – as in my case – being barely able to achieve the quiescent current will eventually lead to noise.
Alex
Every point of the circuit that is required to “move” must be able to move, free and fast; and those points not meant to move must be tied down very tight! (from the Tube Preamp Cookbook)
The RTP3 is the ultimate embodiment of this approach: every stage has a current sink on its cathode and another triode on its anode, locking down the operating conditions of each active device. This means that circuit operation is less sensitive to valve parameters, so even a tired old set of valves (or, in this case, a set with emission well under spec) will give at least good sound. As has been pointed out in this thread, though, low emission will skew the transconductance curves, and so increase distortion, and – as in my case – being barely able to achieve the quiescent current will eventually lead to noise.
Alex
Do you have a theory then, for why you have quiet operation of the other channel? Has there always been noise in the one channel, just low enough to be tolerated?
A small drop in heater voltage may reduce noise, by cooling the cathode and hence the cathode space charge. A bigger drop in heater voltage may push the valve towards the temperature-limited region, where space charge smoothing no longer works so you get the full effect of shot noise - so more noise. 10% is the usual range, and you are well outside that for some valves.
That's a hard one to answer. I can only suggest that both channels were only marginally managing to sustain the intended standing current, and over the course of time natural ageing caused one channel to drift over the edge.
Alex
Guess you're well aware of the heaters' PTC characteristics, so the voltage drop is lower than 5.2V in your case.
Sorry, it can't be calculated, but rather measured - or read from a curve that itself was derived from measurements of a sufficient number of specimen.
Best regards!
OK, I was assuming linearity, which is of course not justified here - my first step is always to start with a linear approximation....
Just a back-of the envelope estimate to see if the filament current in the EH valves might fit some other specification. Of course measuring it would be better, but I'm more of an armchair scientist (in more ways than one) these days.
Alex
Last edited:
Put all the 300mA tubes in one channel and all the 365mA tubes in the other. To have two different tubes with the same type number but different heater currents is totally wrong. It really suggests the odd one is not what the number is. There are standards for typing a tube. To have 365mA for a 300mA tube suggests larger cathode that requires more heater wattage. A Larger cathode suggests more current to larger plates. I would ditch the odd tubes and get your set uniform for whichever type you want.
Perhaps I didn't make myself clear - the heater current appears to be consistent for a given brand, with the EH 6922s appearing to be working at around 365mA, and the Sylvania ones at 300mA.
As I said earlier, I plan to change from current to voltage regulation for the heaters, since this is far better for tube rolling.
Alex
No, each valve has its own current regulator, so twelve LM317s in the box. I certainly wouldn't have liked to assume that all the filaments would have the same voltage drop. Besides, Allen's circuit has the filaments biased at three different DC voltages to control the heater-cathode voltage difference.
Alex
Alex
OK, the earlier statement about not tuning each regulator had me in left field. So if you can satisfy each heater's needs with an adjustment, why not do that and see if it kills the noise? That would seem to be the final test for its source.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.