Davis 201A PA Conversion

[Pdavis68]

Okay, looks like I can just run a resistor from HT to the cathode to provide the extra current to keep the LED conducting in a linear region. Sweet. That's easy.

One of my biggest problems is going to be undoing the bent pins. Since a couple of these tubes are grid leak biased, they bent the cathode pin into the center ground post of the socket and then soldered it to the ground post. So that's going to be loads of fun to undo. I'm not even entirely sure how to do it. I've got a solder sucker and solder wick, but my solder wick sucks (can someone recommend a good solder wick?)

This is actually an area where my knowledge of electronics falls short. So when I do the LED bias, the LED is going to bring the voltage at the cathode up whatever the forward voltage drop is. About 1.8V or so... So if I put a 33K resistor from say 250V to the cathode, why is it that it supplies current but doesn't change the voltage? Is it the LED that forces the voltage to remain at 1.8V or whatever it is?

[Pdavis68]

Well, R23 is definitely 220R. I'm thinking it ought to go back to 1.5K, the way it was. And the NFB tap should go back to 16 ohm. I can replace R29 with a pot and then I can dial in the feedback I want.

I was going to start replacing the grid leak resistors on the inputs tonight and couldn't find any 1M resistors. :-( Got some on the way, but I thought I had some. Hopefully they'll arrive on Saturday.

[Enzo]

The diode has a natural junction voltage. The resistor is there to limit current, not to set voltage.

[Dave H]

Well, R23 is definitely 220R. I'm thinking it ought to go back to 1.5K, the way it was.

220R may well be the best value. I'll have to look at it again. I've been working from the 12AX7 data sheet when the tube is a 6EU7. Doh!

[DRH1958]

220R may well be the best value. I'll have to look at it again. I've been working from the 12AX7 data sheet when the tube is a 6EU7. Doh!

Unless you made a mistake by entering pin numbers for your calculations, you will be OK. The 6EU7 has the same specs as the 12AX7, just a different pinout.

[DRH1958]

2.2k I'd believe
That and the 16R tap will get about 14x more NFB.

Just for my understanding, the amount of 14 times comes from the 1.4 times extra from the change in OT tap and 10 times comes from changing the resistor from 220R to 2.2k? So together it comes to 14?

[Pdavis68]

220R may well be the best value. I'll have to look at it again. I've been working from the 12AX7 data sheet when the tube is a 6EU7. Doh!

Would that be at the 8 ohm tap or 16 ohm tap. Because the original values were 1.5K with it coming off the 16 ohm tap. And given that it's a PA, I would imagine it was intended to clean it up pretty well. I'm not terribly familiar with the direct coupling stuff. I may need to set up a sim of it to get a better handle on what's going on.

[eschertron]

Just for my understanding, the amount of 14 times comes from the 1.4 times extra from the change in OT tap and 10 times comes from changing the resistor from 220R to 2.2k? So together it comes to 14?

Yes. That was my seat-of-the-pants reasoning

[DRH1958]

I was going to say fire it up and see what it sounds like and then read your first post. It breaks up too soon for your taste, so maybe change one input to the standard cathode bias type to get the bias voltage close to center including changing the grid stopper and 1 meg grid leak, decouple the first stage, add your screen grid resistors and see what happens. Maybe the rest of this won't matter if you like the sound. Then take voltage measurements to make sure all is well. Since the NFB is on the 8 ohm tap, you have the option to go up or down on the voltage applied to the loop. Another way to do it is to stack a small resistor, say 150R and then a 1k, or whatever, in series on the cathode of the stage and apply the NFB between those resistors. Then change the values of those to change the NFB. I like your initial changes you proposed in post #39. These seem to me to be the logical changes to get started. You can always adjust the NFB and tail resistor(s) on that stage if you don't like the sound.

[eschertron]

Well, R23 is definitely 220R. I'm thinking it ought to go back to 1.5K, the way it was. And the NFB tap should go back to 16 ohm. I can replace R29 with a pot and then I can dial in the feedback I want.

I was going to start replacing the grid leak resistors on the inputs tonight and couldn't find any 1M resistors. :-( Got some on the way, but I thought I had some. Hopefully they'll arrive on Saturday.

I did a quick load-line chart yesterday and discovered today that I was off by a decimal point. When I thought I was looking at 220R for the operating point, I was really looking at 2200R. I recant my confession (that the confessor already recanted upon), and stand by my claim that 2.2k is probably what you want in there for a 470k anode resistor. The distortion you're hearing may well be coming from that way-too-hot stage.

[Dave H]

2.2k is probably what you want in there for a 470k anode resistor. The distortion you're hearing may well be coming from that way-too-hot stage.

2k2 is fine for the driving stage alone but it is directly coupled to the cathodyne so it has to be biased with its plate at a low enough voltage to bias the cathodyne grid correctly.

[Dave H]

Would that be at the 8 ohm tap or 16 ohm tap. Because the original values were 1.5K with it coming off the 16 ohm tap..

I'd leave it on the 16 ohm tap. If you want the same feedback with 220R cathode as you had with 1k5 cathode and 100k feedback resistors change the 100k feedback resistor to 15k

[Pdavis68]

So here's what I've done so far, just to bring everyone up to date. I only had 1/4W 1M resistors and I don't want those in there permanently (leads are too fragile). I've got some 1W 1M coming today or Monday and I'll replace with those, so I've only worked on 2 of the inputs (the Mic 2 and Mic 3 inputs. I changed them to 1M grid leaks. I got the cathodes unstuck from the grounding post and I have one biased with a red LED with Vf=1.6V. The other is biased with a 1.5K resistor and a 47uF bypass cap (I was looking for a 22uF, but apparently I'm out).

I added another node on the PS with a 3.3K resistor and 10uF cap (I think that'll get me close to a 10V drop). I've moved all the ones that were on the 260V node to the 250V node (because it was a 1 wire change).

So on my plate for today:

1> Add the screen grid resistors (decide to start with 470R)
2> Move the second stage back to the 260V B+ node.
3> Fire it up and see what I have.

I'm probably going to take Dave's advice and move the NFB back to the 16 ohm tap and replace the NFB resistor with a pot (maybe 50K or 100K?)

After that, I'll make decisions about whether or not to change the plate load resistors on the preamps. I'm pretty sure I'll need to replace the tone stack, though I kind of dread that so it may be the last thing I do.

[Pdavis68]

First of all, overall, I think it sounds really good. I've got the 2nd stage off the 5th B+ node and back to the 4th where it was originally and I have the screen grid resistors in place.

But the problem is that frequency-dependent static I mentioned earlier. It's still there. It seems to be only E chords (major or minor) that causes it. So I guess if I just avoid E chords in my playing, I'll be okay. Those aren't used much are they? ;-)


I've attached an MP3 so you can hear what I'm talking about. Because it's recorded with my phone, it's not exactly a hi fidelity reproduction. This is the channel that's biased with the 1.5K resistor and 47uF bypass cap. The LED biased channel has the same issue. And it was there before my changes, so it's probably something I haven't touched yet.


The voltages changed pretty drastically. I added the screen grid resistors and the 5th B+ node so I now have:

284V, 269V, 247V, 220V, 196V

I'm not particularly concerned. Like I said, so far I'm liking the sound I'm getting, other than the static.

Any ideas on the cause?

[Pdavis68]

It seems to get worse as the tubes heat up. Not louder, but just happens more and not just with the E chords. Not nearly as frequency dependent after the tubes warm up. Bad cap somewhere maybe?

I've attached another recording (though you may need to turn up the volume a bit to hear it better).

Is this something I might more easily track down with my scope? I don't have a function/signal generator. Could I just ring out an E note and see the static in the signal? That way I might at least track it down to where in the signal chain it's taking place.

I've never used my scope on anything of this voltage. Anything I need to know? I have an old Tektronix 465 2 channel scope. The probe I normally use is a 10M ohm, 10x, 15.5pF.