screen tap ?

[shooter]

I went back and finished up the paperwork on a SE build and found an oddity.  I'm using and edcor OT that has a 40%tap for the screen but when I measure screen volts I get about 4 more than plate volts.  331 plate, 335 G2.
I pulled the tube, ohmed OT to plate and screen, they matched the tranny numbers, I ohmed the OT windings p1 to p3 = 21.2 and p3 to p5 = 16.2
what am I missing?
thanks

[HotBluePlates]

... when I measure screen volts I get about 4 more than plate volts.  331 plate, 335 G2. ... what am I missing?

The raw B+ is applied to one end of the primary and the plate is on the other end. The screen tap is some midpoint between ends.

The OT primary winding has some amount of d.c. resistance. The total resistance is that between ends, and there is less than total resistance between the B+ end and the screen tap.

Ohm's Law says Voltage = Current * Resistance. Most of the direct current through the primary is tube plate current. There's more resistance from B+ to the plate connection, so more voltage drop than to the screen connection.


(There's more total current from B+ end to screen tap, because there is plate current + screen current; however, the screen current as a percentage of total tube current is small compared to the increase in resistance from screen tap to plate tap, so the numbers don't offset and result in lower voltage at the plate in an ultralinear connection. Always.)

[tubeswell]

If you're concerned about screen dissipation, you can add always add a screen grid resistor in order to limit peak screen current.

[shooter]

Thank HBP, ohms law is what I missed.  I was thinking I should see similar results to taping G2 further down on the B+ rail like is done without a screen tapped OT.  My screen current is good, 7mA, and I do have a 1K installed.  Plate is about 100mA.

So to increase the voltage spread plate to screen, would I abandon the screen tap and use different B+ taps?  I have 394, 366, 337.  I'm trying to gain more headroom in the PA tube, (raising B+ on the plate, adjusting the bias to stay close to 35W plate dissipation ).  what I'm thinking is the closer the screen volts are to the plate volts the closer I get to triode mode vs. pentode mode.

thanks for the help.

[HotBluePlates]

So to increase the voltage spread plate to screen, would I abandon the screen tap and use different B+ taps?

Why would you want to do that? (I think your answer is below; see reply)

what I'm thinking is the closer the screen volts are to the plate volts the closer I get to triode mode vs. pentode mode.

Pentode mode & triode mode are not about the absolute voltages present at the tube elements, it's about the nature of those voltages.

In pentode mode, the screen is pegged to an unvarying d.c. voltage. The exception to this rule is that you might insert a screen resistor to drop screen voltage somewhat when screen current rises (which reaches a maximum at the moment the plate is passing maximum current during the positive peak of the input signal). This mode also has relatively good power sensitivity (how big an input signal is required for a given output power with a given load).

In triode mode, the screen is directly connected to the plate and its voltage rises and falls along with plate voltage during the signal swing. The exception is you might add a screen resistor, which causes a slightly greater drop in screen voltage compared to the plate voltage, due to ohm's law and the screen current passing through the resistor. This mode also has relatively poor power sensitivity (it takes a bigger input signal to create the same output power, with the same load, as you'd get if you used pentode operation).

What makes a pentode (and tetrode) have the characteristic "pentode curves" where plate current changes very little when changing only the control grid voltage is that the screen voltage has more influence than plate voltage on plate current (about 9-10 times more in common output bottles). So when the screen is nailed to a fixed d.c. voltage, you get full pentode flavor. When the screen is connected directly to the plate, the screen voltage is varying right along with the plate, so the tube behaves as if it were a triode (the control grid still has more influence over plate current than the plate/screen, but now the changing plate/screen voltage also causes plate current change).

Ultralinear operation is a mid-point (literally and figuratively) between pentode and triode operation. The screen tap is situated in between the unvarying d.c. voltage at the B+ end of the primary, so there is some middle-amount of varying voltage applied to the screen which is less than what would happen if the screen was directly connected to the plate. You also get more triode-like characteristic curves (with a different flavor of distortion) while retaining some of the sensitivity of pentode mode (harder to drive than pentode mode, easier to drive than triode mode).

So you see, it's not about whether to have your screen a 300vdc or 450vdc, but whether your screen voltage varies, and whether it varies as much as the plate voltage during the signal cycle.

[HotBluePlates]

I'm trying to gain more headroom in the PA tube

... a SE build ...

Define your goal of "more headroom" (a pet peeve of mine because I believe most of this phrase's use in relation to guitar amps & modifications removes almost all meaning from the term).

What I mean is are you wanting more clean power output? More clean volume from the speaker? Maybe you already have a built and functioning amp, but it distorts by "4" on the volume control and you want it to stay clean until "8" on the volume control?

If you want an actual increase of clean power output and/or volume from the speaker, SE was not the way to go. And unless a very-wrong load was chosen for the tube & supply voltage available, you won't get any actual increase in clean volume from the speaker (as measured by a dB meter with a test signal that measures some fixed reference level of THD at the speaker terminals) without changing to a more-efficient speaker.

This is because the output stage, its topology, supply voltage and load, determined the maximum power output (which is why a legit design starts with a specified goal of clean output power, then the output stage is designed, and then everything else working backwards towards the input jack). Once the output stage is settled, there's no significant increase in power or headroom available without changing the output stage (push-pull operation, more or bigger tubes, etc).

Some things give the illusion of "headroom": the relative insensitivity of triode mode requires a bigger drive signal from the driver/preamp, so switching from pentode to triode mode in an already-built amp gives less distortion and the volume control has to be turned higher for the same speaker-volume (which some people interpret as "more headroom" though the amp may not be cleaner and louder). Feedback around the output stage may appear to give more headroom, but it too reduces power sensitivity (how hard you have to drive the input-end of the feedback loop to get a certain power output at the speaker terminals). Preamp feedback may appear to yield more headroom because there's less distortion; same comments apply about sensitivity.

Figure out the output stage based on the clean output power needed. Once the output stage design is done, figure out what drive voltage is required to get your maximum power output. Figure out if you're going to use feedback around the output stage to reduce source impedance, widen bandwidth and/or reduce distortion (until the gain within the feedback loop runs out); determine the amount of that feedback, because it's going to raise the drive requirements of the output stage. Figure out the difference between that signal and the likely average signal at your input jack; this is the gain required. Figure out how many tube stages & what type you'll use (with some total gain in excess of what you just found was the minimum needed to drive the output stage). Add a volume control somewhere so you can turn down too-strong signals but also turn-up for too-weak signals. Figure out if you're going to trim unneeded gain (and maybe increase bandwidth) by using feedback in the preamp. Figure out if you're going to ad other features (like tone controls) that throw away gain in order to work and/or require make-up gain after.

[shooter]

HBP, thank you, I copied your reply's to fully digest.  You nailed what I was missing between triode/pentode!

my reasons/logic for thinking, "more headroom"; If I raise plate voltage, keeping dissipation below max, I would gain "room" for more AC signal to develop at the plate.  I see large voltage on OT's primary equating to OT's secondary yielding more current to drive speaker.  Currently plate is at 337, I wanted to take it to 400ish, I'm not looking for more clean, just a fuller sound, instead of thinner, not sure how to describe the sound yet, the sound is good, I'm just looking to see IF boosting plate volts will boost sound, not necessary volume.

I think all that will be cleared up when I go over the post. 

Thanks again!!

[HotBluePlates]

My reasons/logic for thinking, "more headroom"; If I raise plate voltage, keeping dissipation below max, I would gain "room" for more AC signal to develop at the plate.  I see large voltage on OT's primary equating to OT's secondary yielding more current to drive speaker.

What's the output tube?

[shooter]

It's a winged C 6550.  the JJ data sheet says 35w max plate, which I'm trying to keep around 33w.  G2 6w which I'm running around 2.4w.   the "typical" values on an old data sheet say 400 plate, 250 G2, and 80mA idle.  I'm at 337 plate/g2 and  108mA cathode now. 

[HotBluePlates]

the "typical" values on an old data sheet say 400 plate, 250 G2, and 80mA idle.

That's pentode mode. And with such widely-different voltages, you need a separate supply for the screen (you wouldn't ordinarily just drop B+ down 150v with a resistor to feed the screen, because the screen voltage would sag wildly in use).

What's your intention? do you want to run Triode, Ultralinear or Pentode mode?

Is this amp already built in some form?

[shooter]

The amp is....was done, I was simply filling in my voltage chart n that's how this post began.  I also had 3 gigging musicians evaluate it for me and of course I got 3 different results!  2 of 'em said they thought it should be a little "ballsier" but tonally they liked it.  So like most guys here, I went back looking for things I could "improve" on.  I am gonna try and put it like the "typical" class A from the data sheet so ,plate to 400, G2 to 300ish fixed to pwr rail, not UL tap,(thx for clearing that part up),  (I don't wanna mess with the B+ rail because I think that would mess with the pre-amp)  Hoping that will "make it a little more ballsy".  I've learned a lot here, but have a very long way to go.  I'm trying to go from prototyping to viable, useful builds.  This amp hopefully will be my "basic standard" for SE.  I have 2 more, 35-40W PP amp and a 35-40W hybrid that are still in the tuning stage. 

[HotBluePlates]

2 of 'em said they thought it should be a little "ballsier" but tonally they liked it.  ...  I am gonna try and put it like the "typical" class A from the data sheet so ,plate to 400, G2 to 300ish fixed to pwr rail, ... Hoping that will "make it a little more ballsy".

Well, they didn't do you any favors if they didn't give more info on what needed to change. More bass? More distorted? Louder? Stiffer response?

Anyway, you probably don't want to chase the data sheet condition:
  - 100v difference between plate and screen means you would prefer not to derive the screen voltage by way of a dropping resistor; the data sheet condition for 400v plate, 225v screen shows a >3:1 increase in screen current from idle to full output, which means if you drop 100v by way of a resistor, it will try to drop 300v across the resistor at full output. Actually, it won't get there, because the drooping screen voltage will also choke back maximum output, lowering plate and screen current.

  - The way the tube manufacturer envisioned you'd get the reduced screen voltage is to have a second, separate power supply for the lower voltage so that the voltage stays steady regardless of current demands. That's a pain in a modern amp, unless you're shooting for a screen voltage 1/2 the plate voltage and are using a bridge rectifier with a center-tapped primary on the power transformer (then there is a ready source for 1/2 B+ voltage, when you use series filter caps and connect the PT center to their midpoint).

If you think they were responding to the cleanliness of UL, you might just make an added power supply node with a choke for filtering, as found in most big guitar amps. Your biasing shouldn't change because the screen voltage will be similar to what you already have. Triode and UL mode reduce odd-harmonic distortion, which you'd think would be a good thing from reading write-ups about distortion. The truth is most guitar amps use pentodes, and we're accustomed to hearing its signature; triode and UL mode can sound "dark" or "lacking edge & bite" because of the missing odd harmonic distortion.

If you did the data sheet condition with reduced screen voltage, you'd need to change bias for a lot less G1 bias voltage (because the screen voltage went down). The upside is the output tube would be easier to drive, requiring less input signal to reach maximum output power.

[shooter]

Ya, I thought visual artists were "esoteric" when they talk,  musical artists have the same gene!
Looks like you will save me some soldering time, thx.  I think I will move the B+ tap to the 400, leave everything else alone n just evaluate, but I have a well pump and hard down car to get thru first!
Thanks again HBP

[printer2]

As much as I do not like hum, it might be the "ballsier" sound the two of them said was missing. Not that the hum was missing but the intermodulation distortion (IMD) components. How filtered is the power supply? With lower capacitance you get more IMD and a thicker sound. Not good for speed metal but if you are going for a bluesy sound it may fill things out.

[shooter]

thanks printer2, I did awhile back re-work the PS, moved the CT filament to the cathode of the PA tube, added a 3rd cap so now I have 3 100uf, plus a 47uF near V1.  She went dead silent for hum.  After I try the new B+ tap, I might go back to the original PS setup.