KT66 Rocket Build

[Markusv]

So basically the KT66s are running in triode mode.?

Bob

Seems to be what he is suggesting
Very loud for Triode mode though.

And what voltage separation do you need to go to real pentode mode then?
Or am I missing the point here?

Markus V

[Blind Lemon]

I think I would try to get the B+2 down and see what it sounds like in pentode. You just have to get the screens below the plate voltage.
Bob

[Blind Lemon]

Oh and hey RJ thanks for doing the legwork on this.

BL

[RJ Guitars]

Hey man, this is very interesting stuff... still kinda absorbing it all but glad to know it and pass it on to anyone equally curious.

I am going to think of these amps in the quasi-triode operational mode... not that I have authority to make up such titles, just that it's mostly like a triode amp but still a bit different.

I can't quite convince myself that if we lower the screen voltage below the plate voltage that it becomes a true pentode amp because I would expect that a true pentode amp would have a substantial jump in output power. I wouldn't expect that here but I might be surprised yet again.

One thing that does seem significant to me is a repeat of what my friend Jason wrote earlier, there are several guitar amps that operate in this mode. I've just been oblivious to them. Tweedeluxe had the fortune (or misfortune however you want to look at it) of taking notice of something that just seems wrong.

I am pondering a few things.
1) Is there any reason why we might see this phenomenon in a cathode biased amp and not in a fixed bias amp?
2) How is this part of the formula that produces that superb clean tone that seems to be part of these amps?
3) This seems to be caused primarily by the fact that the KT66's really require a lot of plate current, unlike a 6V6 or an EL84.
4) Other stuff I haven't thought of yet...

Here is a link to an interesting read on KT66's operating in Ultra Linear mode that has a lot of relevant info on triode and pentode operation.

https://tubeamparchive.com/files/ul_100.pdf

It's been quite cool to follow this thread and figure out what those weird observations really are telling us... with some part of me still saying there are parts of this that I haven't quite got a grasp on yet.

rj

[RJ Guitars]

I put a Sovtek 5AR4 rectifier tube in and the Shugang KT-66's then measured voltages.

B+1 = 324
B+2 = 322
B+3 = 256
B+4 = 224
B+5 = 218

V4 & V5 plate voltage = ~318
V4 & V5 screen voltage = ~321

So the same story, not quite as extreme. I am using a Stancor output tranny that is slightly lower DC resistance so it probably all makes sense.

seems like a good place to leave things for now...

rj

[RJ Guitars]

...from my friend Matt of the AudioNerdz group - this is cool!

...As drawn this is
not really pentode operation, the screen and plate of the KT66 are
really triode connected by the 1K resistor plus the DCR of the Hammond
and output coil. There is no AC ground on the screen like there should
be for pentode operation so the screen and plate track B+ together.
Matt

I looked at the schematics to see if I had drawn it the same as the Rocket with EL84's. It looked the same to me. Today I looked further and I see that on an EL-84 tube, pin's 1 & 8 are internally connected. A KT66 has a similar internal connection but they don't connect to the screen so nothing found there.

I've been looking at classic schematics like the Fender 5F6 Bassman and it seems they are the same. As Blind Lemon mentioned earlier, they all go back to a filter cap and ac ground through that.

So I'm not 100% sure if I'm buying into the full notion of triode operation yet... and thus still don't fully understand why the amp would work so well.

rj

I'm going to edit notes into this post until some one else responds...

Here are interesting notes that mention the same topic...
http://music-electronics-forum.com/arch ... -2974.html
http://archive.ampage.org/threads/4/gah ... age-1.html

[skyboltone]

Bob,

As drawn this is
not really pentode operation, the screen and plate of the KT66 are
really triode connected by the 1K resistor plus the DCR of the Hammond
and output coil. There is no AC ground on the screen like there should
be for pentode operation so the screen and plate track B+ together.
Matt

I think this is right. In fact I've been saying as much in the last dozen screen voltage discussions we've had. I think it's worth noting though, that this voltage difference is at idle. I think we get a better idea of what's going on when measuring voltage (scope or VTVM) in a dynamic way. While the amp is working hard. The lack of screen glow would testify to this as well.

[RJ Guitars]

Bob,

As drawn this is
not really pentode operation, the screen and plate of the KT66 are
really triode connected by the 1K resistor plus the DCR of the Hammond
and output coil. There is no AC ground on the screen like there should
be for pentode operation so the screen and plate track B+ together.
Matt

I think this is right. In fact I've been saying as much in the last dozen screen voltage discussions we've had. I think it's worth noting though, that this voltage difference is at idle. I think we get a better idea of what's going on when measuring voltage (scope or VTVM) in a dynamic way. While the amp is working hard. The lack of screen glow would testify to this as well.

Dan,

Thanks for your contribution. I am inclined to believe this occurs but unfortunately it's because it's been said by folks I recognize are much smarter than I am instead of me really understanding it.

Looking at a hand full of schematics today it seems that this may be pretty common. However, at first glance it is counter intuitive to me... verifying my true novice status!

Your confirmation that this is correct adds to the confidence that all is well, I'm just way low on the learning curve for this topic. I recall that your understanding of vacuum tube technology goes pretty deep... I read something about high voltage power systems that you wrote and it was clear to me that you were on a higher tier than the average amp hack like myself. Please feel free to enlighten us with an understanding of why the electrons don't just run away from the plate, hammer the screen with a high current, melt the tube down, game over...

I did get a note from a moderator over on "The Electronics Forum" and he said:

"Higher screen voltage is exactly what I would expect when you have more resistance in series with the plate (and much more current current flowing through that resistance). This is typical of pentode and UL outputs.

On something 6L6-ey like the KT66, extra screen voltage is probably helpful, although that help is only if you're using it (i.e., a somewhat lower plate load). But The Ten Percent Rule states that that advantage is negligible. (It also states that your screen voltage is essentially equal to plate voltage, if you're measuring, like, 320V screen and 305V plate. Not enough to matter.)


Now I understand why we have lower plate voltage but this "Extra screen voltage is probably helpful" and "The Ten Percent Rule states that that advantage is negligible..." does not compute for me, and what is this Ten Percent Rule?

A great exercise for me to work through but a bit of an intellectual stretch for me to wrap my head around. Once I finally get it I know I'll be embarrassed that I was so thick.

Thanks again.

rj

[Kregg]

The very fact that you ask these questions is an affirmation of your intelligence. The worst place an amp dude can end up is with the cork sniffing, comic book guru who thinks he knows everything from the confines of his mother's basement; which he aptly refers to as his lair. Make no mistake, your query has taught me a thing or two and I am enjoying this thread immensely.

[M Fowler]

Kregg,

How did you know I have been cork sniffing.

Mark

[Kregg]

Kregg,

How did you know I have been cork sniffing.

Mark

Ha ...

[RJ Guitars]

Well,

I have a little better understanding after another note from a moderator over on "The Electronics Forum".

Thanks Tim, I am just starting to wrap my head around this... I just hadn't thought about it before but I guess it really has to be this way since the current draw is so high. It just wasn't intuitive to me that the screen can be higher than the plate without blocking all the signal from getting there to the plate and drawing current on the screen from the plate.

Ahh, well that only happens when the plate is around 100V. Imagine you're an electron in a tetrode or pentode. You get burned off the cathode, kind of moping around, then you find a hole in the grid, you feel more and more pull towards the screen grid, then you zip past it. (Some of your electron buddies weren't so lucky and hit a screen wire instead.) As you pass the screen, the electric field reverses and you start slowing down. But you gained so much speed approaching the screen, you're still heading smack into the plate. You get absorbed and become plate current. It's only when you really try to slow down the electrons (when Vp < 100V, so after accelerating fom the cathode and decelerating to the plate, they're going fairly slowly) that plate current drops and screen current shoots up at a dangerous rate (which suggests it's safer to have a lower load impedance than a higher load, although speakers don't allow us to be so picky). There's a little more to it than that, like secondary emission, which peaks around 50-100V in most tubes. Secondary emission is when electrons smacking the plate knock out other electrons (since, being a metal, it's holding a mess of electrons prisoner). These electrons are deflected either by adding yet another grid (the suppressor) or just by design (as in beam tetrodes). Still, it's part of the reason why tetrodes stop working in the 50-100V range specifically.

By the way, what is the 10% rule?

I don't know if anyone else said it, but I always say it.

Note some things about tubes, for instance. Their properties are all consistent within a 10% range. Say you have 300V across a resistor, well then it's drawing 1mA (apparently it was a 300kohm resistor) and dissipating 0.3W. Say it goes up to 330V (a rise of 10%); now current is at 1.1mA. Power dissipated goes up to 0.36W, which is a rise of 20%, but isn't any bother to this 1/2W resistor (oh, and it was rated for 1/2W, too). So do you care? No, not really. Hence, the Ten Percent Rule states that "differences of less than ten percent are negligible". A tube rated for 30W plate dissipation won't at all mind if you run it at 33W (except for cheapass 6L6s that start glowing red at only 25W!), or a tube that's rated for 300V running at 330.

Obviously, this does not extend to cases where you take the difference, like with a long-tailed pair or Wheatstone bridge. If one grid voltage is 100V and the other is 105V, they're close enough, right? Wrong: 105-100 = 5V, enough to turn off a 12AX7. However, it does tell you that your bias resistors (if you were to bias those grids independently, which is a bad idea for exactly the above reason) will have to hit the 100-110V range, but since you need 105.0V to balance it and you just can't guarantee that accuracy, you are going to need more refined control, such as by adding a trimmer pot.

Remember that 10% is a ratio (10% = 0.1), so it also only applies to ratios. Fortunately, lots of things can be compared as a ratio. Resistors are a unit ratio (volts per ampere), but voltage dividers are a unitless ratio (V/V or ohm/ohm) because you can compare output to input voltage, or bottom resistance to total resistance (Vo / Vi = R2 / (R1+R2) after all)....

thank you Tim,

rj

P.S. Tweedeluxe, sorry for the extensive hijack. Can we resume the direction of things toward the status on your build and how those clips are sounding.

[tweedeluxe]

RJ - this isn't a hijack at all! I'm grateful to everyone for helping me figure this out. I should have clips posted later this week.

Tim's explanation makes perfect sense to me, but not knowing enough about electronics, I didn't want to just assume that electrons wouldn't ALWAYS travel to the most positive node. I think it's funny that a pentode sort of "fools" the electrons. And a beam pentode like the KT66 is probably a great prankster.

BL - I tried to run them in pentode mode with more screen resistace. I went up to I think 4k7 as a "shared" screen resistor (so roughly equivalent to 9k4 on each tube in idle - I think), but I couldn't get Vg2 lower than Va. I didn't even try playing it because I thought that was too much screen resistance.

Maybe I should have. As is, the amp isn't extremely loud, but I thought it was due to the relatively low 315V B+ and HOT cathode-biased KT66s.

Skyboltone - thanks for your input. I wish I had a scope! This definitely won't be last KT66-based Rocket-type amp I build, so maybe I can gather that data sometime in the future.

[Structo]

Very interesting topic.

On the small 6V6 amp I just built, I too noticed that the screens were at the same voltage as the plates.
I have 470R screen resistors and at one point I even added more resistance to the B+ line going to the screens and it still didn't change!

I was starting to think it was just the 6V6 tubes since I hadn't used them before.
This amp is also cathode biased. Does that make a difference?

My D'Lite 44 has about a 5v difference between plate and screen. (6L6)

[Structo]

Dang on that first link where Bruce from Mission is helping Jag, he gives him the fix but it must be a schematic or something because it doesn't show.

On my 6V6 amp the Plate voltage comes off the first 40uf filter, then there is a 4H choke, the Screen voltage comes off that.
It tried adding a resistor in series to the screen resistors and it didn't do a thing to lower the screen voltage.


Here's the power supply.