Single Ended EL84 tube amp plate voltage thoughts...

[kevtronic]

Using your drawing and not having the amp on my bench... This was my idea. "I" would start here:

Wow, thanks for the extra effort. I going to download it and check it better in a bit.

[kevtronic]

You may also borrow a tried and true idea, used by none the less than Marshall and Mesa Boogie
Dig and read (you may also post them here for comments) Mesa Boogie Mark 1 and Marshall JCM800.
Both are based on earlier simpler amps, the MB is basically a Twin Reverb channel, the marshall a somewhat modded Plexi, and both have an extra 12AX7 triode added at the front end.
Funny thing is that both amps sport *separate* jacks, the first one going through the added triode, so meant to scream, and the second one *fully* bypassing it and going straight into the next triode grid, for "normal" operation.
You might do the same, so you can both sell your cake and eat it.
Check the way they creatively use the jack internal leaf switch and precautions they take against popping and leaving some triode grid non referenced to ground.

Valve-bunny trail. I like that idea. The single triode did provide a great jazz/blues experience in the original set up. I will be looking at this more for sure...

[kevtronic]

Another thing I'll mention is that by padding down the screen with a 5.1k resistor you will be mitigating excess screen current that DOES happen when the grid is driven too much. Which you can certainly do with the series triodes. That high resistance will be unnecessary when the grid drive is brought down to a lower level and you'll get much better dynamics and more power from the el84. I'd still keep a resistor there, but I might lower it's value to 1k (only after the grid drive is brought down though). 330V is about max for a single ended el84 so I think your 310V should be fine. And I want to reiterate that you DO want to be able to clip the amp at least a little because there may be times when you aren't pushing a full guitar signal into it, but you still want maximum clean output. In order to amplify a lower level signal to full output you need more gain than is necessary for a typical guitar with the volume knob full up. This adds versatility to the amp at the expense of having the amps volume set lower than all the way up for max clean with a typical guitar. That may seem confusing, but it's been explained a couple of other times in the thread. I can understand not wanting the amp to break up at something like 2 on the volume control. It's aesthetically unpleasant. But as mentioned, it IS arbitrary. If the amp is putting out full clean power that's all the clean head room it has even if you adjust the circuit so the amp doesn't clip until set to 10. The only REAL problem with that much gain, as I mentioned, is driving the power tube into over dissipating the screen grid. You're pretty safe with that 5.1k resistor to the screen. But I also mentioned that you can get more and better performance from the amp by reducing both the amount of gain and that resistance value. So THAT would be the real reason to lower the gain. And of course that will also place the point of clipping at a higher number on the volume knob.

Nicely said, nicely said indeed. It kind of puts everything being said into one paragraph. Man, you guys are schooling me here - lol - but, that's a good thing

[kevtronic]

I did get around to doing some voltage testing on the amp this afternoon. The has pretty much been untouched from yesterday. I figured I'd post the voltages I have now, before any changes, as a before and after type thing (hopefully anyways). Kind of a lot of info, but here it is:

Sine Wave @ 1Khz (tenma)
All voltages taken with Beckman 310 Industrial

* Switch Hi/Lo works like Fender style Hi/Lo Input Jacks

Lo (150mV) -------------------------------------------

input jack: 150 mVac
12ax7(a) control grid: 75 mVac
12ax7(a) coupling cap (volume side): 2.37 Vac
12ax7(B) input grid: 0 - 2.3 Vac
12ax7(b) coupling cap to El84 input grid: 0 - 59 Vac

Hi (150mV) -------------------------------------------

input jack:150 mV
12ax7(a) control grid: 150 mVac
12ax7(a) coupling cap (volume side): 5 Vac
12ax7(B) input grid: 0 - 4.6 Vac
12ax7(b) coupling cap to El84 input grid: 0 - 71.3 Vac

Lo (300mV) -------------------------------------------

input jack: 300mV
12ax7(a) control grid: 150 mVac
12ax7(a) coupling cap (volume side): 4.75 Vac
12ax7(B) input grid: 0 - 4.6 Vac
12ax7(b) coupling cap to El84 input grid: 0 - 71 Vac

Hi (300mV) -------------------------------------------

input jack: 300mV
12ax7(a) control grid: 300 mVac
12ax7(a) coupling cap (volume side): 9.5 Vac
12ax7(B) input grid: 9.15 Vac
12ax7(b) coupling cap to El84 input grid: 71.5 Vac

======================================================

Plate to Ground: 309
Anode to Cathode: 300
Screen to Ground: 294
Cathode Resistor: 300R (actual 295 ohms)
Cathode Voltage Drop: 10.96 Vdc

I=V/R=10.96/295=0.03715A=37.15mA
P=VI=(300)(0.03715)=11.145W
11.45W/12W=0.92875 or about 93%

=====================================================

The tube/OT match up for load impedance, etc.
No, negative feedback loop.

The measurements given are from my Beckman 310 multimeter, but I had the scope hooked up to all the same points along with the beckman to view/verify on the scope. All the wave forms looked nice on the scope.

Haven't done much else yet...

It's looking like you've all given me enough ideas for two or three more diy builds. I got enough iron, tubes, etc for a couple push/pull builds with 6V6's or 6AQ5's, but not for another SE at the moment. So, after thinking some on what's been said (design wise) I'll just go for it on this one. The chassis is very small and crowded (but, very quiet humm, etc, wise). But rest assured the ideas given here are being added to my bag of tricks for future use. Thanks, man, seriously, thanks - too cool, too cool...

[kevtronic]

...I'd add a voltage divider between the preamp triodes...

Ok, I got ya, voltage divider - like variable potentiometer (volume control) or maybe static resistor voltage divider. I think I need pepto-bismol for my brain - lol...

[Mike Sulzer]

But if you put a volume control after each 12AX7 stage, you have many more possibilities, and you do not have change from one input jack to the other.

You may also borrow a tried and true idea, used by none the less than Marshall and Mesa Boogie
Dig and read (you may also post them here for comments) Mesa Boogie Mark 1 and Marshall JCM800.
Both are based on earlier simpler amps, the MB is basically a Twin Reverb channel, the marshall a somewhat modded Plexi, and both have an extra 12AX7 triode added at the front end.
Funny thing is that both amps sport *separate* jacks, the first one going through the added triode, so meant to scream, and the second one *fully* bypassing it and going straight into the next triode grid, for "normal" operation.
You might do the same, so you can both sell your cake and eat it.
Check the way they creatively use the jack internal leaf switch and precautions they take against popping and leaving some triode grid non referenced to ground.

[Mick Bailey]

I really like having a volume control after each stage. Some amps use a stereo control in a similar fashion, but the advantage of having dual volume controls is there's a subtle characteristic change when one or the other control is maxed (or not) and the other used to set the desired level. My main tube amp is now an fixed-bias SE KT88 and I got rid of the mid control to add an extra volume control after the second 12AX7 stage.

[Chuck H]

Ok, I got ya, voltage divider - like variable potentiometer (volume control) or maybe static resistor voltage divider. I think I need pepto-bismol for my brain - lol...

Yep. Once I saw your schem I saw that you already had one there. Of course you'd have to set it really low for clean tones. 71V into the grid of the el84 when it can only reproduce 10V clean... YIKES!

[kevtronic]

Ok, I got ya, voltage divider - like variable potentiometer (volume control) or maybe static resistor voltage divider. I think I need pepto-bismol for my brain - lol...

Ya, those three 70 marks are just screaming maxed out. I'm presently looking at the chassis and what is already soldered in versus connected via jumper wires and trying to figure out what to do. I really don't want to rip too much out. But, I could simply parallel two 3K resistors over the two there and two 100K in parallel with the plate load resistors. At the moment I want to solder the .01 cap that goes to the EL84 grid. Gotta free up some jumper wires...

[kevtronic]

Volume control numbers are purely arbitrarily, right? This is the great lesson Spinal Tap taught us.

Ok, great reading there. Being I've been trying to teach myself electronics (mainly music electronics) for a while now (and ever in the process of learning) I like it when things get wrapped up in a nice paragraph or two. Having a degree in computer science and one in computer information systems isn't exactly the same thing as an electronics education...

[Chuck H]

That would be close-ish (about 825r cathode and 50k plate) Still a little cold on the cathode and a small load at the plate. A good bit less gain than the values in the schem I modified. There just can't be very many parts in there to fuss about ripping things out. I make changes like this between cups of coffee But get the right components. Including the 250k pot and the 1k screen resistor. At least 1W rating for the parallel 12ax7 plate and cathode resistors.

Keep in mind also that I just drew that up as a strict interpretation of what you said you were after. You could also do like Juan said and copy a known circuit that has a good clean tone AND a good clipped tone. You'll still have all the clean power, but at a lower volume setting and increasing the volume will add clipping.

[kevtronic]

...Keep in mind also that I just drew that up as a strict interpretation of what you said you were after...

To be honest I'll probably use the schematic you provided for another build. Looks like a nice start and I want to mess first hand with an el84 SE amp with a paralleled preamp tube. It's only a matter of time before I come across the iron for another EL84 SE.

The amp I'm presently working on I've kind of pushed the limits and presently have the output tube with 330 Vp (plate to anode) and biased right at 100%. From what I'm gathering that's top end'ish, so I may back off just a bit. But, this provides me with 235 plate voltage for that 12ax7 as is (haven't changed anything on it yet) and if I back off on plate voltage in the power supply it reduces the voltage available to the preamp tube. And, I just don't have any room in the poor little chassis, so...

To tell you the truth I've never really thought much about biasing a preamp tube. I know they are biased, but the values seem so cut in stone and I've haven't had an issue with before (that I know of anyways). I've repaired many amps (solid state and tube amp). But, this is only my 5th scratch build. Oh, and I did do up a tube driven reverb circuit for my Guyatone GS440. And literally replaced every resistor and cap in the thing. Sad, it had carbon comp resistors and oil caps (except for the electrolytic). The resistors where straying pretty bad and the oil caps were getting leaky. Oh, well. It's one great sounding mojo filled amp any who

[kevtronic]

OK, it's looking like what I am running into here is a lack of understanding/experience in preamp tube biasing and manipulating gain stages.
So, this is where I am going to spend some time.

I found this website: https://www.ampbooks.com/
Has anyone used any of the various tube calculators there?

I figure this simple amp is a good place to bring my knowledge and expertise up a bit
No point in spending time swapping stuff out until I have at least a fair bit of knowledge on WHY I'm swapping it out.

Fer'sur, man, there's lots of opinions out there on preamp tubes the zero's and one's universe.
And, by the looks of it many school of thoughts and rules bending going on...

Like the thoughts I had/have which led to how I originally set up this amp.
1) that a higher cathode resistor value means a more stable bias, but at the cost of a little less gain.
2) that a higher cathode resistor value basically forced the bypass cap to, well, do even more of what the bypass cap does
3) that a bigger bypass cap attenuates less bass (essentially use a cap that allows for the full guitar frequency for a guitar amp)
4) I choose 10uf for a little less bottom end and I've seen the 3K/10uf pair before
5) Also, one of my all time favorite amp used the 3K cathode/1K plate load resistor set up, with a 10uf bypass cap
6) that the bypass cap allows the AC signal to bypass the DC cathode resistor, which helps keep the bias stable
5) that if you remove the bypass cap the bias point will fluctuate allowing for a larger input signal before clipping and adds a bit more compression (good for bass, keyboard, etc)
6) that no bypass cap has been used in great guitar amps like the Fender 5F1, but most amps makers now days choose to use them
7) with that said, using the last preamp voltage gain stage with no bypass cap and injecting the negative feedback is used, which not only does the whole no-bypass cap thing but the negative feedback tends to tame signal spikes

It will be interesting to see how I think about all this over the next days and possibly weeks (probably years - lol).

So, I'm off to try and find some course 101 material on preamp biasing and gain staging (please note - course 101 stuff, not 301 0r 401 - lol)

Thanks again to all and to all a thanks again

[eschertron]

I shouldn't even reply, as I don't have time to do it justice, but...
Some of your assertions are almost word-for-word the technical description of what's going. Very good. Some are more "internet lore -sounding". My suggestion is to explore the physical/mathematical basis for these effects, and stay away from the 'cookbook' solutions. You'll gain a firmer understanding.

[Chuck H]

1) that a higher cathode resistor value means a more stable bias, but at the cost of a little less gain.

The cathode bias resistor will also alter how much signal the gain stage can accurately amplify before clipping AND how much it amplifies it AND (this is something you'll want to know) how that stage clips when it's overdriven. Generally, a higher resistance will allow more input at the grid but less output at the plate. And when clipping a higher resistance will tend to clip in cutoff before saturation making for asymmetrical clipping. The colder you are, the more asymmetrical the clipped wave form will be. This has been used to great advantage in many high gain guitar amp designs. Biasing hot does pretty much the opposite. A center bias will give the greatest clean output for a clean input and clip symmetrically. And on that note... Biasing at 100% doesn't necessarily mean center bias (but it will be close). You actually need a scope to "see" that the tube is not clipping hot or cold for a given input signal.

2) that a higher cathode resistor value basically forced the bypass cap to, well, do even more of what the bypass cap does

Yes, but a bypass cap doesn't render all bias conditions equal. You will have more output at the plate, but you will still have the asymmetry when the stage is clipping because the bias hasn't changed. See 3) for more on this.

3) that a bigger bypass cap attenuates less bass (essentially use a cap that allows for the full guitar frequency for a guitar amp)

The bypass cap does two things. One is to help stabilize DC because it acts somewhat like a charging battery and can give or take DC voltage when there are fluctuations at the cathode. The other thing it does is mitigate localized negative feedback within the stage. The standing resistance at the cathode offers has an "impedance" (AC resistance). The higher the resistance, the higher the impedance and more signal rides on the cathode. More signal causes more negative feedback and reduces output. With a side benefit of improving signal linearity as negative feedback often does. Since the bypass capacitor has a much lower impedance than the resistor it effectively removes signal from the cathode. Reducing local negative feedback and, therefor, increasing gain. The more signal on the cathode (higher resistance) the more effect the bypass cap has on the local NFB.

4) I choose 10uf for a little less bottom end and I've seen the 3K/10uf pair before

With this statement it should be observed that the knee frequency for a given cap value will change with the resistance (impedance) value. The higher the circuit impedance, the more LF a given cap value will pass. I don't know the formula for this because I'm not very tech at all. Someone may chime in about that. I use a lot of on line calculators when I design. ampbooks is a good one. Play with the resistance values for the "bypass cap calculator" and you'll see this effect in action.

5) Also, one of my all time favorite amp used the 3K cathode/1K plate load resistor set up, with a 10uf bypass cap

I think you meant 100k plate load? All of what I said for 4) applies here. So if you ever DO find that you wish to bias a stage differently (for some very good reasons) you can use a formula (or an on line calculator ) to determine your target knee frequency for a give bias resistor value.

6) that the bypass cap allows the AC signal to bypass the DC cathode resistor, which helps keep the bias stable

Nailed it, with caveats. See 3)

5) that if you remove the bypass cap the bias point will fluctuate allowing for a larger input signal before clipping and adds a bit more compression (good for bass, keyboard, etc)

You'll notice the effect more with higher current circuits like using cathode bias on the power tubes. Not so much with lower current circuits where voltage amplification is the goal, like most preamplifier stages. Some of what I mentioned in 3) applies here too.

6) that no bypass cap has been used in great guitar amps like the Fender 5F1, but most amps makers now days choose to use them

There are many reasons design one way or another. Any stage circuit must be taken in the context of the whole amplifier system. ie: the 3k/10uf values obviously worked great for the amp you heard and observed them in, but that doesn't mean your plexi Marshall will be improved by changing any or all the preamp stage cathodes to 3k/10uf.

7) with that said, using the last preamp voltage gain stage with no bypass cap and injecting the negative feedback is used, which not only does the whole no-bypass cap thing but the negative feedback tends to tame signal spikes

NFB systems injected into the cathode circuit can ONLY work if the cathode signal isn't bypassed. So there's that. And as I mentioned you won't notice much or any tonal difference from a fully bypassed stage in a low current voltage amplifier stage other than gain. And gain structuring is usually worked out before that stage to give the desired result. See? Context.

You ask good questions. I was compelled to do my best because it's right here on this forum where I got most of these nuances worked out (to whatever degree I have ). As I said, I'm not an EE or experienced tech and I've never made my living doing anything with electronics. I'm also formula impaired. These are strictly an experienced layman's observations and understanding. Maybe some of the more technically inclined will step in and touch up my response.