SE EL84 amp... other output tube options?

[thelonious]

Hey guys,

I have a single-ended amp with one EL84 as the output tube. I've played around with stuff in the preamp and managed to get some tones I quite like from it. However, I've got a persistent problem with it: when I drive the output tube hard, it always seems to get harsh. I've heard some people say that, all other things being equal, an EL84 in a single-ended amp tends to sound harsher than, say, a 6V6, but that seems like an awfully big generalization to me. In your experience, is that a true statement, even if it is a generalization? Would, for instance, a 6AQ5, 6V6, or 5881 tend toward smoother distortion when driven hard? Or does it all have to do with the operating conditions the tube is placed in? I'd appreciate your helping me to think correctly about this...

[Willabe]

If you could post a schemo with B+ voltages the guys might spot a few things for you to try and get a smoother sound?


                 Brad     

[thelonious]

I will work on that and post! At the moment all I have is pencil sketches (I didn't build this one---bought it a couple years ago when I knew zilch about amps).

But the general statement about EL84s being harsher than other output tubes when driven... is that rubbish?

[Willabe]

Others here have a lot of experience with EL84's and I don't.

It does seem that guys like to run them hard into heavy distro?

OTOH, there were a lot of old tube stereos that sounded good that used them too.


                        Brad     

[DummyLoad]

...it always seems to get harsh.

bypass the bypass cap. nyuck, nyuck... seriously, try lifting the cathode bypass cap and/or add a larger screen resistor. never have cared much for the EL84 because of the sparkle sounding distortion they make when overdriven.

--DL

[jjasilli]

Ditto to willabe.  IMHO el84's sound best clean & overdriven with 300 plate volts of less.  Haven't tried 320 - 350 which is Train wreck territory.

[shortfuse]

I had built a SE EL84 and switched out to a 6V6 after not liking the distorted sound.  It did sound much better with the 6V6.  But I may have had other variables with mine all I had was a volume knob and to distort you had to crank it up to 7-10 on the vol knob.  I also had tube rectification not ss.

[thelonious]

Thanks all!

Will try lifting the cathode bypass cap & messing with screen resistor value as DL suggested. After reading your comments and also this thread from a few years back, I think I might give a 6V6 a shot in it unless the bypass cap lift works a sudden miracle.

[jojokeo]

thelonious (monk - I've always loved that guy's name),

This subject has been discussed many times in other threads of the past so you'd be best served in going through the seach avenue to read some of them. You'd get a lot of information that's been discussed already. This being said, there's quite an opinion on simply the voltage these are driven at, then there's many other issues involved too. Another thing is I've used a number of organ-type amps loaded with el84s and those voltages most of the time always exceeded the 300v rating even before fiiltration. The sound of the old Baldwins and Hammonds were very clean and nice sounding. Think of your local church or baseball park which uses these - for me it's Dodger stadium which has had an organist playing one for decades.

I think many people just stick in the ol' 130r to 150r cathode resistor w/out regard to tube dissipation? If running a lower voltage than spec that's fine but the greater the voltage the more this needs attention. Pushing a tube at or beyond it's limits may get you harsh tones when it's driven to the max in circuit/use, but some may like it that way on purpose? Anyway, try various tube manufacturers, speaker types, cathode (and/or cap) values, plate & screen voltages, output transformers, grid resistors, conjunctive filters, etc...there's a ton of things to try to find your happy place - some more reasonable and practical than others.

[thelonious]

So I reverse-engineered this amp and pencil-sketched the schem about 6 months ago (my first time). Since then I have learned how to use J-schem, but I haven't transferred this yet... my apologies. Hopefully these scans are readable. Also, I discovered that this is basically a boutique-built version of an AX84 Hi-Octane, with a couple mods. Of course I discovered that well after I sketched it all out. So much for hours with pencil and paper. At least I learned a lot while doing it!

My mods: I changed the cathode bypass cap on V1a from 0.68u to 1u. I also changed the 1/2W screen resistor on the EL84 to a 1W because it was looking pretty toasty.

When I can make some noise (baby is napping), I am going to lift the cathode bypass cap like DL suggested and see what that does. I don't want to lose too much volume, though. As I have done more reading at jojokeo's suggestion, I keep coming across people saying stuff like, "I just switched out the EL84 in my SE amp with a _____, and it's much better." You can fill in the blank---6AQ5, 6BM8, 6V6, even 6L6 (if their PT and OT could take it). I'm sure I'd find people saying the opposite, too, but it's enough to make me want to try some different output tubes. I think I'm going to install 7-pin and octal sockets and wire them up so I can swap in a 6AQ5 and a 6V6 and try those.

BTW, I am using a 12AT7 in V1 and a 12AU7 in V2. I've swapped in a bunch of types, and those sound the best to me in this situation.

[Willabe]

Can read them just fine. Nothing wrong with those drawings at all.    

I think it's a good idea on your part to see if you will like a different output tube. Everybodys ears different and play different. 

Sometimes you can tweak an amp till the cows come home and you still wont like it as much as with some other power tubes sound/tone.


                     Brad      


            

[Willabe]

Just thinking out loud;

Since your changing the power tube socket;

I would be tempted to try bypassing the tone stack/V2 just to get it down to the bare bones temporarily.

I'd try to run a wire from V1b's .02 coupling cap right to the 5.6k grid stopper R on the EL84. I'd also bypass the 2 R's and 1 C in between V1b and the 1M pot. If still not right to your ears, I'd try upping the grid stopper, IIRC, guys sometimes go as high as 100K before it's smooth?

I think with that stuff out of the circuit it might be easier to hear the tone of the EL84 by itself.

Or, OTOH, leave it like it is now and just up the EL84 grid stopper and/or add (1 at a time) some smoothing caps across the plate R's or from plate to cathode.

Tubenit has had great resulates with smoothing caps. Here's a link of his with great info, (places, values to try) on them;

http://www.el34world.com/Forum/index.php?topic=14201.0



               Brad        

[DummyLoad]

you're gonna love a 6V6 with about 100uF of bypass cap.

--DL

[PRR]

> when I drive the output tube hard, it always seems to get harsh.

By "hard", do you mean heavy bottom-octave bass?

I would not use _three_ stages of 1uFd cathode-cap, it shaves the bottom octave severely. Pushing will increase the high-end racket but still not get authoritative guts.

I'd try 10uFd at all three AX7 cathodes. That will probably be too much boom. Now cut one or two back to ~~1uFd until it has some bottom but isn't muddy.

[HotBluePlates]

The 470pF cap between V1a and V1b also works to keep highs boosted/bass cut.

I also get the impression that to some extent, the brilliant sound is part of the EL84/EL34 nature. Look up the Mullard EL84 data sheet. You'll notice that your amp lands almost exactly on data sheet conditions for 250v plate and screen. Mullard is kind enough to break down the distortion % into 2nd and 3rd harmonic, and 3rd predominates.

13% distortion total for the EL84, and you amp will get there with ~6v of signal input.

Now look up a G.E. 6V6GT data sheet. Same conditions (250v plate and screen) require ~12v of bias, and therefore ~12v of signal input, double what the EL84 will require to be pushed to its maximum. G.E. wasn't nice by breaking down the distortion into its components, but claims total distortion is 8%; the EL84 had 8% of 3rd harmonic alone under the same conditions.

Since people don't seem to want negative feedback around the output stage in their EL84 amps (thanks to the Vox example), Dr Z seems to use the corrective filter (resistor and cap across OT primary) to tame treble output to some degree. Of course, the Celestion Blue speaker is also not a bright or aggressive speaker, and probably helps tame the treble.

I'd suggest designing/tweaking keeping in mind that all the amp variables interact to give the final result: preamp, output tube, speaker, etc...

[Mats]

Is there a resistor missing at, 12AU7,  V2b-cathode and the tone stack ?
Can one skip 50k-100k resistor from cathode to ground (before tone
stack) ?
/Mats

[jeff]

13% distortion total for the EL84, and you amp will get there with ~6v of signal input.

Now look up a G.E. 6V6GT data sheet. Same conditions (250v plate and screen) require ~12v of bias, and therefore ~12v of signal input, double what the EL84 will require to be pushed to its maximum

It seems to me what you're saying is 6V6 to EL84 with the same preamp isn't a fair fight.
The ratio of pre distortion to power distortion isn't the same.
Is it possible that the EL84 sounds harsher becuase it's being pushed harder?
If you have you're preamp set where it's putting out 6V the 6V6 is going "no problem" the EL84 is going "I'm givin' it all she's got capt'n"
So is working our EL84s harded and then blaming them for being harsher unfair?

What about using a voltage divider before the EL84 to make it a fair fight?
Now you're just comparing the 6V6s distortion to the EL84s distortion. Working them both relatively as hard.

Is this a good analogy?: Give a body builder a 120lb bar bell and a 8yr old boy a 120lb bar bell and then say why's the kid struggling so much? They're both lifting the same weight?

PIN #s ARE WRONG FOR 6V6

[jojokeo]

Lke it. May the best man win. The El84 was designed the way it was for a reason when it was used across the pond back in the day. The mini el84 sounds different to a 6V6 like an EL34 sounds different to a 6L6. We want variety and diversity with which to choose and that's a good thing. Forgot to mention, for EL84 manufacturers there's a pretty significant difference in sound between them, especially going from a JJ to a new Mullard or TAD, then the russian Svetlana, EH, etc...many to try at 10 - 15 bucks a pop or less if you buy used. Check this out for more info: http://thetubestore.com/el84review.html

[HotBluePlates]

It seems to me what you're saying is 6V6 to EL84 with the same preamp isn't a fair fight.
The ratio of pre distortion to power distortion isn't the same.

Well, I think the 1M master volume right before the EL84 levels the playing field. But if you were just doing a cut/paste of a preamp (and phase inverter, for a push-pull amp) with EL84 to 6V6, your impressions might be skewed if you didn't account for the difference in needed drive levels.

I'm basically asserting the data sheet numbers of distortion for essentially the same operating condition indicates the EL84 has more distortion, and probably a greater proportion of 3rd harmonic. That would give the impression of a bright edge to the overall tone.

Is there a resistor missing at, 12AU7,  V2b-cathode and the tone stack ?

Yes, there has to be a missing resistor there; as drawn, there's no d.c. path for tube current.

[jeff]

So a EL84 champ on 10 sounds harsher than a 6V6 champ on 10 not necessarily because the EL84 is a harsher sounding tube than a 6V6 but because it's working twice as hard?
But I guess that's one of the the beauties of using different tubes.

I see what you're saying about the MV, I just thought this would be a good A/B test so you don't have to fiddle with it to get appropriate levels when comparing the two.

[sluckey]

Just some more info... The 6BQ5/EL84 was used in hundreds of big name high quality mono/stereo amps back in the day. They were not harsh or bright sounding in those amps. Distortion was very comparable to 6V6 or 6L6 amps. They definitely sound brighter and distort earlier in most lo-fi guitar amps and I think that's mainly because of the lower drive requirements and the abusive B+ used on these guitar amps. It's all in the circuit design. I personally prefer the brighter sound, but that's because I've lost my hi freq hearing. I also know the bright sound I need/like is too bright for most people and probably drives them crazy.

[HotBluePlates]

So a EL84 champ on 10 sounds harsher than a 6V6 champ on 10 ...

I wouldn't quite say that.

I played a number of early-60's Vox amps at a Nashville vintage shop once, and none of them sounded harsh to me, even at full tilt. Then again, I've played a Matchless Lightning and couldn't understand why anyone would buy it, except that very bright amps sometimes record/mix well.

All I'm saying is that EL84/EL34 has a sound different from 6V6/6L6. I think that might be due to differences in the balance of the harmonics which make up how they distort. Clean, they all seem to sound near-enough the same in amps I've tried.

For what it's worth, I also think the same is true of pentode/triode mode when folks incorporate that switch in their output stage. Most notice that triode mode is quieter, cleaner and somewhat darker. Quieter is due to the triode mode's characteristic curves, which differ from the same tube in pentode mode. Cleaner is somewhat a byproduct of this. I think darker is due to the much-reduced amount of 3rd harmonic distortion, taking the bright edge off the triode mode.

[thelonious]

Thanks all. I'm learning a ton from this conversation. I tried lifting the cathode bypass cap. It sounded quieter, of course, but not significantly sweeter in the highs. On to some of the other stuff:

I would not use _three_ stages of 1uFd cathode-cap, it shaves the bottom octave severely. Pushing will increase the high-end racket but still not get authoritative guts.

Lacking authoritative guts is a fair way to describe this amp right now. The small cathode bypass caps are likely contributing to original shrillness problem, as you said, because I'm cranking the amp in an attempt to get guts, and what I'm getting instead is more high-end racket.

Of course, the Celestion Blue speaker is also not a bright or aggressive speaker, and probably helps tame the treble.
I'd suggest designing/tweaking keeping in mind that all the amp variables interact to give the final result: preamp, output tube, speaker, etc...

Yes, I'm starting to think that I'm going to be doing quite a lot of tweaking to all of those things before I'm finished. It doesn't help that I'm currently playing it through a G12H30.

Is there a resistor missing at, 12AU7,  V2b-cathode and the tone stack ?

Yes! Thank you! I've attached the corrected schem with the 100k resistor added.

It seems to me what you're saying is 6V6 to EL84 with the same preamp isn't a fair fight.
The ratio of pre distortion to power distortion isn't the same.

Well, I think the 1M master volume right before the EL84 levels the playing field. But if you were just doing a cut/paste of a preamp (and phase inverter, for a push-pull amp) with EL84 to 6V6, your impressions might be skewed if you didn't account for the difference in needed drive levels.

I can certainly put together a temporary voltage divider and switch as jeff suggested, but I'm not sure I completely understand the effect of the drive level as it relates to wattage/output. I get that some tubes require much more voltage on the grid than others to be driven appropriately, and that some PI tubes are better able to supply the necessary voltage swing. EL84 and 6V6GT (not JJ) both have a plate dissipation of 12W, so theoretically they should both provide the same output volume under the conditions that make them happiest, right? Is it that being more sensitive, the EL84 takes less grid voltage swing to be pushed to its maximum output power?

[HotBluePlates]

Of course, the Celestion Blue speaker is also not a bright or aggressive speaker, and probably helps tame the treble.
I'd suggest designing/tweaking keeping in mind that all the amp variables interact to give the final result: preamp, output tube, speaker, etc...

Yes, I'm starting to think that I'm going to be doing quite a lot of tweaking to all of those things before I'm finished. It doesn't help that I'm currently playing it through a G12H30.

I've only played through G12H30 speakers in one setting: in an early-70's Marshall 100w cab, and those speakers had the 55Hz cones. Those are different than the Celestion blue, but the ones I played were not bright or edgy. Can't say how different the modern version might be, as I've not played them.

EL84 and 6V6GT (not JJ) both have a plate dissipation of 12W, so theoretically they should both provide the same output volume under the conditions that make them happiest, right?

Yes, exactly. Same B+ and same/similar load, they will each produce about the same output power.

Is it that being more sensitive, the EL84 takes less grid voltage swing to be pushed to its maximum output power?

Right again. You'll get the same ~4w from the EL84 that you'll get from the 6V6. But the EL84 is more sensitive, so with the exact same preamp, the EL84 will be at full output power and beginning to really distort at a lower volume-knob setting.

[thelonious]

Thanks, HBP. I'm going to experiment with an alnico speaker, too, to see what difference that might make.

Got the octal socket installed and wired yesterday. It looks like the tone stack + volume pot provide the grid leak path for the power tube. If I turned all the knobs up to 11, I think the max resistance of that grid leak path would be about 550k plus a 10k grid stopper. That's no prob for an EL84, but 6V6 max grid resistance is 500k. I don't foresee myself ever wanting every knob in the tone stack turned up all the way, but if someone else did it... would the extra 60k be a problem? If so, what should I do to correct it?

Another question - the heaters were DC elevated on this amp by connection to power tube cathode. PT is a Hammond 269EX @ 71mA, so I can't leave both power tubes in and idling... and I can't simultaneously connect heaters to the 6V6 cathode because that connection would put the cathode resistors of the two power tubes in parallel! If I make the bleeder resistor in the power supply the bottom half of a voltage divider as in the modified power supply schem I've attached and elevate the heaters with that, would that work well enough? Any suggestions to improve it?

BTW, there is a great capacitor discharge time calculator for checking bleeder resistor values here if anyone is interested.

[John]

That should work fine for your heater CT, although you might want to increase the value of the top resistor a bit, to bring the voltage to around 70. I only say that because that's what a lot of folks seem to use.

I used the same sort of set-up on mine and it works very well.

[thelonious]

Thanks John. I think I have a 270k that would do the trick there to bring it to about 68-69v.

[thelonious]

This thread is ancient now, but I figured I should post the results. I've been modding this amp on and off for a year and a half, and I think I'm finally happy with it. The solution turned out to be a combination of a lot of the advice you guys gave above. One of the biggest contributing factors was what PRR said about having too many stages with small cathode bypass caps. Another problem was that the builder hadn't shielded the bottom or sides of the chassis. It's possible that some of what I was interpreting as "harsh highs" was actually the side effects of ringing/oscillation, and I just didn't know how to interpret it at the time.

Here are some of the changes that helped:
-Added Full/Tight/Fat switch on the cathode of V1a that switches between .68u, 1.68u, and 10.68u bypass caps (was 1u)
-Removed V2a bypass cap (was 1u)
-Made bright cap across first divider switchable instead of on all the time
-Changed 100k resistors on V2 (direct coupled cathode follower pair) to 33k to be better suited for 12AU7
-Added tone stack lift switch
-Added octal power tube socket with 250R cathode resistor and bypass cap

It sounds a ton better now. I've tested it with EL84, 6V6, and 6L6GB tubes. I didn't want to risk a 5881 or 6L6GC since the power transformer is pretty small. After all of that, the EL84 actually turned out to be my favorite tube.

I've attached the final schem. It doesn't show the octal power tube socket and its cathode resistor/cap. The 820k in parallel with the volume pot gets switched in when the tone stack is lifted. The volume pot and tone stack in parallel provide the grid leak path for the power tube. With the tone stack lifted and the volume pot maxed... an EL84 might handle that much grid resistance, but a 6V6 would probably fry. Since I'm going to be switching tubes in and out, I figured I had better add that resistor.

Thanks again for all the good advice!

[eleventeen]

Very cute! What do read as your highest B+?

And...to nitpick slightly, wouldn't we want to have the stby switch *after* the first e-cap?

[smackoj]

Nice work Theo. You improved this amputator by a factor of 300!

 

[thelonious]

Very cute! What do read as your highest B+?

And...to nitpick slightly, wouldn't we want to have the stby switch *after* the first e-cap?

B+ is right around 255-260V at the first filter cap. Puts the EL84 in class A happy-land, but it's a little low for the 6L6GB I guess.

I'm glad you brought up the standby switch---that's actually an error in my schem. I attached an updated one. I forgot that in this amp, the standby switch lifts the ground from the PT secondary center tap. Since it's before the rectifier diodes, it's switching AC, which is easier on the switch contacts. If it were a tube rectifier, it would definitely be important to have it after the first filter cap (to limit inrush current on the rectifier, according to Merlin?). I was actually thinking about implementing his Improved Standby Switching option with a resistor shorted by the standby switch.

On the amps I built myself, I didn't even bother with a standby switch...

[jojokeo]

Nice design work thelonious. There's a few items I would change but it's your amp and if you're happy then that's all that matters. Have you tried bypassing the CF portion to hear how it sounds? I'm sure it would be nice for a cleaner channel/setting then if you wanted to you could switch this in and out depending how you feel that day or put it on a relay for Rhythm & Lead channels. Or not and just enjoy it the way it is. I'm just throwing it out there...I did something similar to an EF86 design similar to yours here which sounds/plays great w/out the CF too.

[thelonious]

Thanks jojokeo! I will give that a shot and see what happens. This amp excels at dirty tones, but it's tough to get it clean because of the low wattage. Your suggestion might be just the thing...