EL34 vs. KT120

[kward]

I finally got around to trying out a quad of EL34's in my KT120 amp. (Remember when I built my amp, I wired pin 1 to pin 8 (suppressor grid to cathode), so that I could do this).

I ended up biasing the EL34's at 50% of their max plate dissipation, which gives these DC operating conditions:
Queiscent current draw per tube: 25 mA
Plate voltage: 500V
Negative bias voltage: -48 volts

(Remember since I'm using a High Voltage regulator, I still get exactly 400V to the first and second stage audio amps so no rework needed there. )

Max power output is 50 Watts/channel with 0.84V RMS input sensitivity to drive the output stage to full power. (Remember I'm using the Triode Electronics version of the Dynaco A431: 4300 Primary to 8 Secondary).

The 10 kHz square wave at 1 watt output shows about double the initial overshoot of what I finally settled on for the KT120s, which means the amp should start its roll off at high frequencies a bit higher in the frequency spectrum, but it also means they will probably sound brighter or more lit up than the KT120s.

I'm using the New Sensor Mullard EL34 reissue. These tubes sound pretty good, but as expected, the amp does indeed sound more lit up with the EL34s. The KT120 is a really mellow tube which I enjoy immensely, with a very controlled presentation top to bottom and somewhat rolled off highs.

Conversely these Mullard EL34 reissues are brighter (but most of this I expect is my amp, not the tube itself) and a bit hashier in the high registers. Midrange, especially vocals both male and female, are no better in my opinion than the KT120s. Bass is a little muddier with the EL34s.

Nice tubes overall these Mullard reissues, but I do prefer the KT120s. I need to live with these EL34s now for a few weeks to make final judgement.

I am surprised I'm getting 50W/channel out of these little guys and in ultralinear mode no less (at 33% taps with the A431 OPT).

By the way, it was a real pain in the butt to get the EL34s biased correctly. I didn't want to red plate them, so I had to bring the amp to my work bench, flip it over, take the KT120's out, turn the amp on, set the bias to where I thought it would be, then put the El34s in and iterate until i got them were I wanted. Now that I know...I can go much quicker next time I swap.

 

[for_p1]

If you want best sound quality, you need to bias tubes much higher - so that they were in class A most of the time. I would target 90% of dissipation limit. When you have signal to amplify, current increases, but plate voltage drops down. That is why you won't exceed tubes limit. I tested this approach on KT88 based amplifier. Result was a reduction of distortion products by 6 dB at moderate power levels. There is no reason why you won't see the same with EL34.

 

[GordonW]

Though I probably wouldn't go 90% of max dissipation on modern production EL34 tubes- with 500v, there's no reason you couldn't bias them at 40ma each (20w dissipation, or 80% of max). Probably would produce more power, as well as sound better.

If it does decrease distortion, there's a chance your overshoot behavior might be better. Less harmonics to "trigger" the ringing.

Regards,
Gordon.

 

[for_p1]

Overshoot is bad for itself. Sound will be harsh, and amplifier sensitive to no-load condition. I suggest to adjust compensation circuits bringing it to "normal" 10% or lower. Any amplifier with global feedback requires that sort of tuning when active component like tube types are changed.

 

[kward]

Overshoot is bad for itself. Sound will be harsh, and amplifier sensitive to no-load condition. I suggest to adjust compensation circuits bringing it to "normal" 10% or lower. Any amplifier with global feedback requires that sort of tuning when active component like tube types are changed.

Yeah, well, this was just an experiment as the amp is tuned for the KT120s, and I don't want to resolder in new feedback and step network values for EL34s. I guess I could put in a switch that lets me easily swap between output tubes, but that's not something I really had in mind.

Highest I'd probably go in biasing the EL34s is 80%, and that may fix some of the overshoot as Gordon pointed out. I'd like to have the option to run EL34s but not at the expense of the sound the KT120s produce. I'll try it and see, and if I don't like it...oh well no biggie.

Still gotta try KT66s too. :scratch2:

 

[guitbanger]

I'd make 'em pull 50 to 55 ma each. The voltage will drop to around 475? In the Dynaco Mk II (the amp that the A-430/431 trannie's were more or less designed for) specs were 65 ma per tube @ 460-465 volts. Bet the amp perks right up. Listen then measure the rest.

 

[kward]

I'd make 'em pull 50 to 55 ma each. The voltage will drop to around 475? In the Dynaco Mk II (the amp that the A-430/431 trannie's were more or less designed for) specs were 65 ma per tube @ 460-465 volts. Bet the amp perks right up. Listen then measure the rest.

Ooh. (wince). That's 23.5 watts idle dissipation, or 95% of max! Holy smokes (perhaps literally). I guess I will try it. Nothing to lose.

At 65 mA and 460V, that's running the tubes at 29 Watts plate dissipation. I can't see how the tubes could handle that without burning up? Or maybe the EL34 is a rugged bugger and it can handle it.

 

[for_p1]

Ooh. (wince). That's 23.5 watts idle dissipation, or 95% of max! Holy smokes (perhaps literally). I guess I will try it. Nothing to lose.

At 65 mA and 460V, that's running the tubes at 29 Watts plate dissipation. I can't see how the tubes could handle that without burning up? Or maybe the EL34 is a rugged bugger and it can handle it.

Designers of tube amps routinely exceeded tube specifications. But then they required only specific brand of tubes to be used. Of cause you will never pass proper QA with that approach, but who cares if you can sell it.

 

[guitbanger]

The manual actually said not to worry if ya see a little glow on the plates, normal! lol At the 50 to 55 ma setting the plates won't be glowing red and the sound should perk up, bringing up bass content and smoothing out the mids and highs.

 

[GordonW]

If you're running old-stock EL34s (such as Sylvania 6CA7s or Mullard/Amperex EL34s), you can get away with running them right at dissipation limit.

Try that with modern production- and be prepared to swap out outputs on a regular basis. Probably won't blow up, but JJ, EH and other new EL34s just don't hold up when driven that hard. Hence, my recommendation for 80% of max...

BTW: I have lots of experience from what I've seen of those. Lost count of how many JJ and EH EL34s I've seen replaced on Marshall guitar amps, recently (last two years). They just have to be dialed back to not burn them out in a hurry. Didn't seem to be that way before about 2-3 years ago, but it looks that way now...

Regards,
Gordon.

 

[tim614]

cathode bias?
what current did you run the kt120?

 

[hotroady]

If you're running old-stock EL34s (such as Sylvania 6CA7s or Mullard/Amperex EL34s), you can get away with running them right at dissipation limit.

Try that with modern production- and be prepared to swap out outputs on a regular basis. Probably won't blow up, but JJ, EH and other new EL34s just don't hold up when driven that hard. Hence, my recommendation for 80% of max...

BTW: I have lots of experience from what I've seen of those. Lost count of how many JJ and EH EL34s I've seen replaced on Marshall guitar amps, recently (last two years). They just have to be dialed back to not burn them out in a hurry. Didn't seem to be that way before about 2-3 years ago, but it looks that way now...

Regards,
Gordon.

I can confirm. JJ's or EH aren't nearly as rugged as Sylvania/GE 6AC7's. I ordered a new quad of Winged C's made in St Petersburg, Russia. I heard these are the best of the new issue. Hard to get a matched set of fat bottles at a reasonable price these days, but working on it. I'll see soon, how good the C's sound. The glass itself sucks on JJ and EH. JJ EL 34L sounds a little better standard JJ or EH, but would not buy either again at this point in time.:thumbsdn:

 

[kward]

cathode bias?
what current did you run the kt120?

My amp is fixed bias. In the KT120 I was running them at 60 mA per tube @ 460V plate. (28 watts dissipation).

Hence, my recommendation for 80% of max...

I've turned up the juice. Now running the EL34's at 80%. We'll see how they sound... In EL34, I've got a quad of a) New issue Mullards, b) ElectroHarmonix, and c) Winged C. I'm really liking these new issue Mullards at the moment. (Still not as "sweet" as the KT120's though, but I tuned my amp for KT120's so I'm not expecting miracles here.)

Whoever earlier said the EL34s would mellow out at higher plate current...yes, that's what I'm hearing.

 

[guitbanger]

Even old stock EL34's would sometimes exhibit the dull red glow on the plates at those current levels. The USA made 6CA7's will take it, no problem. Keep in mind that's what the designers of these amps intended for them to sound like by running that kind of current. Gordon is right on the money as to his assessment of the currently made stuff. They just won't hold up in a guitar amp or hi-fi. Guitar amps generally don't run as much current as hi-fi stuff, but are subjected to much more mechanical vibration which causes their premature death and microphonics. I'm a big proponent of OS and NOS tubes. Do you know why they (like divorces) cost so much? 'Cause they're worth it.

 

[kward]

Did some measurements today with the EL34s (Mullard new issue) biased at 80%. Overshoot at 1 watt, 10kHz square wave came down a bit, output power is still ~50 watts/channel (49 to be exact), mids and highs are significantly smoother, and (unexpected) the bass cleaned up significantly. They sound closer to the KT120s now, but the KT120's still beat them by a stride, maybe two. IMO of course.

Interestingly, at max power (50 watts) and at 10 kHz, I can actually see phase shift occurring on the scope. Not much, maybe 10 degrees or so, but it is there, and only right before or at clipping. Also, clip is very soft. More rounded and distorted at the tops and bottoms of the sine waves. Unlike the KT120s which don't cause the amp to phase shift at max power at all, and clip in the traditional fashion of chopping off the top and bottom of the sine wave.

 

[for_p1]

Did some measurements today with the EL34s (Mullard new issue) biased at 80%. Overshoot at 1 watt, 10kHz square wave came down a bit, output power is still ~50 watts/channel (49 to be exact), mids and highs are significantly smoother, and (unexpected) the bass cleaned up significantly. They sound closer to the KT120s now, but the KT120's still beat them by a stride, maybe two. IMO of course.

Interestingly, at max power (50 watts) and at 10 kHz, I can actually see phase shift occurring on the scope. Not much, maybe 10 degrees or so, but it is there, and only right before or at clipping. Also, clip is very soft. More rounded and distorted at the tops and bottoms of the sine waves. Unlike the KT120s which don't cause the amp to phase shift at max power at all, and clip in the traditional fashion of chopping off the top and bottom of the sine wave.

When you run K120 try bias them at 100mA. If your output transformer will handle it you should get even better results.

 

[jaymanaa]

I run my PP EL34s at 35ma at just over 500v. I've tried 40ma but couldn't hear a difference so I decided to be kind to the tubes, they are old like me.

 

[guitbanger]

I would be curious to know how the KT120's biased at 40ma. per tube would compare, sound wise, to the EL34's at the same 40ma. per tube.

 

[guitbanger]

Or 50ma. for both tube types? As I stated in my earlier post, MKII A430/431, EL34's, 460/456 plate volts and 65 ma. per tube. MKIII A430/431, 6550/KT88's, 475 plate volts, and 70 ma. per tube. The 10 watt difference in the specs between the two amps and tube types. The choke and GZ34 being the difference and only 5 ma. current difference between tube types. I don't have any KT120's myself, so I can't try this myself (but I might be persuaded to get some in the near future as I have an old school pair of potted A430's for a build) but seeing as how you are using the same output iron, I'd be curious to know what your perception of the sound differences are at approximately the same current levels. I have tried this method using a pairs of MKIII's, with OS Tung-Sol 6550's, GE 6550's, Sylvania 6550's, Phillips/Mullard Xf1's and 2's EL34's, Tesla EL34's, and Sylvania 6CA7's. Biased to the same current (55ma. per tube) I feel that it sure evens the playing field in terms of the "sound" test, especially as I've never listened to any hi-fi at full output (or want to! lol). Thanks for posting your findings and tolerating my ehem, rambles.

 

[kward]

guitbanger,

I'll give it a shot (both biased at 50 mA) and let you know. Give me a few days. I think I'll use my Winged C EL34s for this test though.

Just one caveat though, my amp was tuned for the KT120s. What that really means is at full power in my amp, the KT120s produce 64 watts per channel, so I tuned it for 15 dB feedback @ 64 Watts/channel.

Since the EL34s produce less power (by 15 watts per channel), that means the circuit will provide a bit more feedback than 15 dB when using the EL34s. This could either be good or bad based on your view point. More feedback should extend frequency response just a bit, but it could also have the effect of flattening sound stage and narrowing harmonic content.

What I know for sure is the KT120s sound awfully good in my amp, so it will be tough competition.