Tube amps and speaker ohms

Newbee's first point on the impedance curve of the speaker is taking you to where you really need to be. I'll add in phase angle as well.

One other parameter I think that should be discussed is the crossover topology. How simple or complex the crossover is makes a tremendous impact into whether the loudspeaker presents an easy or difficult load. Tube amps like first order crossovers better than second order crossovers, and second order crossovers better than third order crossover, and on and on. On top of that, things like Zobel networks, trap circuits, and the like can turn things completely on their head. Take some famous loudspeakers, such as Thiels, with first order crossovers that so many compensation circuits - a total bear to drive.

The best generality I can offer is that you want the crossover as simple as can be. The opinions on single driver loudspeakers being a good match with a tube amp is true - because there is no crossover to suck up the power. Of course, you do give up frequency response.

Far too often, sensitivity and nominal impedance give such an incomplete picture that you simply need to walk away from what people will tell you, and give it a try. THAT is the only way you'll know...

Grant, my KT77 comments are based solely on listening to the JJ variant. I'm not sure how the Osram (didn't know they made KT77) and Mullards would sound.

In my Jadis DA60 and Dennis' (Badboss429) Quicksilver, a bunch of us (including Mechans) compared the JJ E34L blue glass, JJ KT77, vintage 1960s Mullard EL34 XF1 (EL34 Holy Grail), and JJ KT88 blue glass. The KT88 came in last. The rest of the tubes had a variety of opinions. Overall, things seemed to split between the Mullard and JJ E34L, but many of us couldn't get past the powerful, intoxicating lows of the JJ KT77.

From what I heard, the first batch of KT77 had issues with the tube pins. The octet I bought in January/February 20076 had this issue corrected. Didn't have any fail, either, but it's a complaint many have put forth with JJ power tubes in general. I have an octet of E34L Blue Glass, and one day while measuring them about three weeks ago, I found one was as dead as a doornail. That came as kind of a shock.

I think the point Newbee made about the EL34 versus KT88/6550/KT90 is something that REALLY needs to be lauded. EXCELLENT POINT!!!

The EL34 is a tube that is going to be in many of these amplifiers. And, while it is my favorite of the traditional push-pull amplifier tubes, it does has its shortcomings. As Newbee pointed out, you are going to give up those low frequencies (and overall power) in exchange for those wonderful mids. Please be aware of this.

Another tube to consider is the KT77 made by JJ. VERY powerful bass, and good power overall. You still don't match those EL34 mids. Consider it a great tube slotting between the EL34 and the KT88. In more than a few amps, we actually heard MORE emphatic bass with the KT77 than the bigger tubes, combined with some seriously better midrange.

I am fast coming to the conclusion to stop looking at "watts per channel" as an efficient means of determining how powerful an amplifier is.

My experience tells me that the power supply and transformers play the biggest role in this, as opposed to whatever reading was obtained when the amp drove a 1 KHz signal through an 8 ohm resistor, or what a company chose to rate an amplifier. Music is not a test tone, it is dynamic, and loudspeakers are not an 8 ohm "dummy load", they are FAR more complex.

Tubes, also seem to have something to do with this whole thing. For example, a 2A3 has less "measured" power (usually, about half as much) than a 300B, but in my experience, the 2A3 does a better job in providing "real world" power to a loudspeaker than a 300B. Similarly, a KT77 looks SO close to an EL34, but they are very, very different tubes, with very different power characteristics. And, for what it's worth, a KT88 is a strong tube. At least as strong as a 6550 or KT90, though the Ei KT90 tube seems a bit more extended in the lows.

In no way am I saying that measurements are something to completely ignore, but clearly, the engineers designing these tests have yet to figure out what is the important criteria, and the means to go about testing it. Until I am told of a better way, playing music through an amplifier and a pair of speakers, and listening with my two ears is going to be the way I test...