Tube amps and speaker ohms


In your opinion , do push pull amps work better with 8 ohms or 4 ohms. .I am under the impression the lower the ohms, the more power is demanded from the amp....Another question, are there low powered SET amps ,and high power SET amps?
I'm looking at a 40 watt 845 tube amp for my 8 ohm, 89 db speaker.. just cked the Thor has a 86 db W18 midwoofers(2 per cabinet) and a 88 db tweeter. Will an 845 amp rated 40 watts be able to drive the 86/88 db speaker? With authority, bass, mids, highs, in dynamic sound stage? Synergy? Or poor match?
bartokfan
+++ Pauly "4 ohms should not be any problem for a tube amp". But in my experience tube amps do not like 4 ohms, especially when the music gets complex, as in orch, even large jazz ensemble is rather thin sounding. +++

All things equal I'd go for a higher impedance speaker, yes, but you are quite mistaken by making an across the board statement that tube amps do not like 4 ohm speakers. Decware has some speaker models rated at 4 ohms that are more suited to tube amps than 99% of all 8ohm rated speakers on the market today, and that no exaggeration.

+++ Push pull's , like SET amps can be picky on what speakers they prefer, for best potential.
I'll stay with 8 ohms. +++

A SET amp by definition is single ended and as such not push pull. I do agree that any tube amp, be they single ended or push pull are picky on what speaker they prefer. Where I disagree with you is that nominal impedance has little bearing whether a speaker is suited for use on tube amps. The actual impedance curve and crossover design are way bigger factors you need to consider.

Regards
Paul
Joe and Ralph,
I have Jm labs nominally rated at 8 ohms. They are 936 Electras, wich have dual 8 inch woofers the impedance will drop to 3.6 ohms. I am not sure which order X-over they have but I think they are 2nd order and probably have a fairly involved construction with a large induction coil etc. I also own speakers which are 16 ohms. I have 8 ohm speakers with rock solid 8 ohm impedance curves. None of these even comes close to the inefficiency of the Von Schweickert 2s in the WAF system driven by 40watt PP amp. (yes I know too many speakers and even more feel free to make offers) . Yet Joe as my witness the VR2s play to levels I wouldn't characterize as eye bleeding in a fairly large room 24X 12 X9 but loud enough that my wife always turns the music down whenever she comes into the room. It makes no sense because an average attack of flatulance will overcome the SPL.
I know you have both spent a good deal of effort, time and money on ensuring the bazzing speed that is the essence of what I like about OTLs.
Yet The new tube amps I have, are output transformer coupled and I credit the size of the trannies in them for making the JMs with a " difficult load" during low frequency passages, sound effortlessly extended and clean As loud as I want. I want a lot sometimes around 95- 100db.
Ralph brace youself- Joe has another 60 watt amp but it weighs about 70lbs mostly trannies. I have not found the trannie compromise you mention, the detail, extension are great. I don't care for bone crushing bass below the 28 cycles the JMs are supposed to make. My 16 ohm speakers are vintage JBLs since you like CAR that should have a good idea of that type sound. They are 1959s still as loud as you can take it not distorted using El84 based low powered integrated amps of that era.
My question is why do amps rated at as little as 9 watts per side but with massive trannies make the speaker sing in all but the worst cases, which are generally planars that drop to an impedance equal to a dead short.
Paul , nice post, and in agreement. FWIW, I demoed a KT88 40 watt intergrated (chinese model/low end lab) just to see how well the 88's work out on orch. The amp is the same weight as my JOR, but offer more tubes in the pre section. Very disapointing. The 86 db Thor's sucked the amp dry. So as you say, impedance of 86 db is something not suited for some tube amps. With a impedance that low, would you say that a important factor is how well the transformers are made/output power?
Tyler's Lin Sig Sys though rated 4 ohms, has a higher db rating at 92, so my little JOR was able to actually work the large speaker into good performance, if not excellent performance. I was very surprised as the LSS has 2 more drivers per cabinet to push vs my MTM design. I was surprised the JOR came through, due to the greater sensitivity of the LSS at 92. (in spite its 4 ohms).
Thoughts?

Bartokfan I think you are confusing impedance with sensitivity. Nevertheless, both are factors that influence how well (or not well) your amp will do.

Rule of thumb is to have a speaker with simple cross over design (or better, no cross over), benign impedance curve, high(ish) impedance and high(is) sensitivity.

I will probably get flack from a couple of SET guys for saying this, but I do not believe any single tube is better per se than another at orchestra (or opera, jazz, rock or anything else.). However, not all amps are equal, and some will perform way over what their rated output would suggest they can.

Don’t get bamboozled into buying a huge tube amp to drive tube unfriendly speakers. Get the right speakers and every tube amp will work well (or work as well as it can).

If somebody tells me his speakers worked fine with his 100 wpc plus tube amp but not with his 50wpc tube amp, I run away. That is the speaker I will never want own, even if I get paid to take them. Proper tube friendly speakers will work well with 2 to 3 wpc tube amps.

As far as orchestra/opera goes, I have had good results with all my tube amps on my speakers. These amps were 300B SET (x 2), 6C33 SET, 6550/KT88 SEP, EL34 PP (x 3), 6550 PP, EL84 PP. I’m probably missing one or two, but the point is they all worked great for orchestra because I mated them with the right speakers.

Once you have tube friendly speakers, you can try any tube type and amp configuration and always end with a smile on your face.

Regards
Paul
A 4 ohm speaker rated at 92db? Quite unusual. Lower impedance speakers are almost always much less efficient. Just a thought...