|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
165.86.81.22
In Reply to: RE: two questions about 45 and 300b..............thanks posted by bouncy ball on April 03, 2014 at 08:52:21
Most of the negative comments about 300B sonics are from folks who have heard them driven by crappy drivers and running rather classic-like operating points and low loads... looking for higher power.Thorsten, Gordon Rankin and Dennis Fraker amongst others have recommended running the 300B at operating points closer to 45-like current:voltage ratios. Gordon suggested it as a way of getting close to the 45 sonics, but with more power and greater dynamic capability (and to my mind, lower distortion for a given output level). Just realised that Thorsten mentioned similar in a thread below.
These a my stream-of-consciousness thoughts about what I would do:
Run the 300B at roughly 50mA bias, 350 volts on the plate for a 1mA:7V ratio (still not quite as high as the 45, IIRC) and low dissipation. Give it a 5K load. You should get a nice 6W or more out of it. I'd also consider using a 6K or 7K load - power would drop back to 5W or less. Try AC filaments first... but you would need a quality hum pot with fine adjustment (not the clunky 100 Ohm POS most use). If this was too noisy, you could try DC filaments with CMC... or even consider a switch-mode supply, which would provide AC heating of sorts at very high frequency (and is easy to implement - though I still have concerns about "noise").
Edits: grammar and a few additions.
Gotta run.
“As long as we have any intention to be right… we should be wary. So long as words have the slightest ego attachment, they are dishonest.” Charlotte Joko Beck
Edits: 04/03/14Follow Ups:
Once every year or two I post something along this line. Here it is again.
It's the relationship among the variables voltage, current, and load impedance, not the absolute values.
I use an arbitrary reference point of RL(ref) = (Vp/Ip) - 2.38*rp. This gives about 3% second harmonic at the edge of grid current, according to the WE table. A higher RL gives less power and less distortion, a lower RL gives more power and more distortion.
Most 300B amps operate at around 70% of RL(ref), giving 5%-6% typically. The traditional max-power point for a 45 is 275v, 36mA, and 4600 ohms. That works out to 1.28 times RL(ref); most people use a 5K transformer which is 1.39 times RL(ref). The WE table does not list such an example, but an educated guess would be less than 2% distortion if the 45 is as linear as a good old-stock WE 300B.
The other side of the coin is that, with a load of 2500 ohms the 45 would probably put out 3 to 4 watts - and sound like a typical 300B but underpowered.
FWIW, the classic 2A3 (250v, 60mA, 2500 ohms) gives 1.10*RL(ref) - also on the high side for low distortion, but not as high as the 45 specs.
For an example at the other extreme, WE lists 400v at 50mA into 3000 ohms. Distortion at clipping is about 10%, output is 11.5 watts, and plate dissipation is only 20 watts. That's 57.5% efficiency, theoretically impossible with a Class A amp - only the high distortion makes it possible.
I offer no interpretation, just the facts.
Is anyone doing this commercially?
big j.
"... only a very few individuals understand as yet that personal salvation is a contradiction in terms."
Dennis Fraker, Serious Stereo 2A3, has been doing this since his first commercial release in 1989. He is the only one I know of. He prefers a 2A3 overall, to a 300B.Jeff Medwin
Edits: 04/06/14
I suspect another reason for Dennis and Don Garber running the 2A3 at conservative OPs also has something to do with 2-stage typology, no?
Observe, before you think. Think before you open your yap. Act on the basis of experience.
No, Dennis always runs tubes in all amps conservatively, often shooting for Golden Ratio proportions of plate dissipation, relaxing current not voltage.He has always explained it to me as higher dissipations cause thermal STRESS and the whole tube sounds STESSED and edgy, not relaxed, when "Hot Rodded" with running it hot. People hot rod tubes because they more often than not have crummy wiring and crummy supplies, and crummy lay outs, so for the initial A-B, the hotter tube dissipation sounds better. Of course, they don't THINK about the tube permanently degrading itself, its' coatings on the metallurgy, etc., after 200 hours, and only being usable at up to 2,000 hours.
Some relaxed 2A3s go 10,000 to 50,000 hours in Dennis' amps, and the performance is consistently high over time. Thermal stress is a key reason says Dennis, its all part of the laws of nature, physics, engineering. Your family car may do 100 MPH, but it will last a LOT longer if you run it down the turnpike continuously at 62 MPH., a Golden Ratio proportion.
Two stage has nothing to do with it.
For the Driver stage, Isamu Asano's L-W amp used a very "relaxed" 0.43 mA. of driver current, Nobu Shishido's L-W amp, 1.0 mA. Dennis is similar I'd say for the Driver stage. But there is NO performance comparison between Dennis' very latest 2A3 amps, and any others in existence. Its simply the best there is IMHO.
Dennis, and my original audio mentor, Mr. Bob Fulton, are smart enough not to build 300B amps. Fulton was HIGHLY incensed, livid, with its popularity in the 1980s, compared to a 2A3. He said to me " in the 1940's, we "all" knew not to use a 300B". Robert liked 45s, 2A3s and 845s. I like 45s and 2A3s.
It MAY be a little different, 300B wise, with VAIC's new design contributions and EML designs now a days, but why bother!! Own an efficient speaker, smell de roses. Later.
Jeff Medwin
Edits: 04/07/14 04/07/14
I just got an email about these EML45s, 300Bs, etc. from none other than Dennis "Tube Wrangler"..... the man himself. Pretty good discussion I say !! Hes something else, plenty smart !!
Here we are, verbatum :
The EML "45" is not really a "45"-- it is built more like the EML 2A3.
Think of it like a "EML 2A3" 45-- just like the JJ 2A3-40 is really a
"300B" 2A3-- (the plate is larger than the tube would normally use).
With the 2A3 we have 2.5 volt filaments and fewer turns on the grids,
compared to a 300B. (BOTH are why the 2A3 is markedly superior to any
300B-- an ordinary "45" falls in between the 300B and the superior EML
2A3, but the EML "45" is a superior tube, just like the EML 2A3 is. There
is actually no performance difference between these two EML types...
assuming either is placed into a circuit that is optimized for IT.
Sure, the "45" has more grid turns-- that makes it a higher-mu tube
(than a 2A3), and that's why the 45 runs at 5K output loading.
Remember-- the harder you load the output tube (up to a reasonable
point), the more IMMEDIATE the sound gets-- the more detailed, deep and
structured. WE LIKE THAT. You also get "more" power out, and more
measured "distortion".
What happens with easier loading? MUCH lower measured distortion AT FULL
POWER. Does this help sonically? Not really-- not ordinarily... because
ordinarily we ARE NOT running full power-- full power measurements, while
interesting, really have nothing to do with musical integrity..... what
we normally appreciate in music.
I tried an experiment with my amps a while back: normally, these are
driving the Great Plains "ALTEC" 604 Duplex into my MLTL enclosure. This
is using the 8 ohm tap on the Magnequest 2.5K output transformer, driving
the crossover network that drives both the woofer and the Mid-Tweeter.
In each channel I placed two amps in series, using both 16 ohm (using the
entire output windings) outputs. Each amp input was in parallel with the
other amp's input (common-drive configuration). Again, the outputs were
placed in SERIES, giving a 32 ohm output tap. This was loaded into the 8
ohm speaker.
I did this with both channels, using 4 amps total.
Well, what happened? It was just AWESOME--! Tons of super-clean,
wideband power. Everything was just plain better.
Why did this work so well? Well, it would not on an ordinary amp-- not
entirely-- it IS a mismatch. Distortion would obviously be higher. But
in the case of these amps, there is plenty of power available to both the
driver stage and the output tube. The power supplies simply drove the
overload cleanly, and the output tube simply accepted the load because it
could get all the current it needed.
It should be mentioned that the 2A3 output tube being run in this case
was operated at 44 ma. plate current at "normal" 2A3 plate voltage. Had
the tube been operated at the usual 60 ma., it would have rejected the
rather extreme mismatch.
All of my new Magnequests are of the ES-025 design as Mikey felt that
this would be rather nice. I agree!
---Dennis---
BTW 45's have LOWER mu than a 2A3, NOT higher as stated, and the reason they're loaded with 5K is that the Rp is double that of a 2A3, 1500 vs. 750 ohms.
Remember-- the harder you load the output tube (up to a reasonable
point), the more IMMEDIATE the sound gets-- the more detailed, deep and
structured. WE LIKE THAT. You also get "more" power out, and more
measured "distortion".
What happens with easier loading? MUCH lower measured distortion AT FULL
POWER. Does this help sonically? Not really-- not ordinarily... because
ordinarily we ARE NOT running full power-- full power measurements, while
interesting, really have nothing to do with musical integrity..... what
we normally appreciate in music.
Does Dennis Fraker have any recommendations for the correct harmonic profile for a SE DC 2A3 amp?
If I set up my FFT with a 1KHz sine wave input, what power output should I be measuring?
How far below the fundamental should each harmonic be in dB's?
This is for his 2.5K plate load recommendation at reduced bias current.
dt 667
Haven't you already demonstrated that lower distortion at full power ALSO EQUATES TO LOWER DISTORTION AT LOWER POWER ? ? ?
"Correct harmonic profile" ? ?? ? ? ? ? Would you believe me if I told you what it is ? I mean, if I made up some numbers and said, "this is it" ??
Considering the numbers of variables involved, is it at all realistic to even try for a specific "profile" ? ? ?
Based on your earlier experiments with a 7B4/12AX7, that you dismissed after measuring distortion and bandwidth issues, how is it that you still find reverence for the same source of this (proven) "less than ideal" advise ?
My question was probably too specific to get any direct answers.
Dennis always seemed to deal with rather esoteric concepts rather than schematics or measurements of any sort.
I never did test the SE 7B4 DC 2A3 circuit on the PC FFT, not yet.
and i hope i am wrong. If so, speak up readers!
Observe, before you think. Think before you open your yap. Act on the basis of experience.
Probably elements of it; probably not all of it.
After my current amp build and speaker crossover upgrade, I'd like to get it on a breadboard and work through some of the options... would probably drive it with an EML 20A or a pentode. Might get to it early in the second part of this year.
Cheers.
“As long as we have any intention to be right… we should be wary. So long as words have the slightest ego attachment, they are dishonest.” Charlotte Joko Beck
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: