|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
76.90.39.24
In Reply to: RE: Maximum Output posted by RPMac on January 27, 2015 at 07:47:30
What CB said.
Also, when you change the idle current you also have to change the voltage and the load impedance to keep the harmonic distortion low.
This is something Dennis doesn't do.
He runs his 2a3 into a 2.5k ohm load, which would be fine for his 250 volts if he kept the plate current at 60ma. but at the 43ma. plate current that he runs, the harmonic distortion would be lower at all power levels if he loaded the 2a3 with 3.9k ohms.
Paul Joppa has a quick rule of thumb that works really well. Load impedance, for a reasonable amount of output power keeping the harmonic distortion reasonably low, = (voltage / current) - (2.38 * the plate resistance)
This follows the Western Electric performance chart and the on line Triode Loadline Simulator very well and is a good rule for all direct heated power tubes.
http://www.trioda.com/tools/triode.html
It's really not about the max power, it's about low distortion at all power levels.
The output stage of Dennis' amp could produce 1/3 the distortion at 1 watt that it does if he just used a 3.9k output transformer instead of the 2.5k output transformer.
But I don't think Dennis understands tubed audio electronics very well so he's gonna do what he's gonna do. It's hard to teach an old dog a new trick.
Dennis is going to point out that a output transformer with higher reflected primary impedance requires more inductance and more wire and there will be more winding capacitance, etc. This is all true but a good winder can wind a 3.9k to 8 ohms output transformer with good frequency response or we can have a custom transformer wound that doesn't have so much inductance. That would keep the winding capacitance low but the bass frequency response would suffer. IMO, a SET shouldn't be playing bass in the first place. That's what Solid State amps are for! :-)
Anyhow, just don't get hung up thinking Dennis is some kind of Guru, he's not.
He's just a builder with a very large ego who thinks he's God's gift to the world, but he's not.
There are a lot of good builders here who know a lot more than Dennis.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Follow Ups:
Obtaining correct Impulse response under changing musical conditions is the electronics that produces the best musical results.
NOTE: The QUALITY of the output trans is a larger factor than its loading-- if loading is already in the Ballpark. Mike and I designed my newest transformers for the 43 ma. load.
Guess what? We decided on 2.5K-- again. We changed a few other things in the transformer formula to get what I want musically. I agree that extra inductance helps "control" woofers. Guess what again-- it's the WRONG way to control woofers! It ALSO slows down the woofer's impulse response, and reduces overall clarity..
I have a proven way to absolutely control woofers-- my design does it to perfection. Build a LSES power supply, and then get an output tube that can be loaded "hard" without distorting music, and then use short, large wiring in the amp to tighten the speaker-to amp's components interface. Sure does work for me!.
I won't discuss whether either of us transformer nuts "doesn't know much about electronics". That depends-- Whose electronics?
Another factor is the 2A3 vacuum tube. If you're running or calculating the NOS RCA-- single-plate or BiPlate-- I don't care-- I've used 'em all-- I do feel for you!
Before you play GOD, why not get some GOOD tubes, and MEASURE the impulse/loading tests on THOSE? I'm thinking EML and JJ-2A3-40, maybe an ELROG some fine day?.
Heck! I could load-up a JJ-2A3-40 at 1.5K, and it would still Kick-Ass as long as the power supply could support it, and distort very little, if run right. No, I'm not that crazy, 2.5K is a terrific compromise with THIS transformer design in THIS amp.
The JJ 2A3 triode-- this isn't Pentode-loading!-- is very tolerant of loading variations, which allows one to "load 'er-up", and get insanely different results than those old tubes (and antiquated published specs) could ever deliver.
What happens if you sock an old 2A3 into one of my amps? Nothing, really. It runs fine at 43 ma. But is it an ideal match for this design? NO. The JJ and the EML are.
The old formulas DO NOT APPLY to modern tubes. Especially the ones from Czech and Germany. SOB!!!!
I want to try some ELROGS! Can anyone talk ELROG into a 2A3 and maybe a 45?
Do I have to step down (musically) to a 300B to get an Elrog?
---Dennis---
"The JJ 2A3 triode-- this isn't Pentode-loading!-- is very tolerant of loading variations, which allows one to "load 'er-up", and get insanely different results than those old tubes (and antiquated published specs) could ever deliver."The JJ 2a3-40 data sheet says the plate resistance is 800 ohms, same as an old RCA.
Are you saying that the JJ tube's plate resistance is wildly different than what is stated by JJ on their data sheet, allowing it to be happy into 1500 ohms at 43ma./250 volts when clearly the RCA would not be?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 01/28/15
Nope! I'm not saying that at all.
The JJ simply doesn't go into minute convulsions at 43 ma. if it's "overloaded"-- (distortion!)--- anywhere nearly as much as old-design tubes-- just because the load changed a little bit.
It's FAR more load-tolerant. It's like a CAT D-9 up against a Bobcat.
I think we both can predict the outcome of that.
Sure-- JJ had to "meet" the old specs-- otherwise, people won't buy the tube and plunk it into their creaky old amps, and yes, the advertised plate resistance is 800 ohms. I said ADVERTISED-- and yes, it will measure-out-- it's not wrong. If one uses the old calcs. for a 2A3, and plunks this tube in, it will operate just like the old NOS tube. But, there, the similarities END. .
This is a new, better, and far more versatile tool, so learn to use it-- which is just what I did. When I get a new tool, I find out what IT does--- I already know what the old tool does-- but I'm going to use the new tool, so I study IT as well as the old tool.
These tubes (JJ and EML, etc.) are designed to operate at far greater dissipation levels than old 2A3's. The JJ-2A3-40 will dissipate up to 40 watts, not crap-out at 15.
At 250 volts across it, and at 43 ma. plate current, one might expect that it would be rugged alright, (it is), but it might not cater well to the low plate current. NOT SO! It sounds clean, clear, fast, ultra-dynamic, super-powerful and has a super-wide bandwidth.
Into 2.5K? That's coasting on a Banana Peel for it. This thing could drive 1K! No, I'm not going to design for 1K, but I'll bet that this tube could drive it.
Since you're a measurements man, I suggest you set up to do what I used to do.
RUN YOUR OWN PLATE RESISTANCE CURVES. Then, you know what is happening for YOUR chosen specs & OP. points, etc.
To do this, I used a variable 0-650 VDC supply for the plate. I tapped a battery set at the voltages I thought would be within the ranges where I might want to operate the tube-- for the applied grid-bias.
At each grid bias point, the variable Plate supply had previously been set at ZERO. At each point, I measured the grid bias. Then, I upped the D.C. voltage on the plate in steps-- At each step, I measured both voltages.
Of course, plate current was being measured all the time.
I drew up a graph, and plotted all the op. points. (use commercial graph paper).
I did this on many tubes. NONE of them were ever perfect, some fell-in close to specs, (At SOME op. points), others did not.
EIMACS, in particular, were off quite a bit. I made about 25 plotted points for each tube.
This was more important when analyzing small-signal tubes, but I did it for all of them.
It's just another item for your basket of calcs. If you're going to do all this calculated stuff, then this will help you do it right.
---Dennis---
Dennis, I'm having trouble understanding what you are saying.
1. You said "...the advertised plate resistance is 800 ohms. I said ADVERTISED-- and yes, it will measure-out-- it's not wrong."
First you say the plate resistance is "advertised" as being 800 ohms.
Now this statement to me means that the plate resistance is something other than 800 ohms. But then you say it's not wrong?
So do you mean that the plate resistance is really 800 ohms?
If that is what you meant, why did you first say "advertised"?
So which is it? 800 ohms or something different?
2. You said "The JJ-2A3-40 will dissipate up to 40 watts, not crap-out at 15." but at 43ma and 250 volts that's only 10.75 watts of plate dissipation.
Why does it matter that the tube will do 40 watts if you are only going to use 10.75 watts?
As the tube operates along the load line, to the left of idle, the current goes up as the voltage goes down.
As the tube operates along the load line to the right of idle, the current goes down as the voltage goes up.
The plate dissipation stays (more or less) the same.
3. You said "Into 2.5K? That's coasting on a Banana Peel for it. This thing could drive 1K! No, I'm not going to design for 1K, but I'll bet that this tube could drive it."
Didn't you already say the plate resistance IS 800 ohms?
If the plate resistance is 800 ohms then a 1k ohm load impedance would cause a very vertical load line and the harmonic distortion would be very high because the operation to the right of the idle point, following the load line, will drop almost immediately into the non-linear region of the plate curves.
The fact that the tube will dissipate 40 watts does not change that. Only a lower plate resistance would change that or a MUCH higher idle current. The angle of the load line is directly related to the plate resistance not the tube's plate dissipation capability.
In all three of the above examples I really don't understand what you are trying to say.
Dennis, I've been to Montana and I know first hand that people from Montana speak English but when I read your posts I don't get the sense that you are speaking English. I don't know what language you are using and I don't understand what you are saying.
Do you?
In all honesty,
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Hey! I get it, TRE!
NEVER challenge a published Dogma. Oh, but I do!
Bring ANY amp that you think is THE best, and I'll run it against this design-- the one that you prefer not to understand-- on our system at RMAF 2015.
This offer is open to anyone who brings a SET that is built right, and looks like it.... something the Demo room can justify showing to the public.
IF all of the people present choose yours, we will run it-- and name it's builder-- to the public.
If that happens, my design will go on the Display table, yours will get to play.
I will not vote unless there's a Tie.
---Dennis---
If there's a tie it means yours is just another 2A3 amp and doesn't require your vote. For that matter, if only 60% of the listeners prefer yours, it's just another 2A3 amp.
!
The Mind has No Firewall~ U.S. Army War College.
The answer is in my last post and is obvious-- a set of dogma proves nothing-- performance and longevity proves everything.
I've had enough of this crap. Good Day.
---Dennis---
this request either."Dennis, do me a favor?
With your amplifier playing music at your normal listening level connect a scope at the B+ end of the output transformer. Adjust the scope for the most sensitivity. Be sure to tell us what that setting on your scope is (volts per division). Do you see any waveform on the scope? If you do, what's the amplitude of the waveform peak to peak?"
Or any of the many other straight up questions I and others have asked you, over the years, where you never gave a straight up answer.
Thanks anyway Dennis. Have a nice day.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 01/30/15
This message has been moved to a more appropriate venue .
Edits: 01/30/15
What does any of that have anything to do with the questions I ask you?
Does the JJ 2a3-40 have a 800 ohms plate resistance or not?
What does the fact that the JJ 2a3-40 will dissipate 40 watts have to do with it's performance in an amplifier where the idle plate dissipation is only 10.75 watts?
If the JJ 2a3-40 has a plate resistance of 800 ohms how can it operate, with low distortion, into a 1k ohm load with the idle current at 43ma?
These are not technically hard questions.
Why can't you just give me a straight answer?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Hey! I get it!
Bring ANY amp that you think is THE best, and I'll run it against this design-- the one that you prefer not to understand-- on our system at RMAF 2015.
This offer is open to anyone who brings a SET that is built right, and looks like it.... something the Demo room can justify showing to the public.
IF all of the people present choose yours, we will run it-- and name it's builder-- to the public.
If that happens, my design will go on the Display table, yours will get to play.
I will not vote unless there's a Tie.
---Dennis---
Interesting that EML 45 data sheet says 3.9k for lower voltage/current operation too.
Doesn't specify for 2A3.
I assume this is all for 8ohm on 8ohm tap.
If you use 16ohm speaker, don't know what Denis uses, on 8ohm tap then that effectively doubles the load impedence seen and reduces power and distortion still further. Am I correct? Need help here.
OOI I use a 5K with 16ohm on 8ohm tap for my 45s. Maybe I should use a 3.9k next time around.
Cheers
Up tapping (16 ohms on 8 ohm tap) does indeed double the reflected impedance seen by the output tube.
Because the impedance has changed but the inductance of the primary has not, the -3db point, in Hz, of the bass response will double.
If, for instance, the -3db point (of the response of the transformer, not the system as a whole. There are many other things involved with that) would have been 20Hz with 8ohms on the 8ohm tap then it will be 40Hz with 16ohms on the 8ohm tap.
There's nothing magic about 3.9k ohms as a load.
If you know the plate resistance, the voltage across the tube and the idle current you want to run......the formula will tell you what a good load impedance, for that tube under those conditions, will be.
3.9k ohms just happens to be a good load for Dennis' 2a3 assuming he's running 250 volts, and 43ma. with a plate resistance of 800ohms as he seems to report.
What impedance would be good for your 45 the way you're running it? I don't know. I don't know the particulars of your output stage but 10k on a 45 is probably pretty good.
Once you do the math you can always go a little higher giving up some max power for even lower HD.
BTW, I also up tap my ElectraPrint PSSS output transformers.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Up tapping (16 ohms on 8 ohm tap) does indeed double the reflected impedance seen by the output tube.
Check
Because the impedance has changed but the inductance of the primary has not, the -3db point, in Hz, of the bass response will double.
This depends if you are using a triode or a pentode. From the simple model the -3dB point is where the inductive reactance is equal to the parallel combo of the Tube Rp and the reflected load so simply doubling the reflected load so keeping the tube Rp constant typically makes the push up in frequency much smaller than an octave for a triode but does approximately double it for a pentode
dave
".... From the simple model the -3dB point is where the inductive reactance is equal to the parallel combo of the Tube Rp and the reflected load ..."Thanks for that information. I stand corrected.
Dave, can you tell me what happens at the other end?
What effect on the high frequency -3db point does "up tapping" the output transformer have?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 01/28/15
What Dave gave you in the table was the "small signal" response. Now, look at the "large signal" response and see what happens when you (via creative tapsmanship) reflect 6400 ohms to the primary as opposed to it's design center of 3200 ohms.
MSL
Builder of MagneQuest & Peerless transformers since 1989
is there a table or formula I can look at?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
for an explanation of the differences btwn small
signal response and large signal response see:http://www.magnequest.com/tech_article2_voltsecond.htm
for further discussion of the effects of L vs "reflected impedance"
see:http://www.magnequest.com/tech2.htm
Builder of MagneQuest & Peerless transformers since 1989
Edits: 01/29/15
"Phase Shift Desired
Multiplier
5 degrees
11.43
10 degrees
5.67
15 degrees
3.73
30 degrees
1.73
45 degrees
1
Here is an example of how to use the above multipliers. Say you have a 5,000 ohm nominal primary impedance and you desire only 15 degrees of phase shift at 100 hertz.
So at 100 hertz we would want (by referencing the table above) an inductive impedance of 3.73 times 5,000 ohms.
Solve for the following;
18,650 equals 6.2834 times 100 times L
L is the unknown variable. In this case L must be 29.6814 henries."
So if the 5K is changed by up tapping to 10k the required inductance for a desired phase shift will double.
So for a -3db point (phase shift of 45 degrees) L = reflected impedance / (6.2834 * frequency). (added to my spreadsheet)
If I double the reflected impedance by up tapping the output transformer the -3db point goes up an octave.
3200:8 with 25.5Hy will do 20Hz at -3db. Connecting a 16 ohm speaker will reflect 6400 ohms and the -3db point will be 40Hz. To maintain 20Hz at -3db would require 51Hy.
Thanks Mike.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Hi Tre:I didn't check your calculations or anything but I think
you got the picture.Just using a small signal model will make some crudballs
look like their ready for prom night...the tougher test.. and the more realistic... is the large
signal test.I found the following text on a Lundahl spec sheet...
to explain large signal vs small signal....FPL= large signal
FRL= small signalWe define Power Low Frequency Limit, FPL, as the frequency where wLP = RLOAD. (The reactive impedance of
the transformer equals the primary load impedance). At FPL, the output power is reduced to 50%.We define Response Low Frequency Limit, FRL as the frequency where a (small) output signal is reduced with
-1 dB due to finite primary inductance. FRL = w / p, if you solve w in
wLP = (RLOAD in parallell with RANODE).
In the FPL when the power is reduced by 3db the phase angle of the load impedance will be 45 degrees. Instead of your loadline being a straight line it's going to begin to look like a NFL football (i.e., ellipse). Your tube generated distortion is going to increase dramatically and that isn't even taking into account that the transformer borne distortion will also be going up at the same time.Plate curves assume a linear with frequency unchanging
pure resistance is loading the anode of the tube. If your instead loading the anode with a transformer... nothing could be further from
the truth. Now you must take into account the characteristics and functioning of the transformer... cause now we have reactive components in a complex relationship with "pure" resistances.Now the effective loadline can look more like a football or beachball.
And this elliptical loadline changes shape with frequency... so that your tube now has to drive an everchanging loadline....and it aint going to be pretty when having to do all that work.So instead of focusing on the "easier" small signal response... looking at and evaluating the large signal response (the decidely tougher test) will be much more predictive of how well your tube and circuit will perform.
MSL
Builder of MagneQuest & Peerless transformers since 1989
Edits: 01/29/15 01/29/15 01/29/15
"It does not matter if your speaker does not go down to 31.25 Hz or not. If you are sending low frequency information to the amplifier, the amplifier will make the plate voltage move. If the plate voltage moves, the tube must spend bias current to charge and discharge the primary inductance as well as the actual primary load impedance. If the plate voltage or current "clips" because of low frequency plate excursions, the sound will suffer. "
"With 27 H, the load line is nasty. With a 27 H primary if there is any low frequencies making it from the preamp to the output transformer, I would expect the sound to be a bit nasty too. "
The practical implications for me; I just placed the high pass filter cap back in series with the input to my SET amp that drives my midrange (and tweeter) that only plays to 200Hz.
Thanks again Mike.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
That's a very important point that Voltsec made--- cause a lot of people will initially say I don't care what my amp is doing down at 50 or 40 or 30 hertz... cause my speakers only go down to X... but as Voltsec points out by then the damage is done....Your response--- using a high pass filter--- it's generally regarded as a taboo approach in hi end circles--- but I think the approach has a lot of initial merit to it. Frank Van Alstine in the mid eighties started using a high pass filter on the input of his ST-70 mod/circuit cause (his reasoning) it did not allow the circuit to be swamped with frequencies which it could not handle.
Or spec and use a transformer with really, really good capabilities in this respect (i.e., large signal goodness).
And to bring this back to "creative tapsmanship"-- i.e., ratioing transformer impedances-- don't do it--- your not likely on the vast majority of transformers to have enough "surplus" L to keep the loadline from becoming a beachball. And your tube really won't be all that happy with that.
MSL
Builder of MagneQuest & Peerless transformers since 1989
Edits: 01/29/15 01/29/15 01/29/15 01/29/15
"Your response--- using a high pass filter--- it's generally regarded as a taboo approach in hi end circles--- but I think the approach has a lot of initial merit to it."If this was a full range amp I might regard it as taboo as well but this is a bi-amped system with a SS amp driving a pair of 15" JBL 2231s using an active 12db per octave low pass filter at 200Hz.
The JBL 2118J midrange speakers roll off on their own and are -3db at 200Hz.
I think keeping some of the LF out of the tube amp with a cap should help the issue that Voltsec points out and I believe this would be true even if my tube amp's OPT had "full" inductance.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 01/29/15
....Paul Joppa has a quick rule of thumb that works really well. Load impedance, for a reasonable amount of output power keeping the harmonic distortion reasonably low, = (voltage / current) - (2.38 * the plate resistance)....Interesting, you are saying that with higher Ia and higher Rp the ER300b should be loaded with about 2k ? Whould it reduce output power even more ?
Edits: 01/27/15 01/27/15
I would have to know more about the ER300b.
How did you get 2K?
Do you know the plate resistance for the ER300b?
When you're dealing with a SET amp (unless your taking about transmitting tubes) power should not be your concern. You should have high sensitive speakers that don't need any real power. The sound (really low upper ordered harmonic distortion) should be your only concern.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
"little less power then the traditional 300B"Where did you read that?
The site I read says it matches the WE 300b
"Conforms to all specifications and tube curves of the Original Western Electric 300B tubes."
Is there a website with complete information?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 01/27/15
There are a few inconsistencies in the trueaudiophile notes; it's a very new tube so give them some time to fix errors and typos.
At VinylSavor
.
Have Fun and Enjoy the Music
"Still Working the Problem"
Impulse response under changing input conditions is the name of the game.
---Dennis---
Of course and operating the tube in a linear way will give that.
How is loading down your output tube to the point where it produces much higher harmonic distortion than it needs to going to give you better "Impulse response"?
You're the one who says you only need less than a watt. The first watt gets a lot cleaner when the output tube is not loaded down.
But I'm just wasting my time. I can't tell you anything, you already know it all!
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
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: