|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
71.199.88.151
I am building a 2A3 SE amp from scratch with the help of a friend since has built over 20 different designs. He recommends using a 5K but I'm not so sure about that yet. I haven't quite decided a front end yet but at first it will probably be RC coupled, maybe interstage later. I built some SE amps about 10 yrs ago and was disappointed probably due to the crappy iron so I want to do it right this time. I would also prefer the iron to be rated at least 100ma in case I decided to use the 300B tube one day. I'm using some highly modified Khorns with constant impedance crossovers. My flavor would to have be an extended bottom end. So with that said could anyone recommend something for this project. Or maybe a custom would be in order with 3.5 and 5K taps??
Follow Ups:
I continue my (apparently hopeless) quest to get designers to understand that you cannot separate the transformer impedance from the operating point (voltage and current) of the tube. :^)
A simple non-dimensional analysis based on the 3/2 power law gives the following relation:
RL = RB - 2.38*rp
RL is a reference load impedance; RB is the "beam resistance" of the tube (plate to cathode voltage divided by plate current); rp is the plate resistance at the operating point.
This reference load impedance provides the same low distortion, independent of the beam resistance. You can use a lower load resistance, as low as 1/2 the reference value; you will get more distortion at full power, but also more power. (Jeff's 250v/42mA/2500 ohm recommendation below is about 0.62 times the reference load, for example.) The distortion reaches the point of diminishing returns for loads much greater than the reference value, while the power continues to drop, so those higher impedances are rarely used.
The things that higher load impedances provide are greater damping and greater efficiency. On the negative side, higher impedances require higher plate voltages which are a factor in tube longevity.
Finally, inductance is important for bass as others have noted. Typical series-feed output transformers run 4 to 15 henries per kOhm impedance, with 8-10 henries being pretty good. For example the Magnequest DS-025 mentioned below is about 20 henries at 2.5kOhms. Note that a 5K transformer would need 40 henries to provide the same bass performance.
You simply haven't a peer to read, understand, and be willing to participate in corroborating your story. Therefore, I suggest warning that your posting is technical so most will not have it fly over their heads and they can skip your post. :-)
As for me, I have not seen the physics you wrote about on that claim myself, so I am not yet there with you. But it sounded good. If you also had an independent internet link supporting your message, I would like to read it. Thanks, Paul.
-Kurt
Well, I did the analysis myself and have not published it. The 3/2 power law goes back to (I think) 1917, also known as Child's Law or the the Child-Langmuir Law. All it says is that the current is proportional to the (plate voltage minus mu times grid voltage) to the 3/2 power.
The analysis itself is very simple. Choose an operating point and plot it on the plate curves (see above), then find the point on the zero-bias curve where the current is twice the quiescent current. A line through those two points is a load line, whose slope is my definition of the reference load resistance. The rest is just Child's Law and some algebra.
Now I see where you're coming from. Thanks, again.
-Kurt
You are most welcome. Some years ago we did a 3-day amp-building class, and I talked on design issues. That slide is from the Friday afternoon lecture, which is why I had it handy for posting.
It would be nice to see a comparison graph of a 300B loaded with a 10K, 5K, and a 2.5K power resistor as "primaries". It would show THD vs. output power at 1 KHz with these 3 lines on the same graph.
And for a more thorough investigation, do it again if we had similar sized OPTs and some kind of minimal speaker load and plot THD vs. power again across the speaker, 1 KHz and 30 Hz.
Then this can show something about where they differ and how much, and about the inductance and so forth.
-Kurt
Hello Paul,
Any chance of you putting up your design notes, so some of us can learn from them?
Thanks.
I'm working on some text to go with the slides, but it's slow work - there are many higher priority things to do, and of course I continue to learn more and have to adjust the notes to correspond!
Of course? You're a much needed exception, and not the rule about the minds of older men! :-(
-Kurt
"The distortion reaches the point of diminishing returns for loads much greater than the reference value,"
As a percentage, can you say when?
Thanks, Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
The WE 300B data above shows that you get about 3% second harmonic and 0.6% third harmonic at the reference load impedance - which I called RL(opt) when I made this plot. Less linear tubes will have more distortion of course; I have seen as much as 5-10% with some tubes such as regulator and TV tubes not designed for linear audio service.
Notice that the distortion values are tightly clustered, while the efficiency cluster is more broad. That's because the actual beam resistance plays a role in efficiency but not in distortion. These plots are just the standard WE table as published with the original 300B data sheets.
I have noticed unloading (higher) plate impedance from manufactures recommended specifications makes for lower distortion. What I find shocking is I hear distortion when many others DO NOT hear it. And, distortion is unpleasent. Anyways, the quality of the output transformer is more important than the 3.5K vs 5K impedance. I will have both high quality audio transformers and the 5K impedance. 2A3s are most linear at 250 volts plate.
I'd recommend the James JS-6123HS, you can go 2.5K, 3.5K, or 5K. I replaced some highly regarded transformers on my 300B amps, and am very pleased with the results. These transformers have very good bandwidth, and really sweet mids, I've got great bass even at 3.5K, my speakers are flat to 30hz. IMO, the best bang for the buck in transformers. Here's a link to a supplier in Canada.
twystd
My friend, Joseph Esmilla of JE Labs, recommended the James to me. Very happy with the results.
Abe
Hi,You state this :
"I'm using some highly modified Khorns with constant impedance crossovers. My flavor would to have be an extended bottom end."
here is how I would do this :
K Horns are about 105 dB, you will operate at about 1/4 to 3/4 of a watt most of the time, so you do NOT need to use or consider the worse-sounding 300B tube.
You can easily use a 2A3, or even a Type 45 on your load. My vote would be to build a JJ-2A3 40 amp with 250 VDC P-K and about 42 to 43 mA. current.
"My flavor would to have be an extended bottom end."
To get that GREAT low end, (and great voltage stability, self regulation), you want to have low DCR iron in the power supply / B+ filters. Do a Pasternak Flywheel L1 using a Triad C-56U with 20 uF C1, and use the new 27 Ohm DCR Edcor 2 HY Ls as L2, with 50 uF as C2, feeding each 2A3 tube.
This will "ROCK your socks off" on the low end.
It would help to have a low DCR power trannie also, about 20 Ohms DCR (or less, I use 9 Ohms) across half the HV secondary (center tapped ground to one end of the HV winding).
Then you are on your way to "fly with the eagles". Use two 5U4s to rectify the B+.
Load the 2A3 hard, use 2.5K to 3.0 K for the output trannie, not 5K !!! No "polite" amps please, we want the K Horns to sound realistic. Lets boggie some cats !!
Jeff Medwin
Edits: 07/11/11
Jeff, I thought the JJ 40 was a 300B with a 2.5V filament.
I agree about the 2.5k primary impedance. I think the Magnequest DS025 is an astonishingly good output transformer at a very reasonable price. I have op transformers from several of the other premium winders and I just can't quit listening to the DS-025s!
Hi,No, its a 2A3, different grid spacing, and best of all, LINEAR cathode emissions, eight vertical, parallel filament strings, unlike most all "vintage" DHTs with non-linear "W" or "V" shaped filament structures.....save for the original RCA 2A3 monoplate.
The tube is actually called a JJ "2A3"-40. Yes, it has a 40 watt 300B plate, but thats all.
At 250 P-K and 41-43 mA., it will sound great and last forever, well, almost forever, tens of thousands of hours. So, it is a BARGAIN... an audio give - away.
Jeff
Edits: 07/12/11 07/12/11
Thanks Jeff, that's the first real explanation of the JJ 2A3 40 that I have heard. I may have to get a pair.
I just got a pair yesterday - No comments yet as they are too new to tell.
I got mine from Eurotubes. Arrived 3 days after I ordered and @ $234US shipped the best price I could find.
Hi,
Once broken in, you may find it needs some help with its balance in the extreme highs. This can be easily adjusted by ear, with multiple film bypasses to the Rk bypass cap, so as to get the tuned response you want to hear up high. Have fun, share your experiences with us all.
Jeff Medwin
That's a great price. The cheapest I could find was Antique Electric Supply at $242.50. I found some of the usual suspects wanted over $500.
5K is often chosen over 3K for more damping control. For the 300B, many people prefer the 5K option over 3K.. Keep in mind that the 5K load also permits the 45 - which you might not want to try with a 3K load.
I've always been impressed with the Parafeed output stage for extended bass performance... In many cases, you can convert an air gapped transformer into duty as a plate choke - and add the parafeed capacitor and transformer after the fact. I've done with with both my 845 amp and an 811A prototype and I did notice in each case that the bass performance was more impactful and controlled.
I have a 6W 845 amp - fixed bias - 6.6K load, front end is IT coupled. I can piss off neighbors will all my 6W in my urban neighborhood on 98dB eff Klipsch RF-3s. I reckon with your K-horns that you would be more than happy with a 300B... After all, the prices on 300B vs. quality 2A3 are not that far away from each other.
I'd def. do fixed bias too....
-- Jim
for low freq amp, try to find an OPT with primary inductance of 20Hy plus (30Hy would be better). whether 2A3 or 300B (I'd opt for the 300Bs greater power)locate your least high order distortion operating point (there are several posts here in AA on that, Paul Joppa's, Thorsten Loesch's etc.., or you can go to WE website for tabulations)and since you're on the DIY route, use the lowest Z out driver/configuration, with enough muscle to drive the final/power tube (1mA, 1M plate out tubes need not apply).
and yes, your intended service (low freq amp), requires a stiff power supply (voltage doesn't waver whatever the material you modulate with)
enjoy.
I'm looking at one of Dave Slagles 5k:8R outputs to drive a 300b tube. I think the inductance will at least 25-30 with the 5K input.
Working on a driver now, may use a passive volume so I may need some gain. What do you think about interstage xformer?
a higher > 5K will increase inductance {READ better bass, all things being equal} @ the expense of power. A fair trade any day for my choice.
The Mind has No Firewall~ U.S. Army War College.
If you want to measure the performance on an 8 ohm resistor for maximum power with minimum distortion into that test load, I would go with 3K.
You have to really look at the speakers you are planning on using and consider the impedance characteristics across the frequency band to determine which impedance might be better. If you have a nominal 8 ohm speaker with a 4 ohm dip in the bass, I would go with the higher impedance. For all practical purposes, it don't matter much. Just get the best iron you can.
I have the Khorns with a constant impedance crossover, the amp always see's a nominal 8 ohm load
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: