Welcome! Need support, you got it. Or share you ideas and experiences.
Return to K&K Audio / Lundahl Transformers
72.131.105.61
In Reply to: Re:300b vs 2a3 posted by Jim Womble on March 23, 2007 at 11:30:28:
What efficiency, more specifically, and what type of volume do you get?
As much as the Karna is the type of design I sooo want to build, practicality tells me I am limited to the 2A3 Amity.
I can tell you that the Karna is much more difficult to build and also much more expensive. More power supply, more DHT filament problems to solve, more interstage iron etc. If you have enough gain you could stick to the Amity which is a wonderfull amp by itself. Or you can build Kevin's design with the 6SN7/6BX7 direct coupled, a design I also would like to try someday.
I wouldn't rule out a three stage amp. You can direct couple the first stage to the second and use current sources for loads for the first stage, not adding more expensive magnetics. It works very well and sounds excellent. The guys here have built several variations of this.
Kevin Carter
K&K Audio
www.kandkaudio.com
Actually, I am well on my way to building a Raven. I decided to start the Raven first, then build the Amity. I read on nutshellhifi that the amp just doesn't sound quite right without a ballsy preamp. Until the Amity is done, I'll be running a triode-connected ST70 driving a set of ME-2's.I hope this path will be sufficient. I like Kevin's input stage, it's just that I'm in the middle of one already.
Kurt, I have used these amps with several different speakers with sensitivity ranges from 88 to 95 db. I don't listen super loud, these amps in either version have had no trouble with any speaker I have used. If I were planning a three stage amp, I would use Kevin's design. The 300b may get a future rework in this direction. I am driving my amps at present with a 12bh7 srpp line stage. It has no problems driving them, but I am building a transformer coupled line stage now (working on power supply today). My only concern with the Raven is it is virtually identical to the input stage of the amity, and as such would appear to have the same issues with source impedance, although I would plan to add series resistance in the grid circuit to allow for a shunt gain control. I am also a little squeemish about back to back stages using the same tube. Some have reported the ecc99 to be a little harsh sounding. My amp isn't, but it is also being driven by the 12bh7 stage.
Hi Jim,I like the ecc99 too, but when I had two in a row the sound got a little, well let's call it 'unfriendly'. I like to suggest the 5687 as a nice sounding alternative. Unfortunately they are not pin compatible.
It looks like the 12BH7A is very similar in characteristics as the ECC99, pinout and all.The Raven is spec'd with an 87 ohm shared cathode resistor and regulated 150V B+. Would this be a good operating point for the 12BH7A? It would be deluxe to just swap tubes and compare results.
The 12bh7 has better than twice the rp, so it may no do as good a job driving difficult loads, the mu is a little lower (16 vs 20), and it can't handle quite the dissipation (if I remember correctly). You would probably do well to plan the cathode r around the 12bh7, you could then drop in the ecc99, assuming you have adequate heater capacity (pulls a little more I think). Some of this I am dredging from memory, as I am at work and don't have any data sheets in front of me. There is another tube that Kevin recommend, once again from memory, a 6n6p, that has higher gain, better section to section matching, low rp, that is pinned the same except for the heater - you would need to be able to switch heater between 12.6 and 6.3 to be about to do in socket swaps - he prefers it in the raven over the ecc99. Another suggestion of his is to run the raven low and hot - 100v @ 20 ma. This is one of the mods I intend to try. These guys have been using transformer coupling for a long time, and know a lot of good stuff.
Jim
I agree, it appears the addition of a Raven basically repeats the first stage. At present I run my stereo out of the iMac, sometimes directly from an iPod. That's my volume control. I also run a powered subwoofer, as I like my bass. So I expect to be needing a pre only to give me volume control and low output impedance to drive potentially a number of amplifier channels. Once the pre is done, I can use my DVD player to run CD's. (Also have been watching the RAKK II, that may be a future project).If I understand you correctly, the Raven burdens the source quite a bit? I have attached my current intent. Impedance varies from 5K to 30K. Am also considering the tribute input transformer. If nothing else, I would like to confirm the preamp section, as that's next in the pipeline. I can always decide on a two or three stage amp design in the future. What I liked about the ECC99 was availability of new product through JJ. Though I continue to hear a lot about the 12bh7, maybe I should consider it. I was avoiding the Karna simply because of the cost and availability of the 45. I want to pick tubes that will be around for a while, and won't cost an arm and a leg.
All comments are desired. I want this to be my last system for a while (wife is getting frustrated with my time allocation- but ME-2's are just about done).
Kurt, you might want to use larger value resistors in series with the grids, that way at lower volume levels the source will see an easier load to drive. The availability of the ecc99 is a nice factor, allthough I have heard the section to section matching can be a little off. The last pair I bought (for the raven) were section matched, costs an additional $4. Compared to the 12bh7, the ecc99 has a much lower rp, making it a better choice for driving tough loads. Having said that, my first version of the raven will probably use 12bh7, as I have a few nos. Nice thing about diy is I can try it and change what I don't like. I have built the power supply as a separate unit with the idea of changing the actual linestage itself.
I actually had increased the value somewhat, based on Kevin's recommendations. I surmised 5K at zero output would be okay, since I'm not hearing anything anyway. At low volumes, I arrive at my 10K target fairly quickly. From there, I was more concerned about putting in too high an ohmic value pot, for high frequency rolloff and noise.Most commercial CD/DVD players should have no problem driving a 10K load, I thought. Or, would it be better in the long run to increase pot value or decrease turns ratio?
I would probably agree with you that a 12bh7 would be sufficient for the preamp stage, and the ECC99 would be a better fit for the driver of the output tubes (or 45 if a three stage amp).
So you are building a Raven? Was your prior post just saying to avoid redundant stages with the same tube, but to still use a pre?
Yes, I am building a Raven, I am just not sure how close to the original design, as I may use different tubes, or op points. As far as the value of the series resistors, you just want to make sure than the source doesn't suffer response issues at lower volume settings. The trade off is you will concede a little gain. Since the shunt pot is in a fairly low impedance loop, I don't think that a higher value would add very much noise unless you went to a very large value of series resistor.
Just to reinforce Jim's point, I have an LL1676 configured 2:1+1 with 5K series resistors and a 100K variable shunt control. The noise added by the line stage is not audible over the residual amplifier hum and noise, which in itself is less than 2mVrms (noise bandwidth 10Hz to 50KHz) at the speaker terminals. The noise from a properly constructed differential tube preamp and 3 stage differential tube power amp, all with AC filaments can be readily made to be plenty good enough for use with speakers having sensitivities in the mid-90sdB. It's not a concern for speakers like the Ariel or ME2 at all.On the subject of 3 stage differential power amps, you don't need a 45 driver stage to have a wonderful sounding amp. There are a host of lower Rp tubes that can and have been used successfully. I have an affinity for the 6BX7GT and the 6BL7GT, both of which I have used in the Art Audio Vivo and in my own amp. Others have used triode-connected 6V6GTs, 6CK4s, the low Rp half of 6EM7s, etc. They are all viable choices.
Kevin Carter
K&K Audio
www.kandkaudio.com
That's probably what I'll end up using then, 2:1+1 with my existing 5K - 100K scheme. Input impedance then varies from 10K to 110K. Any source should be able to drive that.I'll have to think for a while on tube choices. At least that's a few months away.
Thanks guys.
Hi Kurt,It looks like you might be concerned about getting as much gain as possible from the circuit. If you were willing to give up some gain there are a couple of things you could do to raise the input impedance. First would be to use larger series resistors. Second would be to use a different transformer ratio. You could use 1:1 or 2:1 instead of 1:2.
Dave
