ray_moth said:
is this it?
http://www.diyaudio.com/forums/attachment.php?s=&postid=779085&stamp=1133366034
I guess we posted at the same time! 🙂
My measurements indicate this also. And the bandwidth also. It's the 6SL7 that's throwing things off. A little 2nd harmonic. 😉
And cheaper too without a CCS! Remember they could've used a toob CCS, they were available in kit form as well and without a CCS they would be easier to construct. And as a result the company felt that it wasn't worth the extra effort and the resulting complication and money; against what the engineering staff felt most likely! 😀
Not to mention the extra toob. I never did like the sound of an SRPP, IMO of course. And absolutely not necessary. And by going with triode operation you need a driver feeding it with a low output impedance. Though ray_moth's circuit is a very good one, I would try to get rid of the extra stage if possible.
About 39 - 40dB would be ideal. Less NFB would be needed to get around a 775mV input sensitivity (my preference) as you may run into stability problems and a compressed, lifeless sounding amplifier with high amounts of NFB.
Running small siganal tubes at max plate dissipation, very high currents is not necessary or desirable especially in this application.
navin, here's an EL34 PP triode operation with 390V B+ I did in Tube Cad Journals PP Calculator. About 9 watts class A w/6000 ohm P-P load. I wouldn't go with much higher Iq=55mA, 21.5W plate diss per tube. With a 6600 ohm primary you could go with a little higher Iq but a little less total PW but a little more class A output.
Cheers
Wayne
Attachments
cogsncogs said:
the extra stage in ray moth's ckt? I thought you cant get rid of ti thanks for feedback? Ray moth's ckt uses 1 6SL7 per channel your version uses 1/2. I am a bit confused as tto which ckt to use esp since I dont have any 6SL7 yet. I aim to get them though.
9W class A. That would mean 15-18W Class AB right. I was looking at a Iq of 40-50mA.
what about UL? How much power would I expect then? 30W Class AB at 5K?
Also I thought a bit about your escape from reality and realised that the advice and help you offer here must also give you solace. You are a generous and brave man, Wayne!.
If you noticed ray_moths circuit, NFB is taken from the plates of the EL34 and the OPT is out of the FB loop. This has some very clear advantages. Here's where good iron is needed and you don't need NFB to whip 
not so good iron 
into shape! Ringing, what ringing?!! 
😀
Wayne
not so good iron into shape! Ringing, what ringing?!! 😀Wayne
@navin
In ray_moth's circuit the extra stage is needed because of where the NFB is taken from. The pic/chart I posted is for PP-Triode operation. In PP-UL (pentode with plate signal i.e. 43%, injected into the screen grids) mode you will get around 30-40W depending on B+ and the OPT primary load. You could go as low as a 4.3k primary, about 30mA Iq per tube, 450V B+. But with a 4.3k primary/load the output power will be higher but the distortion will be a wee bit higher also.
With a 5k OPT, around 33mA Iq, 450V B+, fixed bias you should get 40W easy. A good quality EL34 will handle this (450V) in UL mode.
You could use tube rectification for the triode/treble version as some will say this will give you smoother highs and SS rectification for the UL version. As you go more towards class AB1 PS demands become greater. As in the chart I posted using this as the treble unit you will be staying more in class A operation with occasional peaks into class AB1. And as this being HF material the peaks will be short depending on what kind of music/noise 😀 you listen to.
Cheers
Wayne 😉
In ray_moth's circuit the extra stage is needed because of where the NFB is taken from. The pic/chart I posted is for PP-Triode operation. In PP-UL (pentode with plate signal i.e. 43%, injected into the screen grids) mode you will get around 30-40W depending on B+ and the OPT primary load. You could go as low as a 4.3k primary, about 30mA Iq per tube, 450V B+. But with a 4.3k primary/load the output power will be higher but the distortion will be a wee bit higher also.
With a 5k OPT, around 33mA Iq, 450V B+, fixed bias you should get 40W easy. A good quality EL34 will handle this (450V) in UL mode.
You could use tube rectification for the triode/treble version as some will say this will give you smoother highs and SS rectification for the UL version. As you go more towards class AB1 PS demands become greater. As in the chart I posted using this as the treble unit you will be staying more in class A operation with occasional peaks into class AB1. And as this being HF material the peaks will be short depending on what kind of music/noise 😀 you listen to.
Cheers
Wayne 😉
The original concept, on which my circuit is based, should be credited to Thorsten Loesch, alias Kui Yang Wang. He has successfully used this technique of cross-coupled NFB in his adaptation of the Shanling amp. The actual component values are my own and they work very well.
This design is easily adapted for UL, adding more NFB to maintain adequate damping and some correction at high frequency to maintain stability.
Gain would be marginal for pentode operation, however, due to the far greater amount of NFB required to attain a reasonable DF. It could be done, albeit with less sensitivity (about 800mv for 45w output).
There's no reason why an SS device shouldn't work as a CCS - I'd be inclined to use a MOSFET such as the IRF820, as in Allen Wright's PP EL34 triode amp. However, I use a pentode CCS because that's what I have av ailable, it keeps things simple (for me, anyway) and there is no risk of damaging the device through accidental exposure to high voltage (I've never had much luck with SS). It needs a higher negative rail than would be the case using SS, of course. The pentode heater is at the neg. rail potential, to avoid any risk of exceeding maximum permitted heater-cathode voltage.
This design is easily adapted for UL, adding more NFB to maintain adequate damping and some correction at high frequency to maintain stability.
Gain would be marginal for pentode operation, however, due to the far greater amount of NFB required to attain a reasonable DF. It could be done, albeit with less sensitivity (about 800mv for 45w output).
There's no reason why an SS device shouldn't work as a CCS - I'd be inclined to use a MOSFET such as the IRF820, as in Allen Wright's PP EL34 triode amp. However, I use a pentode CCS because that's what I have av ailable, it keeps things simple (for me, anyway) and there is no risk of damaging the device through accidental exposure to high voltage (I've never had much luck with SS). It needs a higher negative rail than would be the case using SS, of course. The pentode heater is at the neg. rail potential, to avoid any risk of exceeding maximum permitted heater-cathode voltage.
Aww shucks! Thank you! Many people would say/think that our playing with toobes is unrealistic/fantasy, nostalgic when we have all this new, modern plastic encased SS (chips) with tons of FB stuffed amplifiers/Home Theatre equipment and PLASTIC speaker systems!
Why do we bother with all of this when we can go to Wal-Mart or Rat Shack and buy some plastic boxes? It's because we want to escape from reality! 😎 All joking aside I get something that's hard to describe when by some chance I actually help someone and to help them enjoy something very dear to my heart; those things that look like little light bulbs ---> TOOBES! (warm-glo)
Cheers
Wayne :=B
P.S. I'll post the diagram you want ASAP! 😉
cogsncogs said:
I was looking at a 5K primary for UL and 6.6K primary for triode with 400V B+. Hoping to get 35W and 15W in AB1 respectively.
The reason i am consdiering Class AB is that I understand that in Class A the heat generated would be too high esp since our ambient temps run around 35 deg C.
The other reason I am looking at SS rectification is that I will be using each amp will have one channel in Triode and one channel in UL. SS makes things simpler.
My EL34 are from Svetlana got from Singapore. I dont know how good they are but they cost $1 more than Electroharmonix.
AB1 should have lesser power demands and generate less heat right?
most of what i listen to is rock from the 60s and 70s with some Jazz, Classical and Blues throw in.
ray_moth said:
TL is famous from the days when he built his own version of the WATT/Puppy.
a bit OT:
can you have a look at this and comment please, i wanted to build a simmilar drive with 2 6CM7's which are easily obtained here, i need more voltage output for screen grid drive.
thanks in advance.
can you have a look at this and comment please, i wanted to build a simmilar drive with 2 6CM7's which are easily obtained here, i need more voltage output for screen grid drive.
thanks in advance.
Navin:
: you've got the impedances the wrong way round.
: if these are "Winged C" Svetlanas, then they should be good. If they're not, then they are no better than Sovteks.
: yes, at low volumes. Power demand of AB1 goes up as volume increases.
cogsncogs said:
Hey, Wayne,
We bother with "all this" because we are seeking reality. Realisim in music that is. I live in India where Bose is the undisputed KING of anyone who has made it. A Mercedes Benz in the garage and a BOse HT at home lets people know you are arrived. There are dozens of local Bose immitations and BTW I dont know one person in Mumbai who has DIYed an amp SS or Tube. I know 2 others in India who have but they live very very far (near Calcutta and in Madras if you have a map).
In India if you are rich you buy a HTiB from Samsung, LG, Sony etc...if you are really rich you buy Bose and if you are stinking rich you buy a politician (lord knows they - HTiB, Bose and the politician -all make the same noise).
Thanks for all your help. I hope it makes you feel better too! 🙂 If you are ever in my neck of the world let me know. I owe you big time!
ray_moth said:
Thanks SY about the impedances.
What do C winged Svetlana's look like. The tubes I got say "Svetlana Electron Devices, EL34 Made in Russia 05 06". There is also an "OTK 2" printed on one side.
So at full volume Class AB1 requires more power than Class A? Does that mean that at full volume a 10W Class A amp is more efficient than a 10W Class AB1 (all other factors being as equal as they can be).
cogsncogs said:
Geez you are fast. Where do you find the time?! Which grid leak resistors? R8, R11, 12?
Q: given that the ckt JoJOD818 posted (Post 241 and 242) uses 1/2 6SL7 more what does it gain from this?
navin,
that would be R11,R12, increasing this to 270k will give somewhat higher gain, as for me, i would go with 100k first.
the 1/2 6sl7 is needed to achieve the 48db gain wayne feels is needed for an input sensitivity of 775mV after feedback.
that would be R11,R12, increasing this to 270k will give somewhat higher gain, as for me, i would go with 100k first.
the 1/2 6sl7 is needed to achieve the 48db gain wayne feels is needed for an input sensitivity of 775mV after feedback.
Tony said:
thanks. I think I will live with 100K too. I dont need gain. my sources are a CD player, CD recorder and Ipod.
I guess I did was not clear about the 6SL7....
Wayne's ckt uses 1/2 6SL7 while the schematic JoJoD818 posted uses 1 6SL7. What does JojoD818's ckt gain by this extra triode (1/2 6SL7)?
- Status
- Not open for further replies.