Santa Claus DIY Tube Amp

screamgigi schrieb:

EF86 is inherently a better tube than the octal triodes like 6SL7 or 6SN7 for that position.

Hmmmmmm.....
I'm not a tube guru by ANY means but this is the 1st time I've heard that the EF86 is better than a 6SN7. I cannot make any intelligent comments on this.
I've mentally filed this info off. when I meet some other tube gurus I'll ask them about this.

screamgigi schrieb:

It’s a very linear driver from end to end. We are solving many problems that arise with a Triode in that place, namely Miller Capacitance. Of-course we loose that if we triode wire the EF86.

you've have confused me here! if the EF86 is being wired in triode then why are you losing the benefits?
(BTW, I thought that you were going to wire the EF86 as a pentode for the input stage driver. So, did you just make a typo?)

screamgigi schrieb:

Another advantage is that the high gain of a pentode can be used to apply lots of global NFB to get better damping. All things equalEF86 are sonically a solid tube.

I hope that you are NOT going to put global NFB around this amp you are building? That'll ruin the sound!
If you can help it, use local negative feedback in the input stage or driver stage.

screamgigi schrieb:

Output Pentode or Tetrode Push Pull: I will suggest 6L6-G tubes in UL connection.

This is the other design aspect that will compromise the sound. 2ndly, 6L6 tubes are considered to be "cheap" & "reliable" but are not considered to be low distortion. KT88 &/or 6550 are much better, IMHO, if you want to stick to pentode or beam tetrode output tubes.
3rdly, a triode connected output tube (such as EL34) will sound much better than a pentode. Lower in distortion & you can also get away w/o using global NFB.

I've got a tube integrated that uses about 5dB global NFB & also connects the KT88 in UL mode. It measures superb on the bench but it sounds mediocre at best. Good for my bedroom system where I've got all the components per my wife's requirements esp. speakers pushed against the wall! As far as I'm concerned, I'm not going to make this mistake again! ;).

screamgigi schrieb:

We will now clone the classic Eico HF22 amplifier circuit

I've not heard an Eico amp but I've read several posts by others who have. They claim that it has a classic tube sound from the "glory" days of tubes. I've interpretted that in a negative sense - too syrupy smooth sound where every type of music is coated w/ this tube golden glow sound. Maybe I'm wrong in my interpretation?

Anyway, these are MY thoughts on your design. Please take them in the right spirit. Also, please do not change your design based on my comments as these come very late in the game. thanks!

Scream-gigi,

It is very very nice to see such discussion happening on our forum. I'm not really a valve expert. But i get a lot of help from my uncle, who started his business making valve-radios back in those golden-days. Based on my limited experience coming from my preamp-project...here are some tips - on keeping the noise floor down...

* Please make sure to use the best possible caps all through in power supply. I use Panasonic (Can't get them in Bangalore...i'm sure you get those in Mumbai). I'm not somehow very happy with the Keltron caps you get these days.

* I know silver wiring has its benefit, but also try a good quality two-core shielded cable. It helps to keep in check the noise floor

* In my experience with DACT, i found that it injects noise when kept anywhere within 2" of the power supply PCB. Try to isolate the power supply from the POT & Source-Selector as much as possible.

* Wiring : Component placement/layout, grounding, earth-loop, etc, are other aspects that can also affect the sound. This can only be optimized painfully through trial and error. I think we should discuss this aspect, once the amp is ready and we are able to hear the results...

Regards
Siva.

bombaywalla,
I think there was one cayin amp with EL34's which could switch to UL mode but didnt sound that hot in UL....what was that model again ?

manek.

Hi all,

Can you guys shed some light / discuss the pros & cons of the EF 86 vs the 6SN7 ?

Thanks

a triode connected output tube (such as EL34) will sound much better than a pentode.

I am not sure I agree.

I am using a Prima Luna Valve integrated amp. It comes in 2 flavours :

The PL-1 with EL34s

and

The PL-2 with KT-88.

For my kind of music ( Rock, pop and a touch of light Jazz ) I found the KT88s streets ahead.

I found the EL34s sound like mushy tubes, too euphonic, and much too much like amps for a earlier era...

bombaywalla schrieb:

Hmmmmmm..... I'm not a tube guru by ANY means but this is the 1st time I've heard that the EF86 is better than a 6SN7. I cannot make any intelligent comments on this. I've mentally filed this info off. when I meet some other tube gurus I'll ask them about this.

Actually the Pentode EF86 and the triode 6SN7 are akin to Apple and Oranges. They operate differently, though both can serve the same purposes. In a PP amp I have seldom used or seen 6SN7/ECC82 in the input stage, whereas they make excellent choice for the phase splitter position.

So what is better is purely academic. EF86 will do the job that 6SN7 cannot do and vice versa. But I believe EF86 was developed specifically for audio, while the 6SN7 was developed with RF use in mind. That really doesn’t mean anything actually. 6SN7 is a fine tube and I try to use as many. But not for the 1st hole in a PP amp though.

bombaywalla schrieb:

you've have confused me here! if the EF86 is being wired in triode then why are you losing the benefits? (BTW, I thought that you were going to wire the EF86 as a pentode for the input stage driver. So, did you just make a typo?)

Yes my sentence appears confusing. With a triode we have Miller Capacitance issues. With a pentode we don’t. I should have put it more clearly that we loose that advantage if we wire our input pentode as a Triode. Everything has a flip. With Pentode we need to deal with the partition noise. This is where the excellent characteristics of the EF86 come to play. We will discuss Miller capacitance in a subsequent post.

bombaywalla schrieb:

I hope that you are NOT going to put global NFB around this amp you are building? That'll ruin the sound! If you can help it, use local negative feedback in the input stage or driver stage.

I think gNFB is the most misunderstood aspect of an amplifier design. We have all vague answers on why NFB is bad but not many solid technical reason. Almost 95% of amplifiers, whether DIY or commercial, have some form of NFB. Even some SET amps. As usual, amp builders have a neat divide of opinion on this subject. There are no fence seaters.

It all depends how you implement it. Too much of it is a bad thing. It will deaden the sound although increasing the linearity across the frequency spectrum. In that case I agree with you that it will measure great on bench but sound lifeless.

However, we have many variables in a valve amp. They change characteristics in course of operation. We nevertheless wish to maintain a stable and equalised gain for both the channels to ensure reasonable sound staging. We want as constant gain as possible and NFB is our best bet to achieve that.

So that big question “what if we don’t use NFB with this amp” ?

Yes most certainly we can do without. In fact we will start that way and introduce NFB only if needed and in right measure. Zero or low NFB may increase noise floor and lead to excessive gain. We have means to tackle that too. Bu we have a solid reason to prefer some NFB. Output resistance. Without NFB our output resistance will measure several ohms and that will severely effect the electrical damping factor. Do we want that? We will judge it when the build is finished and the chosen speakers are tried out.

But it’s good that you have raised this issue. We will now incorporate a variable NFB line where you can adjust the feedback from 0 to 20dB. For this we will add a 4 position rotary switch. We will now have greater control on speaker matching. Thanks for this. We shall call it the Bombaywallah mod

bombaywalla schrieb:

6L6 tubes are considered to be "cheap" & "reliable" but are not considered to be low distortion. KT88 &/or 6550 are much better, IMHO, if you want to stick to pentode or beam tetrode output tubes.

Hummm!
This is the most subjective part of tube audio! There are literally 100’s of pages on the net that discuss why 6L6 is good or bad.

Lets put in a real world analogy here.

A preference for 6L6 is like being a Gay person You like it but you want to keep it under wraps. You are forever scared that someone will discover and shout “oh this guy is using a 6L6” !

This nebulous group of shouters are called “audiophiles”. Most of these vaunted folks may have never heard the 6L6 or the EL34 or for that matter any tube in a real world. But that won’t stop them from taking a position. You will find them in web foras like the AA and Diyaudio. I hasten to add that I do not include you in this scheme! I know you are a genuine tube guy who knows and loves his tube I am sure you too have come across people who really make wild claims on tube sound.

But at times some of those brave “gay” folks “come out of the closet” and post about the goodness of the 6L6. Some even make amps ! Like our own Viren-ji. He came out long ago I admire amp makers who equip their reference models with 6L6’s. It shows they are in a much higher plane and for them mere discussions on relative merits/demerits of tubes is as irrelevant as Bluetooth capability for a valve amp. They just concentrate on making a great product. And 6L6 is one of the best tubes you can use. Lets us applaud these guys.

In a bad design even a 300B will sound lousy. It all depends on how you do the implementation. For this build we are using the early 6L6-G that is reported to be one of the best sounding output device ever made. Some the all times classics such as Mcintosh, Harmon Cardon , QUAD, Fisher use 6L6 or its variations.

However, as you have raised your preference for the EL34, we shall now beef up our Power Transformer so that we have a choice of any 6L6 family or EL34 tube. This is the 2nd change we are doing as per your comments. Thanks.

bombaywalla schrieb:

3rdly, a triode connected output tube (such as EL34) will sound much better than a pentode. Lower in distortion & you can also get away w/o using global NFB.

Yes. But this distinction is slightly blurred when you UL wire the EL34. You get triode like performance and Pentode like power. It’s the best of both worlds. Or as many say, worst of both : )

bombaywalla schrieb:

I've got a tube integrated that uses about 5dB global NFB & also connects the KT88 in UL mode. It measures superb on the bench but it sounds mediocre at best. Good for my bedroom system where I've got all the components per my wife's requirements esp. speakers pushed against the wall! As far as I'm concerned, I'm not going to make this mistake again!.

5db of gNFB is very mild. Did you check all the tubes? What is this amp? An UL wired KT88 should sound great.

bombaywalla schrieb:

I've not heard an Eico amp but I've read several posts by others who have. They claim that it has a classic tube sound from the "glory" days of tubes. I've interpretted that in a negative sense - too syrupy smooth sound where every type of music is coated w/ this tube golden glow sound. Maybe I'm wrong in my interpretation?

Well we are building a tube amp. We would certainly like some of the that tube magic or golden glow!. The key is how we adapt various components to strike the right balance between euphony, dynamics and speed. When we finish building we will tweak and roll tubes and hopefully find a nice, well rounded and non-fatiguing sound. A good Class A tube amp is expected to perform somewhat as you describe.

Cloning the Eicho HF22 is just a way of putting it. You have broadly two schools of thoughts when building a PP Tube amp. The Mullard design and the Williamson design. We choose Mullard cause it’s a fine performer and simple to build. The Williamson is reported to be better, but it’s very hard to get stabilised. It requires a very exact OPT which is difficult and expensive today. For you IT guys its like IBM and Apple. In the past most amp maker would follow any one of the design structure or offer both.

Tube audio has certain magic to them. When I come home after a days work, I like to fire up my system, fix a tall drink and put on some Thelonious Monk. I dim the lights, close my eyes and just slide away. Tubes, scotch and Jazz make an irresistible cocktail. I can literally feel Monk playing just for me. It is actually hard to describe in words.

Does it mean we cant have a blast!. It actually depends on what you expect from a blast. If you have twenty guys in a big room who are baying for a loud Led Zeppelin, then some of them are gonna be disappointed. But if you have 20 guys who like rock n roll for party then this amp could be ok. Last Xmas party I used a tiny 5+5 watt SET and we had a houseful of variously un-sober people. So it all depends.

bombaywalla schrieb:

Anyway, these are MY thoughts on your design. Please take them in the right spirit. Also, please do not change your design based on my comments as these come very late in the game. thanks!

Bombaywalla-ji. Based on your input we are making two major changes. I hope you will approve the same. And you know there is nothing like a fixed design. Even after we finish building. That’s the glory of DIY.

sivat-ji

sivat schrieb:

* Please make sure to use the best possible caps all through in power supply. I use Panasonic (Can't get them in Bangalore...i'm sure you get those in Mumbai). I'm not somehow very happy with the Keltron caps you get these days.

I agree with you. I would ideally like to use oil filled caps. Keltron HV caps are still ok. I don’t know about lower voltages though. At that range I uses Elna Starget. RS India sells them quite cheaply. I also have a bunch of cerafins and Rubycons.

sivat schrieb:

* I know silver wiring has its benefit, but also try a good quality two-core shielded cable. It helps to keep in check the noise floor

Actually I have decided against the silver wire for this layout. It will be hard to keep away noise and hum induction. I will use a high quality OFC shielded cable for the signal run from Line in to Vol put. That is the only signal carrying part. Rest we do point to point hard wiring but using a tag board.

sivat schrieb:

* In my experience with DACT, i found that it injects noise when kept anywhere within 2" of the power supply PCB. Try to isolate the power supply from the POT & Source-Selector as much as possible.

In out layout low signals are well away from AC mains area. PSU is well isolated.

sivat schrieb:

* Wiring : Component placement/layout, grounding, earth-loop, etc, are other aspects that can also affect the sound. This can only be optimised painfully through trial and error. I think we should discuss this aspect, once the amp is ready and we are able to hear the results...

I follow star grounding. The mains ground is connected to the chassis. Signal ground is connected to this chassis via a resistor and capacitor network. I can hardly measure any hum.

I notice that you are offering Jensen caps. I have a boxful of Auricaps but certainly like to try the Jensen. Can you tell me the length and dia of your .22uf caps? I need to make space in the layout for those caps.

Amp_Nut schrieb:

a triode connected output tube (such as EL34) will sound much better than a pentode. I am not sure I agree. I am using a Prima Luna Valve integrated amp. It comes in 2 flavours : The PL-1 with EL34s and The PL-2 with KT-88. For my kind of music ( Rock, pop and a touch of light Jazz ) I found the KT88s streets ahead. I found the EL34s sound like mushy tubes, too euphonic, and much too much like amps for a earlier era... :(

well, Amp_Nut it really depends on the implementation. I've heard the Music Reference/Roger Modjeski RM-9 Mk1 & Mk2 both using EL34 tubes. I loved the sound!
it was FAR from euphonic, far from syrupy smooth & far from having a golden glow of the tube glory days. In fact, both version of this amp sounded like a very good solid-state amp - They had very, very good bass slam. I listened to rock/pop & a lot of Jazz on them. simply Superb!

I have a JoLida 502A which is a 60W/ch KT88 connected in ultra nonlinear mode err......... ultra linear mode & it does fairly well. It's not as open as some of the other 6550-based amps I've heard. This JoLida can take 6550 (which is what I was using before I switched) & KT88. SOme people even use Ei KT90 with it but Ei is out-of-business now
This amp has some slam tending more towards a s.s./sand amp but the soundstaging is flat/2D. The EH KT88 has a boom-boom bass. Good for rock but not as good for Jazz where the bass has some very clear definition.
I know that one can do much better w/ a KT88 but not w/ a JoLida-based design - it's quite plebian. Some of the KT88 amps from VTL & VAC & even Art Audio are superior but their price is also superior!

I've just bought a pair of AES Sixpacs. They are breaking in as I write this post. I'll listen to them over the weekend when the EL34 tubes have had about 100 hrs on them.
1st impressions after installing BRAND NEW tubes + removing the amps from their box: Superb midrange clarity. Fantastic hi-freq performance. Fantastic soundstage depth - it was much easier to tell which instrument was in the back & which in front vs. my s.s. amp. Superb & clear bass definition - I was listening to a test CD where Christian McBride was playing the double-bass in "Splanky" & I could hear each bass string being plucked. The amp allowed me to follow every bass line even when he played fast. No blurring of bass sound. What was missing was the bass slam/thump. It'll probably get better as the tubes break in.
Also I'll be using the Sixpacs in 0dB negative feedback mode (-10dB is the other switch option) so I expect to lose some bass control (vs s.s. amp).

So, I have both - KT88 in U-(non)L mode & EL34 in zero feedback mode. Once the Sixpacs breaks in I can compare the 2 better.

Screamgigi,
Thanks very much for the explanations on the EF86 vs 6SN7 & the (gay) 6L6-G tubes!

screamgigi schrieb:

Almost 95% of amplifiers, whether DIY or commercial, have some form of NFB. Even some SET amps. As usual, amp builders have a neat divide of opinion on this subject. There are no fence seaters.

I agree that ALL (each & every WITHOUT exception) amp has negative feedback. I'll even state that there is not a single electronic device that can operate without feedback.
However, the feeback I was referring to explicitly was GLOBAL negative feedback.
All-the-same, I understand you take on this matter. Please go ahead with your scheme.

screamgigi schrieb:

It all depends how you implement it. Too much of it is a bad thing. It will deaden the sound although increasing the linearity across the frequency spectrum. In that case I agree with you that it will measure great on bench but sound lifeless.

you got it exactly!

screamgigi schrieb:

But it’s good that you have raised this issue. We will now incorporate a variable NFB line where you can adjust the feedback from 0 to 20dB. For this we will add a 4 position rotary switch. We will now have greater control on speaker matching. Thanks for this. We shall call it the Bombaywallah mod :

totally flattered! Thank you!

screamgigi schrieb:

I hasten to add that I do not include you in this scheme! I know you are a genuine tube guy who knows and loves his tube

thanks! whew! lagta hai ke baal-baal bache!

screamgigi schrieb:

However, as you have raised your preference for the EL34, we shall now beef up our Power Transformer so that we have a choice of any 6L6 family or EL34 tube. This is the 2nd change we are doing as per your comments. Thanks.

Thanks once again!!
I think that Manek will also be happy as I think that he also quite likes EL34 output tubes, n'cest pa?

screamgigi schrieb:

5db of gNFB is very mild. Did you check all the tubes? What is this amp? An UL wired KT88 should sound great.

Oh yes I checked all the tubes. They are quite stable & hold their bias well. They are brand new matched quad of KT88s. This is a JoLida 502A.
AS I was writing in response to Amp_Nut, the JoLida sounds great in & off itself. When you start making comparison to other 6550/KT88 amps or w/ EL34 amps, you being to hear its short-comings. They become very apparent quickly!
Anyway, like I stated before, my lesson learnt!

screamgigi schrieb:

Tube audio has certain magic to them. When I come home after a days work, I like to fire up my system, fix a tall drink and put on some Thelonious Monk. I dim the lights, close my eyes and just slide away. Tubes, scotch and Jazz make an irresistible cocktail. I can literally feel Monk playing just for me. It is actually hard to describe in words.

I know exactly what you mean!
I also like Thelonious Monk quite a bit. He is one artist that commands that you listen to his music when you play it. It's complex yet simple.
Indeed tubes + scotch + Jazz is a heady cocktail! It's a "Straight, No Chaser" to the head!

screamgigi schrieb:

Bombaywalla-ji. Based on your input we are making two major changes. I hope you will approve the same. And you know there is nothing like a fixed design. Even after we finish building. That’s the glory of DIY. :)

Yes, I do approve of the same. You have been very gracious! Thanks much!

Manek schrieb:

bombaywalla, I think there was one cayin amp with EL34's which could switch to UL mode but didnt sound that hot in UL....what was that model again ? manek.

Manek,
I think that you are referring to the Cayin A-50T stereo P-P amp that uses EL34 tubes & puts out 16W/ch in triode & 22W/ch in UL mode.

Amp_Nut schrieb:

Hi all, Can you guys shed some light / discuss the pros & cons of the EF 86 vs the 6SN7 ? Thanks

Even i'm curious to know about this difference.

Viren, Manek,

You have mentioned that you prefer 6SN7 (infact i've heard this from many folks with lots of experience with Valve..including some very famous names) ...can u please please elaborate. Is it bcos of the euphony of the 6SN7or is it bcos ECC83 (at most times) sounds like solid-stage ??

Regards
Siva.

sivat schrieb:

Even i'm curious to know about this difference. Viren, Manek, You have mentioned that you prefer 6SN7 (infact i've heard this from many folks with lots of experience with Valve..including some very famous names) ...can u please please elaborate. Is it bcos of the euphony of the 6SN7or is it bcos ECC83 (at most times) sounds like solid-stage ?? Regards Siva.

Siva, in a nutshell you have hit upon the correct reason (IMHO) for preferring 6SN7 to 12AX7 - the octal tube 6SN7 has lower distortion than the 12AX7. It seems that the distortion components of the 6SN7 die out faster than the 12AX7. The 12AX7, owing to its higher distortion, tends to sound like s.s. with its odd order harmonics of which several higher orders are present.
I also note that some 9-pin versions like 6CG7 & 6FQ7 are pretty close to a 6SN7 sonically & many people like them over a 12AX7.
FWIW.

bombaywallah....yeah do have a soft corner for EL34's.

Screamgigi, where did you get your six packs ???? from India ?

siva, I feel the 6sn7 used in my preamp has good dynamics, sounds nice and open, yes has that midrange which I adore.

It also seems to be very evenly balanced across the freq spectrum.

manek.

Hi all,

At times its amusing to see these heated discussions as to types and brands of valves to be used. Of course there are differences between valves, and they all leave their signature on the sound. However, use any of the valves made for audio, implement a design properly around them, and you will get a good sounding amplifier.

It's like selecting a wine to go with your meal. There are so many choices and flavours to select from. You've decided on the main course for your meal, select a wine to complement it, and your pleasure is enhanced!

And, there's an element of practicality. I use what's more easily available to me. The designs then reflect that.

Valves now are a limited resource. Let's use them wisely.

Viren.

6SN7 vs 12AX7 (ECC83)

6SN7 and 12AX7 (ECC83) are two entirely different beasts. Lets look at the data: (from Slyvania)

6SN7

Medium Gain Twin triode intended for use as RC amplifier or combined vertical oscillator or vertical deflector amplifier in TV sets.

Typical Operating condition:

Amplification factor: 20
Transconductance: 3,000 mho
Plate resistance: 6,600 ohms



12AX7 (ECC83)

Miniature High gain twin triode intended for use as audio amplifier or phase splitter.

Typical Operating condition:

Amplification factor: 100
Transconductance: 1,250 mho
Plate resistance: 80,000 ohms


6SN7 is closer to 12AU6 (ECC82). 6SL7 with a gain of 70 is closer to 12AX7 (ECC83).

12AX7 can do things which a 6SN7 can never do such as the first hole duties in a Phono or Mic preamp. Similarly, putting a high gain tube like 12AX7 where a low gain tube like 6SN7 is required, such as a Linestage, is asking for trouble.

Both are designed to perform specific task. 12AX7 is an excellent tube. NOS 12AX7's are often more expensive than NOS 6SN7.


DISTORTION

Both will have same level of distortion when operated in there respective loadlines. It al depends on what load line you pick up. 12AX7 is used in Phono stages because of their low distortion at higher gm

I will be more conformable comparing a 6SN7 with 12AU7 (ECC82) or some other similar tubes. And yes 6SN7 is obviously a more linear and pleasing device when compared to others in its family.

I was using 6SN7’s a lot. I still do. But now I have discovered tubes such as 76 and some ancient DHT’s from the 1930’s. My old version of the 845 SE amp had 6SN7 as the input tube. Now I am rebuilding with DHT’s all positions.

Almost all tubes are great. The key is how you utilise them in a particular design.

Manek schrieb:

Screamgigi, where did you get your six packs ???? from India ?

I just pick up a crate of KF from my local daaruwala. Twice a week

viren schrieb:

Hi all, At times its amusing to see these heated discussions as to types and brands of valves to be used. Of course there are differences between valves, and they all leave their signature on the sound. However, use any of the valves made for audio, implement a design properly around them, and you will get a good sounding amplifier. It's like selecting a wine to go with your meal. There are so many choices and flavours to select from. You've decided on the main course for your meal, select a wine to complement it, and your pleasure is enhanced! And, there's an element of practicality. I use what's more easily available to me. The designs then reflect that. Valves now are a limited resource. Let's use them wisely. Viren.

Cannot agree more with you. It was essentially because of this reason, that i raised my earlier question on why 6SN7 is being preffered.

I have tried both the valve(6SN7 Vs 12AX7/12AU7)...there is a difference in flavor to the sound. But otherwise, i found that all these tubes are equally competent, when used the right way.

However, i was alway curious as to why so many folks prefer 6SN7...hence my question. Different people say different reasons.

Manek schrieb:

Screamgigi, where did you get your six packs ???? from India ? manek.

Manek, I think that you might have mis-read: it was me who just acquired the AES Sixpacs (& not Screamgigi) And, it was NOT from my local daaruwalla! ).

hey....bombaywalla...yeah ...i did misread...sorry screamgigi

what tubes are you playing em with now ? Preamp ? source ?

Manek schrieb:

hey....bombaywalla...yeah ...i did misread...

no problems, moderator sahib.

Manek schrieb:

what tubes are you playing em with now ? Preamp ? source ?

using Electro-harmonix EL34 right now as that was what was provided by the seller. it's a used unit as you might have guessed. preamp is a CAT & source is a Wadia + my TT.

sivat schrieb:

Amp_Nut schrieb: Hi all, Can you guys shed some light / discuss the pros & cons of the EF 86 vs the 6SN7 ? Thanks Even i'm curious to know about this difference. Viren, Manek, You have mentioned that you prefer 6SN7 (infact i've heard this from many folks with lots of experience with Valve..including some very famous names) ...can u please please elaborate. Is it bcos of the euphony of the 6SN7or is it bcos ECC83 (at most times) sounds like solid-stage ?? Regards Siva.

Siva,
I came across this article written by the venerable Eric Barbour. Most of the article is NOT for those of us who are convinced about tubes but I think that it is a nice article for those who would like a quick history of tube usage & the various types of tubes.

Anyway, the important part of this article that pertains to your question is about 1/4 of the page down (scroll down). There are some HTML links showing a summary of 2nd harmonic distortion levels for various tubes & transistors. From the graphs it is easy to see that the distortion of the 6SN7 drops very rapidly & besides the 2nd harmonic there doesn't seem to be other distortion.

http://www.milbert.com/articles/CoolSoundOfTubes/

Hi Bombaywalla,

EXCELLENT article. Thankls for the link.

the distortrion figures were Particularly interesting.

The 2 points from the post that I would like to highlight:

1. No other active device possesses both the low distortion products and the low noise floor of the medium-mu triode--albeit at the expense of voltage gain.

This makes the 6SN7 ideal for a pre amp, with no global negative feedback.

Hmmmm it got me thinking. Now to find a source of the 6SN6 in India ...


2. The distortion products of transformers are much lower than those of active devices, yet quite different in character. Note that the odd-order harmonic products tend to be higher in level than the even-order products--exactly the reverse of the tubes and transistors.

VERY interesting. Seems to point that capacitor coupling is better than Transformer coupling ( where ever possible ).

Transformer coupling comes at a price... quite a HIGH price.

Funny that passive Transformers for Volume control applications are so well received. Maybe attention to the core material does the trick ?

I wondering how long till poor Eric Barbour is dragged into it !

6SN7 is quite a venerable tube. No wonder it’s the preferred device for linestage, DAC outputs, Phase Inverter and Active XO’s.

Please note that in our build we are also using the 6SN7 for phase inverter. We were planning to use EF86 as the voltage gain stage.

Since folks here are so vehement on 8 holes , we shall now bring back that 6SL7 in place of EF86. It actually makes the job easier and cheaper. Just one 6SL7 can be shared between the R & L channels. I will save the expensive EF86’s for some other project.

I am sorry but this entire week I was quite tied up in often-mundane issues. I had hoped to put on the schematic and explain the building block. Lets hope we get back to doing that soon.


Amp_Nut-ji

6SN6 ?? I am sure a tube exists but never seen one in India.

Keep it simple and short if you are planning a 6SN7 linestage. A simple cathode follower will do rather nicely. Keep the Rk unbypassed. Or you can explore the Grounded Grid design using a 6SN7.

6SN6 ?? I am sure a tube exists but never seen one in India.

YUP, a typo.

screamgigi-JI

I agree that for the voltage gain stage a tube with higher voltage gain is ideal. That is why I even quoted that point from Eric Barbour.

Anyways, plaese do not consider any of my comments as inputs for this project, as I have said earlier.

I am merely discussing ( and getting to learn ) tubes.

Cheers and best of luck !

P.S: IOf there are any typos above, please excuse !

screamgigi schrieb:

I wondering how long till poor Eric Barbour is dragged into it !

Screamgigi, I REALLY hope that I have not dragged Eric Barbour into this debate merely by sharing an article written by him!
If I did, unintentionally, I hope that I can still so something to keep him out!

I also noticed in 1 of ur earlier posts that you had mentioned Thorsten. I really like Thorsten & I respect his opinions quite a lot. He seems to make & mod equipment for the love of music & real-to-life audio playback. He also seems to think out of the box. He's one of the better fellows to deal w/ on AA.

screamgigi schrieb:

Since folks here are so vehement on 8 holes , we shall now bring back that 6SL7 in place of EF86. It actually makes the job easier and cheaper. Just one 6SL7 can be shared between the R & L channels. I will save the expensive EF86’s for some other project.

please do not use "vehement on 8 holes"!
Many of us like triodes for their low distortion but that does not mean that you cannot use an EF86 in the input stage. As I stated in my original/1st post on this topic, I know very little about the EF86. In fact, I encourage you to use the EF86 instead & show us yet another sonically capable input driver. Speaking for myself, it'll expand my mind/choices for input drivers to beyond 6SN7 & 6SL7s. The article link I cited was to merely answer Sivat's query re. 12AX7 vs. 6SN7 & it was not to change your choice of input driver.

Barbour-ji has an excellent series of articles. I had a PDF of that. I think it can still be found in the net. Makes a great reading. But do make a subjective assessment. He was involved with Svetlana USA business and it is claimed, maybe unfairly, that he brings a certain bias. But of-course we are not dealing with Svetlana tubes here.

We will go with the 6SL7. My schematic is almost complete in PSPICE. I will shortly post some grabs. We have more projects in hand to use the EF86. 300B SE triode comes to mind quickly. Once we are finished with this build we shall go to that.

Thorsten is a quirky nerd. Some years ago I had fixed up a meet with him in UK. He did not show up tough we had agreed by email. I use his idea of using a pentode to drive a triode. Combines the advantage of both.

screamgigi schrieb:

He was involved with Svetlana USA business and it is claimed, maybe unfairly, that he brings a certain bias.

no pun intended here, eh??

screamgigi schrieb:

We will go with the 6SL7. My schematic is almost complete in PSPICE. I will shortly post some grabs.

cool! look forward to that, Screamgigi.

screamgigi schrieb:

We have more projects in hand to use the EF86. 300B SE triode comes to mind quickly. Once we are finished with this build we shall go to that.

ok, chief!

My schematic is almost complete in PSPICE

Just learning.... How well do real world Audio Transformers approximate in PSPICE ?

Good enough to predict amp stability ?

Amp_Nut schrieb:

.... How well do real world Audio Transformers approximate in PSPICE ?

Terribly actually.

This is a grey area. To get real accurate result, you need to measure the OPT at hand and make your own SPICE model. Or if you trust the OPT manufacturers data, then use that. Often that published data and actuall spec varies as much by 25%.

With tube circuit SPICE is fairly correct up to the driver /splitter stage. There after you are sort of at your own. I will agree that your assumption is correct.

Just to illustrate, I am posting the screen grab from of schematic from SPICE. Currently I am just including the Driver and Spillter stage.





I was planning to post these some time later after explaining the PP and SE building block.

screamgigi-ji

What are the key differences between the 6SN7 and 6SL7 ?

Thanks

Amp_Nut schrieb:

screamgigi-ji What are the key differences between the 6SN7 and 6SL7 ? Thanks

Amp_Nut-ji

6SL7 is a high gain triode. 6SN7 is a midium gain triode. Both can be used as AF amplifiers.

A comparision can be found on the post #66 above stacking the 6SN7 against the 12AX7 (ECC83).

ECC83 is almost similar to 6SL7 but in a miniature 9 pin bulb. One vital difference is the gain. ECC83 = 100 while 6SL7 = 70.

Screamgigi, you planning to hit Delhi sometime with this Amp?
A shootout with Virens'Amp would an Audiophiles' mini Nirvana!
We could try it with Virens' Harmony Ones' & also try to set up a listen on Ashish's Amayas!
We delhiwallas can guarantee an enthusiastic audience...
Regards,
Kamal

Sure. Why not
We will work something out. Its not too difficult.
But the comparision maybe unfair

We will now progress to our schematic. Since we are building Push Pull amp, we will just quickly put an overview of how a typical PP topology looks like. Its assumed that one interested in more detailed understanding is accessing the online primer link posted earlier.

PUSH PULL AMP BUILDING BLOCK



In a nutshell the incoming source is voltage gained in the INPUT STAGE. It is then presented to the PHASE SPLITTER which will split the signal into opposite streams at 180deg to each other. Each stream is amplified by the separate OUTPUT POWER DEVICE. This high impendence is then ratio matched to the SPEAKER.

This is a very rudimentary explanation of a Tube Push Pull amp. There are several variations but we are sticking to this basic building block.


SE AMP

For completeness sake we will also put here the Singe End topology as below. Simpler and easy to build. Most believe SE sounds better but you are only limited to a few watts. Any bigger output and you are looking at huge tubes and assorted fittings.



So what are we using to fit into those blocks? Here they are:

Input Stage: 6SL7
Phase Splitter: 6SN7
Output: 6L6-G
Imp matching: Output Transformer Primary 5K anode to anode with Ultra linear taps at 43% of the primary winding. Secondary is 8 ohms.

This page we now have the 1st version of our schematic. We shall later put in Solid State CCS to precisely control the tubes. We will also design and post the Power Supply schemtic.



As you can see audio with tubes is a relatively simple affair. Critical are components. Since you only have a few of them, it is worthwhile to invest in high quality items.

Please don’t skewer me over slow fire if there are minor or even major mistakes on my posting. Its been a real long time since I last put together a Push Pull amp

I was hoping to have a clear stretch this November and do things much more leisurely and carefully. But it seems destined otherwise

Comments pls

This is how it is transposed from the PP building block diagram which is posted above

Lovely and lucid explanations, screamgigi-ji.

No overall feedback, and a clasic, simple circuit.

Lets hope the local iron does it justice.

Cheers !

screamgigi schrieb:

So what are we using to fit into those blocks? Here they are: Input Stage: 6SL7 Phase Splitter: 6SN7 Output: 6L6-G

screamgigi, there seems to be a lot of gain between the 1st 2 stages - >40dB. I know that my JoLida has 30dB+ gain & uses 5dB global negative feedback. In your sch you still need to show the negative feedback ckt, correct?
I suppose that you need the 2 stages of gain? I.E. is there any way to direct connect the phase splitter 6SN7s to the input & make them do double-duty: gain + phase splitting?
Since it's a medium mu triode, you will not have enough drive for the 6L6?

screamgigi schrieb:

Imp matching: Output Transformer Primary 5K anode to anode with Ultra linear taps at 43% of the primary winding. Secondary is 8 ohms.

looks like the 6L6 tube is contructed optimally for UL connection operation (with grid#2 43% the distance from the cathode).

Amp_Nut schrieb:

Lovely and lucid explanations, screamgigi-ji. No overall feedback, and a clasic, simple circuit. Lets hope the local iron does it justice. Cheers ! :prost

as I understand it 1 of the (main) reasons P-P sound inferior to SE amps is the quality of the OPT. since the amp is P-P, it has to go thru the output common-mode voltage (which could/is 0V). usually OPTs have a "dead zone" where a certain amount of output signal energy has to be present to overcome this dead zone. hence there is a loss of low level details & smearing of the images.
In a SE amp, as only 1 tube does the push & the pull, the OPT is biased away from 0V/dead-zone. Thus the output signal does not need to overcome the dead-zone & the results of this a clearly audible - better spatial separation & low level details.

So, I agree with Amp-Nut: get the best frigging OPT you can buy. You will not regret it!
Also, as you know, bass perf is directly proportional to size of the OPT - the bigger the better (of course, without compromising quality). The OPT could set you back quite a bit. The Japs make some the best OPTs, IMHO.
I also believe that some of the amorphous-core Lundahl OPTs (Sweden) are excellent sounding.

Screamgigi ji,

Just a question on the focus of this project. I hope it's to help others understand and build their own valve amplifiers, copies or variants of this one. To that end, it has to be relatively easy to build, with components that will be available to all. It doesn't have to be the best amplifier ever, just a good one!

Some compromises are inevitable, but try not to make the project too expensive or exotic. Let others build the basic amp, and then experiment by themselves upgrading to their hearts desire!

You are doing an excellent job describing the design!

Viren.

viren schrieb:

Screamgigi ji, Just a question on the focus of this project. I hope it's to help others understand and build their own valve amplifiers, copies or variants of this one. To that end, it has to be relatively easy to build, with components that will be available to all. It doesn't have to be the best amplifier ever, just a good one! Some compromises are inevitable, but try not to make the project too expensive or exotic. Let others build the basic amp, and then experiment by themselves upgrading to their hearts desire! You are doing an excellent job describing the design! Viren.

Understand what you are trying to communicate.
However, look at it from the flip side: after going thru all this trouble of designing, compromising & explaining to all the folks on this forum, Screamgigi should eventually land up with a mediocre amp for himself?
what's he going to do with this "just a good one!" knowing full well that its sound can be much better?
He can't sell it (atleast I didn't think that he is in the biz of making & selling amps) & so he'll end up listening to a mediocre amp in his system.
Or, after the lessons are over, should he go back & start upgrading it? Double work, I think.
IMHO, if anyone is to build audio gear, it is worth building correctly & worth building the best possible. Keep the circuit as simple as possible but no simpler.
It might need extra explanations but it'll be worth the time (IF Screamgigi's intentions are to teach tube amp building to the forum members).
Audio is full of mediocre, high-priced gear! we do not need another piece of mediocre gear!
Kuch banaa, kamal banaa!

Amp_Nut schrieb:

Lovely and lucid explanations, screamgigi-ji. No overall feedback, and a clasic, simple circuit. Lets hope the local iron does it justice. Cheers ! :prost

Short & sweet explanation.

Screamgigi, we await similar explanation on the power supply. It also makes a good topic for an argument ...solid state Vs. tube power supply.

Regards
Siva.

Amp_Nut-ji

Yes this is the classic simple Mullard topology. We are just adapting it to our needs which is not a big deal if you can interpret the datasheets.

But I can’t see how we are going to get away without NFB and continuing with the UL connection. We are implementing switchable loop with options of 0, 6dB, 15db and 20dB NFB. That should provide a greater leverage while will match various speakers

Bombaywalla-ji

Yes that’s a little more gain than what we need. But once the NFB loop is in place it should come down.

In this amp we need those two stages. However, if the output device is a pentode like EL84, which requires just a whiff to drive, it is possible technically to use the arrangement you describe. I don’t know what it will do the distortion levels though. As usual this is an interesting idea from you I have seen this elsewhere too and I shall post a schematic later.

The CMV issue that you mention is correct. But only if our output pair is biased too far away from Class A operation. Any audio output transformer is designed to operate with a minimum quiescent current. If the output tubes are not matched then the resulting uneven current draw disturbs the core characteristics leading to smearing and bad low frequency response. A good transformer (read a big core size) will allow a tube mismatch of upto 20% and yet maintain the required performance level.

Transformer winding is an art. There are so many variable that only an experienced winder will know how to go about it. Just using the best material is not enough. I usually buy OPT from Japan or Korea. But I do know some old-timers in Kolkata who are rich in experience and skills. But those poor folks just can’t get high quality laminates and wires etc. If a sufficiently big batch quantity is assured then they can arrange top quality material. I trust them. One of them re-wound a QUAD II OPT for me and that can get technically as advanced as it can.

6L6 (and its variants) sound quite nice on UL connection. But the optimum tube for UL operation is the EL34 and later pentode. Mullard described this is as “distributed load” when the EL34 was launched in the 50’s. But I believe UL connection was first published by Hafler (Dynaco) using the 6L6 tube. This new fangled technology was immediately rubbished by British gurus likes Peter Walker (QUAD) and Williamson who maintained that triode connection was the only route to audio nirvana. However, later in a role reversal Yankee took to triode connection while the Pommies never had enough of UL connection


Viren-ji and bombaywalla-ji

I appreciate the valued opinion of both you gentlemen. We are actually building a very simple amp. We are not trying any exotic tricks. We are not using very high voltages. We are not using tube rectifier.

On the amp you can see that overtly complex things like an extension shaft for Volume pot is eliminated. We have revised our original layout to reflect this.

It’s a fun little project but I can immodestly predict that the performance will not be lacking. I really don’t see any reason why it should. We are using a simple circuit. We are using reasonably good capacitors and resistors. We are using a high quality OPT. And we are building it will a lot of care. I am quite confident that we shall have a good amp in our hand.

And yes bombaywalla-ji I too agree that simple it should be but only as much. I know that what miracles a Solid State CCS can perform. So we shall implement that simple trick. If our original tube gurus has access to SS devices then they too would have done that.

It only requires a couple of transistor and resistors. If some one can build a valve amp, he can surely put in a CCS. Its really a very simple thing to replace that Resistor in the LPT tails.

Sivat-ji
My preferred option is valve rectification. But I am not a sucker for that. This is a Class A amplifier and out nominal current draw will be around 260mA (Channel A+B) and slightly more at peaks. A typical valve rectifier like GZ34 is happy upto 250mA, though NOS examples can be pushed much higher. Anything more than 250mA and you need to parallel two tubes. Besides, I also feel that any amp that needs to delivery more than 20 WPC should have two rectifier tubes

So we will remain on the simpler side and use silicone diodes. At least technically SS is a better option. Your PSU impendence is less leading to better regulation atleast in Class AB situation. You have lower part count and your Power TX saves an additional winding. And you of-course save on lot of heating!

However, had it been a Monoblock we would definitely have gone valve and saved a lot of heart burn

Here is a more complete schematic. PSU and CCS later.





This time with the dreaded NFB

We are making a variable NFB loop. It is not shown on here, because I can't find a 4 pole rotary switch in S-Spice from which this schematic is screen-grabbed.

I dont have a dedicated application for drawing schematics

We now have the first design of our base plate ready.




For the mechanical build we shall follow a the “Sandwich” construction method. We shall mount the valve bases in a “base plate” which is then bolted on to the chassis from inside. The “top plate” which is our chassis will just have openings for valves. This was we avoid the ungainly sight of valves bases and assorted nut/bolts, which are essential to fix our Tag strip and other items.

This is a little complicated method. It would have been easier to mount the valve bases directly on the chassis. But it may not have looked that elegant. We shall follow that industrial build another time

Later this week we should be able start the actual metal work. I will be using CNC wire cutting m/c since I have access to these kind of facilities. But there is no reason why a general metal working shop should not be able to build a chassis. Best bets are electrical panel makers.

Just to visualize the “sandwich”. That pink shaded plate will be inside the chassis and form the bottom plate of the sandwich.




All fittings will be mounted on this base plate so that the chassis top is clean.

This is the general layout of the amp. As usual we shall be doing changes whenever needed.

screamgigi schrieb:

Just to visualize the “sandwich”. That pink shaded plate will be inside the chassis and form the bottom plate of the sandwich. All fittings will be mounted on this base plate so that the chassis top is clean. This is the general layout of the amp. As usual we shall be doing changes whenever needed.

Screamgigi, the choke is a filter for the filament power supply, correct?

Breathless,sreamgigi, at your expertise& knowledge.
More power to you!
Kamal

bombaywalla-ji

No that is the main High Voltage power supply choke. We dont need any chokes to clean the heater supply in this kind of amp. Choke is useful for heater supply of Directly Heated Triodes (DHT) power amps. Some phonestages also benefit from choke in the heater supply

We shall be shortly moving to the PSU.

Kamal-ji

Actually its no big deal. Till now we have not attempted any thing that has not been around for last 50 years. YOu can see that if you are following the thread.

But I envy you! Viren-ji is a certified tube ustaad and you seem to get to meet him quite often for chaai paani biskut