Re: Johan's Square Wave Theory
A diy Radford inspired TL (coincidentily using a kit from the owners of the Radford tradename,
http://wduk.worldomain.net/forum/showthread.php?t=2800
and elctronics are shipping or very near...
http://www.radfordaudio.co.uk/
dave
bulgin said:Hi FD
But of course
bulgin
A diy Radford inspired TL (coincidentily using a kit from the owners of the Radford tradename,
http://wduk.worldomain.net/forum/showthread.php?t=2800
and elctronics are shipping or very near...
http://www.radfordaudio.co.uk/
dave
Why not build the original, upon which all these "aftermakes" are based? The book on building Valve Amplifiers from Mullard is still available, and you could alter them as per Byrith if you so wish.
The best thing about building your own amplifiers is that you can specifiy decent ratings for the mains & audio output transformers instead of the only-just marginal products that the Manufacturers of the time seemed to prefer.
Some good reading from the two articles here:
http://www.lundahl.se/claus_b.html
-F_D
The best thing about building your own amplifiers is that you can specifiy decent ratings for the mains & audio output transformers instead of the only-just marginal products that the Manufacturers of the time seemed to prefer.
Some good reading from the two articles here:
http://www.lundahl.se/claus_b.html
-F_D
Bulgin then appeared to be interested in my heresy of finding something the matter with the Leak Stereo 20 .... but starts wishing for Radford only 3 posts down. Do I rest my case?
But seriously, to my own amazement I found that the Leak Stereo 20 was lacking in design, to the extent that I had to change certain conponent values so as to get optimal performance. I say this with care, and more so as I do not find any other comments on this on several other sites where refurbishing was discussed. But the problem was quite evident on checking with signal generator and scope, on all of 6 amps that came my way for refurbishing (3 stereo units). Also the relevant components were obviously factory fitted and not the result of tweeking. Perhaps others could comment?
The first matter that struck me was a rather starved operating point for the input triode: About 160V over the tube, only 60V over the load resistor R5. (All of the circuit diagrams I could find used the same sketch and component designations.) Also, checking voltage and current values given in the schematic, the anode current (through R5) is 0.6mA, but through R3 it is 0.56 and 0.65mA (with and without pre-amp connected). This was not serious but did set one wondering.
I ended up with R3 = 1K5 and about 120V on the anode, giving some increase in 1st triode gain. (I was not looking for that, but it did seem to indicate that the triode was now on a rather better operating point. Also, the anode signal amplitude is less than 1V for full output so no problem there - but why work in a region of comparatively high 2nd harmonic distortion?)
With R2 = 22K as given, h.f. open loop bandwidth was less than audio; cutting R2 to 5K6 and with R3 = 1K5, moved it comfortably outside 20 KHz.
Another problem to me with the original was some quite over compensation in the NFB line. The square wave rise was severely jagged and a little oscillatory. As a result of the above increase in gain I had to increase R18 from 12K - 18K for the same NFB factor, and decrease C9 down to 68pF before the square wave was almost perfect. (The equivalent capacitor for 12K would have been 100pF, not 200pF.)
Those are the points I am wondering about. Not serious audio-wise otherwise there would not have been so many satisfied owners; but the question does beg: Why less than optimal component values? I am also not commenting on an ECC81 being significantly better in the 1st stage; that would be a design change, although I did change to that in one instance.
Bulgin caught me rather quickly with this thread; I was going to post square wave pictures of what I am talking about but will have to borrow the amplifier (back) to get those done. (As said, not my own.)
So until then, am I missing something? Who else has done this?
(Or am I a compulsive tweeker? No - other amps that I have refurbished as was were Quads, Dynacos and Leak TL12s.)
Regards.
But seriously, to my own amazement I found that the Leak Stereo 20 was lacking in design, to the extent that I had to change certain conponent values so as to get optimal performance. I say this with care, and more so as I do not find any other comments on this on several other sites where refurbishing was discussed. But the problem was quite evident on checking with signal generator and scope, on all of 6 amps that came my way for refurbishing (3 stereo units). Also the relevant components were obviously factory fitted and not the result of tweeking. Perhaps others could comment?
The first matter that struck me was a rather starved operating point for the input triode: About 160V over the tube, only 60V over the load resistor R5. (All of the circuit diagrams I could find used the same sketch and component designations.) Also, checking voltage and current values given in the schematic, the anode current (through R5) is 0.6mA, but through R3 it is 0.56 and 0.65mA (with and without pre-amp connected). This was not serious but did set one wondering.
I ended up with R3 = 1K5 and about 120V on the anode, giving some increase in 1st triode gain. (I was not looking for that, but it did seem to indicate that the triode was now on a rather better operating point. Also, the anode signal amplitude is less than 1V for full output so no problem there - but why work in a region of comparatively high 2nd harmonic distortion?)
With R2 = 22K as given, h.f. open loop bandwidth was less than audio; cutting R2 to 5K6 and with R3 = 1K5, moved it comfortably outside 20 KHz.
Another problem to me with the original was some quite over compensation in the NFB line. The square wave rise was severely jagged and a little oscillatory. As a result of the above increase in gain I had to increase R18 from 12K - 18K for the same NFB factor, and decrease C9 down to 68pF before the square wave was almost perfect. (The equivalent capacitor for 12K would have been 100pF, not 200pF.)
Those are the points I am wondering about. Not serious audio-wise otherwise there would not have been so many satisfied owners; but the question does beg: Why less than optimal component values? I am also not commenting on an ECC81 being significantly better in the 1st stage; that would be a design change, although I did change to that in one instance.
Bulgin caught me rather quickly with this thread; I was going to post square wave pictures of what I am talking about but will have to borrow the amplifier (back) to get those done. (As said, not my own.)
So until then, am I missing something? Who else has done this?
(Or am I a compulsive tweeker? No - other amps that I have refurbished as was were Quads, Dynacos and Leak TL12s.)
Regards.
I get the feeling that I must have wandered into the wrong thread - it seems to have some strange history of which I am unaware. I wouldn't bet on that, Johan. When those things came out, hardly anyone understood much about audio, and not too many Leak customers did any comparisons with other gear to find out what they might be missing.
I doubt if it was considered important - after all, who would know the difference? Like the lack of a PS choke, the overheating power tranny, the . . . I can't go on, I get depressed by it all!
ray_moth said:
Not sure what you are saying.... This was transferred by Bulgin to here because he seems to be interested in the refurbishing I did on Leak Stereo 20s. That was what the first post here implied; after some drift to the Radford, I proceeded with the thread as originally intended.
I would say that was the hey-day of tube audio. After all, it was the time of the Williamson, Quad, Leak TL12, McIntosh, Radford, Dynaco; you name it. In as much as those represented technology of the day, I am only saying that I found it strange that this particular amp (only) would present some design anomoly. I can assure you that in those times there was as much theoretical analysis as today, if not more, and Wireless World thrived on that, etc. I would be surprised if approach was less stringent than today. I doubt if it was considered important - after all, who would know the difference? Like the lack of a PS choke, the overheating power tranny, the . . . I can't go on, I get depressed by it all!
Again, refer to previous commentary. Not to belabour, but in the ones I refurbished the power tranny did not get very hot. How did folks store it, etc. - personal stuff there. Anyway, my purpose was to mention this amp as a particular one that had a worse circuit anomoly than others, and I wondered about it because Leak fell under the upper echelons. (E.g. many broadcasting studios used Leak monitors, as did the SABC.)
I saw that my circuit diagram for the Stereo 20 did not materialise the previous time. Will try again:
Johan,
I remember, in the UK in the fifties and sixties, that a very small percentage of the population knew anything, or cared, about audio reproduction. Those were the days of eliptical speakers driven by ECL82 amplifiers in cramped record players with crystal pickups. If you could afford something more prestigious, you could invest in a slightly better but much more expensive radiogram. In that climate, manufacturers like Leak and QUAD were kings of the hill.
You mention the SABC's use of Leak amps; the BBC also used a lot of Leak gear. To be fair, Leak started off with some quite good products but became rather cost-conscious and this showed in the Stero 20 and Stereo 50. They finally blew it with their abysmal transistor amps (as did several similar companies, on both sides of the Atlantic).
Leak espoused very high degrees of negative feedback. If you read the Leak Point One Amplifiers brocjure, you can see what I mean. Nobody questioned that approach forty or fifty years ago. They also mysteriously referred to a 'triple loop' feedback scheme, which they never adequately explained, IMHO. (I know what I mean by triple NFB loops and I'm darned if I can see them in Leak designs).
Their amps were ridiculously sensitive and their preamps catered for this. For that reason, Leak did not want to sell Varislope preamps for use with amps by other makers, because they wouldn't work with amps having more 'normal' sensitivity. The preamps were also mediocre and tended to be noisy.
Let's face it, would you seriously consider building a Leak Stereo 20 or Stereo 50 with a Varislope preamp today? I certainly wouldn't! I don't want to sound like a crusader but I think Leak had it better than they deserved for many years. However, they were not the only ones.
I remember, in the UK in the fifties and sixties, that a very small percentage of the population knew anything, or cared, about audio reproduction. Those were the days of eliptical speakers driven by ECL82 amplifiers in cramped record players with crystal pickups. If you could afford something more prestigious, you could invest in a slightly better but much more expensive radiogram. In that climate, manufacturers like Leak and QUAD were kings of the hill.
You mention the SABC's use of Leak amps; the BBC also used a lot of Leak gear. To be fair, Leak started off with some quite good products but became rather cost-conscious and this showed in the Stero 20 and Stereo 50. They finally blew it with their abysmal transistor amps (as did several similar companies, on both sides of the Atlantic).
Leak espoused very high degrees of negative feedback. If you read the Leak Point One Amplifiers brocjure, you can see what I mean. Nobody questioned that approach forty or fifty years ago. They also mysteriously referred to a 'triple loop' feedback scheme, which they never adequately explained, IMHO. (I know what I mean by triple NFB loops and I'm darned if I can see them in Leak designs).
Their amps were ridiculously sensitive and their preamps catered for this. For that reason, Leak did not want to sell Varislope preamps for use with amps by other makers, because they wouldn't work with amps having more 'normal' sensitivity. The preamps were also mediocre and tended to be noisy.
Let's face it, would you seriously consider building a Leak Stereo 20 or Stereo 50 with a Varislope preamp today? I certainly wouldn't! I don't want to sound like a crusader but I think Leak had it better than they deserved for many years. However, they were not the only ones.
ray_moth said:
No.
Well, similar topology, but as said before, no ECC83s. My personal choice EF86, ECC81 and 6L6GC for a good 20 - 30W. Not the best, but IMO a balanced topology for a particular niche.
In fact, only later when I came into contact with the Stereo 20, I wondered why not just UL the KT66s, instead of doing the general stretch of the time to pull EL84s up to 14 - 15W. But as you mentioned, economy sharply took over there.
But there; I am cut down to size regarding my own era!
Down here I did not have that much contact with commercial amps (there was not a wide variety of imports anyway); mostly read about them in the then W-W. Thus perhaps I was more used to the category of Leak (TL12), Williamson, Walker, Dynaco, McIntosh and the like. Only a few bought/constructed (loudspeakers) in that class, and of course many of us promptly built/designed our own. (I fortunately owned a Langford-Smith from early times.)
Ah then! So glad you are higher up on the food chain than some promotional writers. Let us descend there for a moment and carefully check: As I see it, loop 1 is the feedback in the ltp phase inverter. (No, no; you must have a ve-e-e-ry liberal view of feedback here to understand!) The second loop is the magical cure for power stages: UL! And the third one we all know. So there (-er - I think).
Regards.
Johan,
The following is an extract from the Leak 'Point One' brochure:
8. "What is a triple loop feedback amplifier?"
By definition, a multiple loop feedback amplifier is one in which voltage can be returned to some of the grids by more than one path (Ref 19). In the TL/12 amplifier both the first and second stages have local returns, or loops, these being enclosed within the major loop made over the complete amplifier. Multiple loop structures will become more common, for it is possible to achieve with them results not available from single loop circuits.
From the above, it is apparent that Leak didn't include UL as one of these three FB loops. I suppose they can be allowed some leeway regarding the splitter, since one half of it does have NFB; however, why they should claim that the first stage has 'a local return loop' is beyond me - unless they mean the unbypassed portion of the IP tube's cathode resistor!
The following is an extract from the Leak 'Point One' brochure:
8. "What is a triple loop feedback amplifier?"
By definition, a multiple loop feedback amplifier is one in which voltage can be returned to some of the grids by more than one path (Ref 19). In the TL/12 amplifier both the first and second stages have local returns, or loops, these being enclosed within the major loop made over the complete amplifier. Multiple loop structures will become more common, for it is possible to achieve with them results not available from single loop circuits.
From the above, it is apparent that Leak didn't include UL as one of these three FB loops. I suppose they can be allowed some leeway regarding the splitter, since one half of it does have NFB; however, why they should claim that the first stage has 'a local return loop' is beyond me - unless they mean the unbypassed portion of the IP tube's cathode resistor!
Yipes!
"Well, well, so now we know." The honourable Harold did not then recognise that UL was also feedback - imagine missing that!
The 100 ohm resistor (R4) will give all of 0.42 dB of local NFB. That is really scraping the barrel for effect. No, I don't believe that bit was well thought out.
Regards.
"Well, well, so now we know." The honourable Harold did not then recognise that UL was also feedback - imagine missing that!
The 100 ohm resistor (R4) will give all of 0.42 dB of local NFB. That is really scraping the barrel for effect. No, I don't believe that bit was well thought out.
Regards.
Leak Stereo 20 - Johan's...
Mmmm
This has turned into a fascinating thread A bit above my knowledge but I'll be back here until I understand
Thank you all
Btw, my S/20's o/p t doesn't get hot - even at high volumes.
And now I will duck incase of some rockets I have the matching refurbished Varislope but it's not in use. Instead, I use my workhorse Kenwood Basic C1 Stereo Control amp with the Leak and with another s/s amp I have here. The result is firstclass
bulgin
Mmmm
This has turned into a fascinating thread
A bit above my knowledge but I'll be back here until I understand Thank you all
Btw, my S/20's o/p t doesn't get hot - even at high volumes.
And now I will duck incase of some rockets
I have the matching refurbished Varislope but it's not in use. Instead, I use my workhorse Kenwood Basic C1 Stereo Control amp with the Leak and with another s/s amp I have here. The result is firstclass bulgin
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.