I rebuilt one of these when I was in grad school; absolutely gorgeous construction and HUGE transformers for the power rating.
In any case, I'd balance the two plate resistors in the long-tail phase splitter, then change out the cathode resistor (the long-tail) for a current source. In my rebuild, I added a fixed-bias supply, took the output stage cathodes close to ground, and converted the output stage to triode.
In any case, I'd balance the two plate resistors in the long-tail phase splitter, then change out the cathode resistor (the long-tail) for a current source. In my rebuild, I added a fixed-bias supply, took the output stage cathodes close to ground, and converted the output stage to triode.
Do you intend to obtain a Stereo 20 chassis and rebuild it, or are you going to obtain transformers etc and use them to build the circuit? All the little capacitors (<1nF) are specific to the output transformers. If you modify as SY suggests, this could also change HF response. Could you obtain access to test equipment (oscilloscope, oscillator)? Do you know someone who could give you guidance in using test equipment?
Hi, thanks for yours guide, i start in this business with the help of a friend , he is an electronical tech and is he first time in tubes, but he knows about instrumental, work with high tension and other stuffs. Your guide is very apreciate, in this country exist the posibility of obtain tubes free or cheap, but the trafos is other history. I am newbie and learn fastly with the help of friends (like both of you). I have the schematic, and trafos, tubes , etc. What are the trafos values? I checkit with the trafos.
Thanks and sorry if my english is no very clear.
Federico
Thanks and sorry if my english is no very clear.
Federico
Hi,
That should read:
....I always liked but utilizing TWO pairs of EL84s per channel.
Note that the suggested improvements by the other members can easily implemented and of course you can drive a single pair of EL84s just the same but the value of cathode resistor in the output needs to go up to about 270 Ohm.
Cheers,
That should read:
....I always liked but utilizing TWO pairs of EL84s per channel.
Note that the suggested improvements by the other members can easily implemented and of course you can drive a single pair of EL84s just the same but the value of cathode resistor in the output needs to go up to about 270 Ohm.
Cheers,
You should also join the Leak Yahoo group http://groups.yahoo.com/group/hjleak/messages
There are some guys there who have done some radical changes.
Oh I see Andy Evans has already replied
There are some guys there who have done some radical changes.
Oh I see Andy Evans has already replied
I'm awaiting help to post the schematic, but in the meantime some of my experiences. The Leak is a Mullard type circuit with feedback. With ECC83s it has way too much input sensitivity, so decision one is low mu valves. Decision two is do you take off the global feedback. I'd say yes - I've tried it both ways and to me it sounds better - more natural, smoother, better mids. I can take some softness in the bass. Decision two leaves you with an easy situation to solve - drive a couple of EL84s with the input of a CD player. If you want balanced, use a diff pair - like the second stage in the existing Leak schematic but with a current source in the tail. Single ended - I'd use a voltage stage DC coupled into a concertina. Both the above solutions only need one coupling cap in the signal path - more than that is silly. Both arrangements are widely used and you just have to choose values. For my concertina the values I have are: voltage stage 47k anode, 500 cathode bypassed with 4.7uF polyprop, concertina 27k top and bottom. With about 275v HT this gives around 90v on the first anode which is what you want to feed the concertina and give it a good voltage balance on anode and cathode, i.e. roughly thirds for cathode/anode/HT voltages. Simple as that. If you impliment this with an existing Stereo 20, you'll have to pull out the circuit board and rewire under it, or make another board out of copper clad fibreglass - you can get the right size cheaply - and put tag strips on it. I rewire the heaters, twisting them properly, and put better wire in for the anode and screen connections of the EL84s at the same time as removing the board. Think about putting in better quality 9pin sockets as well if you wish, though I didn't. change the phono sockets for bigger gold plate ones, and I'd suggest fitting a normal IEC mains socket where the two pin connector sockets are - it fits exactly in the hole. Then better speaker sockets - all this just commonsense. What I did was split the PSU after the first cap (47uF 'lytic) into two chokes 10H at 100mA which I got from Maplin UK. Hammond probably have an equivalent. Anyway, as luck would have it these fit pointing down in the holes of the mains transformer, so you just change the screws for bolts and bolt the chokes on the bottom (orientation N-S). I prefer polyprop caps for the smoothing caps and use 40uF motor runs by LCR or whatever else fits inside and outside once you've taken off the existing two double electrolytic caps. Two of these larger caps go under the OPTs and the other two go on top - easiest solution. And that's basically it. I use Russian Teflon coupling caps, .056 fit the existing board, .1 don't so you have to think laterally. I use holco 1W resistors - a luxury I personally believe in, since there are now so few in the circuit you can chuck money at them, and they do sound better than generic metal film. Cahode bypasses for the EL84s - your choice. Maybe OS-CON, Nichicon ZA or ZL, Panasonic, ELNA Silmic, Nichicon Muse - whatever. Took me years to work all this out!
Part 2 - the valves.
Here are some solutions for the valve stage (V1) DC coupled to the concertina (V2)
Menno van der Veen ECC82, HT280v, V1 Anode 100k, cathode 1.2K bypasses 1,000uF; V2 22K top and bottom a=200v, k=76v
Morgan Jones ECC88, HT 285v, V1 Anode 47k, cathode 830 unbypassed, V2 22k top and bottom, a=187.5v, k=97.5v
Another version - 6N30 HT = 315v, V1 Anode 15k, cathode 300 Vc=4.2V Vp=95V Ip=14ma; V2 6.8K top and bottom a=214v, k=100v, Ip=14.7ma Total unloaded p-p drive is 84V
Morgan uses a higher mu valve, but doesn't bypass, another solution. My experiments show that you need a minimum of mu=6 for V1 (I used a 6EW7 with beefy section first), but that gives you no spare gain at all so not very practical (but no volume control!!). At the same time you don't want - as Menno says - unnecessary gain which you then lose in the volume control. So I think mu=16 up to 25 or so is the right range. This eliminates anything much higher except the ECC88 (6922) which has our other requirement - low anode impedence (Ra). So for double triodes lower mu and low Ra ideally gives us a choice of 6N30, 6N6, ECC99, ECC82 (?), 5687, E182CC/7119, 12BH7, 6CG7/6FQ7. You have another choice since you have three sockets - one double triode and two singles. This brings in the 12B4 which would be my choice. You could use this in either stage - I think it's best in the concertina. Now this brings me to the choice of V1. This to me is the crucial decision. All the above are possibles, but to me lack that ultimate magic. For a while I stuck a vintage 6SN7 in the preamp octal socket on the front. Magic. Not very nice looking. took a deep breath and decided to use two octal sockets on the front end. I bored out V1 socket and put in an octal socket, and made one to fit on the other side where the capacitor is, since it goes in the hole. I now had choice of 6J5, 6C5 and their variants. Serious magic here. That's where I stopped. Nowhere else to go. No more room. Best valves I could think of. So on to other projects. Now, note that the Leak TL12+ has an octal socket neatly on the top ready for use, so that's a better starting point. This amp sounds very good. I suspect there may be some ringing in the OPT - referred to previously - but I can't be more exact. Andy Evans
Here are some solutions for the valve stage (V1) DC coupled to the concertina (V2)
Menno van der Veen ECC82, HT280v, V1 Anode 100k, cathode 1.2K bypasses 1,000uF; V2 22K top and bottom a=200v, k=76v
Morgan Jones ECC88, HT 285v, V1 Anode 47k, cathode 830 unbypassed, V2 22k top and bottom, a=187.5v, k=97.5v
Another version - 6N30 HT = 315v, V1 Anode 15k, cathode 300 Vc=4.2V Vp=95V Ip=14ma; V2 6.8K top and bottom a=214v, k=100v, Ip=14.7ma Total unloaded p-p drive is 84V
Morgan uses a higher mu valve, but doesn't bypass, another solution. My experiments show that you need a minimum of mu=6 for V1 (I used a 6EW7 with beefy section first), but that gives you no spare gain at all so not very practical (but no volume control!!). At the same time you don't want - as Menno says - unnecessary gain which you then lose in the volume control. So I think mu=16 up to 25 or so is the right range. This eliminates anything much higher except the ECC88 (6922) which has our other requirement - low anode impedence (Ra). So for double triodes lower mu and low Ra ideally gives us a choice of 6N30, 6N6, ECC99, ECC82 (?), 5687, E182CC/7119, 12BH7, 6CG7/6FQ7. You have another choice since you have three sockets - one double triode and two singles. This brings in the 12B4 which would be my choice. You could use this in either stage - I think it's best in the concertina. Now this brings me to the choice of V1. This to me is the crucial decision. All the above are possibles, but to me lack that ultimate magic. For a while I stuck a vintage 6SN7 in the preamp octal socket on the front. Magic. Not very nice looking. took a deep breath and decided to use two octal sockets on the front end. I bored out V1 socket and put in an octal socket, and made one to fit on the other side where the capacitor is, since it goes in the hole. I now had choice of 6J5, 6C5 and their variants. Serious magic here. That's where I stopped. Nowhere else to go. No more room. Best valves I could think of. So on to other projects. Now, note that the Leak TL12+ has an octal socket neatly on the top ready for use, so that's a better starting point. This amp sounds very good. I suspect there may be some ringing in the OPT - referred to previously - but I can't be more exact. Andy Evans
Andy, the high input sensitivity was put in because the matching Leak preamp was a rather soggy one tube affair. Your approach is certainly valid, but one can also take the opposite stance, embrace the gain, and use it as an opportunity to increase the feedback, reduce the distortion, and reduce the source impedance. Again, my one-and-only rebuild of the 20 goes back a quarter of a century, but I don't recall any particular difficulty in getting the amp stable with a lower closed-loop gain. As with any other good tube amp, it's just a matter of diddling with the compensation. The guy I built this for still uses the amp and is still driving (happily) a pair of vintage Fulton speakers.
thanks for posting the schematic!
"one can also take the opposite stance, embrace the gain, and use it as an opportunity to increase the feedback, reduce the distortion, and reduce the source impedance."
I built several versions with feedback and the original Mullard type circuit, the best being 6N30 into ECC99. My decision to go to no global feedback was based on the fact that to me it just sounded much better. I put my current circuit head to head with a previous one with the feedback and the decision for no feedback was made by three people in 30 seconds. Once you take off the feedback, the DC coupled concertina is a logical solution. Of course speakers come into it but damping with UL isn't so far from damping with triodes in the output, so we're getting some of the zero feedback SET effect, at least in the sense of smooth treble, liquid mids and warm bass. Going back to global feedback now - for me - tightens up the sound and I don't like it. I guess you get used to the change. Andy
"one can also take the opposite stance, embrace the gain, and use it as an opportunity to increase the feedback, reduce the distortion, and reduce the source impedance."
I built several versions with feedback and the original Mullard type circuit, the best being 6N30 into ECC99. My decision to go to no global feedback was based on the fact that to me it just sounded much better. I put my current circuit head to head with a previous one with the feedback and the decision for no feedback was made by three people in 30 seconds. Once you take off the feedback, the DC coupled concertina is a logical solution. Of course speakers come into it but damping with UL isn't so far from damping with triodes in the output, so we're getting some of the zero feedback SET effect, at least in the sense of smooth treble, liquid mids and warm bass. Going back to global feedback now - for me - tightens up the sound and I don't like it. I guess you get used to the change. Andy
Well, putting aside matters of taste and preferences in signal processing, in my experiments with a variety of tube amps over the years, I came to the conclusion that amps using a properly stabilized feedback scheme (not a big trick, that) did a better job of replicating their input signal. There's no question that eliminating the feedback loop will significantly change the frequency response into a speaker load, which will be clearly audible. It will also significantly increase the distortion, which may or may not be audible. It then becomes the listener's choice as to whether that alteration is pleasing or not.
There's no question that eliminating the feedback loop will significantly change the frequency response into a speaker load, which will be clearly audible. It will also significantly increase the distortion, which may or may not be audible. It then becomes the listener's choice as to whether that alteration is pleasing or not.>
We then come back to the mysteries of audio - what if the alteration is pleasing - what, then, do we measure? And what comes first - measurements or listening? I can see that commercial designs used feedback both because the measurements 'sold' the amp (point one in Leak's case) and because the better damping factor meant that customers with difficult speakers wouldn't complain. Sensible stuff. However, in the case of DIY, which this forum is about, we can choose our speakers or indeed anything provided we can make or buy it. So once we establish that the speakers are suitable for an amp without global feedback, the problem becomes less pressing - this is all written about in Menno van der Veen's book.
We then come back to the mysteries of audio - what if the alteration is pleasing - what, then, do we measure? And what comes first - measurements or listening? I can see that commercial designs used feedback both because the measurements 'sold' the amp (point one in Leak's case) and because the better damping factor meant that customers with difficult speakers wouldn't complain. Sensible stuff. However, in the case of DIY, which this forum is about, we can choose our speakers or indeed anything provided we can make or buy it. So once we establish that the speakers are suitable for an amp without global feedback, the problem becomes less pressing - this is all written about in Menno van der Veen's book.
Hi, remember i have a pair of Tannoy GRF 15", my listen room is about of 40 m2, and play cd, lp, tuner and dvd.
My favorite music is very eclectic ( rush, radiohead, jazz, classical, tango, etnic,celtiac, etc).I plan build an pre with those inputs, then the choice of this amp is well?
Thanks, Federico
My favorite music is very eclectic ( rush, radiohead, jazz, classical, tango, etnic,celtiac, etc).I plan build an pre with those inputs, then the choice of this amp is well?
Thanks, Federico
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.