Hi,
I'm evaluating a SE amplifier, using KT88 tubes.
I found two different, interesting schematics:
One uses a so-called UL schema, the other one is a simple SE (I hope I'm correct), with a local feedback on the secondary of the output transformer.
My questions are:
Which are the difference between these two schematics?
I mean, I can understand how a local feedback works (at the end, I'm an electronic), but not exactly how the UL schema does (I'm a newbie with tubes)... so, why to use a UL tap?
Which are the benefits against a standard local feedback?
The distorsion is lower?
The bandwidth is higher?
Ciao,
Giovanni
I'm evaluating a SE amplifier, using KT88 tubes.
I found two different, interesting schematics:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
One uses a so-called UL schema, the other one is a simple SE (I hope I'm correct), with a local feedback on the secondary of the output transformer.
My questions are:
Which are the difference between these two schematics?
I mean, I can understand how a local feedback works (at the end, I'm an electronic), but not exactly how the UL schema does (I'm a newbie with tubes)... so, why to use a UL tap?
Which are the benefits against a standard local feedback?
The distorsion is lower?
The bandwidth is higher?
Ciao,
Giovanni
Hi Giovanni,
Both schematics are single-ended. The first schematic shows the tube connected in ultralinear mode. The screen is connected to a tap on the output transformer.
The second schematic gives you a choice between pentode mode and triode mode. Pentode mode connects the screen to the power supply. Triode mode connects the screen to the plate, effectively making the KT88 into a triode.
Common wisdom has it that triode mode sounds best, pentode mode gives the most power (and distortion) and ultralinear is a compromise between them. In most pentodes pentode mode will give about 2 to 2.5 times the output power of triode mode with ultralinear between them somewhere. For instance and EL84 can give you 5.7 watts in pentode mode and about 2 watts in triode mode.
In order to use ultralinear mode your output transformer must have an ultralinear tap. Without that tap you are limited to either pentode or triode.
It is possible to install a rotary switch that will let you change your amp between pentode, ultralinear and triode modes. Connect the output of the switch to the screen grid of the tube. On one position solder in a 100 ohm resistor (5 watts would work) and connect the other end of the resistor to the plate pin. On the next position make the connection directly to the ultralinear tap on your OPT. In the third position make a connection to your B+. Now turning the switch changes the mode.
Note that the best operating point for the tube changes with mode so it is unlikely that the operating point chosen for say triode mode will be the best point for pentode. However you can still get an idea of the changes in sound.
My main amp currently is a set of SE KT88 monoblocks running in ultralinear mode and I think they sound great. I am contemplating adding a switch to change modes.
I also just finished a single-ended EL84 amp and have a pentode/triode switch installed. There is no question that triode sounds much better than pentode on that amp even though the power drops from about 5.5 to 2 watts.
Both schematics are single-ended. The first schematic shows the tube connected in ultralinear mode. The screen is connected to a tap on the output transformer.
The second schematic gives you a choice between pentode mode and triode mode. Pentode mode connects the screen to the power supply. Triode mode connects the screen to the plate, effectively making the KT88 into a triode.
Common wisdom has it that triode mode sounds best, pentode mode gives the most power (and distortion) and ultralinear is a compromise between them. In most pentodes pentode mode will give about 2 to 2.5 times the output power of triode mode with ultralinear between them somewhere. For instance and EL84 can give you 5.7 watts in pentode mode and about 2 watts in triode mode.
In order to use ultralinear mode your output transformer must have an ultralinear tap. Without that tap you are limited to either pentode or triode.
It is possible to install a rotary switch that will let you change your amp between pentode, ultralinear and triode modes. Connect the output of the switch to the screen grid of the tube. On one position solder in a 100 ohm resistor (5 watts would work) and connect the other end of the resistor to the plate pin. On the next position make the connection directly to the ultralinear tap on your OPT. In the third position make a connection to your B+. Now turning the switch changes the mode.
Note that the best operating point for the tube changes with mode so it is unlikely that the operating point chosen for say triode mode will be the best point for pentode. However you can still get an idea of the changes in sound.
My main amp currently is a set of SE KT88 monoblocks running in ultralinear mode and I think they sound great. I am contemplating adding a switch to change modes.
I also just finished a single-ended EL84 amp and have a pentode/triode switch installed. There is no question that triode sounds much better than pentode on that amp even though the power drops from about 5.5 to 2 watts.
Thanks for the answer, Sherman
I do not care about power, my "listening room" (my living room, in reality) is small, about 18 square meters...
So I think I do not need to go with power.
I love class A amplifier (up to now I realised only solid state class A amplifiers, like Pass Lab amps), so my first goal is to have the lowest distorsion and the best sounding amp (whatever it means).
I think the power of a single KT88 connected in triode or ultralinear mode will be enough for me, I ws trying to understand the different behaviour of ultra linear vs. triode, etc.
One more question: both the schematics use a feedback from the output, I've seen also sothers schematics without any feedback: do they work fine, better or worse than these two, and why?
Sorry if I'm boring you, but I'm trying to learn as much as I can before to start with building...
Ciao,
Giovanni
I do not care about power, my "listening room" (my living room, in reality) is small, about 18 square meters...
So I think I do not need to go with power.
I love class A amplifier (up to now I realised only solid state class A amplifiers, like Pass Lab amps), so my first goal is to have the lowest distorsion and the best sounding amp (whatever it means).
I think the power of a single KT88 connected in triode or ultralinear mode will be enough for me, I ws trying to understand the different behaviour of ultra linear vs. triode, etc.
One more question: both the schematics use a feedback from the output, I've seen also sothers schematics without any feedback: do they work fine, better or worse than these two, and why?
Sorry if I'm boring you, but I'm trying to learn as much as I can before to start with building...
Ciao,
Giovanni
Hi croccodillo,
I think you have to valutate also more "classical" schematics. For example a single ended of EL34(in triode mode) togheter with a triode signal tube, without FB.
Cheers.
P.s. Detta tra noi le migliori configurazioni sono a triodo, in assoluto. Ancora meglio quelle che usano come valvola finale un triodo a riscaldamento diretto, per esempio la 300B o la 2A3.
I think you have to valutate also more "classical" schematics. For example a single ended of EL34(in triode mode) togheter with a triode signal tube, without FB.
Cheers.
P.s. Detta tra noi le migliori configurazioni sono a triodo, in assoluto. Ancora meglio quelle che usano come valvola finale un triodo a riscaldamento diretto, per esempio la 300B o la 2A3.
Giaime said:
Ahahaahah ma perchè lo dici in italiano? hai paura che qualche "straniero" te lo contesti?
Dai le affermazioni facciamole tutte in inglese per piacere, è diritto di tutti leggerle e replicare.
Ciao
L'ho detto in italiano perchè ad un certo punto mi sono sentito come un cretino a dover parlare in inglese con un italiano.
Paura di contestazioni straniere? naaaaaa!
Anzi, fai pure la traduzione del tutto, così tutti capiscono.
Oh good... nothing worth translating. He said in italian about the superiority of directly heated triodes over other tubes. I said if he was saying it in italian to avoid contestations from other people, and he said nope, he wasn't doing that, he felt just embarassed to speak in English with another italian.
Sorry for the OT...
Sorry for the OT...
croccodillo said:
Giovanni,
A single-ended KT88 in ultralinear mode is capable of about 5-8 watts depending on the operating point you select. In triode mode you might get half of that.
Generally triode mode gives lower distortion than pentode or ultralinear. Feedback is used to help control distortion with the loss of some gain.
Distortion is one of those things that is hard to discuss. With tube amplifiers harmonic distortion is mostly second order and other higher even order harmonics. To the ear that type of distortion usually sounds fine. With solid state equipment harmonic distortion is usually odd order which sounds bad to the ear.
A single-ended tube amp running in triode mode with 4% distortion would probably sound great to most people depending on their preferences.
Here is the schematic I used for my KT88 amps. It was designed by Mikael Abdellah. Do a search here for that name and you can find several threads about the amp with a lot of good info.
Attachments
I loooove Mikael's schematic....
But I'm a little bit confused: how much power can such an amplifier produce?
It is ultralinear, so, I suppose, about 5-8 Watts, as you reported?
And, what is the "operating point" you mention?
Sorry for my stupid questions, but I'm a beginner...
Ciao,
Giovanni
But I'm a little bit confused: how much power can such an amplifier produce?
It is ultralinear, so, I suppose, about 5-8 Watts, as you reported?
And, what is the "operating point" you mention?
Sorry for my stupid questions, but I'm a beginner...
Ciao,
Giovanni
Hi croccodillo,
the operating point is the "way" you choose for make "well-working" the tube. You have to study the datasheet of the tube, seeing the curves, and then choose a load(for output tubes generally it is the output transformer) a B+ and an appropriate cathode resistor(bypassed or not by a cap) for negative V of the grid. Naturally you have to do the same for signal tubes(in this case the load is generally a resistor). This is also a rough semplification of the question and there are many other factors to study ad take care of.
Btw I am not in accord with Sherman, since a KT88 can make more than 5-8W.
It represent a "family" of 35W(plate dissipation) tube, with 6550, and some versions can reach up to 42W(plate dissipation).
So, in triode mode(single ended) the KT88 can reach up to 12W of output power, in tetrode(it is a tetrode and not a pentode) or ultralinear(single ended) can reach up to 20W of output.
In push pull(2 tubes in triode mode) it can reach up to 30W, and in ultralinear mode(push pull) can reach 100W or more of output power.
Generally, an output power of 5-6W is referred to EL34 tube in triode(single ended). Remember that EL34 is smaller tube than KT88. El34 have "only" 25W(plate dissipation).
cheers.
the operating point is the "way" you choose for make "well-working" the tube. You have to study the datasheet of the tube, seeing the curves, and then choose a load(for output tubes generally it is the output transformer) a B+ and an appropriate cathode resistor(bypassed or not by a cap) for negative V of the grid. Naturally you have to do the same for signal tubes(in this case the load is generally a resistor). This is also a rough semplification of the question and there are many other factors to study ad take care of.
Btw I am not in accord with Sherman, since a KT88 can make more than 5-8W.
It represent a "family" of 35W(plate dissipation) tube, with 6550, and some versions can reach up to 42W(plate dissipation).
So, in triode mode(single ended) the KT88 can reach up to 12W of output power, in tetrode(it is a tetrode and not a pentode) or ultralinear(single ended) can reach up to 20W of output.
In push pull(2 tubes in triode mode) it can reach up to 30W, and in ultralinear mode(push pull) can reach 100W or more of output power.
Generally, an output power of 5-6W is referred to EL34 tube in triode(single ended). Remember that EL34 is smaller tube than KT88. El34 have "only" 25W(plate dissipation).
cheers.
Tubes4e4 said:
Remember, it is not only size that matters
Tom
Sure, I agree.
I don't like very much KT88 and I don't need extra-power from amps. The best sound I ever heard was from a 10Y tube amp. Naturally it is only my opinion(or my taste, if you prefer...).
The question here was different: I only tryed to say that when someone choose a KT88, often it is to have power...
Giovanni,
Perhaps a little more on UL (or distributed load, which term I prefer - "the ultra-linear" is a bit over the top. Better linearity is not the greatest advantage, as you will notice from the following).
The UL operation gives pentode efficiency and output power but with almost triode distortion and internal impedance. If one should have an output transformer with taps every 10% and you move the G2 from pentode to triode, you will find that at about 30 - 40% G2 tap, characteristics have become almost triode while available output is still that for pentode - in fact a little more because the G2 is now also contributing to output power. I have posted a table of values on another thread some time ago; cannot recall where.
To quote figures from graphs by GEC (England) for p.p KT88s for the same load and voltage: Max. output power is 47W for pentode, with internal impedance (Ri) = 20K and distortion (D) = 3%. For triode: Max. output power = 23W, Ri = 3K and D = 1.5%. For UL (43% G2 taps): Max. ouput power = 50W, Ri = 4.1K and D = 1.8%. These are for p.p. but the same ratios hold for SE, and roughly for all tubes in general. Also this is not the whole story, but will illustrate the basics.
Further, one must note that D is given as a % of maximum power. This means that for triode operation D = 1.5% of 23W, while for UL, D = 1.8% of 50W. This means that D is actually lower for UL than for triode operation at the same output power.
Some people say that they do not like the sound of an UL stage, but hearing is subjective so one cannot comment on that.
Perhaps a little more on UL (or distributed load, which term I prefer - "the ultra-linear" is a bit over the top. Better linearity is not the greatest advantage, as you will notice from the following).
The UL operation gives pentode efficiency and output power but with almost triode distortion and internal impedance. If one should have an output transformer with taps every 10% and you move the G2 from pentode to triode, you will find that at about 30 - 40% G2 tap, characteristics have become almost triode while available output is still that for pentode - in fact a little more because the G2 is now also contributing to output power. I have posted a table of values on another thread some time ago; cannot recall where.
To quote figures from graphs by GEC (England) for p.p KT88s for the same load and voltage: Max. output power is 47W for pentode, with internal impedance (Ri) = 20K and distortion (D) = 3%. For triode: Max. output power = 23W, Ri = 3K and D = 1.5%. For UL (43% G2 taps): Max. ouput power = 50W, Ri = 4.1K and D = 1.8%. These are for p.p. but the same ratios hold for SE, and roughly for all tubes in general. Also this is not the whole story, but will illustrate the basics.
Further, one must note that D is given as a % of maximum power. This means that for triode operation D = 1.5% of 23W, while for UL, D = 1.8% of 50W. This means that D is actually lower for UL than for triode operation at the same output power.
Some people say that they do not like the sound of an UL stage, but hearing is subjective so one cannot comment on that.
Johan Potgieter said:
A stupid idea:
Use a variac as an output transformer. The wiper is the screen tap, one end of the winding goes to the plate, the other B+. Wind a few turns of new wire to create a speaker secondary winding in a physical location that won't interfere with the wiper. Thus infinately variable triode to UL to pentode range. One serious issue is core saturation, forget if toroid cores hate it or don't mind it....
Even if not a problem it's not likely to be hifi as the variac was designed for 60Hz work.I did not really intend the multitap transformer as a practical idea at all, just to explain how the firm might have experimentally obtained data. I do not think that further experimenting in this way in the DIY fashion would serve much of a purpose; the data seems to be much the same for most power tubes. But your idea of using a Variac is excellent as a first approximation. There are usually a few taps somewhere that could be loaded appropriately. For myself, I use the existing data from manufacturers without re-inventing the wheel.
I'm a little bit confused, but slowly (thanks to all of you) my knowledge is increasing...
So, I think I will try Mikael schematic, mostly because it is so simple (I love simple things).
Now, some other questions:
A shop near me has available the following kins of tubes:
- Electro harmonix 6550-EH
- Svetlana 6550C
- Sovtek KT88
- Svetlana KT88
- Electro Harmonix KT88-EH
Have you ever tried those tubes? Which (in your opinion, of course) sound better, and why?
Again:
I can't found a 6N1P near here, so Ifound something different:
- Electro Harmonix 6922-GOLD
- Electro Harmonix 6922-EH
Are those tubes good?
Which is the best?
Ciao,
Giovanni
So, I think I will try Mikael schematic, mostly because it is so simple (I love simple things).
Now, some other questions:
A shop near me has available the following kins of tubes:
- Electro harmonix 6550-EH
- Svetlana 6550C
- Sovtek KT88
- Svetlana KT88
- Electro Harmonix KT88-EH
Have you ever tried those tubes? Which (in your opinion, of course) sound better, and why?
Again:
I can't found a 6N1P near here, so Ifound something different:
- Electro Harmonix 6922-GOLD
- Electro Harmonix 6922-EH
Are those tubes good?
Which is the best?
Ciao,
Giovanni
croccodillo said:
Hi Giovanni,
take in mind that a Kt88 single ended is expensive. Big output transformers, big supply system(transformer, caps and so on), expensive tubes...
For tubes question, consider NOS tubes, also.
You like this schematic because it is simple? Many other single ended are simple...
Consider this, for example: http://www.audiokit.it/ITAENG/KitElettr/MEGA/Lilliput/Lillman4.gif
It is "only" 2,5W, but it is cheeper, and it will sound better(IMO).
croccodillo said:
Given your choices of KT88 I think I would choose either the Svetlana or the EH. No big reasons, it is just that I've heard them and like the sound. I think the Svetlana (if we are talking about the Winged C logo tubes) has a bit better specs than the Sovteks.
If you can't find a 6N1P you can substitute the 6922 but you will need to change the anode resistor value since the 6922 has a lower max voltage than the 6N1P.
Another and possibly better option than simply putting in a bigger resistor is to decouple the PS. Use a 20K resistor in series from the PS, followed by a 47uF cap to ground and then the anode resistor from the schematic. (Double check that the voltage will be OK on the 6922, I didn't do the math!
)E-Linear
Hey-Hey!!!,
There is another way of getting some more NFB w/o making a loop from OPTx secondary to input cathode...
Peter Millett did it SE just a bit after I did it PP. Same inspiration, just slightly different methods.
Connect the input amplifier tube's plate supply to the g2 tap, instead of another de-coupled B+ supply. Here it is with a PP grid choke on the finals and in PP. Cut it in half for SE. It works best( lowest output Z ) if the amplifier/driver is a pentode, or a faux-pentode cascode.
cheers,
Douglas
Hey-Hey!!!,
There is another way of getting some more NFB w/o making a loop from OPTx secondary to input cathode...
Peter Millett did it SE just a bit after I did it PP. Same inspiration, just slightly different methods.
Connect the input amplifier tube's plate supply to the g2 tap, instead of another de-coupled B+ supply. Here it is with a PP grid choke on the finals and in PP. Cut it in half for SE. It works best( lowest output Z ) if the amplifier/driver is a pentode, or a faux-pentode cascode.
cheers,
Douglas
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.