Hi!
The LL1962a is better suited for lower rp tubes. For higher rp tubes like the 10Y or 26, the 1660 is the better choice.
Regarding low rp driver tubes: Bear in mind that if you choose a 46 as driver, which similar to a 45, you are just shifting the problem one stage upward. Then a input tube is needed which adequately drives the driver
Best regards
Thomas
The LL1962a is better suited for lower rp tubes. For higher rp tubes like the 10Y or 26, the 1660 is the better choice.
Regarding low rp driver tubes: Bear in mind that if you choose a 46 as driver, which similar to a 45, you are just shifting the problem one stage upward. Then a input tube is needed which adequately drives the driver
Best regards
Thomas
I think we have to be careful here as to what we are calling "three stage". What I mean by it is that I have three stages overall, fed from my CD player which is a Rega with no doubt a typical CD output voltage.
I could put one stage in the preamp and two in the amp, or all three in the amp - it's exactly the same, three stages in total.
I think what Thomas is describing is a 2 stage amp driven by a preamp, i.e. three stages again but in a different configuration. Is this right Thomas?
Andy
I could put one stage in the preamp and two in the amp, or all three in the amp - it's exactly the same, three stages in total.
I think what Thomas is describing is a 2 stage amp driven by a preamp, i.e. three stages again but in a different configuration. Is this right Thomas?
Andy
Hi!
That's why I prefer three stage designs. Lynn Olson has a nice writeup on why low Rp tubes are desired for drivers. [FONT=Verdana, Helvetica, Arial, sans-serif[/QUOTE]
Please note that I've been talking 45 output tube! The 300B Lynn is using is a complete different animal and has different requirements on the driver.
I also would not generalize comments on 2 stage or 3 stage. The entire system and it's gain structured needs to be balanced for optimal results.
Also for the two stage approach with input transformer, the preamp needs to be matched! Here real low rp is needed (a few hundres of Ohms).
I think generalisms like tube Y sounds better than tube Z, or 2 stage is better than 3 stage, etc are very short sighted. Things should be matched to each other. Such comparisons are only valid in the given context they ave been tried things might be totally different in another.
All IMHO and IME of course YMMV
Thomas
Yes Andy, that's correct. I keep preaching to consider the entire system. In this case we know that the threadstarter already has a 26 preamp with step down output transformer. So he has a preamp which could drive an input transformer adequately. Hence the proposal of a 2 stage amp with input trannie
Thomas
Hi Thomas! I'm just as keen on the entire system as you are - I just do it a different way!
I've also been saying consistently that tubes have to work into interstages - the sound you get is the combination. No doubt the 10Y in itself is potentially better than the 46. But that's only its potential. In practice who knows - with a LL1692A the 46 may sound better than the 10Y. And so on and so on. It also depends on who is listening and what their priorities are. Mine are tone and clarity, but especially tone.
I was a professional musician for very many years and I have a very good idea imprinted in my memory of what instruments sound like - I had them right in front of me for hours every day. So I try and re-create that. And to do that I make combinations according to the most natural sound I can get, then put the whole system together to get that sound. I've been through years of listening to different combinations of DHTs and circuits, and I must say I'm running out of improvements at this stage. I have a small number of components I use again and again.
I don't think we disagree over the fundamentals, though it's clear that our systems are very different from each other.
The one thing I'll be eternally grateful to you for is introducing me to filament bias!!! That was brilliant.
Andy
I've also been saying consistently that tubes have to work into interstages - the sound you get is the combination. No doubt the 10Y in itself is potentially better than the 46. But that's only its potential. In practice who knows - with a LL1692A the 46 may sound better than the 10Y. And so on and so on. It also depends on who is listening and what their priorities are. Mine are tone and clarity, but especially tone.
I was a professional musician for very many years and I have a very good idea imprinted in my memory of what instruments sound like - I had them right in front of me for hours every day. So I try and re-create that. And to do that I make combinations according to the most natural sound I can get, then put the whole system together to get that sound. I've been through years of listening to different combinations of DHTs and circuits, and I must say I'm running out of improvements at this stage. I have a small number of components I use again and again.
I don't think we disagree over the fundamentals, though it's clear that our systems are very different from each other.
The one thing I'll be eternally grateful to you for is introducing me to filament bias!!! That was brilliant.
Andy
Last edited:
Hi Andy
I don't think we disagree that much and I also don't think our systems are that different. They probably share a lot of sound characteristics !
In case some part or some combination sonds different than another, I try to get an understanding why. Often that is not that easy and one has to just accept the one is better. But also quite often a auick and easy measurement will explain why one sounds different.
Best regards
Thomas
I don't think we disagree that much and I also don't think our systems are that different. They probably share a lot of sound characteristics !
In case some part or some combination sonds different than another, I try to get an understanding why. Often that is not that easy and one has to just accept the one is better. But also quite often a auick and easy measurement will explain why one sounds different.
Best regards
Thomas
No generalizing going on here, Thomas. I fully understand the situation; I think it's a matter of design philosophy rather than not considering the entire system.
Case in point- my WOT preamp has 54 ohm output impedance, driving a power amp with Tribute input transformer, so that is very relevant to what is being discussed. The 46 tube is acting as a driver for the 300B output stage. Ignore the fact there is a 300B output stage, the preamp/input transformer was not suitable for driving the 46 adequately (and the 46 is very similar to the 45). This by ear and measurement- an input transformer does not like heavy secondary loading (they are not designed for it, Lundahl, Tribute, whatever). So an intermediate stage was needed between input transformer and 46. I should clarify that my equipment is all balanced push-pull throughout.
This is where design philosophy enters in. I could have picked for the additional first stage a high gain high Rp tube and an IT designed for that tube. My experience showed me more difficulties lie with that approach, especially if you want your output stage to be the limiting device (again, just my philosophy- I see many designs with a 12AX7 driving a 2A3 or 300B- no thank you). If you get anywhere near zero grid current, your even harmonics kick in. And maybe to some that sounds good. But if I want to have a two stage design with reasonable >1W output, chances are I need to get the most out of each stage's grid range. My measurements showed a high plate resistance source will inadequately drive a grid when requesting significant voltage swing.
So my decision resulted in pursuing low Rp stages and ending up with 3 instead of 2. This is all while considering the whole system, 92dB speakers. It is debatable if the OP speakers of 95dB fit this same requirement. But if I buy a new set of speakers, whether they be 89dB or 98dB, my system will work with great results. I prefer to build my equipment with some flexibility, not constrained for only one speaker system.
My end result that satisfies my personal philosophy is an amp that is solely limited by the output stage. I can drive the amp to 10% THD and every stage of my system is running <0.1% except the output stage. Excess headroom and deep Class A operation results in a very satisfying sound. The next person chooses different priorities. The beauty of this forum is everyone can toss in their philosophy, and the OP can decide what interests him.
I was interested in measurements of other designs to see how they compare. I have never seen measurements of a highish Rp tube into an IT (other than my personal measurements, which were not favorable).
Case in point- my WOT preamp has 54 ohm output impedance, driving a power amp with Tribute input transformer, so that is very relevant to what is being discussed. The 46 tube is acting as a driver for the 300B output stage. Ignore the fact there is a 300B output stage, the preamp/input transformer was not suitable for driving the 46 adequately (and the 46 is very similar to the 45). This by ear and measurement- an input transformer does not like heavy secondary loading (they are not designed for it, Lundahl, Tribute, whatever). So an intermediate stage was needed between input transformer and 46. I should clarify that my equipment is all balanced push-pull throughout.
This is where design philosophy enters in. I could have picked for the additional first stage a high gain high Rp tube and an IT designed for that tube. My experience showed me more difficulties lie with that approach, especially if you want your output stage to be the limiting device (again, just my philosophy- I see many designs with a 12AX7 driving a 2A3 or 300B- no thank you). If you get anywhere near zero grid current, your even harmonics kick in. And maybe to some that sounds good. But if I want to have a two stage design with reasonable >1W output, chances are I need to get the most out of each stage's grid range. My measurements showed a high plate resistance source will inadequately drive a grid when requesting significant voltage swing.
So my decision resulted in pursuing low Rp stages and ending up with 3 instead of 2. This is all while considering the whole system, 92dB speakers. It is debatable if the OP speakers of 95dB fit this same requirement. But if I buy a new set of speakers, whether they be 89dB or 98dB, my system will work with great results. I prefer to build my equipment with some flexibility, not constrained for only one speaker system.
My end result that satisfies my personal philosophy is an amp that is solely limited by the output stage. I can drive the amp to 10% THD and every stage of my system is running <0.1% except the output stage. Excess headroom and deep Class A operation results in a very satisfying sound. The next person chooses different priorities. The beauty of this forum is everyone can toss in their philosophy, and the OP can decide what interests him.
I was interested in measurements of other designs to see how they compare. I have never seen measurements of a highish Rp tube into an IT (other than my personal measurements, which were not favorable).
OT for this thread (except for the gain issue - sorry), but I got a bug about the 6N7 from your projects. Swapped it for the 6SN7 in this 801a amp: http://www.diyaudio.com/forums/tubes-valves/107759-801-amp-power-drive-3.html#post1322965
Works great, with few changes, and I can now drive it to full output with normal sources.
Sheldon
Hi Zigzag,
nobody is talking about a 12AX7 driving a output tube, also nobody is talking about an input transfromer driving a difficult load !??
I think we're not far off with our design philosophies.
I'm curious: What would you consider as a suitable tube to be driven by an input transformer, which is a DHT and low rp ?
BTW: the NC20 has a good bandwidth with tubes like 26 or 10Y. Reaching around 80kHz. Would need to measure again to get the exact results. Good enough in my book anyways.
Best regards
Thomas
nobody is talking about a 12AX7 driving a output tube, also nobody is talking about an input transfromer driving a difficult load !??
I think we're not far off with our design philosophies.
I'm curious: What would you consider as a suitable tube to be driven by an input transformer, which is a DHT and low rp ?
BTW: the NC20 has a good bandwidth with tubes like 26 or 10Y. Reaching around 80kHz. Would need to measure again to get the exact results. Good enough in my book anyways.
Best regards
Thomas
Thanks Thomas, Andy and zigzagflux for all the inputs. its really appreciated. it certainly save me tons of trouble and experiments
i am sure everyone has his personal opinion on this. designing amp and choosing its components is an art. matching it into the systems is another art as well.
thanks Sheldon for the input. 6SN7 is a superb tube. seen it in a lot of SET design and it has current production as well which is a plus.
however i think i am gonna stick to 10y and 45.
would anyone kindly check out the schematic i draw on #38
http://www.diyaudio.com/forums/tube...ll-dht-set-power-amplifier-4.html#post2666874
and offer some advice on its operating points. especially 10Y. i wonder if i could lower its current further. thanks in advance
Erwin
i am sure everyone has his personal opinion on this. designing amp and choosing its components is an art. matching it into the systems is another art as well.
thanks Sheldon for the input. 6SN7 is a superb tube. seen it in a lot of SET design and it has current production as well which is a plus.
however i think i am gonna stick to 10y and 45.
would anyone kindly check out the schematic i draw on #38
http://www.diyaudio.com/forums/tube...ll-dht-set-power-amplifier-4.html#post2666874
and offer some advice on its operating points. especially 10Y. i wonder if i could lower its current further. thanks in advance
Erwin
hi Andy
i think the Vanode, Vgrid and Ianode on the graph should be the same. only that the B+ is different between grid bias and cathode bias. where the cathode bias adds the bias voltage to the B+ required.
is this #46 on you schematic, Andy?
"Grid bias: 174v anode, -24v grid, 10mA as above. This would be as in the datasheet and in the supplied curves. Anode to cathode is 174v since cathode is at zero."
i assume above is correct since its working
Cathode bias: 174v anode, 0v grid, 24v cathode 10mA. B+ is 174v+24v+8v(assume the transformer voltage drop)=206v
correct me please if i am wrong. i am learning here as well
thanks
Erwin
Hi Erwin,
Don't lower the current in the 10Y. If you want to use NC20, raise it to 15mA. Here abozut 350V B+ with a cathode resistor of 2k would work fine
With the Lundahll 1660/10mA you can stick with the circuit as drawn.
Your output stageis fine like this.
I'd start with 1:4 step up at the input. Then adjust up or down if you need more or less gain
Thomas
Don't lower the current in the 10Y. If you want to use NC20, raise it to 15mA. Here abozut 350V B+ with a cathode resistor of 2k would work fine
With the Lundahll 1660/10mA you can stick with the circuit as drawn.
Your output stageis fine like this.
I'd start with 1:4 step up at the input. Then adjust up or down if you need more or less gain
Thomas
Yeah. It's a nice tube. Not many other choices with medium high mu. The single cathode connection means the sections can't be run independently with cathode bias. But since I'm using fixed bias, no problem. I only had to change the socket wiring. No other changes. Just dial in a good bias point.
Sheldon
OK, back to the regularly scheduled topic.
Not intending to deviate from the OP questions, and not attempting to be argumentative, but hopefully this can help define my simple goals when working with transformers.
In principle, I think we are talking about exactly that. The point is that there are a lot of designs out there that run these conditions (high Rp source and heavy grid load), which philosophically I do not prefer. But no doubt there are many that love the sound of these amps, and may not care what the FFT looks like. To each his own. My position is running a high Rp tube with a step up transformer into a 45 is very borderline. Impedance by the square of the turns ratio, after all.
An input transformer's proper application is at the end of a receiving line, close as possible to the next gain stage. It's priorities are common mode rejection, and often times phase splitting depending on amplifier design (like mine). A good transformer will exhibit high impedance to the input voltage, so as not to burden the source. It is difficult enough for a preamp to have to drive a length of cable, after all. These criteria result in a transformer with distributed winding capacitance, often a Faraday shield, and often high leakage inductance. Net result is these transformers do not like to be loaded by anything more than is necessary, and experience shows me they are most unhappy with capacitive loading (a better statement would be they prefer a very specific capacitive load- deviate and you have problems). Proof of this in my setup lied in checking lissajous curves on the output. Significant phase shift differences between the secondaries, even though a simple frequency response test was excellent. Clearly, there was something going on. So, loading the secondary with a power triode (I consider the 45 reasonably high power- it can drive a speaker after all) must be tackled very carefully, and I think measurements can really help validate success or failure. Ears do not always tell the whole story.
Now, make your input transformer a step-up, and things get more complicated. Now the statement 'any tube can run an input xfmr' no longer holds, since I am increasing the load on the source by the square of the turns ratio. 1:4 sounds like a good idea if one needs "free gain", but the source sees 16 times as much capacitive loading, and you may find your high Rp tube just can't do it. Further, one could try to use a lower Rp tube to fix the loading problem, but now you are running into an application somewhat outside the scope of the typical input transformer, and ringing or resonance issues can occur. This is nearing IT use.
An IT is best applied when you need low driving impedance into a difficult load, along with the flexibility of galvanic isolation and low grid circuit impedance. Further benefits could be argued of eliminating capacitors and blocking distortion. Faraday shields are atypical. Capacitive loading caution, though, still applies, and is exacerbated by high source impedances. In my designs, IT's and their driving tubes are capable of sourcing grid current. Not everyone's goal, I understand - but when you want to limit the number of stages, it helps to get the most signal range out of each stage.
So an oversimplification would be to say the input transformer is not intended to transform power, while an IT can.
I have never worked with the NC20, so cannot comment on its performance, but all the iron I have worked with still follows the basics of source impedance and load impedance. Running the transformer in circuit at 20+ kHz and looking at freq. and phase response can be very informative.
With that, I'll just shut up and let the thread continue forward without my interruptions.
In principle, I think we are talking about exactly that. The point is that there are a lot of designs out there that run these conditions (high Rp source and heavy grid load), which philosophically I do not prefer. But no doubt there are many that love the sound of these amps, and may not care what the FFT looks like. To each his own. My position is running a high Rp tube with a step up transformer into a 45 is very borderline. Impedance by the square of the turns ratio, after all.
Actually, I think you can use just about any tube you like to drive an input transformer, since I consider input and interstage transformers to be quite different. My concern with low Rp is centered around IT.
An input transformer's proper application is at the end of a receiving line, close as possible to the next gain stage. It's priorities are common mode rejection, and often times phase splitting depending on amplifier design (like mine). A good transformer will exhibit high impedance to the input voltage, so as not to burden the source. It is difficult enough for a preamp to have to drive a length of cable, after all. These criteria result in a transformer with distributed winding capacitance, often a Faraday shield, and often high leakage inductance. Net result is these transformers do not like to be loaded by anything more than is necessary, and experience shows me they are most unhappy with capacitive loading (a better statement would be they prefer a very specific capacitive load- deviate and you have problems). Proof of this in my setup lied in checking lissajous curves on the output. Significant phase shift differences between the secondaries, even though a simple frequency response test was excellent. Clearly, there was something going on. So, loading the secondary with a power triode (I consider the 45 reasonably high power- it can drive a speaker after all) must be tackled very carefully, and I think measurements can really help validate success or failure. Ears do not always tell the whole story.
Now, make your input transformer a step-up, and things get more complicated. Now the statement 'any tube can run an input xfmr' no longer holds, since I am increasing the load on the source by the square of the turns ratio. 1:4 sounds like a good idea if one needs "free gain", but the source sees 16 times as much capacitive loading, and you may find your high Rp tube just can't do it. Further, one could try to use a lower Rp tube to fix the loading problem, but now you are running into an application somewhat outside the scope of the typical input transformer, and ringing or resonance issues can occur. This is nearing IT use.
An IT is best applied when you need low driving impedance into a difficult load, along with the flexibility of galvanic isolation and low grid circuit impedance. Further benefits could be argued of eliminating capacitors and blocking distortion. Faraday shields are atypical. Capacitive loading caution, though, still applies, and is exacerbated by high source impedances. In my designs, IT's and their driving tubes are capable of sourcing grid current. Not everyone's goal, I understand - but when you want to limit the number of stages, it helps to get the most signal range out of each stage.
So an oversimplification would be to say the input transformer is not intended to transform power, while an IT can.
I have never worked with the NC20, so cannot comment on its performance, but all the iron I have worked with still follows the basics of source impedance and load impedance. Running the transformer in circuit at 20+ kHz and looking at freq. and phase response can be very informative.
With that, I'll just shut up and let the thread continue forward without my interruptions.
Hi Zigzag,
I think there is some misunderstanding. I fully agree to your statements actually. So let me repeat: There was no suggestion to use anything remotely like a 12AX7. Also no suggestion to drive a step up transformer from a high rp tube. Also never suggested to run the step up tansformer into the 45. I don't know where you got that from, certainly not from this thread. I always stated that the step up needs to be fed from a low impedance preamp.
Best regards ... Thomas
I think there is some misunderstanding. I fully agree to your statements actually. So let me repeat: There was no suggestion to use anything remotely like a 12AX7. Also no suggestion to drive a step up transformer from a high rp tube. Also never suggested to run the step up tansformer into the 45. I don't know where you got that from, certainly not from this thread. I always stated that the step up needs to be fed from a low impedance preamp.
Best regards ... Thomas
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.