Johan Potgieter said:
Ac operation of the circuit, you twit. You have not used an LTP for a long time, and looked at cathode coupled theory for even longer - it shows.
OK, OK, Alter Ego, you're correct as mostly ... O-K - always. Don't tell in public. Focusing on a point has its draw-backs.
So fine. Now apologise to Uncles SpreadSpectrum and HollowState (post #28), and see what others have to say about your other remarks.
Sorry, uncle's H_S and S_S. Valve gms need not be equal. I was concentrating on CCS vs resistor. The rest; to repeat: Perhaps old habits die hard, but I usually had a practical reason for figuring those; hope I still do.
Hello Spreadspectrum and Johan
Dear Spreadsp. and Johan,
Hello there - I started this thead to learn more about tubeamplifiers , and up untill now, it has been educational for me and hopefully for others with the same interest. Shall we try to keep our mutual dialogue as tactful as possible?
And Johan - I talked with my audio engineer, and he explained to me the reason why a single cathode resistorand a balance pot. could make the job just as well as a tale CCS. He also had the opinion, that CCS for the driver stage could introduce oscillation( as Rich poited out), if one don't make the circuit layout or hardwiring very carefully.
My point is - if a tale CCS improves sound in any way - we must have it! - If a tale CCS only improves messurement, and it is impossible to detect any advance in the soundstage - we don't need it ! ( but I may use the CCS anyway - If I know the amp messures better). My friend says it don't - but we will see!
I am much closer to finish my AR3a speaker project now, and if anyone could be interested, I could return with specs from two 6550a setups later.
I anyone has a pair of 70W, UL o/p transformers (Dynaco or similar) for the test, I am a buyer.
Keep the good spirit.
Best rgds
Kim
"I never forget a face, but in your case I'll be glad to make an exception." ~ Groucho Marx
Dear Spreadsp. and Johan,
Hello there - I started this thead to learn more about tubeamplifiers , and up untill now, it has been educational for me and hopefully for others with the same interest. Shall we try to keep our mutual dialogue as tactful as possible?
And Johan - I talked with my audio engineer, and he explained to me the reason why a single cathode resistorand a balance pot. could make the job just as well as a tale CCS. He also had the opinion, that CCS for the driver stage could introduce oscillation( as Rich poited out), if one don't make the circuit layout or hardwiring very carefully.
My point is - if a tale CCS improves sound in any way - we must have it! - If a tale CCS only improves messurement, and it is impossible to detect any advance in the soundstage - we don't need it ! ( but I may use the CCS anyway - If I know the amp messures better). My friend says it don't - but we will see!
I am much closer to finish my AR3a speaker project now, and if anyone could be interested, I could return with specs from two 6550a setups later.
I anyone has a pair of 70W, UL o/p transformers (Dynaco or similar) for the test, I am a buyer.
Keep the good spirit.
Best rgds
Kim
"I never forget a face, but in your case I'll be glad to make an exception." ~ Groucho Marx
I've meant no disrespect and have sensed none from others so far. If I have offended anyone, I apologize.
I just turned 28, and have only built one tube amplifier to date, so I am no expert by any means. By the way, I used a 6SN7 LTP with CCS tail and think that it sounds excellent.
This is a good point. You may want some adjustable imbalance to get the least distortion.
Here's how I would put it: If you want a low maintenance amplifier LTP with CCS tail and precision matched plate loads should give best performance, for the signals should be equal and opposite. Since they are equal, output tubes AC characteristics should be closely matched for lowest distortion.
Best, I think, would be to have adjustable imbalance. Then you could adjust out imbalance in output tube AC characteristics. This would require equipment to analyze distortion at the output of the amplifier and give the amp periodic tune-ups. I don't have this much time or energy, so I went the other route. Four kids under five years old will do that to you.
Thanks SpreadSpectrum
I may have misunderstood what Johan wrote back, but this is a great hobby for me, and building tubeamps should be just fun and amusement - I think!
[/B][/QUOTE]
Here's how I would put it: If you want a low maintenance amplifier LTP with CCS tail and precision matched plate loads should give best performance, for the signals should be equal and opposite. Since they are equal, output tubes AC characteristics should be closely matched for lowest distortion.
[/B][/QUOTE]
I may have to use the more simple amplifier circuit without keeping the phasesplitter in a DC/AC " iron grasp", in spite I have access to advanced Brüel&Kjaer equipment( my engineer friend has). If one don't have a tubetester around, we have to make a setup for matching the ouputtubes! As you say: With kids and work we don't unlimited free time to spare.
Thank you for your contribution
Best regards
Kim
SpreadSpectrum said:
I may have misunderstood what Johan wrote back, but this is a great hobby for me, and building tubeamps should be just fun and amusement - I think!
[/B][/QUOTE]
Here's how I would put it: If you want a low maintenance amplifier LTP with CCS tail and precision matched plate loads should give best performance, for the signals should be equal and opposite. Since they are equal, output tubes AC characteristics should be closely matched for lowest distortion.
[/B][/QUOTE]
I may have to use the more simple amplifier circuit without keeping the phasesplitter in a DC/AC " iron grasp", in spite I have access to advanced Brüel&Kjaer equipment( my engineer friend has). If one don't have a tubetester around, we have to make a setup for matching the ouputtubes! As you say: With kids and work we don't unlimited free time to spare.
Thank you for your contribution
Best regards
Kim
Re: Re: Test output trafo!
hey-Hey!!!,
For pentodes, the term 'mu' is a bit...errr, spread out. The pentode is circuit dependant to deliver gain, roughly gm*plate load. This leaves choice of plate load based on g2 voltage and usually restricts options via its small plate dissipation( 12BY7 is slightly better with more available than the EF184 ).
cheers,
Douglas
richwalters said:
hey-Hey!!!,
For pentodes, the term 'mu' is a bit...errr, spread out. The pentode is circuit dependant to deliver gain, roughly gm*plate load. This leaves choice of plate load based on g2 voltage and usually restricts options via its small plate dissipation( 12BY7 is slightly better with more available than the EF184 ).
cheers,
Douglas
Kimjul2005,
I certainly did not experience or intended any disrespect anywhere here.
If my post #42 ....
Please realise I was talking to myself! (italics being my alter ego, as in conscience or such; the sober part of my mind talking back to me, making sure I do not make a fool of myself too often) - as I often do when I have been stupid. My warped way of humor; as in trying to make fun of myself before I become too much of dead-wood.
Yes, a (good) CCS will keep outputs of the LTP equal. My point (perhaps poorly put) was that simply using resistors in a well-designed stage, will probably already get the LTP contribution to distortion lower than other such factors. As mentioned what about the power stage, also the effect at hight frequencies of different Miller capacitance, especially in the case of a 12AX7 (one reason why I do not like that tube in this application). Plus furthermore that most output transformers are not that balanced either (I am talking about a % or 2, as per my previous post). Simply that vs. cost/complexity; not that there is anything wrong with a CCS.
(When testing with a distortion meter in the past and adjustable LTP outputs as in a balance pot, I often found the lowest D figure not with exactly equal LTP output amplitudes, even differing slightly with frequency. But not enough to be of serious concern.)
I certainly did not experience or intended any disrespect anywhere here.
If my post #42 ....
Please realise I was talking to myself! (italics being my alter ego, as in conscience or such; the sober part of my mind talking back to me, making sure I do not make a fool of myself too often) - as I often do when I have been stupid. My warped way of humor; as in trying to make fun of myself before I become too much of dead-wood.
Yes, a (good) CCS will keep outputs of the LTP equal. My point (perhaps poorly put) was that simply using resistors in a well-designed stage, will probably already get the LTP contribution to distortion lower than other such factors. As mentioned what about the power stage, also the effect at hight frequencies of different Miller capacitance, especially in the case of a 12AX7 (one reason why I do not like that tube in this application). Plus furthermore that most output transformers are not that balanced either (I am talking about a % or 2, as per my previous post). Simply that vs. cost/complexity; not that there is anything wrong with a CCS.
(When testing with a distortion meter in the past and adjustable LTP outputs as in a balance pot, I often found the lowest D figure not with exactly equal LTP output amplitudes, even differing slightly with frequency. But not enough to be of serious concern.)
I use a transistor cascode for a CCS for that reason. I have a -15V supply that I run from a voltage doubler off of a spare 6.3V winding using an LM2990-15. There is a little added complexity but not much added cost.
I probably wouldn't do the CCS with a tube due to the complexity/cost factor. More heater supplies, negative HV rail, another hole in the chassis... However, if you have the interest/money/time it would probably be fun and challenging and it would probably work well.
Just my
Hello Johan
I must have misunderstood the sentence and meaning of your reply - For that I apologize!
As for your notes, I understand that the LTP works well without a CCS tale. You are right about the circuit becomes more complex with the mentioned CCS. We have to test the circuit.
Have been thinking about the o/p transf. and the difference in dc ohm between +B center and the plate windings, as the length of the inner and outer windings are different. There could be ac balance in the o/p T. , but what about the difference in ohm? Does this effect distortion? Some bifilar o/p's may have the same physical copper length, but what about standard o/p's? What about the small unbalance in the LTP (without CCS)- how does this effect the o/p output balance dc/ac?
If this is too comprehending for the thread , then please let me know?
Best regards
Kim
Johan Potgieter said:
I must have misunderstood the sentence and meaning of your reply - For that I apologize!
As for your notes, I understand that the LTP works well without a CCS tale. You are right about the circuit becomes more complex with the mentioned CCS. We have to test the circuit.
Have been thinking about the o/p transf. and the difference in dc ohm between +B center and the plate windings, as the length of the inner and outer windings are different. There could be ac balance in the o/p T. , but what about the difference in ohm? Does this effect distortion? Some bifilar o/p's may have the same physical copper length, but what about standard o/p's? What about the small unbalance in the LTP (without CCS)- how does this effect the o/p output balance dc/ac?
If this is too comprehending for the thread , then please let me know?
Best regards
Kim
Hi SpreadSpectrum
Have been thinking the same , but we do not like contaminated silicium - )). I would like to find out whether there is a sound difference, with/without the CCS!
Best rgds
Kim
SpreadSpectrum said:
I probably wouldn't do the CCS with a tube due to the complexity/cost factor. More heater supplies, negative HV rail, another hole in the chassis... However, if you have the interest/money/time it would probably be fun and challenging and it would probably work well.
Just my
Have been thinking the same , but we do not like contaminated silicium - )). I would like to find out whether there is a sound difference, with/without the CCS!
Best rgds
Kim
Re: Hello Johan
Kimjul2005,
No fine! I will limit my funny Afrikaans ways in future!
Regarding your questions:
No, the (usually small) difference in winding resistance is not serious despite some proclamations to the opposite. The total winding resistance is usually quite small compared to the tube plate resistance, even with triodes. The impedance is far more important.
Some transformers do compensate here by the way they place several primary sections, which is actually done more to keep leakage reactance between windings small. (You will be aware that for good h.f. response it is necessary to "interleave" primary and secondary windings, usually to the extent of at least four secondary sections placed in between primary sections in a certain way. In that way one can also cross-couple primary sections.)
In my experience, as said, with good choice of resistors in the LTP that should have negligible effect compared to the normal tolerances in output tubes and even good output transformers. One easily (er - well ...) gets quite low distortion in tube circuits with moderate amounts of NFB. (The distortion in the early Williamson and Quad II was inaudible without extravagant design gymnastics.)
kimjul2005 said:
Kimjul2005,
No fine! I will limit my funny Afrikaans ways in future!
Regarding your questions:
No, the (usually small) difference in winding resistance is not serious despite some proclamations to the opposite. The total winding resistance is usually quite small compared to the tube plate resistance, even with triodes. The impedance is far more important.
Some transformers do compensate here by the way they place several primary sections, which is actually done more to keep leakage reactance between windings small. (You will be aware that for good h.f. response it is necessary to "interleave" primary and secondary windings, usually to the extent of at least four secondary sections placed in between primary sections in a certain way. In that way one can also cross-couple primary sections.)
In my experience, as said, with good choice of resistors in the LTP that should have negligible effect compared to the normal tolerances in output tubes and even good output transformers. One easily (er - well ...) gets quite low distortion in tube circuits with moderate amounts of NFB. (The distortion in the early Williamson and Quad II was inaudible without extravagant design gymnastics.)
Ah, just missed Brandersnatch (he's off-line now).
This is off-topic, with permission from Kimjul2005.
I am not sure I understand Brander's remark. Yes, the Quad II had a cathode tertiary winding. This was in fact an adapted UL arrangement, quite brilliantly done by Peter Walker.
[For those not familiar, looking at UL it is clear that full cathode current flows in the [g2 - B+] part of the primary. Thus putting that on the cathode side still preserves UL advantages, but tend to lower the output stage distortion (where it is most) at the cost of extra input drive. Quad II did not do this fully, using only some 12% of the primary instead of the better 20% - 25% option. (But the latter requires quite a large g1 input swing.) I use that topology myself, as there is no better tube output configuration. The OPT is a bother; fortunately I can have my own wound locally.]
As far as I know this was first done by Walker as stated and not W, but he and Williamson wrote an analysis of this in Wireless World Sept. 1952, replying to the Hafler-Keroes claim that they 'invented' UL (the term is a misnomer). That was in fact first suggested by the brilliant Alan Blumlein of the UK, as early as 1928. Perhaps Brandersnatch has further detail - before both our times, I would think!
(This was dicussed in more detail some time ago under a Quad thread; I don't want to take too much space on that here.)
Thanks Kimyul.
This is off-topic, with permission from Kimjul2005.
I am not sure I understand Brander's remark. Yes, the Quad II had a cathode tertiary winding. This was in fact an adapted UL arrangement, quite brilliantly done by Peter Walker.
[For those not familiar, looking at UL it is clear that full cathode current flows in the [g2 - B+] part of the primary. Thus putting that on the cathode side still preserves UL advantages, but tend to lower the output stage distortion (where it is most) at the cost of extra input drive. Quad II did not do this fully, using only some 12% of the primary instead of the better 20% - 25% option. (But the latter requires quite a large g1 input swing.) I use that topology myself, as there is no better tube output configuration. The OPT is a bother; fortunately I can have my own wound locally.]
As far as I know this was first done by Walker as stated and not W, but he and Williamson wrote an analysis of this in Wireless World Sept. 1952, replying to the Hafler-Keroes claim that they 'invented' UL (the term is a misnomer). That was in fact first suggested by the brilliant Alan Blumlein of the UK, as early as 1928. Perhaps Brandersnatch has further detail - before both our times, I would think!
(This was dicussed in more detail some time ago under a Quad thread; I don't want to take too much space on that here.)
Thanks Kimyul.
It will be decent to ask Kimjul2005 whether I may continue here - not sure it merits a new thread yet, but I do not want to hi-jack.
Hi Douglas,
Firstly, this is not an USA-RSA tiff; I had very nice treatment in the US of A on several occasions!
But I am not sure what you are referring to. What I detected in the mentioned article (which is the only Walker-Williamson-UL response I remember), was that a claim that UL was new and in fact literally ultra-linear, was neither, with all respect. Not going into any possible sarcasm or such in the article - I cannot recall - the scientific fact is that UL combines the 'best of' pentode operation with the 'best of' triode operation. That is, almost pentode efficiency and output capability (in fact exactly the same for KT88), with almost triode low rp and thus tolerance of load impedance - also mainly 2nd harmonic distortion, unlike pentode. There is not significant increase in linearity (according to the literal 'UL'), but that is not the point.
I would be most surprised if either of the 'W's contradicted this; I would like to know where they had such a slip! As said, Walker seemed to have gained his knowledge from the Blumlein, plus then using the cathode winding etc. Again long ago, but in correspondence with him in the 50s he in fact testified to the advantages of distributed load as it was called in the UK, plus giving me some additional information. (I was a student then, and as students will do I asked all sorts of questions.)
To repeat, this is not an argument, just looking for additional info (guess never quitted being inquisitive; after all I later earned my salary doing research!)
Regards
Johan
Hi Douglas,
Firstly, this is not an USA-RSA tiff; I had very nice treatment in the US of A on several occasions!
But I am not sure what you are referring to. What I detected in the mentioned article (which is the only Walker-Williamson-UL response I remember), was that a claim that UL was new and in fact literally ultra-linear, was neither, with all respect. Not going into any possible sarcasm or such in the article - I cannot recall - the scientific fact is that UL combines the 'best of' pentode operation with the 'best of' triode operation. That is, almost pentode efficiency and output capability (in fact exactly the same for KT88), with almost triode low rp and thus tolerance of load impedance - also mainly 2nd harmonic distortion, unlike pentode. There is not significant increase in linearity (according to the literal 'UL'), but that is not the point.
I would be most surprised if either of the 'W's contradicted this; I would like to know where they had such a slip! As said, Walker seemed to have gained his knowledge from the Blumlein, plus then using the cathode winding etc. Again long ago, but in correspondence with him in the 50s he in fact testified to the advantages of distributed load as it was called in the UK, plus giving me some additional information. (I was a student then, and as students will do I asked all sorts of questions.)
To repeat, this is not an argument, just looking for additional info (guess never quitted being inquisitive; after all I later earned my salary doing research!)
Regards
Johan
Hi Johan
Hello Johan,
Just continue,
I try to pick up every piece an put the puzzle together. I have quite another access to the audio field, but math and detalis inside tubeamplification ( not so far away from the three teoretical eye receptors and their electrical way to the optical center in the back of the brain) could bring the understanding to a higher level. I need this to get my brain relaxed.
Go on!
Best rgds
Kim
This was really off the thread. Sorry!
Johan Potgieter said:
Hello Johan,
Just continue,
I try to pick up every piece an put the puzzle together. I have quite another access to the audio field, but math and detalis inside tubeamplification ( not so far away from the three teoretical eye receptors and their electrical way to the optical center in the back of the brain) could bring the understanding to a higher level. I need this to get my brain relaxed.
Go on!
Best rgds
Kim
This was really off the thread. Sorry!
Thanks then, Kimjul.
7n7is,
Thanks for that link. Most informative; wish one can get more of those. The problem now being: Those of us out of the main stream, and perhaps even those in it:
WHAT THE H*** WILL ONE REALLY BE GETTING WHEN ORDERING UNSEEN??
(I disagree here and there, e.g. there is quite a difference between 8417 and 6550; EL34 came from EL37 according to Mullard; but nevertheless.) This is a link every tube user should read. Especially the many photos are most illustrative.
7n7is,
Thanks for that link. Most informative; wish one can get more of those. The problem now being: Those of us out of the main stream, and perhaps even those in it:
WHAT THE H*** WILL ONE REALLY BE GETTING WHEN ORDERING UNSEEN??
(I disagree here and there, e.g. there is quite a difference between 8417 and 6550; EL34 came from EL37 according to Mullard; but nevertheless.) This is a link every tube user should read. Especially the many photos are most illustrative.
A few questions to 6550 PP?
Hi RichW.
Some time ago you explained several details to me from a 6550a PP amp. Ever since you suggested the CCS for the drivertube, I have been thinking a lot about this. The tube poweramp hav'nt been made yet, and the reason is my preamp project with 9 times 6DJ8 ( straight to approx. 1 mHz) - including riaa.
leakstereo20
(On the same site the old 6550a project.)
As soon as the preamp is playing , the work of a 6550a PP amp starts. I have been checking several PP circuits( also Jolida's 6550), and you have a lot of experience with triodes vs. penthodes.
Please check the attached circuit and tell me whether you prefer the STA100 with two penthodes to drive the output tubes or the 12AT7? Untill I read your statements my feeling to get the "best" sonic sound came from triodes!
Second: By many good sounding PP tubeamps the capasitor between the input - and drivertube has been deleted or strapped. Is the "deal" to reduce the amount of cap's to a minimum ( say only one cap between driver and output tube) in order to reach max. clear sound and no other caps in the signal way?
If you prefer triodes: is the 6SN7gt octal tube better sounding than 12AU7/12AT7 noval?
I will try to go for B+ 540 - 550Vdc UL/PP and use Tungsol 6550a reissue , and with the attached circuit as first attempt?
In the attached circuit I like the "double" inputtube because it looks symmetrical to me. I like symmetrical designs. May I have your opinion on this issue as well?
Please notice the changes on the circuit taken from several japanese renoval amplifiers. Only when the amplifier is steady and plays, the CCS shall be connected to the driver. The idea is to use a fet-regulated psu for the pretubes in the poweramp (as the fet-regulators used for the preamp).
Hobe to hear from you.
Regards
Kim
Hi RichW.
Some time ago you explained several details to me from a 6550a PP amp. Ever since you suggested the CCS for the drivertube, I have been thinking a lot about this. The tube poweramp hav'nt been made yet, and the reason is my preamp project with 9 times 6DJ8 ( straight to approx. 1 mHz) - including riaa.
leakstereo20
(On the same site the old 6550a project.)
As soon as the preamp is playing , the work of a 6550a PP amp starts. I have been checking several PP circuits( also Jolida's 6550), and you have a lot of experience with triodes vs. penthodes.
Please check the attached circuit and tell me whether you prefer the STA100 with two penthodes to drive the output tubes or the 12AT7? Untill I read your statements my feeling to get the "best" sonic sound came from triodes!
Second: By many good sounding PP tubeamps the capasitor between the input - and drivertube has been deleted or strapped. Is the "deal" to reduce the amount of cap's to a minimum ( say only one cap between driver and output tube) in order to reach max. clear sound and no other caps in the signal way?
If you prefer triodes: is the 6SN7gt octal tube better sounding than 12AU7/12AT7 noval?
I will try to go for B+ 540 - 550Vdc UL/PP and use Tungsol 6550a reissue , and with the attached circuit as first attempt?
In the attached circuit I like the "double" inputtube because it looks symmetrical to me. I like symmetrical designs. May I have your opinion on this issue as well?
Please notice the changes on the circuit taken from several japanese renoval amplifiers. Only when the amplifier is steady and plays, the CCS shall be connected to the driver. The idea is to use a fet-regulated psu for the pretubes in the poweramp (as the fet-regulators used for the preamp).
Hobe to hear from you.
Regards
Kim
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.