Anyone who has rebuilt a Dynaco SCA35 and reformed this amplifier to a ST35?
What is the experience by use of individual bias adjustment for each EL84?
I have allready one Leak Stereo 20 modified with Kimber caps, and the question is : " will the Dynaco outputtransformers outperform the Leak's in the mid - and high frequence area"?
It is difficult to answer this question - I know - but try to look at my amateur free site called:
www.tubeamp.hobbysider.dk
Some in english - some in danish - some free circuits for tubeamplifiers , but everything free.
Don't hesitate to call me for any information concerning tubeamplifiers mentioned on this free site.
Best regards
Kim Hansen
Denmark
What is the experience by use of individual bias adjustment for each EL84?
I have allready one Leak Stereo 20 modified with Kimber caps, and the question is : " will the Dynaco outputtransformers outperform the Leak's in the mid - and high frequence area"?
It is difficult to answer this question - I know - but try to look at my amateur free site called:
www.tubeamp.hobbysider.dk
Some in english - some in danish - some free circuits for tubeamplifiers , but everything free.
Don't hesitate to call me for any information concerning tubeamplifiers mentioned on this free site.
Best regards
Kim Hansen
Denmark
Attachments
I'd do some reading here:
http://www.diytube.com/phpBB2/viewforum.php?f=1
Those guys have a fair amount of experience with the Z565 output transformers used in the Dynaco ST35 and SCA35. I did a quick search, but turned up nothing comparing the Leak 20 to the Dynaco ST35.
http://www.diytube.com/phpBB2/viewforum.php?f=1
Those guys have a fair amount of experience with the Z565 output transformers used in the Dynaco ST35 and SCA35. I did a quick search, but turned up nothing comparing the Leak 20 to the Dynaco ST35.
Kim,
Dyna's EL84 amps are cathode biased and there is no adjustment, whatsoever. AAMOF, all 4 "finals" share a single RC bias network. That forces the owner to buy tightly matched quads of EL84s. It's much easier for a tube vendor to cull out well matched pairs than it is to cull out good quads. You pay more per tube, when you buy well matched quads. Changing to a RC bias network in each channel is (IMO) a sensible thing to do.
The Z565 O/P trafo is arguably the best Dyna made. Still, I highly doubt that the "iron" Leak used is in any way less good.
I agree with the idea of eliminating the nasty signal shaping circuitry in the SCA35 and replacing the 7199s with 12DW7s. Current production JJ 12DW7s have a satisfactory reputation.
Replacing Dyna's phono preamp circuitry with something based on the RCA passively EQed design is another change to consider. In any event, rig up a 12 VDC supply for the phono preamp heaters.
Dyna's EL84 amps are cathode biased and there is no adjustment, whatsoever. AAMOF, all 4 "finals" share a single RC bias network. That forces the owner to buy tightly matched quads of EL84s. It's much easier for a tube vendor to cull out well matched pairs than it is to cull out good quads. You pay more per tube, when you buy well matched quads. Changing to a RC bias network in each channel is (IMO) a sensible thing to do.
The Z565 O/P trafo is arguably the best Dyna made. Still, I highly doubt that the "iron" Leak used is in any way less good.
I agree with the idea of eliminating the nasty signal shaping circuitry in the SCA35 and replacing the 7199s with 12DW7s. Current production JJ 12DW7s have a satisfactory reputation.
Replacing Dyna's phono preamp circuitry with something based on the RCA passively EQed design is another change to consider. In any event, rig up a 12 VDC supply for the phono preamp heaters.
As usual good post by Eli.
Yes, I cannot imagine what in blazes made Dynaco use common bias components for all 4 EL84s. What could the saving have been??
I have refurbished a few SCA35s and have immediately changed to at least separate bias networks for each channel. Separate bias for each power tube ... I know that topology has a following, but it introduces an extra time constant at l.f. and I have had more problems than solutions from that. Even with 4 non shop-matched EL84s, one usually comes close to 2 suitable pairs.
I found the Leak 20 output transformer to be inferior to the Dynaco - higher leakage reactance, lower inductance - although as said, much depends on the associated circuit. (I am sorry but cannot find my measured figures right now.) Perhaps I should not say, but the 4 Leak 20s I have so far refurbished were not satisfactory. (And all 8 amplifiers were similar.) Thus, as a great surprise (I grew up with Leaks since the TL12), this is a design shortcoming. I found square wave rise times very slow and rounded and had to change several NFB network values to get it on par. But that as an aside.
Regarding substituting the 7199: I am not in favour of a high-mu triode as a power amplifier input stage compared to a pentode (with respect, Eli). My problem is with the dependance of response (stability) on input impedance because of triode Miller capacitance. Because of local availability/economy I did change, but to an ECF80 (6BL8), with slight resistor adjustments.
But that Dyna bias arrangement!
Yes, I cannot imagine what in blazes made Dynaco use common bias components for all 4 EL84s. What could the saving have been??
I have refurbished a few SCA35s and have immediately changed to at least separate bias networks for each channel. Separate bias for each power tube ... I know that topology has a following, but it introduces an extra time constant at l.f. and I have had more problems than solutions from that. Even with 4 non shop-matched EL84s, one usually comes close to 2 suitable pairs.
I found the Leak 20 output transformer to be inferior to the Dynaco - higher leakage reactance, lower inductance - although as said, much depends on the associated circuit. (I am sorry but cannot find my measured figures right now.) Perhaps I should not say, but the 4 Leak 20s I have so far refurbished were not satisfactory. (And all 8 amplifiers were similar.) Thus, as a great surprise (I grew up with Leaks since the TL12), this is a design shortcoming. I found square wave rise times very slow and rounded and had to change several NFB network values to get it on par. But that as an aside.
Regarding substituting the 7199: I am not in favour of a high-mu triode as a power amplifier input stage compared to a pentode (with respect, Eli). My problem is with the dependance of response (stability) on input impedance because of triode Miller capacitance. Because of local availability/economy I did change, but to an ECF80 (6BL8), with slight resistor adjustments.
But that Dyna bias arrangement!
Dynaco ST35
Hi Eli,
Thanks for your reply and answer. I know that the original ST35 has the simple cathode bias for a pair of tubes with one resistor and capasitor . I was wondering whether a single adjustment pot used at each EL84 could keep a closer balance between a pair of unmatched tubes. I have 5 pcs. of nosTelefunken and 4 pcs. of nos E84L Siemens for the project. The all draw slightly different mA from the psu. I have looked at the redesign of the Dynaco ST35 , and they have used the mentioned single tube adjustment. May be I could try to make both setups before finishing the amplifier. The 12DW7 is hard to find nos , so in the first setup I have to use Telefunken ECC82 and ECC83 from my stock.
I have erlier made one tubeamp by use of 6SL7gt (RCA) input - E80CC (Philips rugged) and 4 Sovtek EL84 pr. channel. I used two pots for the negative bias and two balance pots pr. pair (two pots for 4 outputtubes)- after having the tubes burned in, the biasadjustment without signal has moved only a few mA during the last four months.
I really do not expect much better performance from the Dynaco Z565 compared to the outputtranformers made by Leak.
But ones again thanks for your answer.
And to Johan,
Yes - I bought the Dynaco just to check out the better midrange sound from Z565 , said by several local audio-people . Secondly I have only one time used a pentode (EF86) as first stage or input amplifier
- I think I like the sound of triode better!
Hi Ty Bower,
Thanks very much for the link to Tube.DIY - I shall have a closer look.
Thanks
sincerely yours
Kim Jul Hansen. Denmark
Hi Eli,
Thanks for your reply and answer. I know that the original ST35 has the simple cathode bias for a pair of tubes with one resistor and capasitor . I was wondering whether a single adjustment pot used at each EL84 could keep a closer balance between a pair of unmatched tubes. I have 5 pcs. of nosTelefunken and 4 pcs. of nos E84L Siemens for the project. The all draw slightly different mA from the psu. I have looked at the redesign of the Dynaco ST35 , and they have used the mentioned single tube adjustment. May be I could try to make both setups before finishing the amplifier. The 12DW7 is hard to find nos , so in the first setup I have to use Telefunken ECC82 and ECC83 from my stock.
I have erlier made one tubeamp by use of 6SL7gt (RCA) input - E80CC (Philips rugged) and 4 Sovtek EL84 pr. channel. I used two pots for the negative bias and two balance pots pr. pair (two pots for 4 outputtubes)- after having the tubes burned in, the biasadjustment without signal has moved only a few mA during the last four months.
I really do not expect much better performance from the Dynaco Z565 compared to the outputtranformers made by Leak.
But ones again thanks for your answer.
And to Johan,
Yes - I bought the Dynaco just to check out the better midrange sound from Z565 , said by several local audio-people . Secondly I have only one time used a pentode (EF86) as first stage or input amplifier
- I think I like the sound of triode better!
Hi Ty Bower,
Thanks very much for the link to Tube.DIY - I shall have a closer look.
Thanks
sincerely yours
Kim Jul Hansen. Denmark
Kim,
Johan's point about the Miller effect is valid. The 12AX7 section grid leak resistors should not be greater than 100 KOhms.
If you are inclined to experiment, try 12AT7s as the driver/splitters. This ostensibly VHF type doesn't have the unpleasant Miller effect issues and voltage stage gain should be quite satisfactory, as the lower RP compensates well for the lower mu. BTW, high gm types make superior "concertina" phase splitters and 'T7 gm is higher than that of either the 'X7 or the 'U7.
Johan's point about the Miller effect is valid. The 12AX7 section grid leak resistors should not be greater than 100 KOhms.
If you are inclined to experiment, try 12AT7s as the driver/splitters. This ostensibly VHF type doesn't have the unpleasant Miller effect issues and voltage stage gain should be quite satisfactory, as the lower RP compensates well for the lower mu. BTW, high gm types make superior "concertina" phase splitters and 'T7 gm is higher than that of either the 'X7 or the 'U7.
Grid resistors
Hi again Eli,
Ones before I have heard about the miller effect, but did not take any action as I thought, that this problem could not be heard. If you and Johan has experience by use of more than 100 kOhm as grid resistor , I need to try this!
I have several ECC81 or 12AT7 in stock, and I would like to make one original Dynaco setup and compare the 12DW7 with the X and T type.
Thanks for this hint.
Best rgds
Kim
Hi again Eli,
Ones before I have heard about the miller effect, but did not take any action as I thought, that this problem could not be heard. If you and Johan has experience by use of more than 100 kOhm as grid resistor , I need to try this!
I have several ECC81 or 12AT7 in stock, and I would like to make one original Dynaco setup and compare the 12DW7 with the X and T type.
Thanks for this hint.
Best rgds
Kim
Kim,
Not to belabour, but it is not so much that one will hear an immediate difference (and please pardon my scepticism about some reports regarding differences heard when "rolling" tubes).
The point is that a h.f. roll-off (pole) sometimes down into the audio range and dependant on variable input conditions (what source will it be fed from?) compromises NFB design, in that one is forced to accept this roll-off. If it could be higher up in frequency one will have the first pole under one's control (an R.C somewhere), to have it where it is best for the stability of the particular design. Although a low source impedance is usually the case (Rs < 10K), it might]/I] not be; anything is possible these days (especially with some so-called passive pre-amps). This is also the danger when the input is fed from a volume control, often 0,5 - 1 megohm. With the slider in the middle one has an input impedance of several 100K, varying all the time.
The difference is best shown with square wave input and variable series resistance.
Not to belabour, but it is not so much that one will hear an immediate difference (and please pardon my scepticism about some reports regarding differences heard when "rolling" tubes).
The point is that a h.f. roll-off (pole) sometimes down into the audio range and dependant on variable input conditions (what source will it be fed from?) compromises NFB design, in that one is forced to accept this roll-off. If it could be higher up in frequency one will have the first pole under one's control (an R.C somewhere), to have it where it is best for the stability of the particular design. Although a low source impedance is usually the case (Rs < 10K), it might]/I] not be; anything is possible these days (especially with some so-called passive pre-amps). This is also the danger when the input is fed from a volume control, often 0,5 - 1 megohm. With the slider in the middle one has an input impedance of several 100K, varying all the time.
The difference is best shown with square wave input and variable series resistance.
Pole in Tubeamp
Hi Johan,
As my professional field is optic's - contactlenses , eye measuring a.o. - a good friend of mine is audio engineer, and he has informed me of poles in audio amplifiers. I understand some of the theory in tubeamplifiers but not all, but by use of several preamps with different outputimpedances varying from 50 Ohm up to 1 kOhm it became clear, that a low impedance source is one of the most important factors in tube audio design, also corcerning the inner amplification stages of a tubeamp. I have to learn more abount the groundtheory of tubes before I get the full picture.
The copy of the Dynaco Pas preamp I made( output impedance of approximately 1 K), was depending of a high input impedance of the following tube power amplifier - If not I could hear the loss of bass reproduction from about 60 Hz - also messured. I should have made this preamp with a cathode-follower. Next time then....
With my Luxman C03 (output impedance less than 100Ohm) it was difficult to detect any loss at all. We are about to design a preamp with 9 ECC88/6DJ8 - up untill now the frequence responce from the linestage goes beyond 1 GHz - But does it sound good??? I have heard many pre - and poweramp with fine data, but without the sound we all want - wide plasma stereo perspective ,definition of details and "thump" attack in the bass region.
And Johan,
It took me some years before my audio engineer accepted the fact that not everything can be measured. For the moment he works with a russian triode, GM80 I think he calls it. This single triode delivers around 40 Watt RMS with anodevoltage of 1.000 Volt . It is good to have some scepticism , because when we comes to the main issue, and this is not fine or good specs, it is: How does the amplifier sound? In my case how does the tubeamp sound on my AR3a improved speakers?
Best regards
Kim
Hi Johan,
As my professional field is optic's - contactlenses , eye measuring a.o. - a good friend of mine is audio engineer, and he has informed me of poles in audio amplifiers. I understand some of the theory in tubeamplifiers but not all, but by use of several preamps with different outputimpedances varying from 50 Ohm up to 1 kOhm it became clear, that a low impedance source is one of the most important factors in tube audio design, also corcerning the inner amplification stages of a tubeamp. I have to learn more abount the groundtheory of tubes before I get the full picture.
The copy of the Dynaco Pas preamp I made( output impedance of approximately 1 K), was depending of a high input impedance of the following tube power amplifier - If not I could hear the loss of bass reproduction from about 60 Hz - also messured. I should have made this preamp with a cathode-follower. Next time then....
With my Luxman C03 (output impedance less than 100Ohm) it was difficult to detect any loss at all. We are about to design a preamp with 9 ECC88/6DJ8 - up untill now the frequence responce from the linestage goes beyond 1 GHz - But does it sound good??? I have heard many pre - and poweramp with fine data, but without the sound we all want - wide plasma stereo perspective ,definition of details and "thump" attack in the bass region.
And Johan,
It took me some years before my audio engineer accepted the fact that not everything can be measured. For the moment he works with a russian triode, GM80 I think he calls it. This single triode delivers around 40 Watt RMS with anodevoltage of 1.000 Volt . It is good to have some scepticism , because when we comes to the main issue, and this is not fine or good specs, it is: How does the amplifier sound? In my case how does the tubeamp sound on my AR3a improved speakers?
Best regards
Kim
Thanks Kim,
So you can see right through me!
The theory is not so difficult if you are interested. In the beginning certain things took me a while to grasp; now I cannot understand what my problem was. I believe optics are more involved than electronics - well, audio, at least.
Yes, you are correct in the low-impedance feed question. It mostly is so, but then you will not imagine what sometimes happen with so-called passive pre-amps (a contradiction in terms, those things are control units only). They can have uncomfortably high output impedances.
About measurements, I agree that the end goal is the enjoyment of music. But for us engineers technology is important to get there. (No, I do not sit in front of an oscilloscope all day, as the reputation goes - but I have to do that in order to get there!)
Point is that sometimes, in a way, knowledge of electronics "gets in the way" here, if you understand. In pro audio it is generally known and understood what is measurable, and it is often simply lack of knowledge of this that causes the comment "not everything can be measured". Where it gets irritating is when someone without the necessary background tries to convince one that, in a manner of speaking, 3+6=11. I am not trying to start anything here, but I am sure a similar situation exists in optics (well, all other fields). My experience of your field (apart from the specs that I wear) goes up to photography; wonderful measurements and design techniques have opened up there especially with CAD. For me, when I just think of the design of a zoom lens ... brother! Etc. I hope never to give the impression that one knows all - but one does know certain things, as you do in your field.
But apology for having gone off-topic.
Regards
So you can see right through me!
The theory is not so difficult if you are interested. In the beginning certain things took me a while to grasp; now I cannot understand what my problem was. I believe optics are more involved than electronics - well, audio, at least.
Yes, you are correct in the low-impedance feed question. It mostly is so, but then you will not imagine what sometimes happen with so-called passive pre-amps (a contradiction in terms, those things are control units only). They can have uncomfortably high output impedances.
About measurements, I agree that the end goal is the enjoyment of music. But for us engineers technology is important to get there. (No, I do not sit in front of an oscilloscope all day, as the reputation goes - but I have to do that in order to get there!)
Point is that sometimes, in a way, knowledge of electronics "gets in the way" here, if you understand. In pro audio it is generally known and understood what is measurable, and it is often simply lack of knowledge of this that causes the comment "not everything can be measured". Where it gets irritating is when someone without the necessary background tries to convince one that, in a manner of speaking, 3+6=11. I am not trying to start anything here, but I am sure a similar situation exists in optics (well, all other fields). My experience of your field (apart from the specs that I wear) goes up to photography; wonderful measurements and design techniques have opened up there especially with CAD. For me, when I just think of the design of a zoom lens ... brother! Etc. I hope never to give the impression that one knows all - but one does know certain things, as you do in your field.
But apology for having gone off-topic.
Regards
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.