No, it's my impatience. I guess it takes a little longer for links to get in as opposed to the text of the post. I could see the attachments about 30 min after I could read the posts. No biggie.
I.E. 6 and 7 depending on which machine I'm at.
Matt
I.E. 6 and 7 depending on which machine I'm at.
Matt
pothound said:
Yes, you are missing the fact that filament resistance is not constant, it increases as the filaments heat up, which is why you get a higher voltage. The best (and probably only way) is to adjust the resistor in circuit starting with a higher value and lower filament voltage.
Unfortunately, the circuit itself does not lend itself well to use with 6922/ECC88/similar tubes because there are signifficant differences in filament current requirements between versions and manufacturers. Of course, all will be fine if you adjust the circuit for the actual tubes you intend to use, and remember never to power it on unless all 4 are installed.
Hi ilimzn,
I would highly recommend using 6922 types only. Using 6922EH has been very helpful. They are quiet and reasonably consistent. This is more than I can say for many other brands. National 6DJ8's are completely useless. Huge microphonics. I would expect most other real 6DJ8's to act similarly. There are many other numbers that also have plate curves that don't match the 6DJ8 / 6922. Stay with the intended tube types.
So Matt, to give you a mild hint, I use 6922EH for service replacement given any choice in the matter.
-Chris
Thanks for mentioning that. Another "feature" of this amp design.
I would highly recommend using 6922 types only. Using 6922EH has been very helpful. They are quiet and reasonably consistent. This is more than I can say for many other brands. National 6DJ8's are completely useless. Huge microphonics. I would expect most other real 6DJ8's to act similarly. There are many other numbers that also have plate curves that don't match the 6DJ8 / 6922. Stay with the intended tube types.
So Matt, to give you a mild hint, I use 6922EH for service replacement given any choice in the matter.
-Chris
Thanks Chris;
Not sure that a hint is required. I got lucky and the pre and power amps were just re-tubed by the previous owner (in London, Ontario) with EH tubes. Left to my own devices I might have picked them but I can't say for sure. Last time I needed 6922s was to replace some white box Philips (I think) JAN types that had shown themselves to be quite microphonic. I bought Sovteks. They have proven themselves to be very good tubes in the power amp input stage in which they were used (differential cascode). These have low hours at the moment because the amps have been in storage for about 7-8 years (new house, new child) and the living room in the new house wasn't big enough to handle mono-blocks hence my consolidation project mentioned in a previous post.
Referring to the quote, I would never power up a tubed amp without all of the tubes installed unless it had fully regulated supplies or capacitors rated to take the increased voltage at reduced loads although this could still overstress the filaments of the remaining tubes.
Like I said before, I'm not in a hurry to do anything to the SA100. I had listened to an SA100/1000 combo in the 90's and really liked them but didn't have $4000.00. I saw a set for sale on CAM and bought them (1095 CAD for the pair). I Didn't realize that the DIY/professional side of the industry regarded them so poorly in spite of some very good reviews. It wasn't until I got to this website and to a lesser degree on AudioKarma that I saw some fairly negative commets from past users and then of course yourself who has obviously spent a lot of time inside these things making things right. I certainly don't doubt your experiences and after looking at the schematic there are some odd things going there. It's not very elegant. More kind of a kluge. All of the capacitors for the rails are marginal. Certainly if you see multiples of the same failures on the service side, that's a sign of design issues (marginal component selection or susceptibility). To date I have had no issues. It's early days for me with this box and as someone on this forum said, it is in a well ventilated place and I don't leave it on when not in use. The SA1000 I do but not the 100.
As for potential mods, the Aikido looks fairly easy to implement but I'd want some time with a prototype and SPICE to see if the output stage of that circuit could drive 2 pairs of MOSFETs and if the stock Counterpoint transformer could deliver the current on the B+ rail. That was my reasoning for the initial post asking the question about this. Another possibility is the replacement of the output stage with 2SK1058/2SJ162s or some nice ONSEMI bipolars.
All of this has to wait until I have finished a speaker mod (cabinetry ugh!) and a re-build of my mono-block tube amps into a stereo unit (more cabinetry...).
Cheers
Matt
Not sure that a hint is required. I got lucky and the pre and power amps were just re-tubed by the previous owner (in London, Ontario) with EH tubes. Left to my own devices I might have picked them but I can't say for sure. Last time I needed 6922s was to replace some white box Philips (I think) JAN types that had shown themselves to be quite microphonic. I bought Sovteks. They have proven themselves to be very good tubes in the power amp input stage in which they were used (differential cascode). These have low hours at the moment because the amps have been in storage for about 7-8 years (new house, new child) and the living room in the new house wasn't big enough to handle mono-blocks hence my consolidation project mentioned in a previous post.
Referring to the quote, I would never power up a tubed amp without all of the tubes installed unless it had fully regulated supplies or capacitors rated to take the increased voltage at reduced loads although this could still overstress the filaments of the remaining tubes.
Like I said before, I'm not in a hurry to do anything to the SA100. I had listened to an SA100/1000 combo in the 90's and really liked them but didn't have $4000.00. I saw a set for sale on CAM and bought them (1095 CAD for the pair). I Didn't realize that the DIY/professional side of the industry regarded them so poorly in spite of some very good reviews. It wasn't until I got to this website and to a lesser degree on AudioKarma that I saw some fairly negative commets from past users and then of course yourself who has obviously spent a lot of time inside these things making things right. I certainly don't doubt your experiences and after looking at the schematic there are some odd things going there. It's not very elegant. More kind of a kluge. All of the capacitors for the rails are marginal. Certainly if you see multiples of the same failures on the service side, that's a sign of design issues (marginal component selection or susceptibility). To date I have had no issues. It's early days for me with this box and as someone on this forum said, it is in a well ventilated place and I don't leave it on when not in use. The SA1000 I do but not the 100.
As for potential mods, the Aikido looks fairly easy to implement but I'd want some time with a prototype and SPICE to see if the output stage of that circuit could drive 2 pairs of MOSFETs and if the stock Counterpoint transformer could deliver the current on the B+ rail. That was my reasoning for the initial post asking the question about this. Another possibility is the replacement of the output stage with 2SK1058/2SJ162s or some nice ONSEMI bipolars.
All of this has to wait until I have finished a speaker mod (cabinetry ugh!) and a re-build of my mono-block tube amps into a stereo unit (more cabinetry...).
Cheers
Matt
Hi Matt,
Most tube stages do not supply enough current to drive mosfets. That's but one issue, but some people like that sound.
These amps have some issues that will occur due to the component placement. Quite a number actually. This chassis can be rebuilt into something that would sound good. I'm not sure I would use an Aikido to drive the mosfets. It might be better than the existing circuit, but there are certainly better ways to go about this.
Ask yourself what the Aikido design is trying to solve and determine for yourself if this is normally a problem.
-Chris
Most tube stages do not supply enough current to drive mosfets. That's but one issue, but some people like that sound.
These amps have some issues that will occur due to the component placement. Quite a number actually. This chassis can be rebuilt into something that would sound good. I'm not sure I would use an Aikido to drive the mosfets. It might be better than the existing circuit, but there are certainly better ways to go about this.
Ask yourself what the Aikido design is trying to solve and determine for yourself if this is normally a problem.
-Chris
The Aikido would certainly drive the MOSFETs better than the current setup, also, the Hitachi lateral MOSFETs would likely prove to be easyer to drive and set up regarding biassing and paralleling, but I do agree with Chris, some extra work would push the basic setup a few notches forward with comparatively negligible extra investment.
One thig to think about with the Aikido topology would be signifficantly reduced gain compared to the original SA100. In the Aikido, the gain is basically mu/2 of the input tube pair. In other words, you would likely need to select a different input tube, whereuppon you get back into the problem of trimming the heater circuit to it. That in turn points toward the necesity of doing the heater circuit 'properly', perhaps with a regulator or wvwn better, regulators (like 1x LM317 per tube, cheap and distributes the heat evenly, alowes a vast variety of tubes to be used).
One thig to think about with the Aikido topology would be signifficantly reduced gain compared to the original SA100. In the Aikido, the gain is basically mu/2 of the input tube pair. In other words, you would likely need to select a different input tube, whereuppon you get back into the problem of trimming the heater circuit to it. That in turn points toward the necesity of doing the heater circuit 'properly', perhaps with a regulator or wvwn better, regulators (like 1x LM317 per tube, cheap and distributes the heat evenly, alowes a vast variety of tubes to be used).
Hi ilimzn,
My point is that the Aikido seems to be designed more to cancel supply ripple. This should not be a problem with most power supply designs at all. I would look at something different that focuses on reducing distortion on high amplitude signals to replace the existing circuits. Since there is no overall feedback in this design, there are no source resistors and so matching is simply ridiculous! Step one would be to make it more linear and include an overall feedback solution.
Many ways to go, but you are right in that "fixing" the existing errors will get you closer with less effort. One thing you will have right off. Far too much gain once fixed. I even had to pad down the input.
-Chris
No argument there!
My point is that the Aikido seems to be designed more to cancel supply ripple. This should not be a problem with most power supply designs at all. I would look at something different that focuses on reducing distortion on high amplitude signals to replace the existing circuits. Since there is no overall feedback in this design, there are no source resistors and so matching is simply ridiculous! Step one would be to make it more linear and include an overall feedback solution.
Many ways to go, but you are right in that "fixing" the existing errors will get you closer with less effort. One thing you will have right off. Far too much gain once fixed. I even had to pad down the input.
-Chris
- Status
- Not open for further replies.