Randy,
Right. It depends on location. Depends on resistor type. Depends on TSR (taste-system-room).
Different cooking recipes will respond to the sage herb differently, for example.
Like most things in high-end audio this is quite the controversial subject. Some people believe that resistors don't make a difference and others just look for low noise resistors and others believe any differences are subtle. Viva La Difference.
Here is a stream of consciousness take on a few resistor types -- from the point of view that sage can taste different in different recipes, but also that there is a particular sage taste which will usually come through. In my toying around with resistors, I've referenced against a powerful measurement tool I have access to, my live-in audio control subject, Miranda.
Tantalums are one of my favorites. They have a nice clear midrange, are dynamic, good bass, non-glarey shiny treble. In larger doses the midrange can start sounding a bit unnatural and the highs a touch closed in and too shiny/unnatural.
Rikens are warmish and laid back with decent treble. They can sound a bit fuzzy. Too many (and that may include just one) can be boring sounding in tube circuits, imo. Maybe a better match with solid state. Mixing the Rikens with metal films might be a good match in your tube amp.
Caddocks can be very good. I've heard them be a little bright and they can also sound "flat" with reduced soundstage depth. A good match with tants.
Metal films can sound overly bright in some positions but they have good detail, air, bass. I've also heard them where they aren't bright at all. Caddocks are metal film (or similar, I think?) but they don't sound like other metal films I've heard.
I'm not familiar enough with the characteristics of Mills wirewounds to say much. They replaced metal film resistors in my PCM63 Dac and I don't have any impulse to change them out. I also use them in a speaker crossover. Them seem balanced without glare. People seem to like them on tube anodes and power supplies. Success with them as with the others probably depends on, as you say, "where you put them".
I tried Vishay S102 nudes in my phono stage as load resistors. They sounded interesting with a kind of ethereal treble but overall a bit lightweight. Some people like them in serial signal applications like volume controls. I'd like to play around with them more, in very small doses, but they are expensive. I'm guessing a 1:3 proportion with Shinkohs might sound good. But some new recipes just taste bad.
I think mixing resistor types in a circuit is useful as it can balance colorations. Mixing for one circuit element is just an extension of that.
************
I've soldered in the parts of the Aikido PCB which control the lines for the 5687 vs. other 9 pin types since I'm going to stick with the 5687 in position two. I also soldered in the Mundorf ZN tinfoil coupling caps. Those changes got rid of 14 alligator clips.
Right. It depends on location. Depends on resistor type. Depends on TSR (taste-system-room).
Different cooking recipes will respond to the sage herb differently, for example.
Like most things in high-end audio this is quite the controversial subject. Some people believe that resistors don't make a difference and others just look for low noise resistors and others believe any differences are subtle. Viva La Difference.
Here is a stream of consciousness take on a few resistor types -- from the point of view that sage can taste different in different recipes, but also that there is a particular sage taste which will usually come through. In my toying around with resistors, I've referenced against a powerful measurement tool I have access to, my live-in audio control subject, Miranda.
Tantalums are one of my favorites. They have a nice clear midrange, are dynamic, good bass, non-glarey shiny treble. In larger doses the midrange can start sounding a bit unnatural and the highs a touch closed in and too shiny/unnatural.
Rikens are warmish and laid back with decent treble. They can sound a bit fuzzy. Too many (and that may include just one) can be boring sounding in tube circuits, imo. Maybe a better match with solid state. Mixing the Rikens with metal films might be a good match in your tube amp.
Caddocks can be very good. I've heard them be a little bright and they can also sound "flat" with reduced soundstage depth. A good match with tants.
Metal films can sound overly bright in some positions but they have good detail, air, bass. I've also heard them where they aren't bright at all. Caddocks are metal film (or similar, I think?) but they don't sound like other metal films I've heard.
I'm not familiar enough with the characteristics of Mills wirewounds to say much. They replaced metal film resistors in my PCM63 Dac and I don't have any impulse to change them out. I also use them in a speaker crossover. Them seem balanced without glare. People seem to like them on tube anodes and power supplies. Success with them as with the others probably depends on, as you say, "where you put them".
I tried Vishay S102 nudes in my phono stage as load resistors. They sounded interesting with a kind of ethereal treble but overall a bit lightweight. Some people like them in serial signal applications like volume controls. I'd like to play around with them more, in very small doses, but they are expensive. I'm guessing a 1:3 proportion with Shinkohs might sound good. But some new recipes just taste bad.
I think mixing resistor types in a circuit is useful as it can balance colorations. Mixing for one circuit element is just an extension of that.
************
I've soldered in the parts of the Aikido PCB which control the lines for the 5687 vs. other 9 pin types since I'm going to stick with the 5687 in position two. I also soldered in the Mundorf ZN tinfoil coupling caps. Those changes got rid of 14 alligator clips.
Thanks for the info Cappy.
I believe that resistors have sound, and different kinds make a difference.
I will probably keep my current plan, but as time/money allows, try some other things in the future. Tantalum resistors look pretty pricey, but I would like to try them once I get everything working right.
Randy
I believe that resistors have sound, and different kinds make a difference.
I will probably keep my current plan, but as time/money allows, try some other things in the future. Tantalum resistors look pretty pricey, but I would like to try them once I get everything working right.
Randy
Balanced Aikido for Buffalo DAC
Hey,
I'm in the process of assembling a pair of balanced Aikido boards and PSU to use with the Buffalo DAC.
Here is my PSU design, it leverages my existing stock of parts. I'd appreciate any comments.
The transformer is an Antek T150.
The rectifier diodes are UF1007.
Thanks.
Hey,
I'm in the process of assembling a pair of balanced Aikido boards and PSU to use with the Buffalo DAC.
Here is my PSU design, it leverages my existing stock of parts. I'd appreciate any comments.
The transformer is an Antek T150.
The rectifier diodes are UF1007.
Thanks.
Attachments
Hi Boudy,
Good news about the design: low ripple and pretty snappy response.
Bad news: "Lord of the Rings".
More bad news: Since each current tap has differing levels of filtering, with the I2 tap having an additional LC cell, each side of your balanced Aikido will sound different.
You are showing the initial response of your power supply in the first second or so. What you want to do is wait until the supply settles down -- in the first few seconds things are just coming up -- caps are getting charged, etc. The "reporting delay" box allows one to get out of the startup window.
The PSUD chart below shows the ringing of the I2 current tap with a 10% change in current demand. It should be a smooth response but it has a whole lot of ringing.
I'll post more later when I have more time, with some suggestions. You have a lot of parts to work with so I bet you will be able to come up with something good. Also, if you use alligator clips you will be able to move things around to see how changes effect the sound.
I didn't realize that PSUD II supported multiple current taps. That is useful, thanks for the demonstration of that cool feature.
Good news about the design: low ripple and pretty snappy response.
Bad news: "Lord of the Rings".
More bad news: Since each current tap has differing levels of filtering, with the I2 tap having an additional LC cell, each side of your balanced Aikido will sound different.
You are showing the initial response of your power supply in the first second or so. What you want to do is wait until the supply settles down -- in the first few seconds things are just coming up -- caps are getting charged, etc. The "reporting delay" box allows one to get out of the startup window.
The PSUD chart below shows the ringing of the I2 current tap with a 10% change in current demand. It should be a smooth response but it has a whole lot of ringing.
I'll post more later when I have more time, with some suggestions. You have a lot of parts to work with so I bet you will be able to come up with something good. Also, if you use alligator clips you will be able to move things around to see how changes effect the sound.
I didn't realize that PSUD II supported multiple current taps. That is useful, thanks for the demonstration of that cool feature.
Attachments
Thanks for the feedback
Hi Cappy,
The model doesn't accurately depict the planned implementation, the second LC stages would be in parallel, not in series. I don't know of a way to do a parallel model in PSUD.
Did you implement a step in the current source to generate that signal?
Nice psud file name! ;-)
Back to the drawing board...
-Eric
Hi Cappy,
The model doesn't accurately depict the planned implementation, the second LC stages would be in parallel, not in series. I don't know of a way to do a parallel model in PSUD.
Did you implement a step in the current source to generate that signal?
Nice psud file name! ;-)
Back to the drawing board...
-Eric
Yeah, I don't know how to do the parallel supply either. Probably better just to take out that last leg since it is not representative, and put in 100mA for the first current tap.
One problem with parallel legs and LCLC filters is that it becomes a much more complex system, more prone to ringing and resonances. And we don't have psud to help us. Particularly if lowish levels of series resistance is used for damping.
I decided to punt trying a parallel supply. Also, mono vs. stereo supplies is another controversial subject. Some people with a lot of experience believe stereo supplies actually sound better. And that it's better to have "stacked" supplies with a single stereo supply per stage instead of two mono supplies. I'm not suggesting a stacked supply and I don't think it would even work with the Aikido, but I think a pure stereo supply like i am using is worth trying. But with alligator clips you could try different approaches yourself and draw your own conclusions since you have a lot of parts.
Another thing you could try is make the last parallel legs RC cells. This would help with damping and give some isolation between the sides. I don't know how it would sound. So far I'm finding not using too much series resistance in the supply is better. But as usual, there are different schools of thought on this.
-- Bill
One problem with parallel legs and LCLC filters is that it becomes a much more complex system, more prone to ringing and resonances. And we don't have psud to help us. Particularly if lowish levels of series resistance is used for damping.
I decided to punt trying a parallel supply. Also, mono vs. stereo supplies is another controversial subject. Some people with a lot of experience believe stereo supplies actually sound better. And that it's better to have "stacked" supplies with a single stereo supply per stage instead of two mono supplies. I'm not suggesting a stacked supply and I don't think it would even work with the Aikido, but I think a pure stereo supply like i am using is worth trying. But with alligator clips you could try different approaches yourself and draw your own conclusions since you have a lot of parts.
Another thing you could try is make the last parallel legs RC cells. This would help with damping and give some isolation between the sides. I don't know how it would sound. So far I'm finding not using too much series resistance in the supply is better. But as usual, there are different schools of thought on this.
-- Bill
Actually I was proposing just a single supply for everything, not two, like I am doing it.
Lowering the first cap is moving in the right direction. In general you want your cap values to increase as you move from left to right and inductance values to decrease. And not too high a value for the first cap, imo.
I still see some mild ringing with 10 uF though. The B+ voltage is quite a bit lower too.
Lowering the first cap is moving in the right direction. In general you want your cap values to increase as you move from left to right and inductance values to decrease. And not too high a value for the first cap, imo.
I still see some mild ringing with 10 uF though. The B+ voltage is quite a bit lower too.
What do you think of this one?
A single supply for both channels.
There is 6.3v and 12.6v winding available that I can tie in to bump up the secondary voltage as necessary. I'm going to have to use a separate filament transformer(s) to provide the heater current for 8 tubes.
A single supply for both channels.
There is 6.3v and 12.6v winding available that I can tie in to bump up the secondary voltage as necessary. I'm going to have to use a separate filament transformer(s) to provide the heater current for 8 tubes.
Attachments
Your supply looks pretty good -- worth trying, imo.
The fairly large caps seem to subdue the ringing well. It has quite low ripple. The impulse response, while not having the Swenson profile, looks reasonably quick.
Some smart people around here like to see even a slower impulse -- the idea being that a supply that has more "inertia" so to speak is more likely to regulate a particular voltage. You could mess around with the last cap to change the impulse profile.
I've been messing around with choke input supplies so I can't really say how it might sound, though. Given your power transformer you are limited to capacitor input supplies. I don't have much experience building those.
Here is a slight variation on yours that uses your parts and has the quick impulse response with the little bump I've been looking for in my power supply experiments. It has a bigger first cap (47 uF) and a smaller second cap (47 uF). If you use alligator clips you can easily try both supplies to see what you prefer. It gives a few more volts B+ too.
The fairly large caps seem to subdue the ringing well. It has quite low ripple. The impulse response, while not having the Swenson profile, looks reasonably quick.
Some smart people around here like to see even a slower impulse -- the idea being that a supply that has more "inertia" so to speak is more likely to regulate a particular voltage. You could mess around with the last cap to change the impulse profile.
I've been messing around with choke input supplies so I can't really say how it might sound, though. Given your power transformer you are limited to capacitor input supplies. I don't have much experience building those.
Here is a slight variation on yours that uses your parts and has the quick impulse response with the little bump I've been looking for in my power supply experiments. It has a bigger first cap (47 uF) and a smaller second cap (47 uF). If you use alligator clips you can easily try both supplies to see what you prefer. It gives a few more volts B+ too.
Attachments
Here is another one to consider, a CLC. I have some gear with CLC supplies and know they can sound excellent. This particular one has the quick impulse profile with bump. Ripple rejection isn't nearly as good as yours. Still, I think this is worth trying. Simpler sometimes sounds better.
The 6.5 Henry choke is made up of your 5 and 1.5 henry chokes in series.
Putting a 100 uF cap in the first position lowers the ripple and maintains the impulse profile, so you could try that too.
The 6.5 Henry choke is made up of your 5 and 1.5 henry chokes in series.
Putting a 100 uF cap in the first position lowers the ripple and maintains the impulse profile, so you could try that too.
Attachments
I got a pair of Pyramid NOS 12 uF paper in oil capacitors on Ebay about 6 weeks ago. I wanted to compare a single one to the 15 uF Solen in the first cap position after the input choke.
First off, these things are ginormous. They make the 70 uF Obbligato film and oil caps look small.
Second, the sound improved nicely by replacing the Solen with the Pyramid. It was no content -- easily heard in about 10 seconds. So this new (vintage) capacitor becomes part of Configuration M. The upper midrange/treble is fizzy with the Solen, very noticeable on violins, and instrument timbre in general is not as realistic.
Another possibility for folks building choke input supplies who don't want to hunt down vintage capacitors is to try an inexpensive 15 uF ASC polypropylene and oil cap.
First off, these things are ginormous. They make the 70 uF Obbligato film and oil caps look small.
Second, the sound improved nicely by replacing the Solen with the Pyramid. It was no content -- easily heard in about 10 seconds. So this new (vintage) capacitor becomes part of Configuration M. The upper midrange/treble is fizzy with the Solen, very noticeable on violins, and instrument timbre in general is not as realistic.
Another possibility for folks building choke input supplies who don't want to hunt down vintage capacitors is to try an inexpensive 15 uF ASC polypropylene and oil cap.
An externally hosted image should be here but it was not working when we last tested it.
The D1v3 has 4 PCM63K's. Two of them go to the Pass D1 I/V + Jfet buffer - output is around 1.2V RMS.
The other two PCM63K's use a passive I/V of 5 ohms with a 1:10 step-up transformer, then the gain stage I posted (gain=22.)
With the switch I can move the output from tube to D1. I am still waiting on the step-up transormers from Cinemag. When they arrive next week I will post sonic impressions.
If you are wondering why I didn't use the Pass I/V and replace the JFet buffer with the tube, it is because the input impedance of the Pass D1 is 50 ohms. Using the passive + transformer should in theory give less distortion from the PCM63K. I chose 5 ohms because this keeps the ouput under 25 mV which has been reported as the max compliance voltage for PCM dac chips.
The other two PCM63K's use a passive I/V of 5 ohms with a 1:10 step-up transformer, then the gain stage I posted (gain=22.)
With the switch I can move the output from tube to D1. I am still waiting on the step-up transormers from Cinemag. When they arrive next week I will post sonic impressions.
If you are wondering why I didn't use the Pass I/V and replace the JFet buffer with the tube, it is because the input impedance of the Pass D1 is 50 ohms. Using the passive + transformer should in theory give less distortion from the PCM63K. I chose 5 ohms because this keeps the ouput under 25 mV which has been reported as the max compliance voltage for PCM dac chips.
I was playing around with PSUD a month or so ago and noticed I could really bulk up the last power supply cap without slowing or changing the impulse reponse much at all. So I added a 47 uF Obbligato film PSU cap to the 70 uF Obbligato film/oil.
In most ways this Configuration N sounded better -- more perceived extension on the frequency extremes. The highs were a bit ragged, though.
By the way, I got the idea to mix the film and film/oil caps by looking at my DIY Hifi Supply Lux 300b amp where Brian Cherry mixes the very same caps for the last power supply position. Thanks, Brian.
So I tried Configuration O, which was adding a 0.47 uF Vitamin Q paper-in-oil as a bypass on the 70 uF/47 uF combo. Nope, too dark, the sparkle is gone. Configuration P was trying the same thing but with two Vitamin Qs. There are spots on the Aikido boards for bypass caps, so I put them there, near the signal tubes. Nope, even way darker. Configuration Q was trying a single 0.47 uF K40Y Russian cap as bypass. Better, but still too dark. Configuration R was the pair of bypass caps on the Aikido boards. No to that, also.
Finally, I tried a K40Y as a replacement for the 0.47 uF Vitamin Q on the first position before the first choke and no bypass on the 70uF/47uF combo and this Configuration S gave me the sound I was looking for -- nice extended highs and clear and dynamic, non-vintage sounding bass and midrange. Also the tonal palette is wider with the 0.47 uF 630v K40y. I like this cap.
I also played around with Configuration T by raiding my phono stage and pulling out a 0.22 uF VCap teflon and tinfoil capacitor and using this as a bypass on the 70uF/47uF combo. I like this configuration, but so far I've decided I prefer the PSU without the bypass cap. Treble is perhaps a bit unnaturally smooth and there is a very slight darkening to the tone. I could see where a teflon/tinfoil bypass could really work well in some situations, though.
Seeing how well adding the film cap worked out, I tried adding another one for Configuration U. This was too bright and out of balance. So I added another 70uF Obbligato film/oil instead. Nada, too dark and damped sounding. Finally, I tried another 47uF/70uF, so 234 uF total as Configuration V. The tonal balance on this was similar to Configuration N, and at first I thought I might prefer this setup. Noise was down a bit, too, of course. However, listening over time it was obvious this configuration didn't have the life and quickness of the 117 uF combo.
So Configuration S is the winna so far and I've been listening to it for a while now. I'll probably stick with this setup for the unregulated supply. It will be interesting to compare it to a good regulated supply.
In most ways this Configuration N sounded better -- more perceived extension on the frequency extremes. The highs were a bit ragged, though.
By the way, I got the idea to mix the film and film/oil caps by looking at my DIY Hifi Supply Lux 300b amp where Brian Cherry mixes the very same caps for the last power supply position. Thanks, Brian.
So I tried Configuration O, which was adding a 0.47 uF Vitamin Q paper-in-oil as a bypass on the 70 uF/47 uF combo. Nope, too dark, the sparkle is gone. Configuration P was trying the same thing but with two Vitamin Qs. There are spots on the Aikido boards for bypass caps, so I put them there, near the signal tubes. Nope, even way darker. Configuration Q was trying a single 0.47 uF K40Y Russian cap as bypass. Better, but still too dark. Configuration R was the pair of bypass caps on the Aikido boards. No to that, also.
Finally, I tried a K40Y as a replacement for the 0.47 uF Vitamin Q on the first position before the first choke and no bypass on the 70uF/47uF combo and this Configuration S gave me the sound I was looking for -- nice extended highs and clear and dynamic, non-vintage sounding bass and midrange. Also the tonal palette is wider with the 0.47 uF 630v K40y. I like this cap.
I also played around with Configuration T by raiding my phono stage and pulling out a 0.22 uF VCap teflon and tinfoil capacitor and using this as a bypass on the 70uF/47uF combo. I like this configuration, but so far I've decided I prefer the PSU without the bypass cap. Treble is perhaps a bit unnaturally smooth and there is a very slight darkening to the tone. I could see where a teflon/tinfoil bypass could really work well in some situations, though.
Seeing how well adding the film cap worked out, I tried adding another one for Configuration U. This was too bright and out of balance. So I added another 70uF Obbligato film/oil instead. Nada, too dark and damped sounding. Finally, I tried another 47uF/70uF, so 234 uF total as Configuration V. The tonal balance on this was similar to Configuration N, and at first I thought I might prefer this setup. Noise was down a bit, too, of course. However, listening over time it was obvious this configuration didn't have the life and quickness of the 117 uF combo.
So Configuration S is the winna so far and I've been listening to it for a while now. I'll probably stick with this setup for the unregulated supply. It will be interesting to compare it to a good regulated supply.
An externally hosted image should be here but it was not working when we last tested it.
I have a few Hammond 193Ls in my part box, previously used in a now scrapped DIY power conditioner. These are 5 henry chokes with a DCR of 57 ohms, just like the 2H Hammonds that are currently being used.
With the additional capacitance now in the last cap position of the power supply, more inductance can be added without ringing. So I tried the 193L in place of the 157R as Configuration W. This is a pretty large choke, size wise and also by weight, weighing about 8 times more than the 2 Henry Hammond. It has end bells instead of being open frame.
This is one case where bigger is not better. The midrange got hollowed out and the bass was anemic.
Adding more capacitance to the last cap position as Configuration X helped but didn't completely solve the issue.
The possible good news here is that switching back and forth between the two chokes led me to try more capacitance in the last position for the 2 Henry choke in a few different configurations I didn't try before. I'm listening to one of those configurations now, and time will tell if it becomes my new unregulated reference.
With the additional capacitance now in the last cap position of the power supply, more inductance can be added without ringing. So I tried the 193L in place of the 157R as Configuration W. This is a pretty large choke, size wise and also by weight, weighing about 8 times more than the 2 Henry Hammond. It has end bells instead of being open frame.
This is one case where bigger is not better. The midrange got hollowed out and the bass was anemic.
Adding more capacitance to the last cap position as Configuration X helped but didn't completely solve the issue.
The possible good news here is that switching back and forth between the two chokes led me to try more capacitance in the last position for the 2 Henry choke in a few different configurations I didn't try before. I'm listening to one of those configurations now, and time will tell if it becomes my new unregulated reference.
An externally hosted image should be here but it was not working when we last tested it.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.