My PSET 300B amp currently use electrolytics to shunt AC in the cathode bias circuits (both for the driver and 300Bs). Have folks experimented with using film capacitors of similar value in this application, or adding a small film cap in parallel with the electrolytic? What affect does this have on the sound?
For a listening test of electrolytic versus film caps:
Starting . . .
1. With electrolytic caps of good quality,
2. With film capacitors of good quality,
A. With a "double blindfold" test method, Statistically there will be No sound difference.
B. Without a "double blindfold" test method, Statistically there Will be a sound difference.
Take your pick of A. or B.
When making changes to improve the sound, always remember to peel the outer layers off of the onion First; Before you attack the inner layers.
Just my opinions (there will be others).
Starting . . .
1. With electrolytic caps of good quality,
2. With film capacitors of good quality,
A. With a "double blindfold" test method, Statistically there will be No sound difference.
B. Without a "double blindfold" test method, Statistically there Will be a sound difference.
Take your pick of A. or B.
When making changes to improve the sound, always remember to peel the outer layers off of the onion First; Before you attack the inner layers.
Just my opinions (there will be others).
Try it and see? It isn't hard to accomplish.
There are so many things you could do other than bypassing a capacitor to affect the sound of a device. Of course you ought always to be looking for a reasonably good or optimum part for any application, but I think that capacitors tend to be an obsession of those who don't have much technical understanding of their amp/DAC/device in question. It's an easy part to swap in and out and there are plenty of people who will charge you dearly for it.
I speak from experience. I used to be like that. Started out building Bottlehead kits. People are absolutely obsessed with them (and tube rolling) over on the bottlehead forums. Now with a few more years under my belt, a huge number of mistakes made, I feel like urging you to pursue other paths or tweaks. Try building your 300B amp all over again but with a different topology, for example. Or using all "budget" parts. You'll probably be amazed at how little difference it makes.
That said, I found this video series to be pretty interesting. He demonstrates that a small capacitor bypassing a larger one can be helpful--but it can also be problematic. (Not exactly sure which video in the series he covers that.)
It's a little bit like the discussion about whether or not your PSU needs a snubber circuit. There is no "one size fits all" approach--to determine WHAT difference and HOW MUCH difference a bypass capacitor would make, you really do need an LTSpice simulation or an oscilloscope--or both!
Also, I hate to be "that guy," but this is a topic that has been done to death on this forum and elsewhere.
There are so many things you could do other than bypassing a capacitor to affect the sound of a device. Of course you ought always to be looking for a reasonably good or optimum part for any application, but I think that capacitors tend to be an obsession of those who don't have much technical understanding of their amp/DAC/device in question. It's an easy part to swap in and out and there are plenty of people who will charge you dearly for it.
I speak from experience. I used to be like that. Started out building Bottlehead kits. People are absolutely obsessed with them (and tube rolling) over on the bottlehead forums. Now with a few more years under my belt, a huge number of mistakes made, I feel like urging you to pursue other paths or tweaks. Try building your 300B amp all over again but with a different topology, for example. Or using all "budget" parts. You'll probably be amazed at how little difference it makes.
That said, I found this video series to be pretty interesting. He demonstrates that a small capacitor bypassing a larger one can be helpful--but it can also be problematic. (Not exactly sure which video in the series he covers that.)
It's a little bit like the discussion about whether or not your PSU needs a snubber circuit. There is no "one size fits all" approach--to determine WHAT difference and HOW MUCH difference a bypass capacitor would make, you really do need an LTSpice simulation or an oscilloscope--or both!
Also, I hate to be "that guy," but this is a topic that has been done to death on this forum and elsewhere.
I did build these amps myself, although used the basic design and transformers from the ANK C-Core Interstage Monoblocks kit. The main changes I made were in the power supply - using a regulated DC supply for the driver filament and Tentlabs DHT filament supplies for the 300Bs and a few other tweaks. Overall, I'm quite happy with the sound, but was wondering if there was any room for improvement by playing with these caps.
Yes, it's a very easy change, but I'd need to order some moderately expensive caps if I wanted to swap in film caps for the electrolytics which is why I was interested to hear from anyone that has done this experiment already.
I've found very noticeable differences in the sound of caps in my speaker crossovers and for AC coupling applications, even between different film caps, so I don't believe that all caps of the same nominal value will sound the same in all situations. But I don't have any experience with different caps in this application which is why I posed the question.
I'm sure I will get around to building more tube amps with different topologies at some point soon - I have a couple preamp projects to finish first - but this seemed like it would be an easy "upgrade" if others that have tried it found it made a positive difference.
Yes, it's a very easy change, but I'd need to order some moderately expensive caps if I wanted to swap in film caps for the electrolytics which is why I was interested to hear from anyone that has done this experiment already.
I've found very noticeable differences in the sound of caps in my speaker crossovers and for AC coupling applications, even between different film caps, so I don't believe that all caps of the same nominal value will sound the same in all situations. But I don't have any experience with different caps in this application which is why I posed the question.
I'm sure I will get around to building more tube amps with different topologies at some point soon - I have a couple preamp projects to finish first - but this seemed like it would be an easy "upgrade" if others that have tried it found it made a positive difference.
Have you considered trying TentLabs fixed bias module? Comparing that to cathode bias could be interesting. And you may approach the cost and space required to replace existing electrolytics with film caps.
@Thekak - yes, I have looked at that, but switching to fixed bias would require I also change the B+ supply to reduce the voltage unless I'm missing something.
I will probably use fixed bias in my next build though, but I'm thinking of a single tube SET (probably 300B again) with manually set bias. My speakers are pretty efficient (98db/w) and are augmented by separately powered woofer towers covering up to about 180Hz, so even a few watts is probably enough. I am loving the SET sound.
I will probably use fixed bias in my next build though, but I'm thinking of a single tube SET (probably 300B again) with manually set bias. My speakers are pretty efficient (98db/w) and are augmented by separately powered woofer towers covering up to about 180Hz, so even a few watts is probably enough. I am loving the SET sound.
You could always use a small bypass film cap in parallel with the electrolytic. Use a quality film and foil if you want to spend the money. Less than 1uf is usually what is used. It makes a very small difference if at all. The bypass caps are the only electrolytics in my amps. I do film PS caps only. You can go crazy and do a polypropylene bypass and bypass it with a polystyrene across the electrolytic. I’ve heard that you should not use values that are multiples of each other to avoid resonances. So dont use a 1uf and a .1uf across a 100uf.
I keep wanting to believe that there is no difference in sound between different capacitors as long as they're the same value of C and they're good quality. But I keep being proven wrong. I had a phono preamp I'd put together using parts I had in my drawer. I had a pair of good quality polystyrene 0.01uF caps in the RIAA filter. A friend challenged me to switch that out for one of those Jupiter caps (some kind of combination of film types with foil). We made sure the value of C was very close, so that wouldn't be the main difference. Dang it. The Jupiter cap clearly sounded better.
Since a 300B SET is by definition an experiment in subjectivity when it comes to audio amplifier design, I'd say why not use it as a test bed for trying out different types of capacitors, and see if you really do hear a difference? That said, I've never heard a noticeable difference between any two good quality electrolytic caps used for cathode bypass. I've also never been able to live long-term with the sound of something like a 0.01uF or 0.1uF bypass cap across a larger cap. Either I didn't hear any difference (in cathode bypass) or it made the sound too bright (bypassing interstage coupling caps).
Perhaps the best fun with capacitors would be to try to replace the electrolytic caps in the power supply with film types. That's an expensive experiment, though, especially for values of C above 20uF and for voltages higher than 400VDC.
Have fun with it.
Since a 300B SET is by definition an experiment in subjectivity when it comes to audio amplifier design, I'd say why not use it as a test bed for trying out different types of capacitors, and see if you really do hear a difference? That said, I've never heard a noticeable difference between any two good quality electrolytic caps used for cathode bypass. I've also never been able to live long-term with the sound of something like a 0.01uF or 0.1uF bypass cap across a larger cap. Either I didn't hear any difference (in cathode bypass) or it made the sound too bright (bypassing interstage coupling caps).
Perhaps the best fun with capacitors would be to try to replace the electrolytic caps in the power supply with film types. That's an expensive experiment, though, especially for values of C above 20uF and for voltages higher than 400VDC.
Have fun with it.
Yeah, I may just give it a go to see if it makes any difference.
I am using electrolytics in my power supply, although at least they are nice quality Mundorfs. Switching these for something like Clarity TCs might be worth considering as well although I may wait for my next build to try these.
Thanks for the feedback.
I am using electrolytics in my power supply, although at least they are nice quality Mundorfs. Switching these for something like Clarity TCs might be worth considering as well although I may wait for my next build to try these.
Thanks for the feedback.
I have done capacitor comparisons for output stages. In fact, on the ANK 2.1 DAC. Came with the copper Audio Note caps and I swapped in some Clarity caps that I had. The Clarity caps were much cleaner and, to be honest, I preferred the warmer sound of the AN. But I would say it made a very tiny change and even though it was easy to swap, I couldn't do an A/B comparison.
You could always look into the ultrapath approach? Not necessarily an easy road, though.
Personally, I'd leave your 300B as is and build a test amp.
My first from-scratch build was a 2A3 amp. Then I built a 45 amp for my friend. Almost identical amplifiers except for power transformer, input tubes (6J5 vs 6N1P), and power tube coupling caps (Miflex KPCU vs Cornell Dubilier). I preferred the 45. It had a distinctly different tone.
But when I built the 45, I breadboarded it initially and could test different input tubes, bias points, resistor vs bypassed resistor vs LED for (input) cathode biasing. That was very fun and eye-opening.
You could always look into the ultrapath approach? Not necessarily an easy road, though.
Personally, I'd leave your 300B as is and build a test amp.
My first from-scratch build was a 2A3 amp. Then I built a 45 amp for my friend. Almost identical amplifiers except for power transformer, input tubes (6J5 vs 6N1P), and power tube coupling caps (Miflex KPCU vs Cornell Dubilier). I preferred the 45. It had a distinctly different tone.
But when I built the 45, I breadboarded it initially and could test different input tubes, bias points, resistor vs bypassed resistor vs LED for (input) cathode biasing. That was very fun and eye-opening.
I've been doing the same kind of thing on one of my preamp builds, mostly playing around with different topologies and bias schemes. A good learning experience.
I am still on the fence on whether I want to start playing around with my current monoblocks - it's really sounding nice in my system right now, and I'd hate to break something and have to go back to one of my solid state amps to get music. I listen to my system 6-8 hours a day while I'm working from home and love the way the 300B SET amps sound.
I agreed to buy a pair of Monolith SX-11 transformers that were listed for sale here, so I'm somewhat committed to build another SET amp as my next amp project. So this is probably what I should be experimenting with.
I am still on the fence on whether I want to start playing around with my current monoblocks - it's really sounding nice in my system right now, and I'd hate to break something and have to go back to one of my solid state amps to get music. I listen to my system 6-8 hours a day while I'm working from home and love the way the 300B SET amps sound.
I agreed to buy a pair of Monolith SX-11 transformers that were listed for sale here, so I'm somewhat committed to build another SET amp as my next amp project. So this is probably what I should be experimenting with.
I use some polypropylene film caps that I get from ApexJr on the auction site. They are 75uf at 600v and only $5 each. They are great for power supply use. They are not small so you need room to use them or any film cap. Definitely worth a try for the price.
Kirchoff's voltage rule for series circuits.
Please try and follow me around the loop:
+ 370V, the + end of a 100uf filter capacitor
+370V, the + end of the output transformer primary
+360V, the plate end of the output transformer primary
+360V, the plate of the 300B
+60V, the filament of the 300B
+60V, the 1k Ohm self bias resistor; and the + end of the 100uF self bias bypass cap
0V, the other end of the 1k Ohm self bias resistor; and the - end of the 100uF self bias bypass cap
0V, the - end of the 100uF B+ filter capacitor.
And back through the - end of the 100uF B+ filter capacitor, through the capacitor's capacitance, to the + end of the 100uF B+ capacitor
And there you go, two 100uF capacitors in series.
Do you want to use a 100uF film capacitor for the self bias bypass cap, then you seriously need to consider . . .
using a 100uF film capacitor to filter B+.
Just Kirchoff logic.
Please try and follow me around the loop:
+ 370V, the + end of a 100uf filter capacitor
+370V, the + end of the output transformer primary
+360V, the plate end of the output transformer primary
+360V, the plate of the 300B
+60V, the filament of the 300B
+60V, the 1k Ohm self bias resistor; and the + end of the 100uF self bias bypass cap
0V, the other end of the 1k Ohm self bias resistor; and the - end of the 100uF self bias bypass cap
0V, the - end of the 100uF B+ filter capacitor.
And back through the - end of the 100uF B+ filter capacitor, through the capacitor's capacitance, to the + end of the 100uF B+ capacitor
And there you go, two 100uF capacitors in series.
Do you want to use a 100uF film capacitor for the self bias bypass cap, then you seriously need to consider . . .
using a 100uF film capacitor to filter B+.
Just Kirchoff logic.
Last edited:
DC link capacitors is the best buy for both cathode bias and PSU. As usual there are more expensive (and better) ones and less expensive ones but rest assured than even the cheaper ones are better in any technical aspect respect to electrolytic except size. When size does not matter DC-link all the way, except filament/heater DC heating and service power supplies.
Regarding cost, 400-450V DC-high quality electrolytic can cost more than same voltage DC-link ones. Before COVID ,DC-link were very easy to find in large quantities in all sorts of specs. Then I saw shortage, like for everything else. Now it looks like it's slowly going back to normal but still see long lead times on back orders.
Regarding cost, 400-450V DC-high quality electrolytic can cost more than same voltage DC-link ones. Before COVID ,DC-link were very easy to find in large quantities in all sorts of specs. Then I saw shortage, like for everything else. Now it looks like it's slowly going back to normal but still see long lead times on back orders.
Last edited:
If a capacitor is big enough no signal will appear across it, so it can do no harm. OTOH, if one invests enough money and effort to install a boutique capacitor, it will automatically sound "better" to the investor. We're all human.
All good fortune,
Chris
All good fortune,
Chris
My 350V-0-350V and solid state diodes cause the B+ to rise up to nearly 500V; before the tubes start drawing current as they warm up.
An un-loaded supply like that will eventually rise to 500V or More, because the primary DCR and secondary DCR do not drop significant voltage with no load, so the secondary becomes Larger than 350-0-350V
A Capacitor input filter will have more than 500VDC. Expect 1.414 x Vrms
A Choke input filter will Also have more than 500VDC, because there is not a Critical Inductance large enough when the load is zero.
With critical inductance, and loaded, 0.9 x Vrms;
But without critical inductance and un-loaded, 1.414 x Vrms.
Use a 450V Capacitor - OUCH!
An un-loaded supply like that will eventually rise to 500V or More, because the primary DCR and secondary DCR do not drop significant voltage with no load, so the secondary becomes Larger than 350-0-350V
A Capacitor input filter will have more than 500VDC. Expect 1.414 x Vrms
A Choke input filter will Also have more than 500VDC, because there is not a Critical Inductance large enough when the load is zero.
With critical inductance, and loaded, 0.9 x Vrms;
But without critical inductance and un-loaded, 1.414 x Vrms.
Use a 450V Capacitor - OUCH!
I disagree. Signal will appear because they are far from ideal. Losses of all types are there and the difference respect to film is substantial. Hysteresis loss in electronic caps is much worse than that of transformers' cores, in a PSU they might the best tanks for the size but are poor at filtering.
If then the required voltage rating is above 450V, they might not even be economically convenient anymore.
If a 500V cap is too expensive, then stack two 350V caps in series, and put 100k Ohms across each one to balance the leakage current that makes one cap have more voltage across it than the other. And, the two 100k Ohm resistors also act as a bleeder.
But what is more expensive?
A. One 100uF 500V cap
Or . . .
B. Two 200uF 350V caps in series and two 100k Ohm resistors
(you need one or the other of A or B before the choke, and one or the other of A or B after the choke)
Which one is easier to wire?
Which one is easier to mount, 2 or 4 capacitors?
Save Time
Time = Money
Save Money
Just my opinions
But what is more expensive?
A. One 100uF 500V cap
Or . . .
B. Two 200uF 350V caps in series and two 100k Ohm resistors
(you need one or the other of A or B before the choke, and one or the other of A or B after the choke)
Which one is easier to wire?
Which one is easier to mount, 2 or 4 capacitors?
Save Time
Time = Money
Save Money
Just my opinions