Tube DIY Asylum

Do It Yourself (DIY) paradise for tube and SET project builders.

Return to Tube DIY Asylum

Message Sort: Post Order or Asylum Reverse Threaded

audio signal in a 2A3 tube

50.80.146.114
Posted on October 3, 2016 at 11:33:31
banpuku
Audiophile

Posts: 1008
Joined: January 19, 2006
 Newbie question: Does the audio signal pass thru the filament on a 2A3 tube? Or does the audio signal just go thru the grid and plate?

If the audio signal does not pass thru the filament, then how critical are the quality of the resistor and capacitor connected to the filament?

Here is my circuit:



 

Hide full thread outline!
    ...
RE: audio signal in a 2A3 tube, posted on October 3, 2016 at 11:46:28
David McGown
Audiophile

Posts: 584
Location: Silver Spring, Maryland
Joined: September 29, 1999
 For direct heated triodes like the 2A3, the filament is the cathode and therefore in the signal path. The AC current loop includes the cathode bypass capacitor, as well as the power supply capacitor in the B+ supply feeding the output transformer. The current loop (electron flow) passes from ground, through the cathode bypass capacitor, to the cathode (filament), modulated by the grid to the plate, through the power transformer to the + of the power supply capacitor, and then back to ground through the capacitor.

Keep in mind the quality of the B+ (power supply) capacitor is just a critical as the cathode bypass capacitor.

 

RE: audio signal in a 2A3 tube, posted on October 3, 2016 at 12:20:17
banpuku
Audiophile

Posts: 1008
Joined: January 19, 2006
 Very good. Thank you David!

 

RE: audio signal in a 2A3 tube, posted on October 3, 2016 at 13:19:01
Tre'
Industry Professional

Posts: 17297
Location: So. Cal.
Joined: February 9, 2002
 If you want to keep the signal out of the cathode bypass cap and out of the last cap in the power supply, remove the cathode bypass cap and install a UltraPath cap between the B+ side of the output transformer and the cathode connection.

At all frequencies (especially the mids and highs) the reactance of the UltraPath cap will be much lower than the reactance of the total of the other path.

Other path = the reactance of the last cap in the power supply in series with the resistance value of the cathode resistor.

This means that most of the signal travels through the UltraPath cap and very little travels through the power supply cap and cathode resistor.

But to successfully do this the power supply has to have very low ripple otherwise noise will be added to the amplifier's output.



Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"

 

think this is true in any cathode biased amp, posted on October 3, 2016 at 13:20:54
FenderLover
 Triode or not. The bypass cap effects the frequency response.

 

RE: think this is true in any cathode biased amp, posted on October 3, 2016 at 13:45:34
Eli Duttman
Audiophile

Posts: 10455
Location: Monroe Township, NJ
Joined: March 31, 2000
 Uh huh. The bypass cap. combines with the cathode resistor to form a high pass pole. It is necessary to place F3 of said pole below 5 Hz., to avoid phase shifts within the audio band. Guess why I recommend 100 ohms and 470 muF. in combination bias situations. ;>)


Eli D.

 

Ultrapath, posted on October 3, 2016 at 13:48:40
Triode_Kingdom
Audiophile

Posts: 10045
Location: Central Texas
Joined: September 24, 2006
 Tre', I'm bothered by one thing about the "Ultrapath" concept. Lifting the cathode bypass cap off ground moves the AC common for the output stage to the output transformer. That point isn't shared by the driver anode or output stage grid resistor, both of which are still referenced to ground (bottom of the PS caps). Doesn't that mean whatever difference exists between the original output stage AC common (ground) and the new one (cold end of OPT) will now appear across the grid and cathode of the output stage?



 

RE: audio signal in a 2A3 tube, posted on October 3, 2016 at 13:59:55
banpuku
Audiophile

Posts: 1008
Joined: January 19, 2006
 Tre,

So, looking at Lynn Olson's ETF Presentation, he presents the schematic (see below) for the Elliano Ultrapath. Is this what you are referring to as the Ultrapath? If so, it seems like quite a simple change to make.

I currently have a 68uF film cap that I can use as the Ultrapath cap tied to the OPT B+. Is a 68uF cap a reasonable value to use in my circuit (see original post for the circuit)?

Pat



 

RE: Ultrapath, posted on October 3, 2016 at 15:09:38
Tre'
Industry Professional

Posts: 17297
Location: So. Cal.
Joined: February 9, 2002

"Doesn't that mean whatever difference exists between the original output stage AC common (ground) and the new one (cold end of OPT) will now appear across the grid and cathode of the output stage?"

There was always a cap between the cathode of the output tube and ground.

The only change I can see is there are now two caps in series (the PS cap and the UP cap), instead of a cathode bypass cap, between the cathode of the output tube and ground.

Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"

 

RE: audio signal in a 2A3 tube, posted on October 3, 2016 at 15:15:58
Tre'
Industry Professional

Posts: 17297
Location: So. Cal.
Joined: February 9, 2002
 68uf should be fine. It needs to be rated for the full B+ voltage.

The power supply ripple needs to be very low or it will hum.

It's easy enough to try and if you don't like it, put it back the way it was.


Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"

 

RE: Ultrapath, posted on October 3, 2016 at 20:02:36
Triode_Kingdom
Audiophile

Posts: 10045
Location: Central Texas
Joined: September 24, 2006



"The only change I can see is there are now two caps in series (the PS cap and the UP cap), instead of a cathode bypass cap, between the cathode of the output tube and ground."

But that's a huge difference in functionality. Here's what I see taking place in this circuit:

In a "normal" output stage, AC signal current flows through the tube, through the cathode bypass cap, through the B+ cap, and through the transformer primary. Looking at the caps, that's a series circuit. In the Ultrapath design, output stage signal current flows through the tube, through the Ultrapath cap, and through the transformer. It does not - in theory - flow through the B+ capacitor. So, in converting to Ultrapath, we have converted output current from two series caps to a single cap. That's a good thing.

Looking at the input circuit, exactly the opposite has occurred. Input signal voltage was originally impressed on the output stage between grid and cathode through the cathode bypass capacitor. In the Ultrapath design, it's impressed between the grid and cathode through both the B+ capacitor and the Ultrapath cap. Thus, the input circuit has changed from a single cap to two caps in series. That's not so good.

This change to the input circuit creates two potentially negative effects that weren't present in the original. First, the input signal is connected to the cathode through the B+ capacitor, and it will be subject to the same imperfections in that component as was the output signal previously. Second, all the noise from the power supply - which previously could only modulate the cold end of the output transformer - is now directed to the cathode. There it modulates the grid-cathode junction, and will be amplified by the tube.

This goes back to what I was describing earlier. Looking at the Ultrapath circuit above, the input loop and the output loop use different "commons." The input signal is presented between the grid and ground. Ground is its AC common. Output signal current, on the other hand, does not share that reference. Therefore, any AC voltage that appears across the main B+ capacitor will create a differential to ground that will be applied to the input of the tube through the Ultrapath cap.

 

RE: Ultrapath, posted on October 4, 2016 at 08:03:31
Tre'
Industry Professional

Posts: 17297
Location: So. Cal.
Joined: February 9, 2002
 "First, the input signal is connected to the cathode through the B+ capacitor, and it will be subject to the same imperfections in that component as was the output signal previously."

Yes. But will those imperfections manifest themselves in the same way with the voltages and currents of the input signal vs. the voltages and currents of the output signal?


"Second, all the noise from the power supply - which previously could only modulate the cold end of the output transformer - is now directed to the cathode. There it modulates the grid-cathode junction, and will be amplified by the tube."

Absolutely true. Any ripple or noise from the power supply will be applied to the cathode of the output tube.

Like I told Pat, it's really easy to try and if you don't like it, it's really easy to undo.

I'm not necessarily trying to defend the UP connection, I'm just saying it's one alternative.


Tre'


Have Fun and Enjoy the Music
"Still Working the Problem"

 

RE: Ultrapath, posted on October 4, 2016 at 08:30:30
banpuku
Audiophile

Posts: 1008
Joined: January 19, 2006
 I tried the Ultrapath. I got plenty of hum, so apparently my PS ripple is too high. Back to the old method. At least I know that the quality of the capacitor is important, as the audio signal does go thru it.

Thanks for all who helped!

 

RE: Ultrapath, posted on October 4, 2016 at 10:37:55
Triode_Kingdom
Audiophile

Posts: 10045
Location: Central Texas
Joined: September 24, 2006
 The cathode bypass and last PS cap should both be selected for lowest possible ESR, particularly if they're the type that can color sound (such as electrolytic). This is less critical if you're using oil caps or film.


 

RE: audio signal in a 2A3 tube, posted on October 4, 2016 at 16:21:21
dave slagle
Manufacturer

Posts: 5430
Location: NYC
Joined: April 27, 2001
 To complete the whole current loop of a DHT you also need to consider the filament supply as a possible path. AC, DC low Z and DC high Z all offer different possible current loops.

dave

 

RE: audio signal in a 2A3 tube, posted on October 5, 2016 at 06:16:34
dave slagle
Manufacturer

Posts: 5430
Location: NYC
Joined: April 27, 2001
 Here is a treatment of the whole ultrapath from a joenet member.

I haven't even begun to try to digest it but it is on topic...

dave

 

Maybe it should be called UltraNoise, posted on October 5, 2016 at 07:50:56
Triode_Kingdom
Audiophile

Posts: 10045
Location: Central Texas
Joined: September 24, 2006
 He boils it down to this:

"Ultrapath-A: one cap in the signal path, at the cost of µ-tiplied noise;"




 

RE: Maybe it should be called UltraNoise, posted on October 5, 2016 at 10:11:39
dave slagle
Manufacturer

Posts: 5430
Location: NYC
Joined: April 27, 2001
 I think it is fun to think about the various combinations but ultimately have to agree with tre' when he says try it and see. In just about any case a compelling technical argument can be made for or against just about anything and ultimately what really matters is where the rubber meets the road.

dave

 

RE: Maybe it should be called UltraNoise, posted on October 5, 2016 at 10:33:30
Tre'
Industry Professional

Posts: 17297
Location: So. Cal.
Joined: February 9, 2002
 Thank you Dave.

I just want to add that the noise issue isn't an issue if the PS has very low ripple to start with.

PSUD2 shows my LCLC supply has less than 1mV peak to peak of ripple. (.9mV peak to peak)





My DC filament supply (LCRC) for the 300b has more than that (1.2mV peak to peak).





I do concede that with a UP connection the last cap in the power supply (along with the UP cap) is in the input signal path.

I'm not sure what that means in terms of sound. Does that cap do as much damage to the input signal as it would to the output signal?

In the end, a UP connection is not that easy to try. One must first have a very low ripple PS otherwise the noise will swamp any perceived improvement.

Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"

 

RE: Maybe it should be called UltraNoise, posted on October 5, 2016 at 11:15:22
dave slagle
Manufacturer

Posts: 5430
Location: NYC
Joined: April 27, 2001
 I guess I am standing at the top of a slippery slope. The PS ripply isn't an issue since everyone knows it is technically better to regulate all the voltages so of course that means if you try UP and don't like it, it is simply letting you hear the effects of your inferior power supply :-)

dave

 

RE: Maybe it should be called UltraNoise, posted on October 5, 2016 at 11:31:04
Triode_Kingdom
Audiophile

Posts: 10045
Location: Central Texas
Joined: September 24, 2006
 " it is simply letting you hear the effects of your inferior power supply :-)"

Or... It's exacerbating the noise commonly produced by a typical HV supply that's functioning perfectly well. I would add that most designers installing a UP cap to ward off unknown evils otherwise generated in the cathode bypass probably eschew the evils of a regulated HV supply as well. And none of this deals with noise entering from the AC line that can sometimes bypass a supply, regardless of regulation. By all means, build what you want, but any topology that creates an unnecessary vulnerability to noise in the final stage gets tossed off my workbench. :)

 

RE: Maybe it should be called UltraNoise, posted on October 5, 2016 at 21:33:56
Donald North
Manufacturer

Posts: 1296
Joined: February 8, 2001
 To me, regulating the HT is another form of feedback applied to the audio signal. I like no feedback and high impedance inductive loads on triodes with well-passively filtered power supplies.

 

RE: Maybe it should be called UltraNoise, posted on October 6, 2016 at 05:06:05
dave slagle
Manufacturer

Posts: 5430
Location: NYC
Joined: April 27, 2001
 That is pretty much where I find myself too.... from a conceptual level i am also starting to view feedback as pre-distortion in order to compensate for distortion in later stages.

dave

 

Interesting slant, Dave... am reminded of out of band dither noise to increase dynamic range~nT, posted on October 8, 2016 at 00:17:12
Cleantimestream
Audiophile

Posts: 7551
Location: Kentucky
Joined: June 30, 2005
 !
The Mind has No Firewall~ U.S. Army War College.

 

RE: Ultrapath, posted on October 21, 2016 at 09:38:29
banpuku
Audiophile

Posts: 1008
Joined: January 19, 2006



Tre,

I have revisited the Ultrapath yesterday. While I do have some minor hum, the overall sonics is very very good. Removing the bypass cap out of the signal path really makes a positive difference.

So, what to do about that last bit of hum. I have attached a schematic of the amp including the power supply. Is there anything I can do to reduce that last bit of hum? I don't want to go to a DC supply at this point. I was thinking about adding a high quality 4.7uF cap after C3 in the PS (see blue text). Or adding a choke after the 100uF Ultrapath cap (see blue text). Would either of these help reduce the ripple and therefore, reduce hum? Any thoughts?

Pat

Thanks,
Pat

 

RE: think this is true in any cathode biased amp, posted on October 31, 2016 at 17:19:50
banpuku
Audiophile

Posts: 1008
Joined: January 19, 2006
 Eli,

I have been pondering the potential solutions for the cathode bypass cap in my schematic (see the inital post). Please help me understand whether or not I can use a 100 ohm and 470 muF cap in my situation for the 2A3 output tube. If your recommendation can be used, then I could easily get the best 470 muF cap possible, as it won't cost and arm and a leg like a great 100uF cap would. AND, I would avoid phase shifts, correct?

I have been testing 3 different scenarios for the AC filament supply and cathode bypass caps and here is what I hear:

1. 1000 ohm resistor + 100uF ClarityCap TC cathode bypass cap (see schematic in initial post)
- volume output is full tilt (no reduction)
- sound is good, but I can still hear the capacitor vs. no cap at all
- very little hum, but I still would like to eliminate it completely. I used to have a hum pot, but removed it and the sound opened up quite a bit, but now there is a bit of hum.

2. 1000 ohm resistor + 100uF ClarityCap TC cathode bypass cap connected in UltraPath scheme
- volume output is full tilt (no reduction)
- sound is great as there is no cathode bypass cap in the signal path
- hum is a problem

3. 1000 ohm resistor + NO cathode bypass cap
- volume output is reduced significantly.
- sound is clear / clean, but not bottom end umph.
- I forget whether or not there was hum (sorry).

So, I would like to stay with AC filament supply and ideally, eliminate the cathode bypass cap from the signal path, thus my desire for UltraPath or any other scheme that eliminates the cap from the signal path. In your opinion, is there a way for me to reduce the hum when using UltraPath, such as getting better dedicated filament supply transformers (right now I have Hammond 166M2 transformers)? Wiring techniques? Schematic changes?

Eli, any help would be appreciated.

Regards,
Pat

 

Page processed in 0.033 seconds.

[ Top Top of Page ] [ Contact Us ] [ Support/Wish List ] [ Copyright © 2024, Audio Asylum, all rights reserved ]
[ General ] [ Speakers ] [ Tubes ] [ Vinyl ] [ Digital ] [ Hi-Rez ] [ Video Asylum ] [ Cables ] [ Tweaks/DIY ] [ Music ] [ Films ]