Mcandmars ramblings \ build thread...

[Doc B.]

I think you answered your own question. With the series resistors 90% of the residual hum is removed for sensitive headphones. Thus the background is more quiet, which translates to more resolution and the increased load that the amp sees usually creates a sense of flatter, more extended frequency response. Sometimes people will consider this "less tube-like" due to a misguided assumption that tubes must always be noisy and loose in the bass.

[mcandmar]

If anything i am feeling like some of the high end frequency are being rolled off, but i need to spend more time listening to music i know well to say for sure.

Is it worth experimenting with smaller values to find a sweet spot or is there a reason for the 120ohms value?     ...tell me what to google and i'll do my homework

Thank again,

Mark

[Paul Birkeland]

The 120 Ohm resistors used to actually be a standard in the audio industry.

The high frequency response should be identical to what was there without the resistors, provided the resistors aren't abnormally inductive.  (By abnormally inductive, I mean actually using an inductor instead of a resistor, lol)

[mcandmar]

I'm pretty sure they are resistors   Its probobly just my ears that need to recalibrate..

[VoltSecond]

Replace the transformer supply with a switching wall wart.  I have not tried this in this design, but it should work.  I've used this in other low noise designs with success.
1. Get one rated for the expected inrush current for the cold tubes. Designs with a medical rated version tend to be better.  This current rating is so it doesn't wig out during warm up and so it runs cool in normal operation.
2. Use a series wire wound resistor in the output hot and cold to both drop the voltage, limit the surge current and to act as an RF filter.
3. Put a large cap across the filaments (indirectly heated only.) Without the series resistor, the wall wart may oscillate if you add a cap to the output.
4. Bias the filaments above ground.
5. Put grounded copper tape  around the body of the wart (keep far from 115V blades) or a grounded piece of metal between the body and your circuitry for a shield. (You only really need line of site for this, if the wires and parts don't "see" the wall wart and its power cord, its usually good enough.)
6. Buy a ferrite toroid and wrap the output power lead to the wall wart 5-7 times around it (neatness counts, but don't wind so tight the copper in the lead cuts through the insulation around the lead.)

Switchers start out noisier, but they are easier to quiet down than a transformer rectifier circuit.

[mcandmar]

Replace the transformer supply with a switching wall wart.  I have not tried this in this design, but it should work.  I've used this in other low noise designs with success.
1. Get one rated for the expected inrush current for the cold tubes. Designs with a medical rated version tend to be better.  This current rating is so it doesn't wig out during warm up and so it runs cool in normal operation.

Interesting i thought switching power supplies in audio equipment was the work of the devil.  Though my experience is limited to tiny made in China supplies that were absurdly noisy and polluted the entire house with their switching noise.

2. Use a series wire wound resistor in the output hot and cold to both drop the voltage, limit the surge current and to act as an RF filter.

Would that be part of the function of the .1ohm resistor in the existing circuit?  I was reading up on the difference between inductive, non inductive resistors and different winding styles (Bifilar, Ayrton-Perry etc) to cancel out RF.

3. Put a large cap across the filaments (indirectly heated only.) Without the series resistor, the wall wart may oscillate if you add a cap to the output.

I had noticed the Mainline has tantalum caps directly on the tube socket heater pins, but figured i had asked enough stupid questions on said subject to ask..

4. Bias the filaments above ground.

Not entirely sure on what that means.   I have figured out the heater supply winding is center tapped and tied to ground so it swings +/- 3v basically to lower the voltage difference from the rest of the circuit to lower noise.  Or have i got that wrong?

5. Put grounded copper tape  around the body of the wart (keep far from 115V blades) or a grounded piece of metal between the body and your circuitry for a shield.

Do i need a tinfoil hat to go with that?

6. Buy a ferrite toroid and wrap the output power lead to the wall wart 5-7 times around it (neatness counts, but don't wind so tight the copper in the lead cuts through the insulation around the lead.)

Funny you should mention it i was going to start a thread on mains filters to see what other people have done, advise and so on. Will do that shortly as your wisdom would be greatly appreciated..

Cheers,

Mark

[Paul Birkeland]

This all presumes that the noise is in fact coming through the heater of the 6DN7, and not just self noise of the tube, power supply noise, or "other".

[VoltSecond]

Replace the transformer supply with a switching wall wart.  I have not tried this in this design, but it should work.  I've used this in other low noise designs with success.
1. Get one rated for the expected inrush current for the cold tubes. Designs with a medical rated version tend to be better.  This current rating is so it doesn't wig out during warm up and so it runs cool in normal operation.

Interesting i thought switching power supplies in audio equipment was the work of the devil.  Though my experience is limited to tiny made in China supplies that were absurdly noisy and polluted the entire house with their switching noise.

** They start out noisier, but are easier than many linear regulators to make very quiet.  Most of their noise issues come from common mode noise.

2. Use a series wire wound resistor in the output hot and cold to both drop the voltage, limit the surge current and to act as an RF filter.

Would that be part of the function of the .1ohm resistor in the existing circuit?  I was reading up on the difference between inductive, non inductive resistors and different winding styles (Bifilar, Ayrton-Perry etc) to cancel out RF.

** I was thinking on the order of a volt or two to help limit surge currents.  Normal wire wounds are fine for this application.

3. Put a large cap across the filaments (indirectly heated only.) Without the series resistor, the wall wart may oscillate if you add a cap to the output.

I had noticed the Mainline has tantalum caps directly on the tube socket heater pins, but figured i had asked enough stupid questions on said subject to ask..

** Can't speak about the Mainline, but tantalum caps don't like AC voltages and need ~2:1 voltage derating.

4. Bias the filaments above ground.

Not entirely sure on what that means.   I have figured out the heater supply winding is center tapped and tied to ground so it swings +/- 3v basically to lower the voltage difference from the rest of the circuit to lower noise.  Or have i got that wrong?

** I think PJ is a proponent of this.  The filament is biased above the cathode voltage with a large bypass capacitor to single point ground.

5. Put grounded copper tape  around the body of the wart (keep far from 115V blades) or a grounded piece of metal between the body and your circuitry for a shield.

Do i need a tinfoil hat to go with that?

** Tin foil hats are only needed if listening to DEVO. 

Good wiring habits will control inductive coupling. Metal foil (copper is easy to find at craft stores) makes an easy electrostatic shield.  which helps with the other "work of the devil" noise coupling method: Capacitive coupling.

6. Buy a ferrite toroid and wrap the output power lead to the wall wart 5-7 times around it (neatness counts, but don't wind so tight the copper in the lead cuts through the insulation around the lead.)

Funny you should mention it i was going to start a thread on mains filters to see what other people have done, advise and so on. Will do that shortly as your wisdom would be greatly appreciated..

** Inductance changes by the number of turns squared.  Excluding problems from resonant peaks, the inductance in a noise filter usually has to double to be slightly audible. 1 to  2 turns is a big step.  4 to 6 turns is a noticeable step.  7 - 8 turns isn't that noticeable. Second order effects get to be a problem once the winding goes more than one time around the core.

Cheers,

Mark

[mcandmar]

Cap rolling time! 

Having read various capacitor roundups i kept coming back to the Russian Teflons as the best bang for your buck so they seemed like a good place to start. The Teflon varieties FT-1,FT-2, and FT-3 are only available in small capacities under 1uf so i ordered a few FT-3 .1ohm 600v samples for the interstage and burnt them in for a few weeks connected to the speaker outputs of an amplifier playing a selection of frequency sweeps, 0db sine waves, white noise, pink noise, and "How do you say Bass by Bass Mekanik" lol. It was more of a torture test really.

Installation was a bit of a challenge due to the physical size of these caps, but i found with large gauge wiring they stayed in place without moving about. I then added a cable tie to help support the weight to reduce the stress on the solder joints.  The outer casing is bare metal and seems to be insulated from the capacitor but i added a layer of clear heat shrink over them anyway as a safety precaution.

Initially there was a harshness/distortion that seemed most noticeable with cymbals or notes around those frequencies, it's almost as though they are distorting but all the other instruments around it weren't.  It seems this was just part of the long burn in time people say these caps need to settle down, certainly any signs of that disappeared after a week of daily use.

I have been listening to them for about three weeks now and i am still picking up little nuances that have changed. At first i found them very transparent with clearer high end and low end frequencies, but with time i have noticed a couple of things. The tone in the mid-range / low end has much more "something" i can't find the right words for. "Presence and loveliness". For example bass guitars, drums, or any other instrument in that range are much more clearly defined, it's like any piece of music i listen to now i can just concentrate on what the bass guitar is playing without it getting lost in amongst the rest of the instruments. For quite music such as acoustic or Jazz i found wood instruments, double bass or wind instruments, actually sound like wooden instruments, while bass drums/toms have a lovely presence to them. Every now and again listening to music an instrument jumps out at me in a way i have never noticed before. It's all rather fascinating.

Having re-read a few roundups i found people commonly refer to MKP capacitors are being plastic sounding, i understand exactly what they are talking about now. And the FT3 are commonly referred to as being smooth but revealing, have to agree with that too. For me the FT3 are preferred, but i can see how some people might find the greater mid/low end presence not to their liking. And they do make the overall sound very smooth sounding without losing any detail, but for some a more forward trebly presentation may be more their cup of tea.  Having said that i would prefer more high end detail from the S.E.X. amp as i always found it had a very neutral presentation. The FT3s in no way reduced the high end, if anything they unveiled it a little, but it's the magic of the mid/low range is what does it for me. I am very curious to try something better in the future as the difference they have made was very surprising to me, i was expecting a more subtle change. 

Next i am going to look at replacing the Solen output caps, i had a short list of candidates drawn up and then had a moment of weakness in the post holiday sales and ordered a pair of 1.5uf Mundorf Silver/Oil caps.  I was really looking for something with neutral and clarity in its traits, and the silver/oil seem to fit the bill, we shall see..

[mcandmar]

Cap rolling part II: Output caps

This is where things get a little more complicated.  It has become obvious to me that i could re write what i wrote in the above post. A lot of it is still true however it has become apparent that my view of the FT3, or the role of the interstage capacitors in general was viewed through the window of the output capacitors.  In other words replacing the Solen output caps has removed that restriction allowing me to see the full picture.  In hindsight it would have been more enlightening to upgrade these capacitors in the opposite order starting with the output capacitors and then moving onto the interstage caps.  Certainly if it wasn't such a chore to solder/unsolder them i would do exactly that out of interest. If you are thinking about cap rolling the S.E.X amp i recommend starting with the output stage as i suspect that's the stage where the greatest benefit will be found.

My choice for the output caps were 1.5uf Mundorf Supreme Silver/Oil as they have a reputation for being neutral and highly detailed. Physical size is 36mm x 42mm which as you can see i only just managed to squeeze in next to the C4S board. Anything larger would have been a problem.

So how do the Mundorf caps sound?  I don't really know as i am unable to directly attribute what i am hearing to either pair of capacitors so i can only really speak of the sonic difference from my previous post with just the FT3's. In the above post i commented that i wished for more high end detail, and that some might find the mid/low end a little overly pronounced. Those comments have been rendered redundant. It is obvious to me now that what i was hearing was the limitations of the Solen output caps. The overall balance from the lowest of the low frequencies to the highest of the highs is just perfect now, i wouldn't change a thing.

I also don't find it as simple as saying there is more refinement, smoothness, clarity, or high or low frequencies. The change i am hearing is so pronounced and all encompassing all i can say is the sound is superior in every way, or perhaps i have just simply run out of audiophile vocabulary to describe it. If i had to pick just two buzz words it would be clarity and refinement, coupled with the tone and presence i experienced with the above FT3 caps.

The best analogy i can come up with is it's similar to finding that sweet spot in the middle of the room when your head is in that perfect position in relation to the speakers and four walls.  Everything just comes into focus giving you the full sonic picture with all its detail and texture, that's basically the difference i am hearing. I have found that unveiling has brought me closer into the music, and consistently sucks me in well into the early hours of the morning when one more track turns into one more album. Overall the combination of these four caps have brought this amp to a whole other level, and for me well into uncharted territory that i have never experienced before.

Interestingly i have little desire to experiment any further, or upgrade any components, though it does make me wonder where you can go from here with caps like, Jensens, Deulends, Jupiter or VCaps.  For the moment i am more than happy to be sucked in and wowed by this amp, even if the sleep deprivation is starting to catch up with me. And they say crack is highly addictive..

[mcandmar]

In the SEX 2.1 (with PT-7 power transformer) the heater power is only connected to ground for two short periods during the 60Hz power line cycle - just when the diodes are conducting. Most of the time it floats. For that reason, if there is a little leakage between heater and cathode, the voltage will drift away from symmetrical. You can put a couple 100-ohm resistors in series across the DC power and ground the center point if you want to see matched voltages; you are not however going to hear a difference unless the leakage is so bad the tube should be replaced.

The center tap is grounded so that the winding itself is always grounded. This allows it to act as an electrostatic shield for other windings.

This little tip is worth adding to the thread, adding 100ohms resistors from C1&C2 to C3 creating a grounded center point helped to further reduce the harsh bzzzt noise i was hearing, and also allowed me to use sets of tubes i had previously written off as being too noisy.  In my opinion this simple modification made such an effect i would suggest trying it first before experimenting with other modifications, or fitting the 120ohm resistors to the headphone socket.  As always, please report your findings if you do try it...

[Grainger49]

Have you found the source of the buzz?  If you are as handy with a scope as you with assembly you could find if it is of power supply or signal path origin.

[mcandmar]

Not really no. Its not in the audio path so its either HV supply, or heater supply. I had ruled out the heater supply as doubling the capacitance of the filter cap to 22000uf halved the ripple and made zero difference to the output noise.  Having said that the above experiment creating a grounded center point on the DC side of the heater supply did make a difference so i'm not really sure what that tells us.

I cant measure the HV supply directly as my scope is limited to 400v peak to peak. I tried modeling the supply in PSUDII and experimented with difference capacitance and got some very odd results, for example doubling the first two 220uf caps to 470uf actually made the ripple worse at the end of the chain.  However when i replaced the two 680ohm resistors with chokes it did reduce the output noise so i'm almost certain it is coming out of the HV supply.  How to improve it further i just dont know, or maybe this is just the limit of the capability for these tubes.

The caps themselves are as good as it gets for their value and rating, i haven't found anything better than those Panasonics.  I do have two 22uf Panasonics to try at the end of the chain to replace the Lelons as they have better specs, but i dont expect much if any difference. I also have a couple of 2.2uf film caps to try bypassing those but without a way to measure/quantify any difference i am going to hold off.

At the moment i have the noise floor below what is audible so i am perfectly happy with it.  But if i could get it to the stage where i could do without the two 120r resistors in the output path that would be even better as i dont like the high end roll off they caused.  Or maybe i just need to try another brand of resistors.  The real problem is the sensitivity of my headphones, HD650's without the 120r resistors its barely audible, but with the Grados i need the resistors to hide it.

[Paul Joppa]

You can make an L-pad, combining the series 120 ohm resistor with another resistor from headphone tab to ground. This reduces the signal level just the same but presents a lower impedance to the headphone. Try 47 ohms series and 22 ohms in parallel with the headphone - that will give the same attenuation as the 120 ohm resistor (assuming 32 ohm phones) but  a source impedance of 15 ohms instead of 120 ohms.

[mcandmar]

Interesting, i will give that a try.

I was thinking about what resistors to order and realized i never actually tried the amp without the 120r resistors since i installed the 100r resistors into the heater supply. As a test i made up a quick adapter cable from the XLR socket to a TRS socket to bypass them and found its perfectly fine. I had to power down my PC to reduce the ambient noise and really listen to detect any background noise with the Grados, result!

Since i had the cable made up i started switching between the TRS socket (with 120r's) and the XLR without and sure enough the SQ between them is very different. There was a good 1/4 turn of the volume knob between them making absolute a/b comparisons difficult but the conclusion i came to was the resistors make it sound what i would call more tubey. The sound had a softer, rounder edge to it and was much smoother sounding overall.  Without the resistors there is a step forward in clarity and detail, and a proper sharp attack/decay to snare drums and percussion, less of the woolliness basically.

Funny thing is i like them equally as they both have their pros and cons. I'm going to use the adapter cable for a while and see which i prefer over time before i jump in and remove them permanently.

I also rolled a few other tubes into the amp and confirmed the amp was quite with multiple sets so those 100r's really do make a big difference.