[chgolatin2]

For tube audio system what’s better and why? There seems to be a debate about PCB boards and point to point wiring? I feel that point to point provides the end user with a better quality but then again that’s just me. The builder has to take their time to build the amp right, make sure everything is properly solder in, short signal path, and using quality wires and materials. PCB boards allows the manufacturer to produced/build a tube amp in mass produced quantities. So what’s your thoughts on both processes?

[dB Cooper]

Don't think there's actually a performance difference per se. I think pcb's have a bad rep that comes from the old phenolic materials that deteriorate over time from the heat tubes generate. Dyna probably never imagined people would be keeping ST70s running 60 years later. NP with fiberglass boards. PCBs improve repeatability and reduce wiring error potential, considerations for the companies that make (or made)(mostly made) kits.

[audioengr]

There is definitely a performance difference.  The reason is that the dielectric of the FR4 epoxy board is a terrible dielectric for analog signals.  The dielectric absorption is too high I think.

I use to mod a lot of different companies products, including the ML38 preamp from Mark Levinson.  This preamp had literally 15" traces from the front of the PCB to the back to the connectors.  Made it sound very dark.  I cut all of these traces and wired externally with silver wire with Teflon or cotton insulation.  No more darkness.  BTW, ML later came out with a version of the 38 with a Teflon circuit board, so even they knew it was a problem.

For my own products, if there is a trace more than about 1" long, I put vias on the board and wire it externally with silver/cotton twisted-pair.  Makes all the difference in a top-tier component.

Steve N.

[Steve]

I look at the issue differently, and obtained superior performance. I can only speak for my views. I could have done hard wiring, but chose a better route, for me. Caveat: I am retired from manufacturing audio components.

First off, copper will "sound" different than silver.

Ok. Although I could easily performed hard wiring (been at it for 60 years and done it), I prefer epoxy pc boards for several reasons.

1) Via specialized listening test setups, superior, accurate sonics can be had using the board, if properly used. (I did not use a computer program to route my boards). This accuracy can only be matched, not improved upon. However, this statement does not cover subjective qualities one might like.

2) Super wide foils can be used to mimick larger wires but with much less self inductance. Even straight wires have higher self inductance with little change vs gauge. Foils can also be shorter in length with proper layout, thus even less inductance. Epoxy boards are quite alright.

3) Fewer solder connections can also be had for superior reliability and less sonic degradation in reference, superior gear.

4) Designing such results in superior consistency from component to component. Mounting parts on the board
provides more consistent part to ground and part to part orientation. (For better consistency, hand measured
0,5% tolerance parts as well.) This has several beneficial effects.

A) It lessens capacitance from parts to ground and part to part, for better and more consistent
response/performance and...

B) Consistent and improved channel separation, especially as the frequency rises. As such, bundling wires may
look nice, but is an inferior practice. Within a singular chassis, this results in improved frequency response and
imaging. How so you may ask?

Measuring one channels frequency response is somewhat deceptive. The lower the channel separation (which occurs almost always at high and very low frequencies), the more skewed the overall frequency response when both channels are being played. I will try to be very simple so the average consumer will understand.

That is because some unwanted high or low musical is being reproduced in the other channel when it should not. 
This unwanted musical signal adds and makes the sound unnatural because the total response is not flat. This causes unnatural sounding music, but also smearing, mis-positioning, lack of focus of the images.

It is good to check out the channel separation across the entire frequencies.

As such, all the layout, wires, every single part etc are important to accurate, natural musical reproduction. I used pc boards for maximum consistency and sonic performance. (I also used, measured, 0,5% tolerance, parts while in business, for consistency; one percent did not cut it.)

Hope this helps.

steve
 

[FullRangeMan]

Some hi end preamps use hard copper wire and hard silver wire to obtain the best sound for each circuit position, hard to do a silver track PCB.

[A_shah]

Some hi end preamps use hard copper wire and hard silver wire to obtain the best sound for each circuit position, hard to do a silver track PCB.

Pretty technical stuff most of it seems to go over my head, however my humble experience seems to tell me I like the Sonic quality of my Point to Point wiring no pc board far better than SS( which seem to have a slight hiss with high efficiency speakers)  even tube amps with a PC board----- The present dual mono I have among other amps that i own is amazingly quite ! similarly, the headphone amp from the same manufacturer also is amazingly quite
Love the simplicity of the wiring , here are a couple of pictures of the guts of my Mono and the newly acquired HP amp.

Mono-60 watt amp point to point wiring




3.5/4  watt Headphone amp with 2.2 output impedance



The older the irons get the sweeter they seem to sound
can somebody explain to me why do these amps have white  LED's/diodes ? inside the circuit ? 

Asghar








   

[Freo-1]

Point to point, no question.  PCB's are done for cost, not for sonics. Think about this:  Audiophiles spend scads of cash on so called "high end" interconnects, yet, the real limiting factor in performance for tube gear are the  PCB's.


Here are pictures of my 6AH4 Clone preamp and 1625 power amps.  No PCB's in the audio signal whatsoever.

[Steve]

Point to point, no question.  PCB's are done for cost, not for sonics. Think about this:  Audiophiles spend scads of cash on so called "high end" interconnects, yet, the real limiting factor in performance for tube gear are the  PCB's.

If that were the case, then my 11A line preamplifier would not be perfectly accurate, in absolute terms. Listening testing via special proprietary setups, more than one method, however demonstrate otherwise by many over the decades in my lab. I suspect the problems you and others encountered were due to:

1) The layouts of the boards was done by a computer program, not by an individual. If that is the case, maybe in certain cases. But "no question" is quite an over statement.

2) The part's sonic quality is inferior, which is almost always the case; parts, tubes, can be a a large source of noise as well. Price does not dictate accuracy of parts (capacitor, resistor, tube) quality.

3) The overall design of the auditioned component, in general, was compromised. From the designs I have seen, virtually no analog audio circuits are designed properly. Wrong value and quality of parts (from point 2), poor layout, virtually guarantee compromised performance, both sonically and quality wise.

I have designed and specially listening tested many many components in my 50 years in the lab. More than likely, the points listed above were probably the culprit in your experiences.

As such, your definite conclusion just is not accurate.

Super high prices does not dictate quality of sound, instead, usually cosmetics.

Cheers

steve

Caveat: I am not manufacturing electronic components anymore. I continue to do research on my
personal components, particularly test speakers.

[Freo-1]

No piece of audio gear is perfectly accurate (but one can get pretty close).  I'll stand by the assertion that point to point for tube audio is a better option overall as opposed to PCB's.   


I do agree that price does not automatically translate to better sound. 

[Steve]

No piece of audio gear is perfectly accurate (but one can get pretty close).  I'll stand by the assertion that point to point for tube audio is a better option overall as opposed to PCB's.   


I do agree that price does not automatically translate to better sound.

If individuals, over the years, cannot tell a sonic difference via preamp in or out of the system, or secondly, through specialized listening testing procedures, well............. How can one get better than that.   

As alluded to in my previous post, I find very very very few circuits that are properly designed in the first place, which includes parts quality, lay out, HW or PC. All affects quality. I guess we will just have to agree to disagree.

I like the fact that we both concur that price doesn't determine overall musical quality.

Cheers and all the best.

steve

[Freo-1]

If individuals, over the years, cannot tell the difference via preamp in or out of the system, or through specialized listening testing procedures,  well.............   I have yet to have a hardwired component pass the same type of specialized listening tests for musical accuracy as the 11A did. My previous post's points are still totally valid.

I find very very very few circuits that are properly designed in the first place, which includes parts quality, lay out, HW or PC.

I like the fact that we both concur that price doesn't determine overall musical quality.

Cheers and all the best.

steve

Not quite sure I understand what you mean by specialized listening tests.  Audiophiles tend not to be able to tell the difference between CD Redbook and Hi-Res versions of the same recordings.  Not sure how one can validate based on subjective listening tests.  Besides, one thing I have learned over the years is that the speakers are the key to detecting deltas with hardware playback setups.


You may have a point about circuit designs.  Tube electronics tend to be more forgiving  than SS designs.  One person who's designs I personally like a lot is Thomas Mayer.  I had a clone of his 6AH4 preamp made, and that piece was a "a-ha" moment for me with tube audio.  The noise floor from that design is very good, and the holographic quality from the preamp is very nice. 


I think the larger point I was trying to make is that most commercial tube products use PCB's for reasons other than sound quality. For the DIY tube crowd, no reason to not use point to point.  If I recall correctly, the HK Citation I/II used point to point wiring. 

[Steve]

Not quite sure I understand what you mean by specialized listening tests.  Audiophiles tend not to be able to tell the difference between CD Redbook and Hi-Res versions of the same recordings.  Not sure how one can validate based on subjective listening tests.  Besides, one thing I have learned over the years is that the speakers are the key to detecting deltas with hardware playback setups.

The listening tests I developed are proprietary, and the results are in absolute terms, not subjective as simply auditioning.   As very very simple examples, is the preamplifier bright and the speakers full sounding to compensate? Maybe it is the amplifier that is bright? Are the speakers zinging on the highs, while using a 12AX7, which rolls off the highs?
In reality, one has to address many multiple problems in designs such as focus, dynamics, soundstage width, soundstage depth, smearing, grunge etc. etc.

One thing important is take your time. I took years, when listening testing, and lots of musical selections, not just three or four. 

I personally found that if one sets up for red book, high resolution can sound bright. Conversely, when one setups for high resolution, red book can sound lacking highs. But this is just a general impression as recording quality/mixing varies.

You may have a point about circuit designs.  Tube electronics tend to be more forgiving  than SS designs.  One person who's designs I personally like a lot is Thomas Mayer.  I had a clone of his 6AH4 preamp made, and that piece was a "a-ha" moment for me with tube audio.  The noise floor from that design is very good, and the holographic quality from the preamp is very nice.

I agree many/most tube designs tend to be more forgiving. Reviewers see this as well. The designs often follow a similar pattern, few filter stages with small electrolytic capacitors, chokes believe it or not (SET designs almost require chokes however), small value coupling capacitors and other problems. One will never get truly accurate sound from such designs.
Tube components can be thin as well, and also right on the money rather than soft, forgiving.

I think the larger point I was trying to make is that most commercial tube products use PCB's for reasons other than sound quality. For the DIY tube crowd, no reason to not use point to point.  If I recall correctly, the HK Citation I/II used point to point wiring.

I can partially agree, especially with economical designs from large companies.

Cheers and all the best.

steve

[Ericus Rex]

I'll let others argue about the sonic differences between the two.

I will say that tube sockets soldered into PCB boards can be a tube roller's nightmare.  Tube sockets are often so tight the pulling out a tube can flex and damage the PCB and wreak havoc on the solder joints.  PCBs also don't like the heat of tubes.  If these issues are considered in designing the amp's layout than PCBs can be used without reliability issues.

[Freo-1]

The listening tests I developed are proprietary, and the results are in absolute terms, not subjective as simply auditioning.   As very very simple examples, is the preamplifier bright and the speakers full sounding to compensate? Maybe it is the amplifier that is bright? Are the speakers zinging on the highs, while using a 12AX7, which rolls off the highs?
In reality, one has to address many multiple problems in designs such as focus, dynamics, soundstage width, soundstage depth, smearing, grunge etc. etc.

One thing important is take your time. I took years, when listening testing, and lots of musical selections, not just three or four. 

I personally found that if one sets up for red book, high resolution can sound bright. Conversely, when one setups for high resolution, red book can sound lacking highs. But this is just a general impression as recording quality/mixing varies.

I agree many/most tube designs tend to be more forgiving. Reviewers see this as well. The designs often follow a similar pattern, few filter stages with small electrolytic capacitors, chokes believe it or not (SET designs almost require chokes however), small value coupling capacitors and other problems. One will never get truly accurate sound from such designs.
Tube components can be thin as well, and also right on the money rather than soft, forgiving.

I can partially agree, especially with economical designs from large companies.

Cheers and all the best.

steve

Thank you for responding.  Good points about filtering, etc.  The DIY gear shown here did attempt to ensure that those issues were addressed.  For example, the 6AH4 preamp clone has a lot of filtering, which is why it is so quiet.





To keep the cost in line, had Heyboer make all the iron for the preamp, and the power amps audio and power transformers.  Used NOS chokes on the power amps. 


Could you post some more information on your preamp?  Looks very interesting. 

[Freo-1]

I'll let others argue about the sonic differences between the two.

I will say that tube sockets soldered into PCB boards can be a tube roller's nightmare.  Tube sockets are often so tight the pulling out a tube can flex and damage the PCB and wreak havoc on the solder joints.  PCBs also don't like the heat of tubes.  If these issues are considered in designing the amp's layout than PCBs can be used without reliability issues.

Excellent point.  The heat from power tubes on a PCB is a non starter, IMHO.  I struggled with that issue on a ARC D115 I owned years ago.  I swore, never again.

[audioengr]

Some hi end preamps use hard copper wire and hard silver wire to obtain the best sound for each circuit position, hard to do a silver track PCB.

My PCB's are all silver-plated, not gold-plated like other manufacturers.

Steve N.

[audioengr]

Point to point, no question.  PCB's are done for cost, not for sonics. Think about this:  Audiophiles spend scads of cash on so called "high end" interconnects, yet, the real limiting factor in performance for tube gear are the  PCB's.


Here are pictures of my 6AH4 Clone preamp and 1625 power amps.  No PCB's in the audio signal whatsoever.

Agreed.  Certainly complex layouts like digital must be done with PCB however.  Analog is a different matter.

If these were my components, I would use pure OCC silver sire with cotton insulation for the signal paths.  Silver-plated copper wire twisted pairs with Teflon insulation for the power delivery paths.  This is what I did in my own SET monoblocks.

[Steve]

Thank you for responding.  Good points about filtering, etc.  The DIY gear shown here did attempt to ensure that those issues were addressed.  For example, the 6AH4 preamp clone has a lot of filtering, which is why it is so quiet.

Could you post some more information on your preamp?  Looks very interesting.

Sure. As you mention, the filtering performs functions. Several in fact.

1) As you mentioned, hum is dealt with, noise and rfi as well. The noise becomes basically related to the tube type/manufacturer resistor etc in a properly designed filtering system, proper physical circuit shielded as well.

2) Proper filtering isolates the signal circuitry from the rectifier, power transformer, power line etc. I see so many discussing the need for some sort of "conditioner". This would not be the case if proper filtering were used.

3) I used 5 stages of filtering in my 10A and 11A line  preamplifiers when I was designing and manufacturing. Besides the above pluses, the size and type of power transformer is inconsequential since in class A, the average current is virtually constant. If the power transformer makes a sonic difference, then the power supply filtering system is inadequate.

4) I use no chokes in either my preamplifiers or push pull monoblock amplifiers. Single ended is another story
however. The problem with chokes in power supplies is that the isolation between filter capacitors varies with
 frequency. This is especially critical between the decoupling capacitor and previous filter capacitor. This has sonic consequences.

Not only does the power supply filtering stages have to deal with ripple voltage, nonsense in the line etc, but from the other end, the musical signal.

Let's investigate a typical line stage's output signal at the plate of a vacuum tube.

We have a (musical) signal at the plate of the tube (same as with collector of a transistor, but let's stick with tubes). Musical signal current flows through C coupling capacitor and through Rg grid resistor to ground. However, we also have musical signal current flowing from the plate through RL, plate resistor, and through the C1 decoupling capacitor to ground.

Let's pretend we have a choke between power supply C1 decoupling capacitor and the previous C2 filter capacitor. Let's say the choke is 5 henries inductance.

At 20 hz we have 628 ohms of reactance between the decoupling capacitor and previous filter capacitor.
At 200 hz we have 6280 ohms of reactance between the decoupling capacitor and previous filter capacitor.
At 2k hz we have 62800 ohms of reactance between the decoupling capacitor and previous filter capacitor.
AT 20 khz we have 628,000 ohms of reactance between the decoupling capacitor and previous filter capacitor.

At 20 khz we have near complete isolation between filter capacitors C1 and C2, but at 20 hz we have a close
relationship between C1 and C2. To be very simplified for ease of understanding for newbies, as the frequency lowers, C1 interacts with C2 more and more. In a sense, C1 and C2 are almost combining to form a larger capacitor. There is also a little phase shifting.

If C1 is 30 ufd, the reactance at 200 hz is 26.5 ohms. Signal voltage and current will be present. (In a perfect
power supply, the reactance of the decoupling capacitor should be zero ohms.)

At 20 hz, the reactance of C1 is 265 ohms. Some of the signal at C1 is coupled/present at C2 as well since the inductive
reactance is quite low.

At 20 khz, the choke's reactance is so large that C2 sees virtually no signal voltage or current.

What we have is a non linear situation in the power supply. 

This is just a little aspect of what one has to contend with when delving into high end reference component design.

I hope I have helped, without scaring anyone.

Cheers and all the best.

steve (retired from component manufacturing)

[FullRangeMan]

A doubt: for power amps laminated power transformers offers best sound or best electrical performance than Toroidal ones?
Thanks