[Steve]

Actually no to both for several reasons. Tube circuits are expecially prone to
hardwiring problems due to higher impedances and physical layouts.

1) Consistency from unit to unit is lacking. Stray capacitance to ground varies.
Thus high frequency response is inconsistent.

2) Even at fairly low impedances, frequencies as low as 5khz are "transmitted" to other parts
of the circuit. If the adjacent wires are even remotely in close proximity, 3/4" or even farther depending upon
impedance, wires of the same channel will act as an antenna to "pickup" the signal. This
will affect the musical signal, and in an inconsistent manner due to variations in component wiring layout.

3) Because of midrange frequency transmitting problems, channel to channel separation will
also suffer. At 1khz, one may have 90db separation. However, at 10khz, the separation may only
be 50db. This means that higher musical frequencies will also occur in the other channel, thus not
only affecting soundstaging etc, but also perceived brightness of musical material because two channels
are producing the same higher frequencies. Again, this will cause an inconsistent measurement due
to variations in wiring layout.

4) Bundles, although cosmetically looking nice, will pass signals to adjacent wires because of super close
proximity.

5) Layout of wires usually means variation in layout of parts as well. This also presents all
the problems in points 1-4.

Cheers.

[JakeJ]

Great topic, Steve.  I have first hand experience with bundled wiring causing electrical problems so I can relate.

Is there any chance you could provide some oscilloscope screen shots to demonstrate what you are talking about?  We do have some members that steadfastly believe point-to-point wiring is the only way to wire an amp yet there are so few mfrs that do this in production.  Personally I can't think of one right now.  Anyone that can please share that mfrs name.

[Steve]

Great topic, Steve.  I have first hand experience with bundled wiring causing electrical problems so I can relate.

Is there any chance you could provide some oscilloscope screen shots to demonstrate what you are talking about?  We do have some members that steadfastly believe point-to-point wiring is the only way to wire an amp yet there are so few mfrs that do this in production.  Personally I can't think of one right now.  Anyone that can please share that mfrs name.

Hi Jake,

No, I  don't have any scope photos. Been along time ago. I do have two examples that will provide some information.
Some years ago I performed an experiment and will attempt to relate
to the best of my knowledge.

I used two wires, 6" long, 5/8" apart. Capacitance between wires "A" and "B" measured approximately
1pf. The scope ground and signal generator ground was located at the signal generator ground
terminal, approximately midway between "A" and "B" grounds. Thus there was not common ground
 current, or signal, common to both "A" and "B", except possibly internally in the generator cabinet.

The probe of the scope was connected to the signal generator ground terminal to double check,
and no signal was measured. The signal generator's signal wire clip was placed on the end of
"A" wire and the lead placed as far as possible from the end of the 6" wire, and at right angle for
minimial interference.

An HP signal generator source provided a varying frequency signal current through the source 
6" wire "A" to a load. I cannot remember the load value, but I believe at least a few ma of signal
current was provided.

"B" 6" wire had one end grounded and the other end a 2.2k ohm load to ground. The scope probe 
measured the amplitude and phase at the 2.2k ohm load point. At 5khz, a signal was detected by the scope approximately
40db down. Phase shift was very very minimal indicating the signal was not capacitively coupled.
As the frequency was increased, "B" wire's signal continued to increase, which is logical and scientific in nature.

As one lowered the frequency to 1khz, the signal on wire "B" lowered as it should.

The second example occurred when testing wire bundles. With a tuner connected to input "tuner",
selector switch set to "CD" input, music from the tuner was clearly perceived from my listening
position, some 8 feet away. Simply moving the "tuner" wire away from the others, some 3/4", reduced the
crosstalk to near zero with no music heard.

It is quite apparent that care should be taken when laying out, as well as wire placement in
any signal device.

Cheers Jake.

[Captainhemo]

Very interesting  guys....
Here is   what may be a dumb question  but I'll ask it anyhow.

Obviously  the PCB is going  to solve the consistant layouot and  component placement issues of a point to point  wired amp  but what  stops the  above issues relating to  "cross talk" or  "transmission/reception" by  wires in a point to point wired amp  when using a  circuiit board ?
Are the signal paths   positioned far enough  apart  that  it isn't an issue with a PCB .  I sem to recall seeing some pictures of boards where the   paths  are laid out pretty close to each other   but I could be mistaken.

just wondering if there is more going on   than just consistant spacing of the signal pathways

-jay

[pehare]

Ralph Karsten, Roger Modjeski & Gordon Rankin are a few that come to mind that use point to point.....these guys seem to know what they're doing

[Ericus Rex]

Are you specifically talking about amps?  Preamps?  or both?

Blair at Niteshade is another designer who I think only uses P2P.

[Freo-1]

For power amps, I think point to point is the way to go.  For preamps, it's a bit tougher call.  There are some very good DIY clone boards available today that far exceed most commercial boards.  One can stuff the DIY boards with top shelf components, and obtain some great performance as a result.

[Steve]

Very interesting  guys....
Here is   what may be a dumb question  but I'll ask it anyhow.

Obviously  the PCB is going  to solve the consistant layouot and  component placement issues of a point to point  wired amp  but what  stops the  above issues relating to  "cross talk" or  "transmission/reception" by  wires in a point to point wired amp  when using a  circuiit board ?
Are the signal paths   positioned far enough  apart  that  it isn't an issue with a PCB .  I sem to recall seeing some pictures of boards where the   paths  are laid out pretty close to each other   but I could be mistaken.

just wondering if there is more going on   than just consistant spacing of the signal pathways

-jay

Hi Jay,

Good question. Layout, layout, layout. You are quite correct, not all layouts are the same. In fact,
most designers don't understand the problem because they have never performed any testing. (Now they
probably will jump on the bandwagon.)

Amplifiers can also take advantage except in a slightly different way since filaments draw alot of current. If I go deeper, it
gets real complicated.

Cheers and hope this helps.

[Steve]

Are some preaching such? IF so, did any of them list the weaknesses of Pt to Pt wiring? I have
not heard of any.

The factual test information in previous posts on this string should help one to
be aware and help to obtain optimum results when designing equipment.

Hemo, you are correct, very few designers/manufacturers have used PC boards in an optimal manner.
Usually it is computer driven program with very thin foils.

(With different audio designs, others will have to perform tests on PC boards to improve channel separation
etc.) Hopefully, the issues will be addressed, but in any case, knowing the weakness of Pt to Pt wiring increases 
the public's knowledge.

Preamplifiers can almost always be optimally designed with PC boards. Exception might be filament circuitry.
Amplifiers, because of high filament currents, probably best to go hybrid, partial hardwire and part board.

Cheers.

[Docere]

Hi Steve,

I have not posted here for a while, but a few glasses of red and, well...

Interesting thread. I agree with some of your premises and arguments in principle... but not your conclusions/ solutions.

There are countless examples of shoddy P2P wiring. But, would it be honest to compare crap examples of P2P with the finest examples of PCBs... then claim PCBs as inherently superior?

I see some convincing arguments for not bundling cabling, but little to suggest - apart from repeatability - to support the use of PCBs over well-executed hard-wiring. I'd rather have a simple(ish) tube circuit well hard-wired and performing from very, very fine to excellent versus a (merely) fine sounding PCB. Put differently, I think hard-wiring has greater potential, and if well-executed - although still slightly variable - will outperform a well laid-out PCB. However, digital and other types of circuits are likely a different matter... I am not familiar with high-gain tube pre-amps that use lots of little tubes, so I have no idea what might be preferred for these.

There are methods that can maximise the benefits of both PCBs and P2P without incurring the disadvantages... but that is a separate discussion.

Steve, I notice you are an industry participant; do you sell products that utilise circuit boards?

Cheers.

[Quiet Earth]

Some companies use traceless circuit boards to consistently place the parts in the same location, and then they hardwire the leads. This seems like a good compromise to me. At least there is more consistency in manufacturing this way, regardless of how it is layed out.

I'm not sure if the audio band needs the same attention to layout as the Ku band, but its definitely worth considering.

[Steve]

Some companies use traceless circuit boards to consistently place the parts in the same location, and then they hardwire the leads. This seems like a good compromise to me. At least there is more consistency in manufacturing this way, regardless of how it is layed out.

I'm not sure if the audio band needs the same attention to layout as the Ku band, but its definitely worth considering.

 

Quite true, Quiet, except one difference. I believe the audio band definitely needs to consider layout.
Actually, I coined the term (and copyrighted the term "lead to lead") wiring some 17 years ago,
and for good reason.
 
"Lead to lead" connecting eliminates wires and solder connections vs hardwiring and typical PC board layouts/foils. Fewer wires
and solder connections generally means improved sonics for a given circuit. In fact I have a couple of pages pertaining
to hardwiring, pc boards, and "lead to lead".

http://www.sasaudiolabs.com/theory4a.htm

Good to see some others have upgraded their designs.

Cheers.

[Steve]

Hi Steve,

I have not posted here for a while, but a few glasses of red and, well...

Interesting thread. I agree with some of your premises and arguments in principle... but not your conclusions/ solutions.

My premise is absolutely correct based on factual data. Hardwiring and bundling has inherent problems. Solution is also straight forward.

There are countless examples of shoddy P2P wiring. But, would it be honest to compare crap examples of P2P with the finest examples of PCBs... then claim PCBs as inherently superior?

I already addressed that point in a previous posts. The best P2P wiring has problems, which is what I addressed.

I see some convincing arguments for not bundling cabling, but little to suggest - apart from repeatability - to support the use of PCBs over well-executed hard-wiring. I'd rather have a simple(ish) tube circuit well hard-wired and performing from very, very fine to excellent versus a (merely) fine sounding PCB. Put differently, I think hard-wiring has greater potential, and if well-executed - although still slightly variable - will outperform a well laid-out PCB. However, digital and other types of circuits are likely a different matter... I am not familiar with high-gain tube pre-amps that use lots of little tubes, so I have no idea what might be preferred for these.

I have already presented factual data that disagrees with your opinions about hardwiring. Please produce some facts that support your opinions, especially that the best hardwiring sounds better than the best PC boards or hybrids.

There are methods that can maximise the benefits of both PCBs and P2P without incurring the disadvantages... but that is a separate discussion.

QuietEarth has already presented such. I have just as well in my response to Quiet. Please present such information if you have some.

Steve, I notice you are an industry participant; do you sell products that utilise circuit boards?

Absolutely, with "lead to lead" connecting when possible. My preamplifiers are almost all PC board, properly laid out, with a little hardwiring, and my amps are hybrid Pc board and hardwiring (filament current, high voltage, output transformers etc). 

Doesn't your Metaxas amp use a large PC board?

Cheers.

[Docere]

Steve,

I think we are probably more in agreement than not. I also apologise for the pernickety tone of my post; it is counter-productive for me to post when in a particular fame of mind. I try to impose a limit of two posts per thread, so I be specific in this final post.

You have presented some opinions; some I share (especially the bundling point), others I don’t. I did not read any references to “facts” that clearly indicate the superiority of PCBs over hardwiring. I saw conjecture that could be based on some understanding of electrical theory and physics and simple experiments. If you can point me to some peer-reviewed articles that that document experiments – with strong internal and external validity - comparing well implemented PCBs to hardwiring in a range of tube circuits, I will reconsider my position.

rant
People, me included, confuse opinion with facts continually; it is our way. Opinions are important, but, we should not accept opinion, no matter how expert, as fact. Perhaps I have put too much emphasis on this point – it is what first irked me about the thread; best to pull my head in and move on…
/rant

Although not mentioned specifically, I had three things in mind when posting my comment about methods having the benefits of PCB and hard-wiring 1) an amp I am currently planning the layout for, 2) an article in SP about using Teflon sheet for mounting components, and 3) an article about AN build methods from many years ago. QuietEarth subsequently posted some specific details of a similar approach used by manufacturers. I agree with your comments in response to Quiet’s post. My relatively simple build will use 2mm Teflon sheet, finely drilled to mount components and connect component-to-component where appropriate or using good quality wire where not. I like this approach for the technical reasons discussed, but also for its simple practicality: easier to work on; allows concurrent build of different parts of the amp; promotes tidiness and logical layout; and is very suitable to optimising layout.

My point about you being an industry participant was in relation to competing interests: potentially promoting your products on a public – not manufacturer – forum. I was not suggesting that you were not walking the talk, which might be what you thought? Anyhow, the information you have shared will be helpful for many.

As for the Metaxas, of course it uses a PCB. I don’t understand your point: The Metaxas is fine for what it is and is best built on a PCB, but I don’t see how that is relevant to our discussion about rather different types of circuits.

All that aside, I think we are largely in agreement and I apologise for being argumentative in my previous post.

Cheers.

[Steve]

Docere

Steve,

I think we are probably more in agreement than not. I also apologise for the pernickety tone of my post; it is counter-productive for me to post when in a particular fame of mind. I try to impose a limit of two posts per thread, so I be specific in this final post.

Apology accepted.

You have presented some opinions; some I share (especially the bundling point), others I don’t. I did not read any references to “facts” that clearly indicate the superiority of PCBs over hardwiring.

Please do not misrepresent my data presented as opinion. It is measurable, factual data that is meaningful in every single audio design, whether hardwiring or pc board. Afterall, all designs use parts, connections, layouts. I have some 45 years worth of lab experience, including in the college lab with not only helping students but also professors.

All one has to do is check out the photos of hundreds if not thousands of designs (over the years) of hardwiring and PC board layout to see the obvious flaws. Proper layout is expecially important to hardwiring since almost all hardwiring is in close proximity to ground, and both hardwiring and PC board layout need attention to parts layout.

Hardwiring is not superior to Pc board wiring in anyway, unless the Pc board is improperly designed/layed out. In fact, you are following my example in your own design. See below. Am I marketing because other manufacturers and diyers will read and learn? I made this information public 17 years ago.

It is almost as easy as checking out schematics for circuit quality.

I saw conjecture that could be based on some understanding of electrical theory and physics and simple experiments. If you can point me to some peer-reviewed articles that that document experiments – with strong internal and external validity - comparing well implemented PCBs to hardwiring in a range of tube circuits, I will reconsider my position.

First, please do not misrepresent factual measurable experiments as conjecture. The points I addressed above are of concern to all analog designs, which, which contain parts, wires, layouts, and you yourself address below. Of course each layout depends upon the particular circuit.

People, me included, confuse opinion with facts continually; it is our way. Opinions are important, but, we should not accept opinion, no matter how expert, as fact. Perhaps I have put too much emphasis on this point – it is what first irked me about the thread; best to pull my head in and move on…

That is why I perform measurable experiments, to present factual data on subjects, which apparently no one else is performing.

Although not mentioned specifically, I had three things in mind when posting my comment about methods having the benefits of PCB and hard-wiring 1) an amp I am currently planning the layout for, 2) an article in SP about using Teflon sheet for mounting components, and 3) an article about AN build methods from many years ago.

Yes, I already did the research and have white papers addressing the problem, and solution, more than 17 years ago. Each layout will depend upon the actual circuit. I still do not understand why you are agruing against my points, when you are addressing them.

My point about you being an industry participant was in relation to competing interests: potentially promoting your products on a public – not manufacturer – forum. I was not suggesting that you were not walking the talk, which might be what you thought? Anyhow, the information you have shared will be helpful for many.

Notice I did not mention or hint of my company when posting the information. The information is factual, needs to be addressed
in any circuit design.

QuietEarth subsequently posted some specific details of a similar approach used by manufacturers.
I agree with your comments in response to Quiet’s post.

I am sure my work has gotten around since I first posted the information. Other
issues have also been copied by other manufacturers.
Evidently, other manufacturers consider the issue important.

My relatively simple build will use 2mm Teflon sheet, finely drilled to mount components and connect component-to-component where appropriate or using good quality wire where not. I like this approach for the technical reasons discussed, but also for its simple practicality: easier to work on; allows concurrent build of different parts of the amp; promotes tidiness and logical layout; and is very suitable to optimising layout.

Thank you for positively considering the points I have previously mentioned. One help, layout parts for minimum capacitance not only between parts and ground, but also between said parts, separate channels as much as possible.

One further point, cost of hardwiring vs proper Pc board.

Cheers.

[JakeJ]

Gents,

This has been a most interesting thread so far and is an important subject to electronics in general but especially in audio electronics.  Things have been discussed in a most civil manner and I know you will keep it that way.

To Docere's credit he made the simple point that we all make mistakes in interpretation of facts and opinions.

To Steve's credit he has presented his information without any callouts for his company or equipment.  It would be nice to see some hard evidence such as links to published articles and/or comparative images of oscilloscope readings of each design compromise.

Anyone engaging in this discussion is also subject to his or her beliefs.  Lets try to remember that and allow some flexibility where it's needed.

Thanks and carry on.

[Steve]

Gents,

This has been a most interesting thread so far and is an important subject to electronics in general but especially in audio electronics.  Things have been discussed in a most civil manner and I know you will keep it that way.

To Docere's credit he made the simple point that we all make mistakes in interpretation of facts and opinions.

To Steve's credit he has presented his information without any callouts for his company or equipment.  It would be nice to see some hard evidence such as links to published articles and/or comparative images of oscilloscope readings of each design compromise.

Anyone engaging in this discussion is also subject to his or her beliefs.  Lets try to remember that and allow some flexibility where it's needed.

Thanks and carry on.

Thanks Jake. The RCA Radiotron Designers Handbook, written by at least 26 engineers, only covers the problem in terms of
minimizing hum. Otherwise I have not seen any indication that anyone else has pursued the issue at all. Having also worked in the RF field
for a number of years has given me an advantage in that it is clearly understood that even at 50khz or less, wire becomes an "antenna".

The only unknown is how low of frequency is "transmitted" by a wire. I have provided factual data that even at 5khz, a 6" length wire "A"  will "transmit" a signal and be received by a wire, 6" long "B", with a load impedance of only 2.2k ohms to ground. With a higher load impedance, say, 10k ohms or 100k ohms, "B" signal received will be much greater in voltage. With a lower load impedance, the signal received by "B" will
be less.

Cheers. 

[JakeJ]

You're welcome, Steve.

To all parties interested and reading this thread.  Steve Sammet does have more detailed information on his website under the "White Papers" link with some diagrams and data provided.  The articles are fairly short and easy to understand.  I understand why he doesn't provide a direct link as that would raise some eyebrows and cause the possible deletion of the thread.

Everyone here is plenty web savvy enough to find the information and I'd be willing to bet there are other articles out there, although access may not be free.  If anyone can find more written information please post a link to it.

[Docere]

Hi all,

I am breaking my two-posts rule; here goes:

As stated, I have generally agreed with Steve, but disagreed with one of his conclusions and ideology. With my background in clinical/ health research, I reviewed Steve’s posts with the intention of setting a few details straight. What I read surprised me, not because of what Steve stated – he has presented his information in an entirely appropriate way for a forum – but my interpretation and reaction to it. With a jaundiced eye I had initially interpreted his posts as initially claiming that PCBs are inherently, absolutely superior to hardwiring. This line of though also just happened to threaten my hidden beliefs and sense of Self. However, Steve wasn’t. Even if he were, my reactions and strategy would have been counter-productive. This situation is entirely My Bad. Not one of my finer moments, but hey, these past few weeks are littered with such, so what is another?

I sincerely apologise to Steve for assuming his motivations were less than helpful; I also apologise to others in general for (unintentionally) muddying the thread and impairing discussion.

I will be heavily limiting my posting to forums for the foreseeable future and intend to be especially aware of my thinking when I do post. I will leave it there.

Regards.

[dB Cooper]

Great topic, Steve.  I have first hand experience with bundled wiring causing electrical problems so I can relate.

Is there any chance you could provide some oscilloscope screen shots to demonstrate what you are talking about?  We do have some members that steadfastly believe point-to-point wiring is the only way to wire an amp yet there are so few mfrs that do this in production.  Personally I can't think of one right now.  Anyone that can please share that mfrs name.

Seems to me that this is mostly seen in musical instrument amps. 'Point to point' seems to be shorthand for 'We put a lot of TLC into building our amps'. I have even seen cotton insulated wire used as a selling point.