Ping Caelin Gabriel RE: Design of Experiments for audio applications

[PHC1]

Originally Posted by Puma Cat

Well...I think you're gonna like it!

I am definitely curious as to the outcome.

[Puma Cat]

You can write up another one of your great reviews!

[crwilli]

Well done Stephan. I used to use Taguchi methods and mini tab software to perform DOE in the lab. Wow, that was many years ago and Would never have thought to apply it to audio questions.

Perhaps I can apply it to setting up my new room. # and location of panels, Location and settings of sub... hmmm

[mchydro]

Originally Posted by PHC1

I am definitely curious as to the outcome.

Serge....I'll be very interested in your opinion. Please share your thoughts when you have a chance. [emoji106]

[PHC1]

Originally Posted by mchydro

Serge....I'll be very interested in your opinion. Please share your thoughts when you have a chance. [emoji106]

To be honest, I already use the Guerrilla and Zebra power cords that are equipped with the oversized Wattgate terminations and are the size of a healthy snake specimen. They are also well shielded and sound great FWIW and as far as power cords go. Back in the day I exhausted myself comparing back and forth and at the end of the day these two stood out as absolute bargains. Sadly neither company is around because the audiophiles do not take things seriously unless they write a check with a more substantial amount.








I hate comparing power cords as it goes against my inner logic and I was happy to hear what I thought was an adequate improvement at that time and for the money. Was it real or imaginary, I can't tell you because switching power cords in and out and comparing is the silliest thing in my mind. The moment you switch your amp or preamp off, the thermal stability it lost and in the time it comes back up to optimal performance, the memory of what you heard 10-15 minutes earlier can be very deceiving.

Having said that, I will give it my best try. I simply needed another power cord and the Shunyata is no doubt just as worthy as anything else.

[CGabriel]

Thank you, Puma Cat. The DOE is indeed an interesting tool that can help to objectify and correlate multiple parameters. I would say that this is similar to how I design products except the data points are in my head based upon many years and thousands of experiments.

In the specific case of power cord design, there are several parameters that affect the final performance. But the two primary elements are DTCD and EMI/RFI effects. There is a direct correlation between DTCD measurements and the listeners perception associated with dynamics and bass power. Of course, there is a point of diminishing results as DTCD approaches specific levels. Conversely, most filtering methods have a negative impact on DTCD. But, improved EMI/RFI performance has perceived sonic benefits associated with terms like “grain, grit and harshness”. So you can see that there is a “push-pull” relationship between EMI/RFI (noise reduction) and between DTCD (instantaneous current).

Of course, there are other design parameters that must be taken into account for a specific end product. But these are two that are easily measured and have direct correlations to what people hear.

[Puma Cat]

Originally Posted by crwilli

Well done Stephan. I used to use Taguchi methods and mini tab software to perform DOE in the lab. Wow, that was many years ago and Would never have thought to apply it to audio questions.

Perhaps I can apply it to setting up my new room. # and location of panels, Location and settings of sub... hmmm

Ah, yes, Tacuhi methods...they are very good for doing "Robust Design" DOEs, that is, making a functional response immune to the influence of noise. Yes, I've used MiniTab, but JMP has far exceeed when it comes to performing DOEs.

And yes, you can do a DOE on setting up your room if you can get ahold of a calibrated mike and REW (Room Eq Wizard). If so, I can help you design the DOE and it would be fun to do the analysis once you have the data.

[Puma Cat]

Originally Posted by CGabriel

Thank you, Puma Cat. The DOE is indeed an interesting tool that can help to objectify and correlate multiple parameters. I would say that this is similar to how I design products except the data points are in my head based upon many years and thousands of experiments.

In the specific case of power cord design, there are several parameters that affect the final performance. But the two primary elements are DTCD and EMI/RFI effects. There is a direct correlation between DTCD measurements and the listeners perception associated with dynamics and bass power. Of course, there is a point of diminishing results as DTCD approaches specific levels. Conversely, most filtering methods have a negative impact on DTCD. But, improved EMI/RFI performance has perceived sonic benefits associated with terms like “grain, grit and harshness”. So you can see that there is a “push-pull” relationship between EMI/RFI (noise reduction) and between DTCD (instantaneous current).

Of course, there are other design parameters that must be taken into account for a specific end product. But these are two that are easily measured and have direct correlations to what people hear.

Thanks, Caelin for the reply.

From your description of filtering for RFI/EM, it seems there is a "negative interaction" between DTCD and filtering functionality. DOE is a nice approach to characterize this because it will tell you the degree of impact of the interaction as you can look at the "sums of squares" in the variance components analysis to assess the degree of impact of each. The prediction profiler can then tell you where to set the "knobs" of your control factors to maximize both.

Cheers,
Stephen

[PHC1]

Originally Posted by CGabriel

Thank you, Puma Cat. The DOE is indeed an interesting tool that can help to objectify and correlate multiple parameters. I would say that this is similar to how I design products except the data points are in my head based upon many years and thousands of experiments.

In the specific case of power cord design, there are several parameters that affect the final performance. But the two primary elements are DTCD and EMI/RFI effects. There is a direct correlation between DTCD measurements and the listeners perception associated with dynamics and bass power. Of course, there is a point of diminishing results as DTCD approaches specific levels. Conversely, most filtering methods have a negative impact on DTCD. But, improved EMI/RFI performance has perceived sonic benefits associated with terms like “grain, grit and harshness”. So you can see that there is a “push-pull” relationship between EMI/RFI (noise reduction) and between DTCD (instantaneous current).

Of course, there are other design parameters that must be taken into account for a specific end product. But these are two that are easily measured and have direct correlations to what people hear.

For many years I used various monoblock amplifiers that were close to speakers in which case the EMI/RFI has little to no concern to the rest of the components at a significant distance. In that case a monoblock power cord optimized for current delivery would be better in my mind. The rest of the power cords for the other components can then be optimized for shielding. Just a thought.

[CGabriel]

Originally Posted by PHC1

For many years I used various monoblock amplifiers that were close to speakers in which case the EMI/RFI has little to no concern to the rest of the components at a significant distance. In that case a monoblock power cord optimized for current delivery would be better in my mind. The rest of the power cords for the other components can then be optimized for shielding. Just a thought.

Good observation. Power amplifiers predominate in the need for DTCD since they are the last link in the signal amplification chain. The signal levels are high, compared to something earlier in the signal path (think phono pre). Further, amplifiers must transfer “power” (voltage x current) not just amplify a signal voltage.

A source component, like a CD player may only need one amp of continuous current. However, it has many circuits that are susceptible to noise while at the same time it “generates” noise that may be conducted out the power supply onto the power line. This means that DTCD would be less important while CCI and EMI immunity will predominate. Noise levels in source components will be amplified by components later in the signal chain.

This is why the ideal power cord for an amplifier ‘may’ not be be best for a source component or digital component. And vice-versa. Also if you are testing or evaluating a new power cord, it is best to test it in with a the specific component it will be mated to. Not quite as dramatic an interaction as between amplifiers and speakers but similar in concept.