advertisement


PCB track shapes?

Not too sure about shapes, but composition of the PCB can make a big, big difference. Recently we got two sets of otherwise identical boards made for a power-amp. One was conventional fiberglass and low cross sectional area tracing...the other was Teflon with 3-oz copper traces. Having two amps side by side with the different boards revealed pretty big differences...the Teflon/3 oz copper traced one sounded much more open and organic, in addition to having a better sense of grip and LF dynamics...the fiberglass board had a subliminal but pervasive hashy background sound. The problem is the fiberglass board cost around $60, the Teflon one was closer to $800. BUT...this was the same order of magnitude as a black-box power supply upgrade. Note that the ground traces of the fiberglass board were reinforced with hotwire links to give a lower impedance, while the 3 oz copper traces on the Teflon boards were left as-is.
Teflon boards are mil-spec as they resist corrosion and degradation due to UV/cosmic rays (hence they are found in satellites). But effectively a PCB acts as an interconnect...and there are no credible interconnects made with low purity metals that conventional PCB traces use, or employ high dialectric fiberglass (which is largely responsible for capacitive coupling between PCB traces). Whereas any self respecting interconnect uses at least high purity copper conductors, with the best using Teflon dialectrics.
 
Was this a mega power kW amp or a serious bit of uhf kit?

Teflon is used in satellites because it doesnt gas out like fibreglass.
 
I think there is something else at work, fibreglass boards are good to at least 10Gbits/sec on a transmission line, maybe you have marginal stability.
 
russel said:
I think there is something else at work, fibreglass boards are good to at least 10Gbits/sec on a transmission line, maybe you have marginal stability.
Click to expand...
I don't think its a matter of stability....anymore than stability is a factor why a cable with PVC dialectric sounds different/worse than one with Teflon. I once got a spool of MIT3300 interconnect wire at a time when it was held in high regard. I found out that it sounded much better completely removing the fawn colored dialectric that enshrouded it.

The Teflon PCB was a one-off experiment-or so I thought. Now we are building a preamp using a Teflon-3 oz traces PCB as well. For a production run of 2 boards it cost almost $1000 per board, but initial listening suggests it was money very well spent.
 
3oz copper is 100um and 3x the normal thickness, this will make an improvement if your layout puts class AB switching current into the tracks.
The difference between FR4 and a good Rogers RF material at 100MHz is only the wider track for 50 Ohm impedance, thanks to the lower dielectric constant. The lower losses are barely measurable below 500MHz
 
image000000.jpg

I wonder how many people have actually A/B compared otherwise identical PCB varying only composition. I know objectively you would expect differences in the RF spectrum. But still........
 
You haven’t done a comparison with composition difference only because you say the fibreglass board has a low cross sectional area copper track, so the track resistances will be different.
 
What's dialectric? Something to do with dialects?

Maybe you should try FR2:
https://db-electronic.com/en/pcb-manufacturing_s35.htm

On a more serious note - the design of the PCB will have vastly more effect than the materials at frequencies that matter for audio:
https://www.edn.com/electronics-blogs/bakers-best/4426162/PCB-signal-coupling-can-be-a-problem
Keep your impedances low enough, use the correct spacing between tracks, keep VHF out. Be cautious with 2+ layer boards. Have a look at (top-end) Naim preamp pcbs, they do all of these things for a reason.
 
russel said:
You haven’t done a comparison with composition difference only because you say the fibreglass board has a low cross sectional area copper track, so the track resistances will be different.
Click to expand...

Yes, but I did give the fiberglass board the benefit of extra hotwires of maybe 16ga wire for the important ground paths. Combined with the 1 oz traces this is probably even lower impedance than the 3-oz traces in the teflon boards.
And besides, the cost is already prohibitive enough to allow trying every little variation.

As far as top-end Naim preamps go......I owned a 552 for years and this preamp is leagues ahead of it.
 
sq225917 said:
Through hole none soldered it's all the rage.
Click to expand...
Via hole plating thickness is another variable, normal decent quality FR4 PCBs can be as low as 20um plating, cheap boards less than this, as plating the holes is a slow process and time costs money.
 
rontoolsie said:
image000000.jpg

I wonder how many people have actually A/B compared otherwise identical PCB varying only composition. I know objectively you would expect differences in the RF spectrum. But still........
Click to expand...

It also looks as if you have not star earthed the PCB which will have a highly detrimental effect on sound quality. You need to get the basics right before you start faffing about with esoteric materials.
 
You must log in or register to reply here.


advertisement


Back
Top