DIYHiFi.org

For the sake of audio
It is currently Sun Dec 17, 2017 2:20 am

All times are UTC




Post new topic Reply to topic  [ 12 posts ] 
Author Message
PostPosted: Wed Oct 15, 2014 3:15 am 
Offline
Benjamin
User avatar

Joined: Mon Dec 13, 2004 3:29 pm
Posts: 4333
Location: Somewhere other than here.
This has nothing to do with our usual nonsense, but I am adding it, as an instructive rant, to the folks who may be reading this crap, and wondering how any of it makes sense.

I can not go into details (may not be NDA, but I still won't give any details), but a bidnis associate sent me something to measure. Actually, two of them. Identical, whatever they are.

One sounds great, and the other much less than great. He is frustrated and annoyed. So, he sends both of them to me, hoping with my different type of gear, and different ways of looking at stuff (digital vs RF), I can help him out.

Nope.............can't help him out. They measure identical. At least from any way I can think of looking at them. My measurements can also verify component values (he was concerned the folks who do the assembling may have screwed up), but all the components are the right value, and in tolerance.

So..........what to do?

I dunno..............except maybe send them to some hobbyists, who can more accurately describe how they sound different........and then make a standard distribution curve.............and that will tell us..............

Uh, that they sound different!

We already know that.

If we just had an F-word, we could solve this problem..............................

Actually, we do have an F-word. It is the one everyone uses, and it is appropriate at times like this.

Translation: if you have ever done this for a living, you will have gone through this, many times. And understand how frustrating it is.

Measurements are just like any other tool. They are useful for the right job, and at the right time. And not, at most all other times. If you don't believe me.....................trying fixing a leaky toilet with a spectrum analyzer some time. I would like to see that.


Top
 Profile  
Reply with quote  
PostPosted: Wed Oct 15, 2014 11:04 am 
Offline
Dog
User avatar

Joined: Tue Dec 14, 2004 12:26 pm
Posts: 4981
Location: Lisbon, Portugal
Jocko, I'm intrigued.
Are both circuits exactly the same?
I mean, same PCB, same passive and active components, etc?
Because one different cap in the right place can have some sonic impact.
Some small SMD caps have no identification, so you don't exactly know what's there.
But the thing is, although it can have some sonic impact, the whole thing may measure the same...
A leaky toilet seems like an easier problem to solve. :D

_________________
Carlos Filipe

"Any intelligent fool can make things bigger, more complex, and more violent. It takes a touch of genius - and a lot of courage - to move in the opposite direction." Albert Einstein


Top
 Profile  
Reply with quote  
PostPosted: Wed Oct 15, 2014 1:38 pm 
Offline
Goat

Joined: Sat Mar 04, 2006 5:26 pm
Posts: 576
Location: Nova Scotia
I'd start swapping major parts between the two units and see if you can "swap the sound" in the process.


Top
 Profile  
Reply with quote  
PostPosted: Wed Oct 15, 2014 6:26 pm 
Offline
Benjamin
User avatar

Joined: Mon Dec 13, 2004 3:29 pm
Posts: 4333
Location: Somewhere other than here.
It only has passive components. All the parts came from the same "bag o' stuff" and both were assembled at the same time. And same PCB, in case that was not stated.

We just need an F-word, because through the collaborative process, us professionals (who are in fact idiots), the solution will emerge.

Right?


Top
 Profile  
Reply with quote  
PostPosted: Wed Oct 15, 2014 8:29 pm 
Offline
Muriel
User avatar

Joined: Thu Jul 26, 2007 5:45 pm
Posts: 1167
Jock,

Jocko Homo wrote:
One sounds great, and the other much less than great.

If we just had an F-word, we could solve this problem...


Problem is probably the PCB itself... Run into that a few times. No way to confirm, but we moved ALL parts from a "bad" PCB to a "good'un" (no fun for the junior doing it with around 800 SMD parts) and the problem remained with the PCB itself.

Just make sure QC can weed them out and change the PCB vendor.

Ciao T

_________________
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." Richard Feynman


Top
 Profile  
Reply with quote  
PostPosted: Wed Oct 15, 2014 8:44 pm 
Offline
Pig

Joined: Wed Jun 29, 2005 6:34 pm
Posts: 1068
Location: East Coast USA
I, ahh, can't tell you why I know this, but...

Basic FR-4 boards are not all the same. The ratio of glass to epoxy, the orientation of the glass fibers, and all the things you might expect from making a stew of various stuff means that the dielectric variation across a board as well as between different boards of even the same lot, will be non-trivial. This is especially true for frequencies above several MHz.

This is one of the reasons why some companies choose to invest in more exotic base materials for their pc boards. Consistency matters.

For the same reason, buried traces for rf signals may not always be a great idea.

Bad chemistry.

I should add that most of the measurements made by pc board suppliers that are supposed to ensure goodness in their products may not be entirely applicable to what you are doing.

Now, if they had an f-word of some kind, things might be different.


Top
 Profile  
Reply with quote  
PostPosted: Thu Oct 16, 2014 2:49 pm 
Offline
Muriel
User avatar

Joined: Thu Jul 26, 2007 5:45 pm
Posts: 1167
Hi,

CG wrote:
Basic FR-4 boards are not all the same. The ratio of glass to epoxy, the orientation of the glass fibers, and all the things you might expect from making a stew of various stuff means that the dielectric variation across a board as well as between different boards of even the same lot, will be non-trivial. This is especially true for frequencies above several MHz.


Yup. Also, don't forget FR-4 is hygroscopic and manufacturing processes involve liquids...

This is one of the reasons why i really do not like to use striplines on PCB's, just move the components so they so close you don't need no steekin' stripline.

Then again, if you look at PC Motherboards and the routing there, holy moley. If you have a bad board the PC keeps crashing and no-one is any smater why.

Could just go for Aralon, but a little common sense in design can help tremedously

Ciao T

_________________
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." Richard Feynman


Top
 Profile  
Reply with quote  
PostPosted: Thu Oct 16, 2014 7:11 pm 
Offline
God

Joined: Thu Dec 30, 2004 10:50 pm
Posts: 998
Location: Minnesota
More PCB stories;
Made the mistake of ordering a batch of FR4 boards from Malaysia. I used a vendor on the advice of a friend. These were for a preamp design.

Built the boards and some of them just did not work; oscillate, not stable, signal dropping in/out. All the nasty stuff you can think of.

Last time I make that mistake.

This was only 4-5 years ago.

We take the board for granted way too many times......

Mike


Top
 Profile  
Reply with quote  
PostPosted: Fri Oct 17, 2014 12:16 pm 
Offline
Goat

Joined: Sat Mar 04, 2006 5:26 pm
Posts: 576
Location: Nova Scotia
I've designed plenty of things with high speed "transmission-liney" interfaces on FR4 - RF, high speed ADC/DAC, SDRAM, etc... and have had no issues at all that could be blamed on impedance tolerance. Usually when a board doesn't work and the PCB is to blame, it's a hard fault - open vias/traces, a flake of copper connecting two adjacent traces, solder mask faults leading to opens/shorts, etc.

Main thing is to specify the material's dielectric constant. On boards I've designed where it's critical (directional couplers, RF splitters/combiners, amplifiers, etc) and I've got room to, I'll put down a dielectric test pad. If the capacitance of the pad doesn't meet spec, get mad at the board house.


Top
 Profile  
Reply with quote  
PostPosted: Fri Oct 17, 2014 9:47 pm 
Offline
Sheep

Joined: Fri Dec 12, 2008 2:05 pm
Posts: 82
The scenario gmarsh describes sounds more plausible to me than the field effects of an inconsistent dialectric but in high speed circuits with micro-currents, I can believe almost anything. I think you guys finally scared away yldouright or maybe you just beat him into submission but speaking from my own point of view: I've built amps. More than fifty, less than a huindred but still...some of my builds sound righter than others no matter how closely I match the parts and no matter what time of day I play them so it isn't just a quieter mains or something like that. To me, this is a solvable mystery but access to the gear to snoop down the differences is a problem for the majority of diy guys. Maybe it doesn't matter for you guys but I would benefit from someone who could show me a repeatable pattern of what electrical attributes matter the most to what I'm hearing. Not looking to make enemies, just sayin'....


Top
 Profile  
Reply with quote  
PostPosted: Fri Oct 17, 2014 10:30 pm 
Offline
Benjamin
User avatar

Joined: Mon Dec 13, 2004 3:29 pm
Posts: 4333
Location: Somewhere other than here.
Those of us who do have the gear have looked. Really, we have. No one has found it.

Maybe, some day, someone will. I do know who it won't be, though.

Of course, he is free to look. Hope he won't blame us when he doesn't find it. (Blaming us for somehow stopping him from finding it, since we won't join it, is where I draw the proverbial "line in the sand".)

But even when someone (like me) can show a direct correlation between measured jitter, and anything in the realm of perceived sonic benefits is also roundly criticized. I don't let it stop me, though.

Doesn't stop them from trying to replicate what I can do. Since they can not replicate it, that means I am full of prunes.


Top
 Profile  
Reply with quote  
PostPosted: Sat Oct 18, 2014 2:22 am 
Offline
Pig

Joined: Wed Jun 29, 2005 6:34 pm
Posts: 1068
Location: East Coast USA
gmarsh wrote:
Main thing is to specify the material's dielectric constant. On boards I've designed where it's critical (directional couplers, RF splitters/combiners, amplifiers, etc) and I've got room to, I'll put down a dielectric test pad. If the capacitance of the pad doesn't meet spec, get mad at the board house.

I've done that, and more.

At one point I was so frustrated that I put down two microstrip traces the length of the board, one on the "left" side and one on the "right", nominally identical, with SMC connector pads at each end of both. Just to make measurements.

A network analyzer sweep from 10-1000 MHz usually gave different results between the two identical traces. And, not consistent from board to board. The board fab house's measurements demonstrated conclusively that the traces were identical. So much for universal testing frameworks...

The more knowledgable people in the company didn't want to use a different board material - even at close to the same price - because they didn't believe it to be necessary. This despite the data sitting before them. After all, they all had MBA degrees. So, I changed the design to put a coaxial connector at the back end of the board and we used a coaxial cable instead of a board trace. (This was used for a front panel test point, so it needed to be reasonably accurate.). Big commercial success.

So, if you haven't had a problem like this, consider yourself fortunate and stick with that vendor. Or use a better dielectric that isn't sourced by a thousand different chemists. The BOM cost might go up some, but the yield does, too. (Funny how the design engineer is always responsible for the cost, but the yield is always to the operations guys' credit...)


Top
 Profile  
Reply with quote  
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 12 posts ] 

All times are UTC


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB® Forum Software © phpBB Group