smd vs through-hole

[pucho812]

What is wrong with a modern SMD circuit board?
Modern printed circuit boards are made with SMD components. The problem with a SMD board is that the sound quality of the tiny SMD components are much worse than the old
through-hole components. Plus the SMD board is cheap an tiny, so its now possible to use much more components
and that's what manufacturers do.
Result: a circuit board with to much components which has a bad sound quality.

I found this on the web today.  I'll just leave it here.

 

[john12ax7]

There is a legitimate tech discussion to be had with SMD vs through hole. Some criticisms of SMD are valid,  others not so much.

 

[iampoor1]

There is a legitimate tech discussion to be had with SMD vs through hole. Some criticisms of SMD are valid,  others not so much.

And most of the compromises have reasonable solutions, and/or are easy enough to design around.

The whole smd vs through hole debate is sort of like someone with bad handwriting blamining the pen.

People blaming technology for sound quality need to get their objectivity meter checked out.  8)

 

[JohnRoberts]

I found this on the web today.  I'll just leave it here.

You could have just left it there.... ;D

JR

 

[pucho812]

You could have just left it there.... ;D

JR

But there is no fun in that

 

[john12ax7]

There are people with good ears who may not necessarily understand the tech side of things.  But the part about small SMD sounding worse is actually true,  and measurable. Tiny package thick film resistors with tiny package ceramic (not npo/c0g) capacitors is the recipe for cheap and poorly performing audio circuits.

 

[pucho812]

I would love to see those measurement

 

[john12ax7]

I remember Audio Precision having some nice articles on analog design and component selection. Some component manufacturers have data as well

Here are some real world measurements for resistors

https://groupdiy.com/index.php?topic=9035.0

http://www.davehilldesigns.com/smt_resistror_distortion_rev1.pdf

 

[ruairioflaherty]

I remember Audio Precision having some nice articles on analog design and component selection.

You need to sig up for an AP account to view this but it's well worth it, lots of good info there on Measurement. The Audio Measurement Handbook is in there somewhere too.

https://www.ap.com/download/designing-for-ultra-low-thdn-2/

Saw Hofer give this presentation at AES and it was excellent.

 

[iampoor1]

There are people with good ears who may not necessarily understand the tech side of things.  But the part about small SMD sounding worse is actually true,  and measurable. Tiny package thick film resistors with tiny package ceramic (not npo/c0g) capacitors is the recipe for cheap and poorly performing audio circuits.

Right, but this is easy to measure and easy to design around. Douglas Self covers the resistor issue (albeit briefly) in Small Signal Audio Design as well.

 

[benb]

You need to sig up for an AP account to view this but it's well worth it, lots of good info there on Measurement. The Audio Measurement Handbook is in there somewhere too.

https://www.ap.com/download/designing-for-ultra-low-thdn-2/

Saw Hofer give this presentation at AES and it was excellent.

It indeed says "You must be logged in to download this file." but I clicked DOWNLOAD and it did, without asking me for anything.

 

[john12ax7]

Right, but this is easy to measure and easy to design around. Douglas Self covers the resistor issue (albeit briefly) in Small Signal Audio Design as well.

It's certainly a solvable problem. But good design is not necessarily "easy" 

 

[JohnRoberts]

I would love to see those measurement

For one look at component data sheets..

When I was designing my outlet tester I experienced issues with physically small SMD resistors not delivering rated resistance at high voltage. I had to test at 500V.  But this is not news to real design engineers. Even some through hole resistors have been known to suffer voltage coefficient (resistance changes with voltage) at voltages encountered in the feedback path of a high power audio amplifier.

Surely this phenomenon was known to tube designers (which I am not) who dealt with hundreds of volt rails, if they were looking closely enough (might be hidden by other larger distortion mechanisms).


JR   

 

[QUEEF BAG]

heat dissipation seems like it would be a problem

 

[JohnRoberts]

heat dissipation seems like it would be a problem

The smd resistors are rated for power dissipation just like TH resistors.

Of course higher power resistances are silly using smd.

JR

[edit] I've seen some power amp emitter degeneration resistors use smd...  they tend to vaporize and blow a hole in the PCB when power devices fail short circuit to the rail. [/edit]

 

[iampoor1]

It's certainly a solvable problem. But good design is not necessarily "easy" 

Oh I agree.  I would never imply that a good design is easy. In fact, even a bad design can be hard!  ;D

But I think the issues with SMD components in audio circuits are fairly well documented and shall I say, simple enough to design around? If anything, I think the biggest limitation with SMD is good quality coupling capacitors. Not many film options avaliable. ANd the ones that are....not cheap to put it nicely!

 

[JohnRoberts]

Oh I agree.  I would never imply that a good design is easy. In fact, even a bad design can be hard!  ;D

But I think the issues with SMD components in audio circuits are fairly well documented and shall I say, simple enough to design around? If anything, I think the biggest limitation with SMD is good quality coupling capacitors. Not many film options avaliable. ANd the ones that are....not cheap to put it nicely!

You can get COG/NPO smd caps that are arbitrarily good.  SMD film caps are questionable.

Back a couple decades ago I had smd film caps melt in the reflow ovens. 

JR

 

[zamproject]

Back a couple decades ago I had smd film caps melt in the reflow ovens. 

outch !
so far I don't take the  time to read literature about smd vs TH for audio.
But what about stress and potential chemical/substrate part degradation due to reflow ?
With TH you only heat about 2 second the leg at opposit layer (a little more with wave soldering), but with (no hand soldering) SMD you heat the whole materials for long seconds.

Best
Zam

 

[JohnRoberts]

outch !

The data sheet said the (plastic) film caps should survive the process temperature our contract mfr was using, but in fact I found lots of capacitance value drift and some physical deformation. Luckily now we can buy NPO/COG ceramic caps that are very linear and robust. Easier than arguing with CM about their process.  :

This is not a new problem, back in the 70-80s I loved polystyrene TH caps but they were likewise not very robust, and had trouble surviving mass production processes.

so far I don't take the  time to read literature about smd vs TH for audio.
But what about stress and potential chemical/substrate part degradation due to reflow ?

This is actually a mature technology (SMD has been used for decades). If you think about it the thermals stress from a reflow oven is probably more gentle than running through a wave solder machine where the board and parts go from room temp to hundreds of degrees in an instant. 

With TH you only heat about 2 second the leg at opposit layer (a little more with wave soldering), but with (no hand soldering) SMD you heat the whole materials for long seconds.

There's a reason polystyrene caps used such wimpy wire leads, to not conduct heat into the cap and melt it.  :

Best
Zam

Again this is well understood and with the exception of those problematic plastic film SMD caps, (that may not even be sold anymore), SMD is very robust. To flip this and talk about positives the relatively large solder pad joint vs the much smaller mass component actually reduces physical stress from vibration and normal wear and tear on components. The smaller physical packages allow tighter PCB layouts in some cases reducing critical loop area for noise sensitive circuits (like mic preamps in mixers).

I'm old so the tiny size of modern SMD can be a PIA, but overall the benefits deliver a net positive.

JR 

 

[zamproject]

Thanks for those precisions JR
Best
Zam