R/E/P Community

Please login or register.

Login with username, password and session length
Advanced search  

Pages: [1]   Go Down

Author Topic: Surface Mount components  (Read 5431 times)

gregdubuis

  • Newbie
  • *
  • Offline Offline
  • Posts: 17
Surface Mount components
« on: September 21, 2006, 09:20:15 pm »

Hi everybody!

What do you think about the Surface Mount components?

We can find this components in SSL AWS900/ Neve summing box/ Dangerous Music

I'm asking this because i' ve "talk" with Chris Muth and he give me is version of the SMC stuff.

At the begin this technology was for portable stuff but right now they are using this into High End equipment...
Unfortunately all my friend who buy some of this type of gear have ton of problem (Mic preamp, bus output out of order).

This technology can really replace "normal" discreet components?

Could you please share you're experience into this new way of design.

Best regards
Logged
Greg Dubuis
Sound Engineer & Electrical Engineer
Route de Genève 17
CH-1003 Lausanne (VD) Suisse
Phone:   +41 21 323 08 80
Email:   greg@studioduflon.com

Visit Studio du Flon at http://www.studioduflon.com
And at http://www.myspace.com/studioduflon

Jakob Erland

  • Full Member
  • ***
  • Offline Offline
  • Posts: 203
Re: Surface Mount components
« Reply #1 on: September 22, 2006, 09:16:56 am »

An interesting aspect of this - about resistor noise:

http://www.prodigy-pro.com/forum/viewtopic.php?t=9035

Jakob E.
Logged
..note to self: don't let harman run your company..

Jon Hodgson

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1854
Re: Surface Mount components
« Reply #2 on: September 22, 2006, 09:52:35 am »

gregdubuis wrote on Fri, 22 September 2006 02:20

Hi everybody!

What do you think about the Surface Mount components?

We can find this components in SSL AWS900/ Neve summing box/ Dangerous Music

I'm asking this because i' ve "talk" with Chris Muth and he give me is version of the SMC stuff.

At the begin this technology was for portable stuff but right now they are using this into High End equipment...
Unfortunately all my friend who buy some of this type of gear have ton of problem (Mic preamp, bus output out of order).

This technology can really replace "normal" discreet components?

Could you please share you're experience into this new way of design.

Best regards


The first thing to understand what you are comparing against what, the second to understand what the differences might be.

It isn't a question of discrete vs surface mount, because they aren't mutually exclusive terms, you could have an all discrete circuit which used nothing but surface mount components!

Discrete basically refers to the transistors and other associated components being seperate, rather than all integrated onto one piece of silicon, in simple terms we'll say seperate transistors rather than chips.

Surface Mount refers to the form of the package, be it for a resistor, a capacitor, a transistor, an IC, etc... a surface mount component is designed to sit on the surface of the board, as opposed to traditional components which have wires or pins on them which are poked through holes in the board.

So I guess what you are asking is not about Surface Mount versus traditional through hole packaging.

Well the first thing to understand is that in many cases they are the same device, just packaged differently. Though sometimes something may only be available in one package type, and so if he wants to stick purely to one form (surface mount is better for automatic assembly, through hole tends to be more practical for manual assembly and maintainance for example), then the designer might be forced to change or compromise his design accordingly.

Surface mount offers the possibility of packing things far more densely, not only are the packages themselves generally smaller, but also not having holes going through the hole board every place you need a component gives far more flexibility on routing of pcb traces.

This is generally speaking a good thing, but I can imagine that the closer proximity of components and tracks also generates possibilities for problems due to them affecting each other. If it is the case then this doesn't make surface mount in any way inferior, it just makes care and knowledge in pcb design more critical.
Logged

danlavry

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 997
Re: Surface Mount components
« Reply #3 on: September 22, 2006, 03:36:28 pm »

gregdubuis wrote on Fri, 22 September 2006 02:20

Hi everybody!

What do you think about the Surface Mount components?

We can find this components in SSL AWS900/ Neve summing box/ Dangerous Music

I'm asking this because i' ve "talk" with Chris Muth and he give me is version of the SMC stuff.

At the begin this technology was for portable stuff but right now they are using this into High End equipment...
Unfortunately all my friend who buy some of this type of gear have ton of problem (Mic preamp, bus output out of order).

This technology can really replace "normal" discreet components?

Could you please share you're experience into this new way of design.

Best regards


There is a lot of wrong information floating around in audio, and the through hole vs surface mount discussion is not spared.

Generally, a good way to view components is "one at a time", without any pre conceived notions. The comparisons between parts should be done AFTER you study the parts, and it does depend on a specific use, and on a case by case basis.

There are few "formulas" or words of wisdom that hold a real examination. For example, You will probably no find a surface mount resistor that can handle say 10 watts of power. Power means heat generation, and dissipating 10 watts in a tiny volume will cause huge temperature rise. But how often do you dissipate that much? Each resistor in a circuit dissipates some known amount (it depends on the specific circuit). Say you have a 5V on a 10KOhms, then the current is .0025 watt. Even the tiny surface mount resistors can handle much more then that with no noticeable change in behaviour. Aside from resistor power, there is current handling capabilities, maximum operating conditions (such as temperature) and more.

Similarly, capacitors should be viewed according to their capabilities. From questions about voltage, ripple current, internal impedance to the more subtle issues such as dielectric absorptions, frequency characteristics and more... At the end of the day, you find the same various materials in both through hole and surface mount.

With semiconductors, there is some wide spread but misguided belief that the bigger through whole part will run cooler. It is not always the case. The semiconductor chip itself is "wired to the package pins via very thin(!) wires and the length of the wires does make a difference, the same way that the length of the leads of a transistor does. Often, a surface mount "pin" can be coupled to some copper area (pad, or ground plane or power plane) to conduct heat better then the old through hole...

Surface mount certainly has the advantage in high speed applications, where reducing size amounts to much lower capacitance and inductance (which tend to stand in the way of fast voltage and current changes).

I could go on and on. But to simply state it: I would strongly suggest to avoid making judgements based on the choice of packaging. True, you are unlikely to find a GHz system based on through hole, and unlikely to find a power amp speaker drive based on tiny surface mount transistors, but there is a lot of middle ground in between. The choice of each component in each individual circuit is based on numerous factors, from material to placement, from reliability to environmental conditions, from distortions to price... This is the domain of the designer, and coming at it from the outside with general statements is going to get you into wrong conclusions any day of the week.

Design of gear requires following and forcing "signal flow" in circuits. Doing the job well means that there are no shortcuts to studying and understanding the various factors that impact the signals. That requires knowledge about which factors apply in each and every case. For example, when working with say 15V range, I already know that I do not need to care about 100KOhms resistor heating up. But if I wanted to use 100 Ohms value, then I need to address the issue of heat...  

One can not "shortcut" years of study and practice with a simple statement such as "through hole" sounds better. Such statements are NOT scientifically based, and the "data" supporting it is no better then "I hear it" with no real ABX tests to support it, and certainly with no measurement data to support it.

I am not aware of a single case where a proper repetitive ABX was administrated to show that through hole is better or worse then surface mount. Doing such a test is ridicules. Such a test could not be done by just replacing through hole with surface mount parts. The decision to use one type or the other may be based on good rational that applies to that specific design. A house meant to be made out of wood may not stand up when you make it out of rocks...

Regard
Dan Lavry  
http://www.lavryengineering.com


Logged

gregdubuis

  • Newbie
  • *
  • Offline Offline
  • Posts: 17
Re: Surface Mount components
« Reply #4 on: September 23, 2006, 05:29:55 am »

Many thanks for all your answer.

I'm sure that it's possible to do a really good circuit with only
SMC components because quality SMC exist.
But in fact the "problem" is that the designer can be tempt to use
standard or low range SMC components...

Like i say in my first post, when you buy an AWS900 and then three month later ten mic preamp and bus output are out of order, it's not really normal! And I have other example to share.

But anyway thanks again for sharing your experience.
Best regards.

Greg

Logged
Greg Dubuis
Sound Engineer & Electrical Engineer
Route de Genève 17
CH-1003 Lausanne (VD) Suisse
Phone:   +41 21 323 08 80
Email:   greg@studioduflon.com

Visit Studio du Flon at http://www.studioduflon.com
And at http://www.myspace.com/studioduflon

Jon Hodgson

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1854
Re: Surface Mount components
« Reply #5 on: September 23, 2006, 06:50:03 am »

gregdubuis wrote on Sat, 23 September 2006 10:29

Many thanks for all your answer.

I'm sure that it's possible to do a really good circuit with only
SMC components because quality SMC exist.
But in fact the "problem" is that the designer can be tempt to use
standard or low range SMC components...


As he can with through hole components, there is really no difference in this respect
gregdubuis wrote on Sat, 23 September 2006 10:29


Like i say in my first post, when you buy an AWS900 and then three month later ten mic preamp and bus output are out of order, it's not really normal! And I have other example to share.



Dan and I were talking about circuit quality in terms of audio quality, not longevity. If you have circuits actually failing then this is a manufacturing problem, and is independent of the choice of package type. Something went wrong in the factory where the boards or possibly a batch of components, were made. If the boards have been manufactured correctly then their life expectancy would be pretty much identical.
Logged

danlavry

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 997
Re: Surface Mount components
« Reply #6 on: September 25, 2006, 04:17:25 pm »

Jon Hodgson wrote on Sat, 23 September 2006 11:50

gregdubuis wrote on Sat, 23 September 2006 10:29

Many thanks for all your answer.

I'm sure that it's possible to do a really good circuit with only
SMC components because quality SMC exist.
But in fact the "problem" is that the designer can be tempt to use
standard or low range SMC components...


As he can with through hole components, there is really no difference in this respect
gregdubuis wrote on Sat, 23 September 2006 10:29


Like i say in my first post, when you buy an AWS900 and then three month later ten mic preamp and bus output are out of order, it's not really normal! And I have other example to share.



Dan and I were talking about circuit quality in terms of audio quality, not longevity. If you have circuits actually failing then this is a manufacturing problem, and is independent of the choice of package type. Something went wrong in the factory where the boards or possibly a batch of components, were made. If the boards have been manufactured correctly then their life expectancy would be pretty much identical.



I agree. Components falling off sounds real bad. I do not recall any surface mount part falling off on my products. Of course, a poor design can cause a failure of a device device to heat up to the point of causing solder meltdown, or "burning" a PC board or a mechanical breakdown... But that is not restricted to surface mount.

As I said, each part has to be designed with a lot of care and attention to the details that count in the SPECIFIC case. A lot of surface mounts are very light weight and the solder is more then enough to hold them even with many G's of acceleration. Of course I would not consider a 1 lb transformer with 4 tiny surface mount pins... I do not think you can find such device...

Also, a through hole has it's own set of reliability issues, especially when the PCB is multi-layer. Each hole has to connect to other traces, often in an inside layer, thus a multilayer through hole should not be flexed too much... Surface mount does not have that issue. (Do not flex any boards, surface or through hole, there are other issues about flexing...).

There are times when I choose through hole. All my switches and connectors are through hole, because unless one can securely fasten a surface mount part to a front or rear panel, I would get nervous about the user applying too much mechanical pressure on a trace. That does force me to have a "mixed" board (both through hole and surface mount). A mixed board cost more to produce, you "run it twice", once for surface mount, once for through hole. So there may be a "temptation" to use only one type packaging.

It is a fact of life that anyone interested in staying with through hole technology, is hanging on to old ways, fighting a losing battle. The world has been moving to surface mount. Show me an IC such as an AD or a DA or a sample rate converter in through hole, and I will show you a poor outdated converter...
And as a converter designer, I can tell you that the small size of bypass capacitors is of major important, and that some linear regulator IC's run cooler with surface mount package, due to better coupling to a copper pad or a ground plan.

When surface mount first came in, I was a little unhappy, because it made prototyping and bread boarding less accessible to the little guy. This still the case, and I try to get through hole parts for breadboards. This days, there are a lot of surface to through hole adaptors, and it is still a pain to deal with. But trying to hang on to the fading past is a serious limitation on a designer options.

And since it is the audio industry, you will always find someone telling you that the old stuff sounds better, with no data, no double blind ABX. And again, a listing test in this case is totally inappropriate, because it compares a design that is optimized for one technology against itself using a technology it was not designed for...

The decision of packaging should be left to the designer. It is a decision best made on a case by case basis!    

Regards
Dan Lavry
http://www.lavryengineering.com

Logged

gregdubuis

  • Newbie
  • *
  • Offline Offline
  • Posts: 17
Re: Surface Mount components
« Reply #7 on: September 25, 2006, 05:05:59 pm »

So many thanks again!

You have conciliate me with SMC components.
I'm also a electrical engineer so i can understand all the quality of SMC!
Like no wire so less inductance, low temperature so less thermal transient distortion etc.

So the problem is only due to the topology of the board... I hope they gonna do something for that.

I would like to thank also Dan Lavry because we have one of your incredible converter.
You do really a great job, we had make a lot of blind test and he win all the test.

Best regards.

Greg

Logged
Greg Dubuis
Sound Engineer & Electrical Engineer
Route de Genève 17
CH-1003 Lausanne (VD) Suisse
Phone:   +41 21 323 08 80
Email:   greg@studioduflon.com

Visit Studio du Flon at http://www.studioduflon.com
And at http://www.myspace.com/studioduflon

Jon Hodgson

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1854
Re: Surface Mount components
« Reply #8 on: September 26, 2006, 08:17:37 am »

gregdubuis wrote on Mon, 25 September 2006 22:05

So the problem is only due to the topology of the board... I hope they gonna do something for that.



Bad sound quality I would put down to probably being topology or poor component selection.

Circuit failure I would say is probably a fault in manufacture, maybe a dodgy batch of components, maybe the solder was contaminated, maybe the temperature control wasn't working correctly... you can't know without finding out exactly what the failure was.

The first requires a redesign, the second better quality control (and occasionally a redesign if it turns out that the placement of components compounds the manufacturing problem). Of course no quality control can every be 100% successful (for example if you test every unit for 100 hours before sending it out, sooner or later you'll encounter a unit that fails after 101 hours), so you may just have been unlucky.

Have you had any report on what the actual failures were in this SSL unit you refer to?
Logged

danlavry

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 997
Re: Surface Mount components
« Reply #9 on: September 26, 2006, 02:30:06 pm »

gregdubuis wrote on Mon, 25 September 2006 22:05

So many thanks again!

You have conciliate me with SMC components.
I'm also a electrical engineer so i can understand all the quality of SMC!
Like no wire so less inductance, low temperature so less thermal transient distortion etc.

So the problem is only due to the topology of the board... I hope they gonna do something for that.

I would like to thank also Dan Lavry because we have one of your incredible converter.
You do really a great job, we had make a lot of blind test and he win all the test.

Best regards.

Greg




Thanks for the kind comment.

I would like to mention a couple of other points here from the perspective of a printed board design and manufacturing.
Most electronics designers pass on the schematic to a person doing printed circuit layout. I do not have a PCB layout person, I have done 100% of the printed circuit layout. Layout takes a lot of time, but I keep my hands on both schematic and layout because it yields best results.

I started laying board around 1990, after about 20 years of passing a schematic to a PCB layout person. I find that the amount of "specialized instructions" needed for layout is so large, that it is simpler and better to just do it myself. What I mean by "specialized instructions" ranges from where to place the parts and traces relative to each other due to unwanted crosstalk, heat considerations, trace length and width, grounding and power distribution... and so on... Layout is particularly important in mixed signal applications (both analog and digital on the same board (or chassis).

I mention the above to give my next comments the credibility of hand on experience, so here we go:

When I started doing layout, most of the parts were still through hole. Like many others, I was under the somewhat misguided "gut feel" that the much smaller parts could yield a much smaller board. Yes, surface mount yields smaller boards, but the size reduction is not really proportional to the size of the parts. With parts on one side of the board, and traces on both sides (or more in the case of multi layer board), one needs to have a lot more feed through holes to connect between planes.
Say you have a 50 pin part in through hole technology, and a double sided board. One can both "start and stop" a trace on either side. In the case of multilayer, one can put traces on additional planes, and keep them there.
But with surface mount, a trace that is not on the top layer (component layer) needs to be "sent down" via a connecting plated through hole, and then "sent back up" through another hole.
Also, while surface mount drove printed circuit technology to higher precision (narrower traces and separation between traces), one can not always take advantage of such capability. Many traces needs a lot of separation (for cross talk), some traces need more width (for higher current) and so on.
I agree that a surface mount enables smaller space, but it is less so then expected.

Regards
Dan Lavry
http://www.lavryengineering.com      
Logged
Pages: [1]   Go Up