Hey,

I was just wondering if there are sound quality difference between the same spec's surface mount components as opposed to their through-hole counterparts? Is there a preference towards through-hole?

Through-hole packages may apply a bit less stress to the die resulting in slightly lower noise. Metal can packages tend to isolate the die quite well so you'll see some seriously high performance op-amps in a metal TO-something case.

Also, through-hole packages allow a bigger die so high-performance BJT's tend to come in through-hole packages - except RF, where SMT is pretty much dictated by the frequencies in use. SMT resistors and capacitors perform better in most ways - especially in the high frequency regions of the spectrum (above about 40 or 50 MHz they start to shine). But the lowest noise resistors and caps in the audio region appear to be through-hole in most cases.

i've always dismissed surface mount audio as cost-cutting inferior quality, but at the same time have worked on some very high end surface mount rf products.

dale brought up a lot of good points. in the case of a resistor noise will be less in thru-hole (noise goes down as they get bigger, i.e a 1 watt resistor has less noise than 1/2 watt) but at the same time a thru-hole component will be unusable at rf frequencie due to the parasitics.

never actually did an a/b listening test, but now i'm curious to try.

I have found the availability of my favorite film capacitors dielectrics harder to find, as some plastic films are not robust enough for reflow temperatures, but in general the electrons don't know the difference. There are probably good and bad parts in SMT just like thru hole.

In the early days of SMT you couldn't get some of the obscure parts in SMT packages but it seems much better now.

Noise sensitive circuits, like mic preamps seem to benefit from the tighter component layouts possible using SMT.

JR

Surface mount metal oxide resistors are noisy and have poor linearity. They also drift with temperature. Monolythic ceramic capacitors have high DA and also have temperature drift problems. Both sound like crap.

There are good metal film 1202 surface mount reistors available for low cost. Good luck getting audio designers to use them.

Caps are different, there are NO GOOD surface mount audio grade caps available. All the surface mount films are mylar or another fomula that can tolerate the soldering heat. There are no surface mount polystyrene or polypropylene film caps available. Wima tried and failed, the plastic melted during reflow soldering.

Vishay makes the best resistors, the bulk foils. They are also available in surface mount at a cost of $15 each. Now you know why you havn't seen those yet.

Jim Williams
Audio Upgrades

That depends on if you use the wider bandwidth to push the noise up in frequency, though I think you'd have to change your design approach quite a bit. You have to become an RF designer instead of an audio designer up there, and my RF skills are a bit spotty. Right now I'm doing a layout of an all-SMT engine management computer, and there's a bit of RF management you need to do but it's not that bad.

No I am talking about small consoles/mixers where the input circuitry can be located literally underneath the input jack. Larger thru hole components end up on the same side of the PCB and next to the jacks since two objects can't occupy the same physical space.

This is perhaps more of a benefit to a certain class of product, but most SMT layouts will be smaller and tighter. Even just the parts being smaller, means they make smaller antennas for noise pickup.

Like i said before I doubt the electrons can really tell a difference, so it will hinge on more practical concerns like the quality of components you can source in the different technologies.

JR

It's not the technology that's used that makes the difference- it's the designers. SM is not inherently inferior or superior.

A good example is microphone buffer circuits. These tend to be very sensitive to component issues due to the low signal level, among other things. Almost all capacitor mics these days use SM. There are plenty of great sounding mics using SM.

Really interesting. Thanks. Any other comments? What about

Surface mount designs are needed for rf work these days. Through hole parts add lead inductance to the component. That can add to create resonance at rf frequencies. Add pcb traces and their parasitic capacitance and inductance and that spice design can fall on it's face.

For audio, those are not serious enough to overcome the sonic penalties from using surface mount passives, at least the standard metal oxide resistors, chip ceramic and electrolytic caps. Surface mount silicon is usually a better high frequency performer at the expense of power dissapation.

One design that used both to their advantage was the original Rode NT-2 and 1 mics. The audio parts were through hole transistors, roederstein mk-3 resistors and Wima fkp/mkp polyprop film caps. The mics had rf pickup problems in Canada. Turns out it was the lead inductance of a pair of Wima mkp-2 .015 uf caps from pin 2 and 3 to ground. This is a usual practice to prevent rf pickup. The capacitor leads caused rectification of rf and radio could be heard. When those caps were changed to surface mount mono ceramics, the rf problems went away.

Jim Williams
Audio Upgrades

Just to clarify for those playing along at home.. rectification only occurs across some nonlinearity, like a semiconductor junction. The capacitor lead inductance merely prevented it's capacitance from being effective at attenuating the RF interference down to below the rectification threshold voltage.

I guess SMT parts are better in the regard that we don't have to deal with the non ideal characteristics of through hole leads and excessive packaging. From an engineering perspective SMT is a far more elegant solution to the problem of delivering simple Rs, Cs, Ls to where they are needed.

I regret that there isn't enough demand for sweet cap dielectrics like Polystyrene to justify someone delivering a SMT solution, but that was barely robust enough for thru hole manufacturing. Sounds like a business opportunity for somebody.

We''ll have to struggle along with NPO/COG for SMT in the meanwhile... My one attempt at a supposedly process tolerant film cap ended badly... they didn't survive the process, with opens circuits and values shifting all over the place. Not mechanically stable.

JR

Hi
Quote:
[Just to clarify for those playing along at home.. rectification only occurs across some nonlinearity, like a semiconductor junction. The capacitor lead inductance merely prevented it's capacitance from being effective at attenuating the RF interference down to below the rectification threshold voltage.]

Thanks JR, another thread has an audio input transformer being suspected of 'causing' RF intrusion into a preamp.
I am not aware of a line input transformer having that good a response in the VHF radio band.
Matt S

No, that's a USER error. Those of us who use spice all the time know the caveats that "new users" fall into. Part of the learning is in school, the real learning begins when you get a job. That and I've used 3d solvers, trace modeling, etc. When you use something like Genesys, it's almost impossible to not incorporate parasitics in your sim.

There are a couple ex-silicon designers here. They know some very elegant solutions to problems using SMD parts. I mean heck, they designed them!

It's my belief that SMD is just as good as any other part. It's what you choose and how you use it in combination more than what you use on it's own. you have to know the beast so to say.

I've seen designers using top-of-the-line parts that get foiled by simple mistakes or lack of testing/understanding. I've also seen guys who've used the cheapest parts and built novel solutions that others would swear couldn't have worked.

Surface mount has been a boost to rf and high speed computer design. It allows better performance and speeds, without it we would not be typing here today.

That being said the readers should also understand that the switch to surface mount for audio applications was not done to increase audio quality, unlike what some advertize, like Rode mics.

That is a perfect example as the original NT-2A designs were well received as a breakthrough in price/sound quality in 1995. They were designed using top quality through hole components like German Roederstein metal film resistors, Wima FKP-2 polyprop film and foil caps, etc. When I did that design I used and specified those parts. Rode tried to sub them out using cheaper metalized ERO caps and asian resistors. They called me and told me the mics didn't sound as good. They elected to use the good parts and raise the price $50. That, however, is rare.

Once all the competitors got into the chinese mic game, Rode elected to cut manufacturing costs to compete and converted those mics to all surface mount. To this day customers call me to ask which are the original version as they seem to have a better reputation. One can ruin the sound of a quality mic by just using one ceramic cap off the capsule. This is why new Neumann U-87's do not have the sound quality of their mics made just 10 years ago. The only thing that has changed are the surface mount component swaps. The design and circuit is the same. Readers can do their own test by comparing a new U-87AI to a U-87AI made before 2002. That shows what a NPO mono ceramic cap and metal oxide resistors sound like next to a polystyrene cap and metal film resistors.

I also do surface mount design here. It's not a problem if you know the pitfalls and limitations. This is not rf design, it's audio where component choices are critical. There are excellent passive surface mount resistors available, not too expensive either. It's rare to find these as most manufacturers elect to use metal oxide formulations to save a few pennies. Those parts also increase the design's THD as those resistors have poor drift vs temperature. These are measurable results and are audible as well. Those resistors are noisy.

Surface mount silicon offers better lead inductance and usually better stability and phase margin due to less lead inductance and capacitance. Where they fall short is package heat dissapation (I heat sink some of these parts) and servicability. Fried surface mount silicon can damage the pcb and cause rework headaches. There are benefits to be able to yank a fried DIP 8 part from a socket instead of pcb rework. Surface mount pcb's are not made to be serviced, but disposed and replaced. When the surface mount electrolytics dry out it's landfill time. I don't like that, we have enough electro waste already there. When through hole electrolytic caps dry out, they are easily replaced. Landfill avoided. Gear keeps working, for decades in some situations. I expect my through hole audio analog gear to live far beyond my lifetime.

The bottom line why you see more audio surface mount gear is the bottom line. Loading up an auto insertion machine and stamping out finished pcb's instead of paying some bored English housewife to stuff a console pcb is why you see it. It's about cost cutting, no more. High end designs will continue to use through hole technology only because for audio, performance is superior.

Jim Williams
Audio Upgrades