The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Discreete vs SMT. Noob wants info. Keyboard Synthesizers
Old 21st July 2007
  #1
Lives for gear
 
jupiter8's Avatar
 

Discreete vs SMT. Noob wants info.

This has been bothering me a bit lately.I just want some info on the subject in general for no particual reason other than i'm curious. I'm talking about small series stuff,mainly synths but also other stuff like mic preamps etc.

I know the difference (or just about) between SMT and discreete. However can discreete be fully automated ? You could possibly build a multimillion $ machine with optical hooplas but is it economically viable for small series stuff ?

Take the Neve 1073 reissues for example. Are those hand soldered ?

I'm curious about the price of the componenets as well. Take the Moog Little Phatty. I imagine the actual component cost of the voice chip being next to nothing. Is that the case ? I cannot imagine the components beeing that expensive.

Difference in sound. I imagine a discreete being able to run on higher voltages and that it should therefore be able of a a higher quality sound. It is of course impossible to give an exact answer on the differences but a ballpark figure would be nice. I know MOTM changed their cards from discreete to SMT. Would the difference in sound be:
1. Night and day.
2. Small but easily dectable.
3. As small as makes no difference.
4. No difference ?

Thankful for any enlightment on these matters.

EDIT: More of a semantics question i guess. Discreete means something like "Made up of individual components instead of ICs",does'nt it ? Which means a SMT device could be either discreete or use ICs. But you call it SMT either way to distiguish between the two manufacturing processes,right ?

Last edited by jupiter8; 21st July 2007 at 12:20 PM.. Reason: More questions.
Old 21st July 2007
  #2
Lives for gear
 
daniel c's Avatar
 

Quote:
Originally Posted by jupiter8 View Post
EDIT: More of a semantics question i guess. Discreete means something like "Made up of individual components instead of ICs",does'nt it ? Which means a SMT device could be either discreete or use ICs. But you call it SMT either way to distiguish between the two manufacturing processes,right ?
I have always heard components referred to as being "Through Hole" or "Surface Mount". Through hole because you put the pins/legs/wire through a hole in the circuit board and solder it in place. Surface mount because they sit on the surface of the circuit board, affixed with a conductive paste.

As for the difference in sound, you probably wouldn't hear much difference between a circuit built with surface mount components, and the same circuit built with through hole components. YMMV
Old 21st July 2007
  #3


There are still through-hole automated stuffing machines in the world. Though they are not common anymore.

The big SMT automated "pick and place" machines with a reflow oven on the other end are pretty easy to reconfigure. It is ecconomical to use them on smaller runs of, say, 1000 boards.

The thing to watch out for is the RoHS bu115hit. I'm not sure those will be very reliable for a while.




-tINY

Old 21st July 2007
  #4
Lives for gear
 
JohnRoberts's Avatar
 

Indeed, through hole components have been automated for decades, but you would have to run panels across one machine to insert dual inline ICs, a different machine for resistors and axial capacitors, a third for radial caps and transistors. If volume justified it, odd component inserters (glorified robots), could insert pots, jacks and other "odd" stuff. Not very efficient.

SMT is a newer generation than conventional leaded components where a great deal of the waste is eliminated and the manufacturing process is simultaneously cheaper and quite robust.

I too share concern about the ROHS initiative as causing more problems than it solves, but that's typical when the government gets involved in technical issues. "We're from the government and we're here to help you"

JR
Old 22nd July 2007
  #5
Lives for gear
 
nofi's Avatar
 

A "discrete circuit" refers to a circuit that is made up of discrete components... ie each component in the circuit is a basic, simple, atomic building block - resistor, inductor, capacitor, transistor, diode, etc. In audio electronics it basically means no opamp components. (but you can make an opamp out of discrete components - this is fine, and indeed respected, because rupert neve does it)

The basic idea of SMT vs through-hole is already described. as far as differences in components go, there are discrete SMT equivalents of most discrete through-hole parts you would ever use, but SMT parts tend to be smaller, and hence lower power (but they don't HAVE to be) so I think that a SMT equivalent of a known through hole circuit will often end up being designed to be lower power, and hence be a different circuit, and hence sound a little bit different...

Any opamp that has a through hole and an SMD package is going to have an identical bit of silicon in each, so someone who can hear the sound of an opamp circuit that's using a DIP package vs a SO package, is on very shaky ground, in my opinion...


Also, people seem to say they hate SMD because they can't work on them, but this is kinda stupid if you ask me. SMD boards are actually simpler to work on than through-hole boards if you have the right (very basic) tools, and it's just a case of learning the techniques to deal with it, rather than stamping your foot and whining.



oh yeah.. to answer your question on the MOTM, not having hear 2 different pieces side by side, I pick answer 3.
Old 22nd July 2007
  #6
TMI
Gear Head
 

SMT vs. etc etc

OK,

Across the street from Studio 54, Eventide Clockworks had a through hole pick and place machine stuffing boards in the lobby, 27 years ago and it was not new then.

SMT Technology is far superior to through hole PCBs.

BUT

Not all components are available in SMT. In particular, those used in exotic audio components.

When a component is available both ways, the SMT component is nearly always superior. The available routes when designing a surface mount board are far more flexible than a through hole board. The lead inductance or even capacitance is always less and most passives are lead-less.

Having said that, there is a third option, point to point wiring. Old tube components were wired this way and every tech building or repairing the unit changed the layout from the original design. Still, being a 3D construction in primarily high impedance circuits, it did have some advantages and was far more durable in say guitar amp applications.

In something like a 1073, the magic is in the type of capacitors and the core material / winding of the inductors and transformers. It is a simple thing to get a ferrite core SMT inductor of the same value, it is quite another to make it sound the same. You can build a PCB or even SMT version to sound the same as the free wired version IF you take your head out of the sand when specifying the parts and doing the layout.

A word about machine VS hand soldering. Machine soldering is always better, provided the machine is set up properly which is not always the case. Hand solderers tend to get trained or fired before they can ruin as many units as a poorly set-up machine can.

Lead free solder and new fluxes are horrible. They make every operation hypercritical. They have not got all the lead out of the water system plumbing yet but they are hell bent to remove it from circuit boards NOW! MCI had enough trouble soldering MOLEX headers with lead. Today I am seeing shiny new boards that you can just lift the components off. Joint looks fine, but no alloy attachment to the component or lead wire.

Tom Maguire
Old 23rd July 2007
  #7
Lives for gear
 
ulysses's Avatar
"Discreet" means you're careful not to let a secret out. Madames and assassins need to be discreet. API 2520 op amps are "discreet" in that they're potted in epoxy so nobody can see what's in there.

"Discrete" refers to circuits made up of individual components, as opposed to an "integrated circuit" (or IC for short) which puts multiple components on a single piece of silicon. The Jensen 990 discrete op amp actually contains an Integrated Circuit, a chip that contains multiple matched transistors, because it's easier to get them matched well when they're on the same chip.

"Surface Mount" components, or SMT (Surface Mount Technology) or SMD (Surface Mount Devices) are components that do not have pins which are designed to protrude through a hole in the circuitboard. They get soldered onto a copper land on one side of the board. SMDs can be discrete components or integrated circuits.

Any components that mount on a circuitboard by way of pins that protrude through holes in the board can be referred to as "through-hole".

Then of course there are components that aren't intended to be soldered onto a circuitboard. Switches, potentiometers, lamps, tube sockets, relay sockets are all commonly found with "solder eyelets" which are designed to have wires soldered to them. They're all available in PC-mount versions as well, even surface mount.

No one type of technology is inherently superior to another. People often try to argue the universal superiority of one type of components or another, and those people are jackasses. Context is everything. Good design means using a circuit that does what it's supposed to do and choosing components that are appropriate for the application. A surface-mount component doesn't have to be a lower-power device, but sometimes they are and sometimes that's okay. The trick is to know which components need to handle more power and which don't. There's usually some math involved. That's why they call it engineering.
Old 23rd July 2007
  #8
TMI
Gear Head
 

Jackass? Really now

Dear U,

The electrical superiority of surface mount packages cannot be disputed. The electrical superiority of dense SMT layout, likewise is beyond dispute. There have been poor results but they have nothing to do with SMT. Poor results are related to ignoring analog rules in layout and ignoring audio rules in component selection. This is compounded by a natural predisposition toward digital components and layout tools in this relatively new technology.

I took the SMTPLUS PCB design courses over 20 years ago.

Welcome to SMT Plus, Inc

James Blankenhorn actually understands packaging technology at the root level. The advantages of shorter trace lengths, more direct paths and smaller reactive values result in more stable designs, period. It's an engineering fact, not an opinion or conjecture. With modern, high speed op-amps the difference can be substantial.

Having said this, I still do a lot of through hole work simply because it is practical, the parts are available off the shelf and it's simple to assemble. In a large scale production with cooperative vendors the surface mount versions are better.

Tom
Old 23rd July 2007
  #9
Lives for gear
 
JohnRoberts's Avatar
 

There are numerous benefits to SMT but the discussion is not totally one sided. The compelling factor pushing the industry toward SMT is cost. No leads to cut off and throw away, no holes to drill to stick those leads through, smaller PCB, etc. $$$

I won't argue the specific downsides of SMT as they are few and manageable. To Ulysses point, there is far more involved in the performance of any circuit than the lead technology used.

JR
Old 23rd July 2007
  #10
There are some fantastic quality surface mount passive components out there. Problem is, you are not likely to ever see them in audio gear, besides the exotic super expensive stuff.

Most do this on the cheap. Metal oxide resistors at 1/10 watt sound like crap next to a high quality metal film resistor. Yes, there are some low cost 1/8 watt surface mount metal film resistors, Mytek is a good example.

You will not find Vishay bulk foil surface mount resistors in audio gear, they run around $10 each! These are the best resistors in the world, but for $10 they better be.

Capacitors are another story. Sad, but there are not any high quality audio purpose surface mount film capacitors available. Polystyrene and polypropylene cannot take the heat so they are not made in surface mount. I have had numerous discussions with Wima, Germany about this. Even Sennheiser's top-o-line MKS rf condensors use surface mount passives as does new U-87's, not a good idea unless you want to destroy a reputation.

There are great electrolytic caps in surface mount, the newer tantalum replacements are a good choice as long as the equipment can be disposed of in around 20 years without regret.

I prefer to design exclusively with thru-hole passives. I want the gear to last and I want it to be able to be repaired, don't we have enough of this stuff in landfills already?

That's what led to this RoHS crap in the first place.

Jim Williams
Audio Upgrades
Old 23rd July 2007
  #11
Lives for gear
 
jupiter8's Avatar
 

Some good info in here. Thanks a lot everyone.
Old 23rd July 2007
  #12
Lives for gear
 
JohnRoberts's Avatar
 

I did experience some trouble with SMT film caps.

I encountered some value drift and outright failures due to process time/temp issues. While on paper these caps should have survived the factory, some factories are not as tightly controlled wrt their process.

Through hole is kinder and gentler to fragile film caps.

JR
Old 23rd July 2007
  #13
Quote:
Originally Posted by Jim Williams View Post
Capacitors are another story. Sad, but there are not any high quality audio purpose surface mount film capacitors available. Polystyrene and polypropylene cannot take the heat so they are not made in surface mount. I have had numerous discussions with Wima, Germany about this. Even Sennheiser's top-o-line MKS rf condensors use surface mount passives as does new U-87's, not a good idea unless you want to destroy a reputation.

There are great electrolytic caps in surface mount, the newer tantalum replacements are a good choice as long as the equipment can be disposed of in around 20 years without regret.


MLCC caps are pretty good, but not available in bigger values. Maybe it makes sense to rethink the impedance points we design around....



-tINY

Old 25th July 2007
  #14
Gear Addict
 

So to sum it up, I think the OP was confusing two different concepts:

1. Through-hole versus surface mount
2. Integrated circuits vs. circuits built from discretes

I seriously doubt that #1, in and of itself, is ever going to make a difference to the sound of an audio device, provided the specified components are equivalent. Unavailability or quality problems that occur with some SMT caps has already been noted by the above posters, but those things are more artifacts of poor design and/or poorly controlled manufacturing than SMT per se.

#2 can indeed make a difference. The design of almost any IC is in some senses a compromise, because the designer generally wants the part to be able to cover a range of potential applications in order to improve sales. When designing with discretes, it is possible to tailor the circuit more closely to the specific application by careful choice of values and specs of components used. The big problems with going all-discrete are: (1) cost. You can cram an awful lot of components on a die, at little additional $/component. Take the extreme case of a modern microprocessor -- doing that with discretes would cost, well, about as much as an IBM 360 cost back in the day. (Read: You can't afford one. Don't bother asking.) (2) Size and weight. Discretes take up a *lot* more board space than an equivalent circuit on an IC. With SMT, that's becoming even more true. Keep in mind that the bulk of electronics manufacturing is in consumer electronics, and the fashion in consumer electronics these days is: no matter what it is, make it smaller! (I'm not sure why we keep going down that road; I can barely press the buttons on my cell phone as it is...) We had a related thread on the electronic music board a few months ago, about where all the good polyphonic synths have gone. The answer is that the IC's that the classic polysynths like the Prophet-5 and the JP8 used are no longer available -- Doug Curtis RIP, and SSM is now owned by Intersil and they have dropped all but the most broadly applicable IC's in that line. Re-creating the JP8 in all-discrete circuitry would probably result in something that weighs twice as much as a CS80, and is almost as stable in tuning.

Which leads around to the other topic that was mentioned: MOTM going to SMT on some of its designs. A couple of things about that. First, it's not unique in the modular market; Dotcom, for one, has been doing some SMT for several years now. (Roger doesn't like for his circuit boards to stick out the back...) The second thing is that, oddly, it's actually been motivated by further discontinuation of IC's. For example, Paul is faced with the task of redesigning the MOTM-190 VCA to eliminate the now-discontinued CA3280 OTA before the available stock runs out. Its replacement might very well be an all-discrete circuit (the only other choice would be the LM13700, which would require some redesign of the supporting circuitry, and it doesn't have specs as good as the 3280). That will take up more room on the board, and so Paul might choose to go to an SMT design so that the board doesn't get larger.
Old 25th July 2007
  #15
Lives for gear
 
JohnRoberts's Avatar
 

I don't know whether it was your intent but you have brought up one of the areas where integrated circuits improve upon discrete (device matching, thermal tracking, etc).

If one were forced to make a discrete VCA or OTA they would surely use precision arrays which are just very simple integrated circuits.

JR
Old 27th July 2007
  #16
Lives for gear
 
jupiter8's Avatar
 

Quote:
Originally Posted by cornutt View Post
We had a related thread on the electronic music board a few months ago, about where all the good polyphonic synths have gone. The answer is that the IC's that the classic polysynths like the Prophet-5 and the JP8 used are no longer available -- Doug Curtis RIP, and SSM is now owned by Intersil and they have dropped all but the most broadly applicable IC's in that line. Re-creating the JP8 in all-discrete circuitry would probably result in something that weighs twice as much as a CS80, and is almost as stable in tuning.
I sure got a couple of issues mixed up and thanks to you all i'm now a bit wiser.

I was one who argued in the poly thread you're refering to. I'm not looking for a fight,i just want some info. That CS 80 example makes no sense to me.

The Oberheim OBx was made without ICs and so is the SE Omega 8. True the Omega has software envelopes and LFOs but still. Both the OBx and the Omega are'nt very heavy and at least my OBx was very stable tuningwise. I've never tried the Omega but i can't imagine it being very unstable tuningwise.

And was'nt the weight of the CS80 mainly because of a lot of other stuff ?
The OBx for sure is a lot lighter.
Old 27th July 2007
  #17
Lives for gear
 
daniel c's Avatar
 

Quote:
Originally Posted by jupiter8 View Post
The Oberheim OBx was made without ICs and so is the SE Omega 8. True the Omega has software envelopes and LFOs but still. Both the OBx and the Omega are'nt very heavy and at least my OBx was very stable tuningwise. I've never tried the Omega but i can't imagine it being very unstable tuningwise.

And was'nt the weight of the CS80 mainly because of a lot of other stuff ?
The OBx for sure is a lot lighter.
Hmmm I would have to disagree about the Oberheim OBX being made without IC's. I've been looking at the voice card schematics and it would appear that it's only the oscillators that are made up of discrete transistors (mostly), but the filter, VCA , envelopes are all IC's, with just a couple of transistors here and there. The signal path after the oscillators is mostly IC based.

You can download the service manual from this site;

http://www.bluesynths.com//index.php

And here is a close up of a voice card. I'm afraid that there are a lot of IC's in there.

The Oberheim SEM is a different story. There are still IC's in there, but more sections of it are made up of discrete components.
Old 27th July 2007
  #18
Lives for gear
 
jupiter8's Avatar
 

Quote:
Originally Posted by daniel c View Post
Hmmm I would have to disagree about the Oberheim OBX being made without IC's. I've been looking at the voice card schematics and it would appear that it's only the oscillators that are made up of discrete transistors (mostly), but the filter, VCA , envelopes are all IC's, with just a couple of transistors here and there. The signal path after the oscillators is mostly IC based.

You can download the service manual from this site;

Bluesynths

And here is a close up of a voice card. I'm afraid that there are a lot of IC's in there.

The Oberheim SEM is a different story. There are still IC's in there, but more sections of it are made up of discrete components.
Not just the oscillators - the low pass filter is discrete as well, a filter design borrowed from the Oberheim SEM. The charismatic multimode-filter design is unfortunately missing. Each voice has its own separately allocated voice card, letting it go about its business on its own. The only integrated circuits on the voice cards despite the OP-amps are the envelope generators made of Curtis CEM3310 chips.

From the bluesynth site.

Sure there are ICs in there. I was talking about the voice chips which i thought was 100 % discrete but turns out not quite 100 % but almost.
Old 27th July 2007
  #19
Lives for gear
 
JohnRoberts's Avatar
 

Quote:
Originally Posted by elrushbo911 View Post
SMt is not as reliable, but that is getting anal.
just my .02
I have found SMT (before ROHS) to be quite reliable, and ROHS is not limited to SMT.

What is the nature of poor reliability you attribute to SMT? I am only aware of a few specialized part issues (like low temp film caps).

JR
Old 28th July 2007
  #20
Lives for gear
 
daniel c's Avatar
 

Quote:
Originally Posted by jupiter8 View Post
Not just the oscillators - the low pass filter is discrete as well, a filter design borrowed from the Oberheim SEM. The charismatic multimode-filter design is unfortunately missing. Each voice has its own separately allocated voice card, letting it go about its business on its own. The only integrated circuits on the voice cards despite the OP-amps are the envelope generators made of Curtis CEM3310 chips.

From the bluesynth site.

Sure there are ICs in there. I was talking about the voice chips which i thought was 100 % discrete but turns out not quite 100 % but almost.
That statement from the bluesynth site is a little strange, and I think that they have lifted it from someone elses site, it looks very familiar.
You might want to compare the schematics of the OBX and the SEM. I know that they say the OBX filter is discrete, but in the schematics there is one discrete transistor, 3 x CA3080 IC's, 2 x TL082 IC's, and a TL081 in the filter section. These IC op-amps are doing the work in the filter, along with the capacitors and resistors of course, like any filter circuit.

When they made the OBXa they used Curtis chips, in place of the discrete oscillator circuit, and in place of the IC op-amp based circuit they used in the filter.

The SEM has a mostly discrete filter. A Oberhiem 8 voice made up of SEM's is rather huge compred to a 8 voice OBX.

Okay, I will let it go now heh
Old 28th July 2007
  #21
Lives for gear
 
jupiter8's Avatar
 

Quote:
Originally Posted by daniel c View Post
That statement from the bluesynth site is a little strange, and I think that they have lifted it from someone elses site, it looks very familiar.
You might want to compare the schematics of the OBX and the SEM. I know that they say the OBX filter is discrete, but in the schematics there is one discrete transistor, 3 x CA3080 IC's, 2 x TL082 IC's, and a TL081 in the filter section. These IC op-amps are doing the work in the filter, along with the capacitors and resistors of course, like any filter circuit.

When they made the OBXa they used Curtis chips, in place of the discrete oscillator circuit, and in place of the IC op-amp based circuit they used in the filter.

The SEM has a mostly discrete filter. A Oberhiem 8 voice made up of SEM's is rather huge compred to a 8 voice OBX.

Okay, I will let it go now heh
Oh, don't quit on my account. I'm learning stuff all the time here. This thread has been very helpful. Thanks again all.

I know some can get really defensive and personal. Not me. I'm in it to learn. So when i argue with you it's more "But i've heard this and that,isn't that true?" than "Why are you so mean to me?"

And when i question stuff you say it's more "Really ? Interesting. Care to elaborate on that?" then "You're an idiot. Everyone knows it's this way".


Could well be i'm dead wrong on the discrete polysynth part. It's no secret that i'm no expert in these matters. I'm just a very curious guy who wants to learn. So when i see stuff that does'nt make sense to me i ask. Simple as that.

Peace
Old 30th July 2007
  #22
I used to work on Stevie Wonder's CS-80's. I installed balanced output circuits and other stuff into them. They had around 40 pcb's on end with a million miles of shielded cables back and forth. A couple had some chips, but the remainder were discrete transistor pcb's made out of the tan crap that cracks real easy.

Not fun, don't miss them. I don't see how that could not be made today. As to the Curtis chips, etc, it's much cheaper to model them into digital than blow more new silicon for them, no market for that anymore.

Jim Williams
Audio Upgrades
Old 2nd August 2007
  #23
Gear nut
 

There is nothing wrong with through hole or smt, if done properly. SMT allows large scale manufacturing but we don't play there, so, use what you like the best.

ROHS is going to bite us big time. The reliability issues are just now starting to surface, after products have been in the field for a year or so. The other issue with rohs that hasn't been discussed much in the audio press is that the soldering temps are higher than they used to be. This means, and I have seen this, that if the components are over heated at assembly they are degraded. You may not see or notice it for months then little by little they start to drift out of tolerance. The higher soldering temps also mess with the transistors in the front end of opamps, the opamps become noisier than they are supposed to be. The list of issues goes on....
Old 2nd August 2007
  #24
Lives for gear
 
JohnRoberts's Avatar
 

Quote:
Originally Posted by mikeg View Post
There is nothing wrong with through hole or smt, if done properly. SMT allows large scale manufacturing but we don't play there, so, use what you like the best.

ROHS is going to bite us big time. The reliability issues are just now starting to surface, after products have been in the field for a year or so. The other issue with rohs that hasn't been discussed much in the audio press is that the soldering temps are higher than they used to be. This means, and I have seen this, that if the components are over heated at assembly they are degraded. You may not see or notice it for months then little by little they start to drift out of tolerance. The higher soldering temps also mess with the transistors in the front end of opamps, the opamps become noisier than they are supposed to be. The list of issues goes on....
I share your concern about ROHS and have experienced temperature issues with film caps at pre-ROHS temperatures.

I am not aware of temperature mechanism that makes transistors noisy. Annealing junctions (by heating them with current) AFAIK reduces noise. Could you expand upon that degradation mechanism?

JR
Old 3rd August 2007
  #25
Gear nut
 

I don't know the mechanism. I have seen it. In a coulpe different places of employment when the chips went through the wave solder the chips would fail a 1/f noise test. Solder the chips in by hand and kep the temp down and all would pass.
Old 3rd August 2007
  #26
Gear Addict
 
CoteRotie's Avatar
 

Quote:
Originally Posted by mikeg View Post
There is nothing wrong with through hole or smt, if done properly. SMT allows large scale manufacturing but we don't play there, so, use what you like the best.

ROHS is going to bite us big time. The reliability issues are just now starting to surface, after products have been in the field for a year or so. The other issue with rohs that hasn't been discussed much in the audio press is that the soldering temps are higher than they used to be. This means, and I have seen this, that if the components are over heated at assembly they are degraded. You may not see or notice it for months then little by little they start to drift out of tolerance. The higher soldering temps also mess with the transistors in the front end of opamps, the opamps become noisier than they are supposed to be. The list of issues goes on....
There were some issues with RoHS in the beginning (tin whiskers, higher temp profiles) but those have largely been addressed and overcome. (I work as an engineer for a large semiconductor company and I deal with this every day, so I speak from experience.) We're not going to see problems due to RoHS in equipment that is currently being manufactured.

Having said that, I think RoHS was a really bad decision. I don't think anyone looked at the big picture and decided that RoHS was the best environmental use of the XX billions of dollars that it cost everyone. There were likely far better ways to spend that money, but no one did the analysis to my knowledge.

Regards,

John
Old 3rd August 2007
  #27
Lives for gear
 

John,

My question would be, your experience seems to be very large scale production, do you see this as true with small run, and handbuilt production?

I don't have budgets for studies like this, obviously.
Old 3rd August 2007
  #28
Lives for gear
 
JohnRoberts's Avatar
 

Quote:
Originally Posted by mikeg View Post
I don't know the mechanism. I have seen it. In a coulpe different places of employment when the chips went through the wave solder the chips would fail a 1/f noise test. Solder the chips in by hand and kep the temp down and all would pass.
Thanx, most of the 1/F noise problems I've seen were process related (at the silicon process level). I remain interested...

JR
Old 3rd August 2007
  #29
Gear Addict
 
CoteRotie's Avatar
 

Quote:
Originally Posted by JohnRoberts View Post
Thanx, most of the 1/F noise problems I've seen were process related (at the silicon process level). I remain interested...

JR
Me too. As far as I know popcorn noise IS process related and has a 1/f power spectral density. (As does flicker or contact noise which is related to fluctuation contact between two materials.) Maybe some accelerated diffusion could occur at high temp that could change things, but it's usually not the case. We haven't had any complaints from our customers about this, but applications where it would be noticed are probably few.

Regards,

John
Old 4th August 2007
  #30
Gear nut
 

The chips I saw this in were TLOxx units. At the time they were used in the front end of a biomed piece of gear where noise was an issue. (I know, I know, why use a TLO in the first place) We did multiple tests; 100 pieces through the wave and test, failure rate over 70%, solder the chips in by hand, failure rate of 3%.

At my last job we were seeing motor controllers failing after 6 months in the field, we were finding a lot of cracked joints and bad joints in general. Now, with that said; process does have a lot to do with it and I don't know that the stuffing house knew what they were doing with rohs parts.
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Similar Threads
Thread
Thread Starter / Forum
Replies
noon / Rap + Hip Hop engineering and production
7
orksnork / So much gear, so little time
2
imlikeajungle / Music Computers
8
jlearman / So much gear, so little time
5
jdunn / Geekslutz forum
7

Forum Jump
Forum Jump