John Curl's Blowtorch preamplifier part II

Status
Not open for further replies.
Sometimes the grizzled manager is incorrect. ;)

jn

You really wouldn't want to tell him that! By mistake he thought one of my guys had caused a problem, after he was done expressing his thoughts, my guy tried to quit.. on the spot. This was even though he was right! Took four guys to convince him to stick it out.

Seems to have worked out. He continues to be in pro sound doing well enough his wife the medical school graduate started the family before doing an internship!
 
Given the fact that auplater and SY are here, maybe they can weigh in? My experience is based on specifying, purchasing, testing, and using in production the different plating schemes..they could speak on the chemistry aspect..

jn

Our optical physicist tried to explain reflection on the atomic level to me, at the very least I gathered only a few atoms and the sublayer does not participate.
 
scott wurcer:
Our optical physicist tried to explain reflection on the atomic level to me, at the very least I gathered only a few atoms and the sublayer does not participate.

Did he explain that the same atoms can be used to form both smooth and rough surfaces? ;)

Producing a rough surface will take pretty large values of "few".
 
Last edited:
ENIG is one of (if not the most) the most popular finishes for SMD boards these days, du to both shelf life and flatness. Gold embrittlement isn't seen as a big problem these days, but it is barly an atom thick layer of gold over the nickel. Though ENIPIG is becoming more popular, they can use even less gold.
I cant see any benefits of using teflon or any other pre-preg other than one of the many FR4's out there for audio, its good up to the gigaHzt.
Ands as to slots in ground planes, the best advice is dont do it.
 
I use 2% silver bearing solder, and the temp is kept at room temperature, approximately.

So far as I know, my solder fillets look great. No complaints as of yet, but it has only been 24 years. Maybe, in future, somebody will complain. I doubt it, however, because although you make a good point, it takes a certain percentage of gold to get there and there is just not that much gold on the surface when the solder joint is made. I do believe, that the 2% silver in the solder also limits the scavenging of other materials into the solder.
The 2% silver only drops the eutectic from 183C to 179 C. Not a biggie when it comes to preventing scavenging. The gold is still going into solution despite the silver, nothing you could do about that. Your product keeps on ticking because you've kept the design cool.

Most gold plating isn't 24K. A little of some other metal is added for hardness. Its called a hard gold plating in the industry. It usually makes the surface shinier. A pure 24k surface is usually duller and less durable. Its rarely used in electronics.
Matte gold is best when you have low inventory and a good JIT system. If you have to pass salt spray or elevated storage, it takes about 50 microinches to pass mil spec solderability..and that will require a two stage dip process to pull off the gold. 2 to 5 microinches served well when stock dwell was a month or so. Oh, and always over at least 10 microinches of nickel, a flash as it were. Most can use ENIG, that is well established now.

edit: hard gold is great for insertion/abrasion resistance and contact pressures.

Our optical physicist tried to explain reflection on the atomic level to me, at the very least I gathered only a few atoms and the sublayer does not participate.
Talkin to them physicists again Scott? Haven't you larned?;)

ENIG is one of (if not the most) the most popular finishes for SMD boards these days, du to both shelf life and flatness. Gold embrittlement isn't seen as a big problem these days, but it is barly an atom thick layer of gold over the nickel.
The last ENIG paper I reviewed in 2008 wasn't internally self consistent. First, it states that immersion gold is self limiting at .8 to 1.3 microns (um), then two paragraphs later states that immersion gold is self limiting at .05 to .1 microns.

The atomic radius of gold is .14 nanometers, .05 microns is about 350 atoms thick. .8 microns is about 5700 atoms thick.

We're using ENIG here, and for reliability concerns, I "assisted" upper management in seeing the error of their ways with regard to the elimination of lead in the soldering of about 5 thousand pieces of hi rel equipment. (yes, sometimes throwing tantrums, writhing on the conference room floor, and holding one's breath is indeed an effective argument technique...) :eek:

cheers, jn
 
Last edited:
I spoke to my layout designer, Carl Thompson (the T in CTC). He did remind me that the design is between 24 and 28 years old, and we do things somewhat differently today, but not any better, in reality. We prefer not to use a nickel interface if we can avoid it. Our best connectors, for example, are direct gold on copper. We used them in the Vendetta, where we could afford them.
 
>Replacing all of the Supertex mosfets essentially
>fixed the amp up to new-like performance.

I at one time revamped a preamp circuit from ua709
(in cans) to DIP LF356. After a year the preamps
started sounding sorta muted. The Signetics guys
said it was leakage in the plastic. I assume the plastic
is better now ?

There have been significant advances in the plastics. Lower ionic contamination is one, better TCE another, better glob top material evolved.

The manu's became more technical with regard to the molding design, providing better physical catches on the leadframe for the plastic to grab.

Passivation techniques and concerns also changed.. better oxide layers, sometimes silicon nitride, getting rid of lateral pnp's, better guard rings on the silicon.. I've had problems with ionic contamination creeping in and compromising chips back in the day..twas when I larned about Scanning Ion Microscopy.

The wirebonding guys also got better..you've no idea how many times the silicon under the bonding pad ended up cratering, but some manu's just did a second bond next to the crater and shipped it.....:eek:

Soldering process control also helped.

Cheers, jn
 
www.hifisonix.com
Joined 2003
Paid Member
I never knew Signetics did the LF356. I thought that was National. I also used them in my first preamps. They were ok for the price, but by the early '80's there some other really better options, one them the NE5532/4 but of course these were bipolar inputs with high Ib.

Molding compounds are a very specialist area and there have been significant improvements over the last 20 years. One of the reasons metal and ceramic packages are no longer needed.
 
Talkin to them physicists again Scott? Haven't you larned?;)

No, in 1982 I tried in vain to find anyone who understood the implications of thin metalizations on the Seebeck effect. I was trying to verify my empirical numbers for aluminum to gold plated Kovar.

John - We're making and offer tomorrow on a place in Bayview SF so I won't be far, Berkeley didn't work out this time around.
 
Last edited:
I never knew Signetics did the LF356. I thought that was National. I also used them in my first preamps. They were ok for the price, but by the early '80's there some other really better options, one them the NE5532/4 but of course these were bipolar inputs with high Ib.

Molding compounds are a very specialist area and there have been significant improvements over the last 20 years. One of the reasons metal and ceramic packages are no longer needed.

http://home.earthlink.net/~dimlay/sitebuildercontent/sitebuilderfiles/seriesiiopamp.pdf

We were hooked up with Signetics some way.
Probably habit.
 
Status
Not open for further replies.