On Wed, 25 Feb 2004 07:51:15 +0000, "Ian White, G3SEK"
wrote:

[snip]

|Oh yes, and lay off the coffee and Coke for several hours beforehand!

No kidding!

During the development of the Phoenix Missile Solid State Transmitter,
which used 16 IMPATT diodes in a cavity power combiner, we were
serializing every diode.

The state of the art in diode manufacturing was such that getting 16
matched devices was almost impossible, so we tested hundreds and then
picked those that most closely matched. These were a microwave pill
package with a 3-48 screw and a flange about 2mm in diameter. We
scribed about a 10 digit serial number right into the gold on the
flange.

I was good for about 30 diodes each morning until the coffee and
shakes set it.

"Ian White, G3SEK" wrote in
:

Even with an 'essential tremor' in my soldering arm
I manage down to 0.7mm pitch SMD's.
Indeed, magnificatrion does the job!

Sam

Tim Wescott wrote:

In the case of RF devices like DDS's and high-performance
microprocessors (you think 1GHz isn't RF?) the surface mount package
significantly enhances performance.

Re-mounting an SMD onto a DIP header guarantees minimum ground-lead
lengths of about a quarter-inch, so watch out for some "interesting"
changes in performance.

I haven't found them to be that bad to work with; you just need good
eyes and a steady hand.

I don't have notably good eyes or hand co-ordination, but have done
plenty of SMD work using an extra pair of strong reading glasses (on top
of my normal pair) and a Weller TCP iron with the finest-pointed conical
tip.

It's remarkable how your hands become much steadier and more controlable
when viewed under magnification. The other thing is to rest your elbow
or forearm on the bench, so you only have to move your wrist and hand.
Oh yes, and lay off the coffee and Coke for several hours beforehand!

Given those aids, anyone who doesn't have a clinical condition affecting
hand co-ordination should give SMD a try. (Note: "I don't want to" is
not a clinical condition :-)

Other essentials:
* a good work light
* thin silver-loaded solder (20g maximum)
* a flux pen or a supply of liquid flux
* thin fluxed desolder braid (1/8in maximum)
* a good pair of tweezers, that won't cross over and flip the SMD to
who-knows-where
* a clean workspace, because you *will* drop things

With only that equipment, I've mounted AD DDS chips (and much else) on a
board and they worked first time.


SMD work *is* do-able... so just do it!

On Wed, 25 Feb 2004 07:51:15 +0000, "Ian White, G3SEK"
wrote:

[snip]

|Oh yes, and lay off the coffee and Coke for several hours beforehand!

No kidding!

During the development of the Phoenix Missile Solid State Transmitter,
which used 16 IMPATT diodes in a cavity power combiner, we were
serializing every diode.

The state of the art in diode manufacturing was such that getting 16
matched devices was almost impossible, so we tested hundreds and then
picked those that most closely matched. These were a microwave pill
package with a 3-48 screw and a flange about 2mm in diameter. We
scribed about a 10 digit serial number right into the gold on the
flange.

I was good for about 30 diodes each morning until the coffee and
shakes set it.

Yes Tim, I agree. Unfortunately, the physics of RF favors "small" and
"compact". This is the price we pay for such high performance. It's not yet
to the point where it's a 'show stopper' for experimenters, but you have to
be increasingly resourceful and persistent. Many have dropped out of
hombrewing at a time when the possibilities are most exciting, merely
because they can't or won't adapt to the new packaging technologies and the
need to learn a little about firmware programming in order to make the new
generation of parts work. It's not that they "can't" learn, it's mostly that
they "won't" learn.

Joe
W3JDR



In the case of RF devices like DDS's and high-performance microprocessors
(you think 1GHz isn't RF?) the surface mount package significantly
enhances
performance. I haven't found them to be that bad to work with; you just
need good eyes and a steady hand.


--------------------------------------
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Tim Wescott wrote:

In the case of RF devices like DDS's and high-performance
microprocessors (you think 1GHz isn't RF?) the surface mount package
significantly enhances performance.

Re-mounting an SMD onto a DIP header guarantees minimum ground-lead
lengths of about a quarter-inch, so watch out for some "interesting"
changes in performance.

I haven't found them to be that bad to work with; you just need good
eyes and a steady hand.

I don't have notably good eyes or hand co-ordination, but have done
plenty of SMD work using an extra pair of strong reading glasses (on top
of my normal pair) and a Weller TCP iron with the finest-pointed conical
tip.

It's remarkable how your hands become much steadier and more controlable
when viewed under magnification. The other thing is to rest your elbow
or forearm on the bench, so you only have to move your wrist and hand.
Oh yes, and lay off the coffee and Coke for several hours beforehand!

Given those aids, anyone who doesn't have a clinical condition affecting
hand co-ordination should give SMD a try. (Note: "I don't want to" is
not a clinical condition :-)

Other essentials:
* a good work light
* thin silver-loaded solder (20g maximum)
* a flux pen or a supply of liquid flux
* thin fluxed desolder braid (1/8in maximum)
* a good pair of tweezers, that won't cross over and flip the SMD to
who-knows-where
* a clean workspace, because you *will* drop things

With only that equipment, I've mounted AD DDS chips (and much else) on a
board and they worked first time.


SMD work *is* do-able... so just do it!


--
73 from Ian G3SEK 'In Practice' columnist for RadCom (RSGB)

http://www.ifwtech.co.uk/g3sek

"ChipS" wrote in message
...

"W3JDR" wrote in message
...
Take a look at Analog Devices' line of DDS chips. Pretty much a one-chip
(needs ucontroller to drive it) digital solution to turn your 10MHz
clock
into virtually any frequency below 5 MHz (in theory; below 4MHz in easy
practicality). The AD9834 only consumes 20mw at 3.3 VDC. There are many
other devices in the product line. Many have built-in comparators to
produce square-wave output.

While blasts from the past, let me remind you of the CD4059 and the much
faster 74HC/HCT4059. The 74-series parts guarantee operation on inputs as
high as 20 MHz. "typically" as high as 50 MHz. These will divide by any
binary value from 3 to 2^16 or even as high as 21327 with some gaps.

The division ratio is set by 16 external pins plus some configuration pins
(24-pin DIP package).

The rub is that the part is very rare these days. I suppose discontinued.

Thanks to all for the replys. Thanks particularly to Joe and Mike W. for
the good info.on the daughterboard and soldering service.

--
Chip
KC5UES
real e-mail address:






-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 100,000 Newsgroups - 19 Different Servers! =-----

"ChipS" wrote in message
...
Thanks to all for the replys. Thanks particularly to Joe and Mike W. for
the good info.on the daughterboard and soldering service.

Read the latest QEX. The frequency synthesizer article contains hints on
soldering 0.6 mm spaced surface mount leads.

"ChipS" wrote in message
...
Thanks to all for the replys. Thanks particularly to Joe and Mike W. for
the good info.on the daughterboard and soldering service.

Read the latest QEX. The frequency synthesizer article contains hints on
soldering 0.6 mm spaced surface mount leads.

Thanks to all for the replys. Thanks particularly to Joe and Mike W. for
the good info.on the daughterboard and soldering service.

--
Chip
KC5UES
real e-mail address:






-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----== Over 100,000 Newsgroups - 19 Different Servers! =-----