Ian White GM3SEK wrote:
Jim Barber wrote:
I've been working on a high-power HF autotuner design that I'd just
about shelved because of the necessity of soldering the DDS chip. (it
uses an internal RF source to power the phase detector)

This thread has helped a lot. When time permits I may order up some
parts and give it a try. I have a good-quality magnifier lamp and a
fine-point temp-controlled pencil, what I don't have is OEM eye lenses
and flexible finger joints... ;-)

Most of my SMD work is done using half-moon reading glasses - the
strongest and cheapest available from the drugstore - in front of my
normal glasses. When not in use, they hang from a neck cord.

After publishing this suggestion, a very generous person *gave* me a
professional-quality binocular microscope. This is wonderful for big
jobs like assembling a whole new board... but to be honest, I still tend
to use the double glasses more, because the "setup process" is much
quicker.

You will find that the tip of the soldering iron becomes much steadier
when seen under the magnifier. This feedback loop is a wonderful thing,
especially if you give it the best possible chance to work:

* "Use the scope, Luke!" - learn to concentrate exclusively on what
you're SEEING. This is harder than it seems, because for normal
soldering we also rely a lot on our sense of touch, and automatically
tend to press the tip of the iron against the parts being soldered. With
SMD this is a disaster - it simply pushes the parts out of position - so
you need to break that habit and concentrate on using ONLY your eyes.

The problem with a 10x microscope is first FINDING the soldering iron
tip in the field of view. If I am looking into the microscope and
trying to place the iron on the work often the iron ends up way in left
field (or I push it into my other hand OUCH!) TOO much magnification
can be TOO much of a good thing. I wish I had another set of eye piece
lenses (or an other objective) so I could go down to 5X sometimes.

My microscope is a true binocular with poro prisms and dual objective
and eyepiece lenses. It was a 'gift' from a former employer (they were
cleaning up the lab and threw it out). I did have to jury rig a stand
since only the optic head was found in the trash, the unit originally
mounted on a long spring loaded 'gooseneek', like those circleline
magnifier lamps. (One of the focus knobs is cracked in half, but I can
live with that).

* Steady your arm against the edge of the bench, so you're not trying to
stabilize your whole body.

* Stay off the caffeine - it creates a fine tremor that the brain cannot
stabilize (too much "noise" in the feedback loop).



Damn Cuban Coffee (Well, I work in Miami!)

ken scharf wrote:

I bought one of these kits to make use of a sample AD9851 chip I got
from Analog Devices a few years ago.

http://www.amqrp.org/kits/dds60/index.html

I knew it was hopeless to bread board something with this chip, but even
with a PC board soldering something with such close lead spacing is a
challenge (especially when you're past 50 with failing close in vision
and less than rock solid stable hands). Still I figured I'd give it a
try. Armed with the smallest soldering tip available for my Weller
PES51 soldering station, a good magnifier lamp AND a binocular
microscope I gave it a try to solder the chip to the board.

Well there is good news and bad news. The bad news is that it is
impossible to solder the chip by hand without creating solder bridges.
The good news is that I did a good enough job to get the chip 99%
perfectly centered on the solder pads, and you can remove the solder
bridges with solder wick without removing the chip from the PC board.

It would have been easier with thiner solder (I had .021" dia solder)
and a thiner soldering iron tip, but inspection with the microscope
shows no shorts, and it looks like all the pins are properly soldered.
Naked eye it doesn't look pretty but it should work.

Now to solder those chip caps and resistors! (They should be easier,
the AD9851 was the worst part to place with the tight spacing, all the
other parts have lead spacing at least twice as wide).


Be sure to more or less "drown" the pins to be soldered with flux. If you
get solder bridges, you probably have to little flux applied - really slob
the flux on the pins. When you think you have enough of flux, add more...

Any bridges are corrected with solder wick.

The AD98xx can be soldered with a 0.5mm ceramic tip. Get one of these cheap
( 10$ ) magnifying visors, especially if you are half blind as I am...

The AD9851 is one of the easiest SMD IC's to solder as there are pins only
on two sides.

Besides... We have been using SMD's since the early eighties now so stop
whinging over it, adapt! They wont go away and they actually improve most
RF designs!

//Dan, M0DFI

On Dec 29, 3:37 pm, Dan Andersson wrote:

Besides... We have been using SMD's since the early eighties now so stop
whinging over it, adapt! They wont go away and they actually improve most
RF designs!

I've seen some suggestions go around that are very similar to hot-air
rework. I just wanted to mention that SparkFun electronics sells some
rework stations cheaply, and has useful tutorials on soldering SMD
parts. Their site: http://www.sparkfun.com

wackyvorlon wrote:

On Dec 29, 3:37 pm, Dan Andersson wrote:
Besides... We have been using SMD's since the early eighties now so stop
whinging over it, adapt! They wont go away and they actually improve most
RF designs!

I've seen some suggestions go around that are very similar to hot-air
rework. I just wanted to mention that SparkFun electronics sells some
rework stations cheaply, and has useful tutorials on soldering SMD
parts. Their site: http://www.sparkfun.com

When you have a proper pc board smt is easy (once you've mastered the
new soldering techniques. I had no problem getting the AD9851 onto the
board. I think the microscopic hairs I'm seeing are a result of using
the solder braid wick. I may have not heated it enough and pulled it
off too soon leaving solder hairs behind. Flux sure does help, and I'll
use plenty when I solder the next few smt ic's down.

I have some smt mounting boards that allow mixing smt parts with through
hole on 100 mil grid proto boards. I fear that the extra long leads
this adds to the smt parts will make rf performance a problem,
especially with 100-400mhz clock signals to the dds chips (AD9951).
I've heard of people designing their own pc boards using laser printer
output and iron on toner for resit. I've tried this before but with
'bleeding' of the toner during application I don't think I can get
better than 50 mil trace separation. Also I've had bad luck etching
boards with very thin traces, the traces get etched away before larger
areas of copper are finished etching. If I try designing boards for SMT
parts I'll probably have to farm them out to a professional house, but
this can be expensive for making but one board. (Unless you plan on
writing a QST article and selling the extra boards....).

All of my first year students can hold 20 mil traces and spaces, the C
students can hold 15 mil, and the really good ones can hold 10.

Jim


I've heard of people designing their own pc boards using laser printer
output and iron on toner for resit. I've tried this before but with
'bleeding' of the toner during application I don't think I can get
better than 50 mil trace separation.

RST Engineering wrote:
All of my first year students can hold 20 mil traces and spaces, the C
students can hold 15 mil, and the really good ones can hold 10.

Yeah, but I'll bet your first year students' eyesight is 30 years
younger than that of the average participant in this NGgrin!

--
Doug Smith W9WI
Pleasant View (Nashville), TN EM66
http://www.w9wi.com
(I was a first-year student 29 years ago)

On 2006-12-30, ken scharf wrote:
I've heard of people designing their own pc boards using laser printer
output and iron on toner for resit. I've tried this before but with
'bleeding' of the toner during application I don't think I can get
better than 50 mil trace separation.

Print your design on some kind of clay coated paper (people have
used everything from magazine pages to photo paper intended for
inkjets) and look at the print with a loupe. Cheap laser printers
will produce a result full of pinholes. A nice Xerox will make a
good, solid black. I haven't had any trouble with toner melting
or running. In fact, once you transfer it onto the PCB it's hard
to get off!

If I try designing boards for SMT
parts I'll probably have to farm them out to a professional house, but
this can be expensive for making but one board. (Unless you plan on
writing a QST article and selling the extra boards....).

You've got to pick the right board house for each order. Some, like
batchpcb.com, are particularly cheap if you want few, small boards.
They don't have setup fees or even per-board fees, only a per-order
charge. But their $2.50/in^2 grows faster than some other places
which have bigger minimum size boards/minimum orders.

--
Ben Jackson AD7GD

http://www.ben.com/

On Sun, 31 Dec 2006 04:22:42 -0600, Ben Jackson wrote:

You've got to pick the right board house for each order. Some, like
batchpcb.com, are particularly cheap if you want few, small boards.
They don't have setup fees or even per-board fees, only a per-order
charge. But their $2.50/in^2 grows faster than some other places
which have bigger minimum size boards/minimum orders.

I have been using an off-shore fab for years and they seem to do just what i
want, but have setup fees that make protos expensive. So I went and looked at
batchpcb.com but they do mention a $US10 setup fee in their FAQ.

http://www.batchpcb.com/faq.php?osCs...a389ad75decda8

Do they actually charge it?

Ben Jackson wrote:
On 2006-12-30, ken scharf wrote:
I've heard of people designing their own pc boards using laser printer
output and iron on toner for resit. I've tried this before but with
'bleeding' of the toner during application I don't think I can get
better than 50 mil trace separation.

Print your design on some kind of clay coated paper (people have
used everything from magazine pages to photo paper intended for
inkjets) and look at the print with a loupe. Cheap laser printers
will produce a result full of pinholes. A nice Xerox will make a
good, solid black. I haven't had any trouble with toner melting
or running. In fact, once you transfer it onto the PCB it's hard
to get off!

If I try designing boards for SMT
parts I'll probably have to farm them out to a professional house, but
this can be expensive for making but one board. (Unless you plan on
writing a QST article and selling the extra boards....).

You've got to pick the right board house for each order. Some, like
batchpcb.com, are particularly cheap if you want few, small boards.
They don't have setup fees or even per-board fees, only a per-order
charge. But their $2.50/in^2 grows faster than some other places
which have bigger minimum size boards/minimum orders.

My problem with the iron on transfers wasn't being able to print thin,
close spaced lines on to the transfer medium. The problem was that when
ironed on the toner lines 'spread out' so adjacent lines touched. Also
a problem was timing just how long to leave the board in the soup to
etch. I did heat the etchant up first and used an IR lamp to try and
keep it warm, but even so the thin close spaced traces were close to
being over etched long before larger and wider spaced traces were
'done'. (maybe a problem with 'Radio Shack' etchant?)