Construction Techniques
Using SLIMs

6-13-08,  This page is in progress.  Much to be added.
11-22-08,  Updated with more good poop.
11-29-11,  Added more good poop on how to fence a module.
    I offer this page as a general guide for the builders of the SLIM MSA.  It is relevant for building any circuit, not just the SLIMs.  I will add items to this page as time goes on. 
Topics Covered on this Page
Prevent Trouble
Static Electricity Precautions
Wiring Interconnections
RF Connector Installations
Direct Coaxial Cable Connections
Fencing and shielding
Component soldering
Component Cleaning
Component removal
Measuring Power and Voltage

Nice to know Links:
Types of RF Connectors with pictures.

Prevent Trouble
    When a proven design is built, there are only two reasons why it won't work.
One or more components are bad from the manufacturer.  Very unlikely!  Or, the builder screwed it up.  Highly likely!  To prevent this, I will discuss a few things that will save you much grief.
Here are the most common construction errors and how to avoid them.
1.  Bad solder connections
    Almost all of the parts used in the SLIMs are surface mounted devices, SMD's.  When they come off the assembly line, they are perfect and ready to be used.  However, they can sit on a shelf for a long period of time before you use them.  Their solder connections can become corroded, or tarnished, and will reject a good solder flow.  Use a good magnifying glass when installing components.  Look closely at their solder connection points.  If there is any doubt as to their cleanliness, use some type of burnisher to clean them until they shine.  I like to use a pencil erasure.  Do not use any chemical that may leave a residue.  I like at least 90 per cent isopropyl alcohol.  And for goodness sakes, DO NOT blow on an alcohol-wet part to dry it.  It will absorb the moisture in your breath and leave water marks.  This includes circuit boards. Check out your local pharmacy. I found 99% isopropyl at Safeway, and Tom Thumb supermarkets.
2.  Wrong part in wrong position
    A good carpenter will tell you, "measure twice and cut once".  I suggest you make doubly sure you are installing the correct part.  Make sure the polarity is correct.  And certainly, make sure the component is installed on to the correct pads.  The SLIM's pwbs do not have a silkscreen layer with markings for components.  However, each SLIM has a document called LAYSLIM-xx-yy.  It is the only document to be used for placing parts in the correct position.
    Make absolutely sure the part is the correct part.  Almost all chip capacitors and inductors look alike.  Unless you trust your parts vendor (I don't), measure these parts with your trusty AADE L/C meter.
3.  Destroyed parts
    SMD's are not only small, but usually quite brittle.  Do not use excessive force when installing or removing these devices.
    Also, electrostatic discharge (ESD) can destroy a component without you knowing it.  Install your components in a good ESD environment.  See paragraph relating to Static Electricity Precautions.
4.  Solder bridges
    Too much solder is just as bad as not enough solder.  Usually, solder bridges between component pins are very obvious.  But, if signal traces are running under a component, solder bridging can occur without them being seen.  Once the soldering process is completed, use a good magnifying glass or microscope to inspect each and every solder joint.  You cannot inspect a joint too many times.  Believe me, multiple inspection is not a waste of time.

 Static Electricity Precautions
    I won't beat this subject to death.  I will just give you a hint for a simple static free work area.  For a very easy and cheap working surface, use an aluminum cookie sheet.  It must be untreated and unpainted.  (not teflon).  Fabricate a length of hook-up wire with alligator clips on each end.  Clip one to the cookie sheet, the other to the ground of your test equipment or soldering iron base.

Wiring Interconnections
    Power and Signal Interconnections between SLIMs are multipin wiring.  I standardized the multipin connectors as .1 inch spaced, vertical pin headers.  They can be purchased in strips of 36 pins, and cut to the preferred length.  As an option, headers can be deleted, and point to point wires can be directly soldered.  For direct soldering to the header pins, the Tin plated headers (that are called out in the parts lists) are fine.  If gold plated female mating connectors are preferred, then substitute gold plated header pins for the modules.  I do not recommend using any mating connector that is not gold plated.

RF Connector Installations for SLIMs
    If an RF connector is installed, the center hole must be drilled out to the size of the center post of the RF connector.  This hole size varies among connectors, but .035 inch is common.  The bottom center pad is allowed to be destroyed, since the 0.080 pad on the top component layer will be soldered to the center pin.  On the bottom, if there is little, or no, clearance from the center hole to ground, the clearance can be widened by grinding the ground area away, using a large (sharp) drill bit.  This makes a "V" shaped gouge around the center hole, without widening the via hole on the top surface.  If the RF connector has ground lugs, they must be cut off flush to the base, with a Dremel type tool.  The RF connector is then placed flush to the ground plane and the base soldered around it's perimeter.  This is actually a better RF ground connection than with ground lugs and mounting post holes.
connector installation slim/conninst.gif
In this diagram, notice the comment "Cut dielectric flush to base".  When the connector is heated during the solder process, the dielectric will swell and expand.  This will cause the connector to push itself away from the pwb.  It is a good idea to pretreat the connector before installing on the pwb.  Do this by heating the connector and allowing the dielectric to expand.  While it is still hot, use a sharp blade to cut the dielectric flush to the base.  When it cools, the dielectric will retract, leaving an internal gap.  Then, when the connector is soldered to the pwb, the dielectric will expand again, but only to the plane of the connector's base.

Direct Coaxial Cable Connections
    I have created a somewhat, unorthodox way of coaxial cable interconnections.  Normally, RF connectors are mounted on the bottom of the SLIM.  However, due to limited real real estate, many RF pcb connectors need to be modified.  And, since RF pcb connectors are a major expense, I have created a way to connect coax directly to the board, while retaining excellent RF characteristics.
    Every SLIM module with an RF connection has been designed and layed out to accomodate a direct coaxial cable connection.  The top layer connection (signal pad) is a .08 inch diameter round pad, with a hole diameter of 0.020 inch.  The bottom layer is a .041 diameter round pad with the .020 hole.  It is surrounded by a ground ring that has an inside radius of .030 inch and an outside radius of .090 inch.  The ground ring is not noticed since it is connected to the bottom layer ground plane.
  This allows a clearance of 0.010 inch between the center pad and the ground ring.  However, this is a very narrow clearance, and it is advisable to cut away the bottom center conductor pad, using a sharp drill bit.
  The coax cable is connected on the bottom layer with its center conductor passing through the 0.020 hole and its outer shield soldered to the ground ring.  Both RG-188 soft coax and RG-085 hard pipe have center conductor diameters of 0.020 inch.  Sometimes the pwb house will over-plate, and the .020 holes will "shrink" to .019 or less.  If this happens, simply insert a steel pin (steal one from your wife) and "wiggle" the hole a little larger.

direct cable connection slim/coaxconn.gif
   When the cable is heated during the solder process, the dielectric will swell and expand.  This will cause the cable to push itself away from the pwb.  It is a good idea to pretreat the cable before installing to the pwb.  Do this by heating the cable and allowing the dielectric to expand.  While it is still hot, use a sharp blade to cut the dielectric flush to the outer shield.  When it cools, the dielectric will retract, leaving an internal gap.  Then, when the cable is soldered to the pwb, the dielectric will expand again, but only to the plane of the outer shield.  Be sure to cut off the excess center pin on the top side of the pwb.  At high frequencies, excess pin length acts like a tiny antenna.

Fencing and Shielding
    Most SLIM modules are totally shielded to prevent RF interference.  I use a 1/2 to 1 inch tall "fence" cut from a coffee can lid (tin plated steel), placed around the perimeter of the module's pwb and soldered on both sides of the board.  A cover is formed from the same material, so that it fits within the top of each module's fence.  After testing is complete, the lids can be fully soldered to the fence.  Copper or brass is optional.  The lid can be installed externally (on the top of the fences) or internally (recessed within the module). Either is acceptable. Internal lids are easier to install, but more difficult to remove.
slim/fencing.gif


Component Soldering
    Much has been written on soldering techniques.  Many builders prefer to mount SMD's using solder paste and reflowing all connections at one time.  Personally, I still use the old fashon method.  Ye olde soldering iron.  All components have two or more solder point connections.  When one point is connected to a ground pad, there is usually a via very close, connected to the ground plane on the bottom of the pwb.  This pad really "sucks up" the heat of the soldering iron, and poor solder joints can occur.
    Here is how I do it, using a chip capacitor as an example.  Look at the receiving pads on the pwb, and determine which pad is going to be the least difficult to accept a good solder joint.  Pretin this pad with your soldering iron using flux and solder, or multicore flux solder.  Do I need to say, 60/40 solder?  For goodness sakes, DO NOT USE lead free solder.  Use it for fishing weights.  Pretin the correct end of the capacitor for this pad.  Do not clean the flux from either the pretinned capacitor or the pwb pad.  There will be enough flux remaining for the next step.  Now, hold the capacitor on the correct pads using your favorite tweezers.  The capacitor will not be flush to the board, due to the pretinning.  Now, apply the soldering iron tip to the pretinned areas of both the capacitor and pad, and allow the solder to flow.  The capacitor will now sink flush to the pwb.
    Repeat this process for all the components to be mounted on the pwb.  Some components will have multiple pads that will very easily accept solder.  Go ahead and solder these joints, leaving the "heat sucking" pads for the next step.  If a component is fully soldered, take a black dry marker and mark the top of the component, as if to say, "nothing else needs to be done to this component".
  Now, for the final soldering step.  We are going to heat the pwb to about 100 degrees C, or 212 degrees F, and solder the remaining pads.  For a heater, I use the base plate of my coffee maker.  It will actually get much hotter than needed so precautions must be maintained.  Place the coffee maker on the static free work area.  I turn it on and allow it to warm up, and occasionally, touch the base plate with a wet finger.  When the finger moisture begins to sizzle, it is 100 degrees, C.  Remove power from the coffee maker.  Set the pwb on the hot base plate.  Begin soldering the remaining "heat sucking" joints.  They will accept solder quite well.  After about 1 minute, remove the pwb from the base plate.  Re-power the coffee maker and allow it to build it's heat back to 100 degrees, C, using the wet finger indication.  Remove power, reposition the pwb, and continue the solder process.  Use this technique until all components are soldered.  The black marks on the previously completed components will save you time by not searching for components that don't need soldering.
    When completed, wash and clean the board with isopropyl alcohol (90 % or better).  The black marks will wash away with alcohol.  Some black permanent markers will not clean away.  Test your marker to make sure it will clean up with alcohol.  If it doesn't, don't use the marker.  Any residue is conductive and will make your life miserable with unknown or "weird" electrical problems later.

Component Cleaning
Here is the best method for cleaning.
    Go to the hardware store, plumber's section and ask for "acid brushes". Also called "flux brush". They are used to apply flux to copper tubing before soldering. They are cheap, about 20 cents. Get several. The bristles are about 1 inch long. Cut them back to 1/2 inch, making the brush stiffer. Also, go to your pharmacist and ask for "pure" isopropyl. He may direct you to 99 or 100 %. I get my 99% at Kroger (aka Safeway or Piggly Wiggly).
Lay out 3 small containers to hold the alcohol. I use the little plastic containers that hold the icing in the cinnamon rolls. About a teaspoon in the first. Just a few drops in the other two. Use 3 acid brushes, one for each. This prevents cross contamination.
1. Use enough alcohol from the first to well saturate the components. Drown them. Let the unit sit for a minute to let the flux soften. Then scrub. Use a paper towel to absorb the "wash". Repeat twice.
2. Use brush 2 and container 2 and scrub again. Use clean area of paper towel to blot up the "wash".
3. Use brush 3 and container 3 and scrub again. Use clean area of paper towel to blot up the "wash".
Let the unit dry naturally. You can "fan" it, but do not blow on it. After drying hold unit and look for "stains", "rings" around components, etc. If you see anything, it is not clean. Visual inspection under components is impossible, but if there is any flux under them, the "rings" are a giveaway to leftover flux.

Component removal
    Removing surface mounted devices (SMD's) from a printed wiring board is an art.  An improper removal technique can result in destruction of the SMD, traces, pads, and even neighboring components.  I will give a few techniques that have proven well over my many years of experience.  I'll split the discussion into two parts, dual lead SMD's and multi-lead SMD's.
    Dual lead SMD's are very easy to remove and I have developed a technique that is a guaranteed winner.  I call it the "Third Hand Technique".  What you will need: ESD workstation, two soldering irons, sharp pointed tweezers, rubber bands, string, and an adjustable, overhead attachment point.  The idea here, is to heat both SMD leads at the same time and have a "third hand" remove the component.
    To create the "third hand", place a rubber band through the tweezers.  Attach a length of string to this rubber band.  Wrap the tweezers with another rubber band to make the tweezer points come together.
ssapics/tweezers.JPG ssapics/lamp.JPG
    Attach the string to the adjustable, overhead attachment point.  I use my goose neck lamp.  Adjust the overhead attachment point up and down until the tips of the tweezers are about 1/4 inch (6mm) above the SMD.  Pry open the tweezers and place over the body of the SMD.  Release the tweezers.  It will now grip the SMD and want to pull it up off the pwb.  Using one soldering iron in each hand, reflow the solder on both SMD leads/pads.  The SMD will "jump" off the pwb, remaining in the grip of the tweezers.  See the following pictures:
 ssapics/twepos.JPG ssapics/twegrip.JPG
For picture clarity, I do not show an ESD work station.  Please use ESD precautions.

    I must admit, removing multi-lead SMD's are troublesome.  If the leads are close together and all the leads can be heated with two soldering irons, use the "Third Hand Technique", described above.  Use soldering irons with very wide tips.  If the multi-lead SMD has been originally soldered to the pwb using solder paste and hot reflow method, there is something many people don't know about.  In most cases, the SMD leads are not touching the copper pad of the pwb.  There will be a very small space between the SMD and the pad of the pwb, filled with solder.  This is due to the fact that the SMD is lighter than molten lead and will "float" during the reflow process.  The following is a good way to remove these SMD's.  First, remove as much solder from the leads/pads as possible, using the solder wick method.  It is quite possible that all of the solder can be removed.  Then, clean the area with alcohol, removing all flux residue.  The next step is my "trick".  I use a .001 inch feeler guage.  For those not familiar with a feeler guage, it is a small sheet of stainless steel, used for adjusting the gap area of a spark plug.  They usually come in various thicknesses, all on a single tool.  Available at any automotive supply, such as Auto Zone.  Stainless steel does not accept solder very well and can be used to "move" molten solder.
  Push the feeler guage into the area between the SMD's lead and the pad of the pwb.  It will probably not slide in, due to the "leftover" solder bridging between the lead and pad.  Using a cleaned soldering iron (no solder or flux), heat the lead/pad until the feeler guage slides between the lead and pad.  Let cool, and continue to the next lead.  This "feeler guage" method works quite well on integrated circuits that have semi-flexible leads, such as the DDS chip.  It does not work well with SMD's that have "hard contacts", such as the Minicircuits, VCO.

You can email me at:   wsprowlsatyahoodotcom
  I will answer your specific questions or comments as soon as possible.  For general suggestions, post them on our Yahoo Spectrum Analyzer Builders Group Page so others can comment.
(end of page)