• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

SMD Parts and the DIYer

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
SMD parts require tooling not found on a "typical" tube/valve DIYer's bench. The soldering equipment we use rates to destroy the parts, before pigtail leads are attached in a satisfactory manner. :( OTOH, some useful items don't come in leaded versions. I've found myself wishing for leaded versions of parts, like this Schottky diode, on any number number of occasions.

How to escape from "Frustration City"? My thinking is to attach pre-tinned, solid, copper leads to desirable SMD parts with silver bearing, electrically conductive, epoxy. Please notice that curing time decreases and conductivity increases with the application of heat. The hot air guns we use with heat shrink tubing, set on the low setting, appear to be a good fit for the elevated temperature "requirement". Given the high cost and 2 year shelf life of the conductive cement, working on batches of parts that exhaust a given epoxy supply seems to be proper procedure.

Have at it, guys! Do my cogitations have merit?
 
Member
Joined 2009
Paid Member
Depends... If your workbench contains a half decent soldering iron like a Weller TCP or WTCP and you have a fine tip for it (I recommend a 1.6 mm chisel tip), then SMDs are actually not that bad to hand solder. I think a lot of the apprehension that people have stems from fear or intimidation and "I can't possibly do that" mind sets. And some of it is just lack of experience.

Here is my suggestion:
Make a prototype board that has a bunch of 1206 footprints. Solder some resistors onto those. Then do the same for 0805 and 0603 if you want to push it a bit. You'll probably have to run some traces away from the footprints to avoid lifting the pads. When you master that, move on to actual circuits.

Much of SMD soldering by hand comes down to tricks of the trade. Here are mine:
1) Flux is your friend. Get yourself a flux pen. Flux increases the surface tension and makes the solder wick to the pins and not make shorts between pins. This works particularly well if you have a board with solder mask (the green coating you see on my boards for example).
2) Use a fine tip soldering iron. I use a 1.6 mm chisel tip. I also have a 6.3 mm chisel tip for the tabs that are used on power components (TO-263 packages).
3) Use thin solder. Get some that's 0.5 mm in diameter. It really makes it easier to get the right amount of solder onto the joint.
4) Another friend is solder wick (aka desoldering braid). Get some fairly thin stuff - maybe 2-3 mm wide.
5) Get a pair of non-magnetic tweezers. Don't go cheap here. I think I paid about $40 for my Techni-Tool titanium tweezers. They're worth it!

For resistors, tin one pad, then place the resistor and hold it with tweezers while reheating the solder. Let the resistor sink into the solder, remove the soldering iron, allow to cool. Then solder the other pad. Go back and retouch or clean up with solder wick if necessary.

For ICs without tabs (like SO-8 packages), I normally tin pad 1 and use a technique similar to what I describe above. I solder the corner pins first, then the rest. If you get a solder bridge, clean it up with solder wick.
For bigger ICs like QFP packages, I solder the corner pins using a similar method and then solder the rest of the pins. It's rather hard to not get a solder bridge, so I often just put globs of solder on and clean it all up with solder wick after. Even after using the solder wick, there's usually enough solder left to from a good bond.
For power devices and regulator ICs with tabs like the LT3080 I use in my 21st Century Maida Regulator, I tin the thermal pad and the tab on the device first. Then using the 6.3 mm chisel tip, I heat the tab and the PCB footprint simultaneously and let the IC settle in the solder. Then I solder the pins. If the tab is not exposed - like on the LM22673 in my Universal Filament Regulator, I heat from the bottom of the board through the thermal vias. I have pictures of this process on the Universal Filament Regulator site (see my website below).

The key with all of this is to apply flux to the PCB footprint before soldering. It's not so critical with resistors, but for anything with closely spaced pins, it really helps out a lot. As does having a board with solder mask.

If your eyesight makes it hard to see the components, I suggest using either a dissecting microscope (5~10x magnification should do it) or a magnifying hood. They're actually not that expensive. I don't recommend the lamps with magnifying glasses. They often don't have enough magnification. For some, a pair of reading glasses may be enough to allow your eyes to focus on the components.
I find that I'm able to solder 0805 without magnification. I can do 0603 as well, but I really prefer to have a microscope available for that. Especially if the board goes out to a customer.

Hope this helps...

~Tom
 
Last edited:
Tom,

Thanks for the reply. What you said makes lots of sense, when working with PCBs. My ideas are oriented towards P2P construction. The Schottky diode I linked would be damned useful when bridge rectifying "120" VAC, with either 4X SS or 2X SS and a vacuum rectifier. Zero SS diode switching noise is a wonderful thing.

BTW, I do own a Weller temperature controlled soldering station and a grounded, cheapy, LOW wattage "Rat Shack" pencil. I remain very leery of parts without leads.
 
You could make a very small rectifier board by etching, grinding, or cutting the copper from a single-sided blank. Leaving large area pads for soldering the SMDs and drilled holes for the connecting wires, or using cut component leads to solder it right onto the main board as a homemade module.
A clean board and flux ("Use it like your brother-in-law sells it," as the saying goes) is half the battle; just getting used to the new dimensions is most of the other half. The rest is just making a good fillet.
 
I have to use a magnifier, but find the work not too difficult. Temp controlled iron with chisle point, fine solder, a wick, and a little patients goes a long way. Solder masks are a must, but that is from my experience so far. Only made a few prototype boards for SMD.
Cleaning the flux is an absolute must.
Just my 0.02.

I figure if I can do it anyone can
 
I use surface mount schottky bridges routinely in P2P construction. I simply flatten out the "gull wing" leads and solder them to wires I've previously mounted on perf board; said wires then make the needed connections in the circuit. I collect lead snipping from my leaded components in a little plastic ziplock bag and use those to make p2p connections.
 
For more complex smd parts like opamps I find the best solution is to stick the smd part to the board with a tiny piece of Blu-tac or mortite on the body (I mean TINY, jusy enough so it doesn't move) then flood the leads on with solder and wick off the excess. Works like a charm.
 
A past roommate of mine who was in school to be an electronics engineer at the time (now does R&D at a well known consumer electronics company) would do surface mount small chips like a QFP-48 with a heatgun and solder paste, and would then use a fine tipped iron, fine wick, and a fine hobby knife type razor blade to clean up any excess, but he was pretty good at it and rarely had to do any retouching from what I remember. I also remember him reflowing a board he etched in a old frying pan on the stove once...seemed to work! I haven't tried his heatgun method (haven't needed to do small chips), but will be soon, so I may not be remembering something important about how he did it. Secure the board to something, and hold the chip in place with tweezers so the air doesn't blow it away, I think he would also use a dab of some fairly sticky flux to hold the chip if the solder paste itself wasn't enough.

Fiddly chips are probably best to put on a breakout board, many different packages are available online for hobbiest fairly cheap with solder mask, less tiny chips like to-252's and most surface mount opamps, resistors and caps can be done by hand with a soldering iron with a fine tip.

My hands aren't that steady and shake a bit when i try to do fine work so I have a block of high temperature plastic of some sort (I think it's teflon, it was an off-cut from a different engineer friends project, either way it's white, shiny, and my 25w iron doesn't melt it) with a notch filed in one side to support my soldering iron. I rest the thick metal body of the iron above the tip on it, about mid way up. I got the idea shooting with a rifle rest one day. If I can do surface mount chips this way pretty much anyone should be able to as well.

That diode would be easy to do with an iron in the method I've described above, again, secure your board to something, use tweezers, and solder paste makes it easy. Lead based solder melts at a lower temperature so I find it makes surface mount easier, and would probably be a good idea with the heat gun. I'm sure lead free would work fine too though.

I think I've seen clamping "sockets" like the ones used for computer CPU's for opamps and other similar surface mount components, these are an option if you want to be able to easily replace or change a chip, however I think they were pretty expensive and I'm not sure how available they are or for what packages.

Good luck! Surface mount really isn't as hard as it seems like it would be, it's way more DIY friendly than some would have you believe!
 
SMD parts require tooling not found on a "typical" tube/valve DIYer's bench. The soldering equipment we use rates to destroy the parts, before pigtail leads are attached in a satisfactory manner. :( OTOH, some useful items don't come in leaded versions. I've found myself wishing for leaded versions of parts, like this Schottky diode, on any number number of occasions.

How to escape from "Frustration City"? My thinking is to attach pre-tinned, solid, copper leads to desirable SMD parts with silver bearing, electrically conductive, epoxy. Please notice that curing time decreases and conductivity increases with the application of heat. The hot air guns we use with heat shrink tubing, set on the low setting, appear to be a good fit for the elevated temperature "requirement". Given the high cost and 2 year shelf life of the conductive cement, working on batches of parts that exhaust a given epoxy supply seems to be proper procedure.

Have at it, guys! Do my cogitations have merit?

Eli,

I have never tried your method. My greatest worry would be leads breaking or you may not get a good electrical joint.

I do a lot of SMT hand soldered boards for my business. Typically I favor 0805 footprints and I do a lot of hand soldered fine pitch leaded ICs all the time. I use a pencil iron from Pace with a tiny hooked tip. The bent tip helps a lot because I can work at odd angles and I can present as much of the tip to the component as I need.

As stated before, flux is your friend. I use copious amounts and wash it with acetone.

I also use PC boards from Express PCB. They have a simple 4-layer CAD program for free that works extremely well. I do a lot of small run boards and use their production 4-layer service that costs about $110 for three 2.5" by 3.8" boards with solder mask.

You layout one 2.5" by 3.8" board any way you want, divide ground and power planes if you want, and subdivide the boards to get multiple boards out of one, if you want. Express PCB then makes three of those boards for that one price.

There are other services they offer, so check them out.

My point is that PCBs make life very easy and SMT doubly so if you choose to go that route. I like SMTs because you can cram parts together and reduce noise and board space. You can also put through-hole parts in there for a mixed topology board, which we do a lot.

Once you start doing SMT you will really not want to do much through hole work any more. It is just that good.

For the well heeled, Pace makes a killer solder/desolder station with a hot air reflow iron and a desoldering iron. It also can use solder tweezers to pull off SMT parts. It is model MBT 350. Expensive, but if you do a lot of soldering it is the cat's meow. I use the hot air reflow iron only for shrink tubing.
 
Last edited:
For more complex smd parts like opamps I find the best solution is to stick the smd part to the board with a tiny piece of Blu-tac or mortite on the body (I mean TINY, jusy enough so it doesn't move) then flood the leads on with solder and wick off the excess. Works like a charm.

If you apply a good bead of solder flux to the leads before soldering you can literally wet the solder tip and roll it along the IC legs.

The flux will automagically wick the solder along with the tip so all you need to do is wash the board in acetone or alcohol to remove the excess flux.

I routinely do .65mm pitch ICs this way without solder bridges or solder wick, which adds more heat stress to the part.
 
Tom's techniques are spot on. I use all of them myself.

For hand soldering SMD parts I think the biggest aid other than the proper iron is the small solder. Picking up a roll of Kester .38 mm 63/37 with 245 no clean flux was the best investment made for DIY SMT construction.

Another source for flux is the no clean flux for the CHIPQUIK removal system. I purchased the SMD291 10CC syringe of flux from Mouser.

On some occasions I've used 1206 resistors as grid stoppers. One end gets soldered to the tube socket then I connect the other end to the terminal strip with 30ga magnet wire. Only down side is if you are not careful when probing/poking around you can break the resistors off the tube sockets.

I have a Leister hot air SMD rework station. It mostly gets used for part removal when working on PC boards. Putting the parts back on I use the hand soldering methods above.

I was working in the board build operation at Tektronix in a technician position during the early part of the transistion from through hole to SMD. Did a lot of board repair and helped train the SMD repair people on some of the more advance techniques.

Solder paste is great to use and makes working on SMD devices much easier. The downside is the cost of the applicator tools and short shelf life of the paste, makes using solder paste at home for hobby uses impractical.

For removing IC's the safest thing to do is use fine cutters and clip off each lead at the body of the IC. After the body is out of the way you can use your fine tweezers and soldering iron to remove the leads. Clean the pads with the fine solder wick, clean the board and solder in the new part using the techniques discussed above.

Remember, a high quality temp controlled iron and a light touch will take you a long way.

Has anyone tried the SMD tweezer soldering irons?
It looks like it could work great for chip parts like resistors and capacitors.

Gary
 
What you said makes lots of sense, when working with PCBs. My ideas are oriented towards P2P construction.

That dawned on me about 2/3 of the way through my write-up. I figured I was past the point of no return...

I remain very leery of parts without leads.

I was too until I actually tried it. If you can find an old PCB with some decent size surf mount parts on it, you can always desolder those and try soldering your own parts on there. Just an idea...

A past roommate of mine who was in school to be an electronics engineer at the time (now does R&D at a well known consumer electronics company) would do surface mount small chips like a QFP-48 with a heatgun and solder paste

Yeah... I wasn't going to bring that up... :) One of the guys I work with is able to solder LLP/QFN packages that way. I think he was getting 80+ % yield on the heat gun process. Now we have a reflow machine and the yield is closer to 100 %. Note that the LLP/QFN packages have no leads sticking out. Only pads on the bottom of the part.

Another guy at work is capable of soldering a single strand of wire to a pin on a mounted LLP/QFN package!

Oh, and then there's Matt who can solder 01005 components (0.01" x 0.005" size) by hand.

Oh, yeah. SMD work by hand is possible!

I also remember him reflowing a board he etched in a old frying pan on the stove once...seemed to work!

Lot of people use toaster ovens for the reflow.

You could make a very small rectifier board by etching, grinding, or cutting the copper from a single-sided blank.

For SMDs you really need the footprint to be right. Or at least close. Taking a Dremmel tool to a piece of scrap board is not a recipe for success in my opinion. Either lay out a small board and do a toner transfer or use one of the prototype boards available. SurfBoards are quite nice for this. They're a bit spendy for what they do, but very convenient to use.

While I'm at it... For desoldering SMD parts, hot tweezers are really nice. That's the pro and expensive solution. On resistors and capacitors, I find that a wide tip on the soldering iron works better. Just heat the entire component, let it stick to the soldering iron, and pull it off. Of course, the component is toast at that point. Don't plan on reusing desoldered SMDs anyway.

~Tom
 
I LOVE SMT.

It didn't exist when I started doing electronics. Now I use it whenever possible.

Smaller.

Cheaper.

Less holes to drill if you make your own PCBs.

Less parasitics.

One-shot soldering in a (dedicated) toaster oven. Which, BTW cost less than a classy soldering iron.

I still use leaded components, although sometimes I surface mount them. I even surface mounted a tube socket for prototyping.

If I want to remove a chip, I cut it off, leg by leg, using a scalpel. Pick up the legs with a soldering iron, clean up the pads with solder wick. You can't re-use the chip...

Solder paste is entirely practical for home use, buy it from dealextreme and get some flux jelly to mix in with it, so it's easier to use with a syringe. You need a coarse, blunt needle, such as used for refilling printer ink cartridges. DX used to carry plastic nozzles for this purpose, they may still. You can etch a solder stencil in copper or brass sheet using toner transfer. It'll probably be a bit crude, but it doesn't need to be any better, because you can get away with using a syringe to run a bead of paste right across the pads for a 44-pin TQFP.
 
I prefer SMD components, too.

I started to do it with an iron, in the same way as Tom, but then I read about the toaster-oven method, and from there to the Sparkfun tutorials on reflow. They really wrote the book on it:

https://www.sparkfun.com/tutorials/59

And I absolutely concur with their winner: the 1500W electric hotplate.

Yes, just a cheap (GBP 18 = Euro 20 = USD26) 1.5kW electric hob.
Other tools:

- pointy-tip tweezers;
- EDSYN syringe of lead-free solder-paste
- IR thermometer
- PCB for preference, or get some plain FR4 and cut a matrix of tracks into it with a hacksaw.

The sweet thing with the hotplate is that you just:

- dab paste on the PCB pads (don't need to be too careful);
- use tweezers to mount all the parts;
- meanwhile preheat the plate to 200 deg C
- place the PCB on the plate & turn it up to 11
- monitor temperature with the thermometer and tweek your method to meet the reflow profiles given in many data sheets.

The investment is very small, and the results are excellent (yes, you can solder 0,5mm pitch chips this way). Why hesistate?


.
 
This guy has the toaster oven reflow process down:
Gbcart rev2: Assembly & Reflow - YouTube

A friend of mine knows him and says he claims to have 87 % yield on soldering 200-ball BGAs that way. From the video, it looks like he's using a stencil made from mylar or similar plastic film. Pretty neat. I bet you could hook up with the Fabber culture and find someone with a laser cutter who could make the stencil.

~Tom
 
For SMDs you really need the footprint to be right. Or at least close. Taking a Dremmel tool to a piece of scrap board is not a recipe for success in my opinion. Either lay out a small board and do a toner transfer or use one of the prototype boards available.
For a rectifier board?
It sounds like you're leery of well-known DIY techniques. I would use a resist pen before I attempted a toner transfer on such a little board. Or paying for an overpriced SMT prototyping PCB.
I agree with you if we're talking about a full circuit, but for four diodes or less you're just wasting time and money IMO.
 
- place the PCB on the plate & turn it up to 11

"Turn it up to 11". ROFL! I love your scientific methodology there, Rod. But, seriously, that's all that's needed. If need be, you can hit the top of the board with a heat gun while it's on the hot plate.

I've heard about people using a regulator stove burner for reflow as well. Just put a piece of aluminum on top of the burner as a heat spreader.

I think, personally, I'd go for a toaster oven reflow as the cool-down is easier to control.

~Tom
 
"Turn it up to 11". ROFL! I love your scientific methodology there, Rod. But, seriously, that's all that's needed. If need be, you can hit the top of the board with a heat gun while it's on the hot plate.

I've heard about people using a regulator stove burner for reflow as well. Just put a piece of aluminum on top of the burner as a heat spreader.

I think, personally, I'd go for a toaster oven reflow as the cool-down is easier to control.

~Tom

Yes, I am oversimplifying a little, but if you try it with a 1500W hotplate, and a heat spreader (eg a 3mm thick Alu extrusion of Area slightly above the PCB's), and the method:

- preheat to 180 .. 210 deg C, eg for 5-7 minutes
- place the PCB
- turn it right up! 1500W like this is perfect for PCBs around 100 x 100mm and probably anywhere around that size. with the jobs I have worked on, the temperature slopes have been nearly perfect.

.. and monitor the temperature with the IR pistol-thermometer - you can get very close to the soldering profile in the data sheets. Varying the preheat temperature allows tweeking of the profile to get in within the recommendations.

Connectors, big fat SMC diodes, chip caps and 100-TQFPs can all be reflowed in one pass, and the joints are - and look - just like a full factory job.

The cool-off is easy - just pick up the board with pliers and put it down (CAREFULLY!) on four stand-offs (eg M6 nuts) and watch it cool within spec.

The reason for preheating is that the profile requires a dwell at 150-200 deg C to activate the flux and control the main profile. Don't start the hotplate on full - the 1500W type can get above 300 deg C, and we must keep components to 250 deg C max., and even there, only briefly.

for bigger boards, look out for the 1800W electric Teriyaki griddles, or the skillet in the Sparkfun guides.

Toasters are not so good as they look, in practice - it's harder to see when everything has soldered, and noncontact heat is not so even (see the Sparkfun guides).

Recommeded lead-free paste in a syringe:

Edsyn CR88

CR 88 - EDSYN - SOLDER PASTE, 96.5/3.5, 221 DEG | Farnell United Kingdom

Recommended flux:

ELECTROLUBE - SMF12P - FLUX, PEN, 12ML

SMF12P - ELECTROLUBE - FLUX, PEN, 12ML | Farnell United Kingdom

I don't bother with a Stencil, even for making 10 boards, even with 100-TQP on 0,5mm pitch. Just dabbing with the syringe of paste is reliable, when you get the hang of quantity (very little!). The clean-up with stencils is more hassle than it is worth, and you waste a lot of expensive paste that way.

Either way you reflow, this is a great fun way to make stuff - and hugely quicker, and neater, than doing the joints one-at-a-time. Highly recommended to all!
 
The reason for preheating is that the profile requires a dwell at 150-200 deg C to activate the flux and control the main profile.

Yep. The recommended profile calls for a relatively long preheat period as I recall. I know the $20k METCAL reflow/rework machine we have at work takes about 15 minutes to go through the full profile. (The 15 minutes is my estimate, I haven't actually timed it as I don't normally use that machine).

Toasters are not so good as they look, in practice - it's harder to see when everything has soldered, and noncontact heat is not so even (see the Sparkfun guides).

That's been a concern of mine as well. The uneven heat is evident when using the toaster oven to toast bread. The center gets burnt and the edges barely tan.

One note on this reflow (regardless of process): Do it in a well ventilated room. It does tend to stink up the house... It's probably no worse than regular soldering, but all the soldering happens within a 10-minute window.

~Tom
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.