Soldered 0805 vs 0603 vs 0402 vs 0201 vs 01005 vs 008004 today :)

[Pedro147]

A quick question - your tweezers are non-magnetic, right?

Yes they are, thanks though    It is just the flux sticking to them.

[analogo]

Hol' up. There are two 0603s, depending on which side of the pond you are?

There are also two 0402 and two 1005

https://en.wikipedia.org/wiki/File:SMT_sizes,_based_on_original_by_Zureks.svg

[khs]

If you want to use tiny (ceramic) capacitors, take into account the capacity decreases with increasing voltage.

This effect increases with decreasing size, so very small capacitors may not work.

https://www.maximintegrated.com/en/app-notes/index.mvp/id/5527

shows the effect for 1812 vs. 0603.

0603 is not very small.

 

[alexg]

Hi All, just accidentally found out today that this post has been replied recently, I did that experiment back in February and not it is October, I am still learning, however I've been hand soldering under microscope every couple days day since then and now looking back I'm thinking it is not as scary as it seems all the way down to 01005, 008004 would still require individual attention but also very doable.
If I had to give an advise to anyone doing exactly same experiment for fun, I'd say you need:

 - Good tweezers, that one on the picture I bought 2 years ago for about $6 - $7, I also have couple sets of those relatively cheap ones, the one you can buy off ebay for like few dollars for bunch, those are good for nothing, they flex too much, they look sharp at first glance but under 10x you can see that tips are uneven, no feedback at all, I only use them to pick dried silicone glue off a tip of nozzle. Those that I have in the picture have nice sharp even tips, non magnetic, made from some kind of nice steel, I think all small footprint SMD rework starts with good tweezers.
 - Do not drink coffee, lift any weights, move lawn, or do any physical labor before this project, this will make your hands shaky and won't get anything right, you can pick a component from the tray, etc, but once you place it on board and need to hold it steady or nudge it a little your movements won't be as precise as needed.
- Get PCB pads size right, as you can see in my experiment pads for 01005 and 008004 spread apart more than they suppose to so those components we constantly pulled to one side more than the other so I had to constantly correct it into place bumping off center, etc. I think with right pad sizes it would have been much simpler.
- Do not heat it with hot air from the top, even at slowest airspeed it will blow those 01005 and 008004 right off, apply heat from the bottom under components.
- Do not try applying solder paste to 01005 and 008004 pads, its just too small, apply flux, then drag some solder with iron over pads it will make a solder bump on each pad, then place your component there and when you reheat solder from the bottom it will grab your component.
- And of course you need a microscope, non of those head mountable magnification glasses will do the job in my opinion, I used SE400Z with 20x for this project and that was just right.

My three favorite sizes now are 0603, 0402 and 0201 with most favorite being 0402
0603 is nice and juicy package with, resistors have printed values, capacitors have nice large values, easy to grab, easy to place, easy to flip if its upside down right on the spot, pad is big enough so applying solder paste is easy, etc
0402 is my favorite, easy to maneuver  under microscope, pad is still big enough to apply solder paste, it is very space saving, so when I design circuits I can fit more parts per square inch so to speak, love it.
0201 I like very much too for its space saving abilities, but unfortunately moving it from tray to board, grabbing with tweezers, etc, is more challenging when you need to do bunch of them, one or two, okay, but when you need to populate a full board it will be annoying, also flipping them right way is a nightmare, they stick to tweezers etc,

Those are my personal thoughts and observations, it may or may not work for others but hopefully it can help you a little.
I have to redo this experiment again now as a skills refresher

Hi All, just wanted to share cool experiment I did today, I soldered few capacitors today on same board starting from 0805 down to 008004

Just curious: What size of tip, and what diameter of solder did you use? Did you add extra flux?

[Howardlong]

I was intending to do a video of making something useful at the WLCSP close pitch, 0201 and 01005 level, but haven't got around to it.

In the meantime here's a couple of pics of how to flash an LED from a single 1.5v cell.

There are some design considerations here; first you need something off the shelf that will work with 1.5v, and preferably down to about 1v. Second, your LED needs 1.6v to show any signs of life.

This design uses some 0.8V logic, two schmitt trigger inverters in a single package. One oscillates at about 2Hz to flash the LED, but to get enough voltage to light the LED you need to up the game a bit, so the second schmitt trigger runs at around 10kHz and operates a single stage charge pump. This won't work properly above about 1.9v input or so as the schmitt trigger's output clamp diodes will start to conduct so you'll get dim/bright rather than on/off. Even the diodes are very carefully chosen for minimum Vf and Ir.

Now the dual Schmitt trigger, an SN74AUP2G14, is available in a number of package sizes, but the smallest is a 0.4mm pitch BGA, so I went for that. None of this SC70 rubbish. The LED still flashes (just) down to 0.95v.

Board is BusBoard SP3T-50x50-G.

I found out a while back when soldering BGA/WLCSP packages in this way is that you need to use lead free solder, the connections become very brittle and fragile otherwise. Also for these reasonably high impedances you need to put the board in an ultrasonic bath: when I was debugging this board, and couldn't get it to work, I could see the ionic reactions bubbling away under the microscope. At that point I chose to put it in the bath, rinsed and dried it off,  and ten minutes later it worked perfectly.

As soon as you apply any flux again though, you need to re-clean.

For normal stuff, 0402 and 0603 are my typical targets: 0201 and smaller are typically more expensive, so for volume production it doesn't make sense unless size is a key design parameter.

Note this is the package top view, when flipped on the board to break out, pinout is mirrored.


Penny, SOT23 and 8 pin DIP for comparison




FWIW, a few months ago I did do a video soldering a 0.5mm 20 pin WLCSP to a breakout board.
https://youtu.be/e9mdGqu1aZ0

[Pedro147]

Those are my personal thoughts and observations, it may or may not work for others but hopefully it can help you a little.
I have to redo this experiment again now as a skills refresher

Thanks for the great insights and tips alexg. 

Nice work on the BGA breakout and LED circuit too Howardlong

[adras]

Impressive, well done! Too bad all these magnified pictures don't really show the real size.

I recently failed building this kit: https://www.aliexpress.com/item/Free-Shipping-Factory-Price-DIY-Spherical-Rotating-LED-Kit-POV-Soldering-Training-Kit-blue-and-red/32755860635.html?ws_ab_test=searchweb0_0,searchweb201602_1_10152_10065_10151_10068_10344_10342_10343_10340_10341_10193_10194_10304_10307_10060_10302_10155_10154_10056_10055_10054_10059_10534_10533_10532_100031_10099_10338_10339_10103_10102_10052_10053_10107_10050_10142_10051_10320_10321_10322_10084_10083_10080_10082_10081_10177_10110_10111_10112_10113_10114_10180_10312_10313_10314_10184_10319_10078_10079_10073_10186,searchweb201603_14,ppcSwitch_5&btsid=0c5f1b34-f2ed-4ea9-9ec2-c59d0a27e281&algo_expid=efa01a5c-a789-4d7b-bbb9-9d83de61949b-1&algo_pvid=efa01a5c-a789-4d7b-bbb9-9d83de61949b

Soldering the LEDs is a huge pain in the ass. So if you're looking for a challenge try that kit. And send it to me when you're done

[Howardlong]

More nerdy micro soldering. This time a 12 ball WLCSP ATTINY20, I believe to be the world's smallest MCU (before you go to bare die), set up as a blinky. There's a SOT23 and DIP8 there for comparison.

The LED is 0603, about the same size as the MCU, and you might just be able to make out the 01005 current limiting resistor. Pitch is 0.4mm offset, so a bit more dense than a straight 0.4mm pitch, makes soldering the middle balls interesting, both of which are used.

The programming header is 0.05" pitch.

[Pedro147]

Howardlong you have far too much time on your hands 

But no seriously this is fantastic work. A video of your creation in action would be awesome if possible one day.

[nanofrog]

But no seriously this is fantastic work.

+1 

If you hadn't made the above couple of posts, I'd never have thought doing something that small possible without an adapter board or custom PCB.

Even using my microscope, I'm still not sure I could manage it though (essential tremor in my hands and arms).  At least my neurologist said I don't have Parkinson's. 

[MT]

Close to masochism..     For the sake of fun i would like to see a similar video but by someone who has parkinson!  
Howardlong deserves  on his shoulder for his soldering achievements!

[Vtile]

When 1206/0805 surface mount PCBs first started to appear the old 'gray beards' complained that 'electronics just wasn't serviceable any more'

Now we have 008004 and people are still trying to solder it at home

Well yeah, it is not serviceable any more. Why there is a such thing as a board swapping if it would be truly serviceable.

[Vtile]

I think I need these tools to enjoy the future of electronics as a hobby:

https://images-na.ssl-images-amazon.com/images/I/61vilOmXMCL._SX522_.jpg
https://www.motoman.com/hs-fs/hubfs/Blog_Robots/Application_Robot_Page_Images/SDA10_600px.jpg?t=1508273115461&width=300&name=SDA10_600px.jpg

[nanofrog]

Well yeah, it is not serviceable any more. Why there is a such thing as a board swapping if it would be truly serviceable.

Even if a board is technically capable of being reworked (and may be at some point after an initial board-swap to get the customer up and running as quickly as possible), the labor rate usually makes it more expensive.

[Pedro147]

Even if a board is technically capable of being reworked (and may be at some point after an initial board-swap to get the customer up and running as quickly as possible), the labor rate usually makes it more expensive.

A very valid point. Aside from the cost constraints, there is also a lack of knowledge when it comes down to diagnosis and component level repair of some of today's devices. People like Louis Rossmann, Ipad Rehab, Chris Long and STS Telecom among others on Youtube are leading the way in showing what is possible particularly in the area of Apple products.

I suppose because the cost of these devices is high, it can make it more attractive to have them repaired seeing as Apple refuses to do nothing but the most basic repairs to their products 

[bson]

For really small parts, say 0402 or less, just the tinning from a HASL coat is enough to solder.  Dot pads with a flux pen, place the parts, and apply heat - hot air, plate, oven, etc.  I generally personally prefer ENIG plating (mainly so boards don't corrode just sitting around), but sometimes HASL is pretty handy.  On the down side it's less flat for larger ICs, and makes them every so slightly harder to align (the tinning bulges), but not a big deal.

Wiha bent tip ESD safe tweezers are pretty good.

Edit: I'm a fan of Lindstrom hand tools in general, and decided to buy a set of their TL ESD tweezers to try them out.  So now have these on order: http://www.techni-tool.com/844TW030

[Pedro147]

Interesting take on HASL which I normally use. Also an informative write up on surface finishes http://www.epectec.com/articles/pcb-surface-finish-advantages-and-disadvantages.html

Some nice tools there and yes good tweezers are so important aren't they.

[nanofrog]

Wiha bent tip ESD safe tweezers are pretty good.

Edit: I'm a fan of Lindstrom hand tools in general, and decided to buy a set of their TL ESD tweezers to try them out.  So now have these on order: http://www.techni-tool.com/844TW030

VOMM is the ODM for Wiha tweezers. That particular shape is a #7 on the Swiss profile list.

Ideal-tek is the ODM for Lindstrom's tweezers (Bernstein, TDI, and others), and the vast majority of mine are ideal-tek (brands vary). FWIW, I've a pair of TA w/ conductive ceramic tips, NC (#6), but the vast majority are SA.

• TA = titanium, softest spring force; these can handle 1400C continuously without issue
• SA = stainless, anti magnetic, and anti-acid; medium spring force
• NC = nickel-chromium, strongest spring force IME; to put this in perspective, it's similar to a decent ball point pen

All of these are chemical and non-magnetic. Temperature tolerance varies (particularly TA with ceramic tips), as does hardness and weight of the alloy. TA would be the lightest, and NC the heaviest for a given shape for example.

Erem (105pgs., .pdf) makes excellent tweezers, as well as Dumont, Regina**, Rubis***, and Sipel (all are Swiss made, and Sipel is the least expensive of these brands IME; do note they're not finished as well as the others).

There's another company by the name of Viola* (COO=Italy), whose tweezers aren't bad either. These are almost always rebranded; mine are stamped Excelta (doesn't look like they carry Italian made any longer), but TDI also offers them (look for SA-I at the end of the P/N). To put these in perspective, Swiss made tweezers are usually spec'd at 0.001", while the Italian made are 0.003". Finish is nice IME.

* Link is for a retailer as I wasn't able to find the company's website. They're less expensive than Swiss made as a general rule.

** Be careful, there's fakes out that are made in China; they're easy to spot as the last I looked, they were going for ~$6 per pair (Vetus?). To put this in perspective, the genuine Swiss made Regina ~$30+ per pair.

*** French company, but their tweezers are made in Switzerland. Example.

[Howardlong]

Interesting, I didn’t know Ideal-Tek were the ODM for Lindstrom: I have a few of each, and Ideal-Tek are usually a bit cheaper.

I have to say I am a complete noob about the grading and numbering of tweezers, and how it works across brands (or not). #7 seems to mean bent tip, which I like. Then there are different grades of sharpness, but I don’t know what the difference is in practical terms between, for example, fine, extra fine, and needle, but I know what I like.

Is there a noob guide?

[nanofrog]

Interesting, I didn’t know Ideal-Tek were the ODM for Lindstrom: I have a few of each, and Ideal-Tek are usually a bit cheaper.

I have to say I am a complete noob about the grading and numbering of tweezers, and how it works across brands (or not). #7 seems to mean bent tip, which I like. Then there are different grades of sharpness, but I don’t know what the difference is in practical terms between, for example, fine, extra fine, and needle, but I know what I like.

Is there a noob guide?

Yeah, I figured out a long time ago to find the ODM, as there's no other company's markup added to the price. 

Another that might be of interest, would be Schmitz, who only make cutters and pliers and is the ODM for companies like CK Tools, Bernstein, and NWS for example. Really good stuff IME, and typically 50%, sometimes less, than brands like Erem or Tronex's street prices (both make really good stuff too, but Schmitz offers a LOT more value due to their cost; performance is definitely on par). The fast majority of my pliers are Schmitz, and I've one of their cutters (CK Tools labeled). Most of my cutters I found used, usually in mint condition, on eBay for ~$10 each (Excelta branded Tronex, Erem, Swanstrom, and CK Tools).

Regarding Swiss patters, take a look at any of the other Swiss tweezer makers, Ideal-tek in particular, or for a shorter list, Grobet. FWIW, these patterns were originally designed for watchmaking, the shapes have applications with electronics as well. But instead of INOX used in watchmaking (stainless steel that's magnetic), they're TA, SA, or NC (there are other alloys, such as one that contains Cobalt).

[Howardlong]

I did a video of soldering up a complete circuit from beginning to end with a 0.4mm BGA package, and 0201 and 01005 passives, including the good, the bad, and the downright ugly. I need to remember to stop drinking a pint of coffee before doing this stuff!

https://youtu.be/JWS2NWK3D4g

Edit: this is my first attempt at a multi-camera setup, which kicks in at 07:20 after an introduction. I was running four cameras, and switched between them all in post. It definitely adds a lot of time for editing.

[Pedro147]

Wow thanks a lot Howardlong I for one appreciate you taking the time to do this.

Just about to sit down and watch your handiwork. It is going to be good 

[ElektroQuark]

A lot of time but a lot of fun too.
I have three videos on my Youtube channel right now (just started), but I'm really  considering a multi-camera setup, even a multi-angle setup.

[Howardlong]

Rather than doing the switching completely manually, I have a multicam plugin that comes with the software I use (Powerdirector) which supports up to four cameras. I can't particularly say one way or another whether this software is any better than any other software, as the only other video editing I've ever done is with Moviemaker.

The plugin allows me to select up to four media files, and allows you to sync them all up in the plug in by a number of means, I use an audio analysis. You can then compare and choose a default audio channel. You then play all four channels together, selecting which camera view you want with the mouse or the number keys 1, 2, 3, 4 corresponding to the camera view in "real time", but you can also go back and change. This is obviously far faster than having to completely manually edit dozens if not hundreds of separate clips.

On playback I noticed that some of the focussing isn't quite right, so I need to pay more attention to that next time. Getting a decent image through the camera port of the microscope is tedious, as it requires fine and repeated adjustment, and is of course co-dependent with the eye piece focus which is what I'm seeing.

Three of the cameras I use have a half hour cut off, so I have to remember to take a break to re-start recording. Thankfully I did remember!

One of the cameras will do 4k/50p, and you can pretty easily tell from the video which one that is. The other cameras will also do 4K but only at 25fps. I prefer 50fps, I find it's smoother, so I run them in 1080p, all with a shutter speed of 1/100th across all cameras to reduce motion blur. I also ran some of the cameras with the "ex tele conv" which crops the frame to native pixel-for-pixel resolution increasing zoom without deteriorating quality.

There's a lot to learn.

[TheDane]

** Be careful, there's fakes out that are made in China; they're easy to spot as the last I looked, they were going for ~$6 per pair (Vetus?). To put this in perspective, the genuine Swiss made Regina ~$30+ per pair.

Besides the price, what should one be on the look out for to detect the fakes?

Are they total crap, just the best of el-cheap'o tweezers - or are they actually usable in a pinch?
Thanks