-
EEVblog #689 - How To Rework Solder SMD Chips - BTTF Time Circuits Repair!
Posted by EEVblog on 30 Nov, 2014 21:38 -
Dave shows you how to rework and replace a blown SSOP surface mount SMD chip with ChipQuik, solder wick, and drag soldering. And also mentions other methods using a hot air gun and pre-heater.
Can he fix the Back To The Future Time Circuits and restore the timeline to it's original order?
Bonus rant about the lack of PDF schematics in Github projects.
The BTTF Time Circuits were designs and made by Shackspace
https://github.com/shackspace/bttf
http://shackspace.de/?p=4343
Video of Dave blowing up the Time Circuits:
-
Has the wafflemaster replaced the 74 logic?
-
Just wondering, was there any particular reason why cutting off the pins of that chip was not an option?
-
Is this # 688 or 689?
-
Just wondering, was there any particular reason why cutting off the pins of that chip was not an option?
IMO it has a much greater chance of ripping a pads off from mechanical stress. -
Instead of use "patented" ChipQuick kit you always may use https://en.wikipedia.org/wiki/Rose%27s_metal
Retail price for this alloy is about $3 for 100g pack (Russia). -
Just wondering, was there any particular reason why cutting off the pins of that chip was not an option?
I forgot to mention that's another option. Have mentioned that in previous videos.
It's possible, but because of the fine pin pitch you'd have to cut them with a knife and that can put stress on the pads. -
Just wondering, was there any particular reason why cutting off the pins of that chip was not an option?
I forgot to mention that's another option. Have mentioned that in previous videos.
It's possible, but because of the fine pin pitch you'd have to cut them with a knife and that can put stress on the pads.
Dremel + diamond disc + very gentle force applied will do. -
Dremel + diamond disc + very gentle force applied will do.
Good idea! but the standard abrasive cut off discs will cut the leads much better than diamond plated discs. -
Don't forget the enamelled wire trick!
-
Regarding git missing the PDF.
This is a common missunderstanding is what git actually is. Git is a source control tool, it is not designed to track binary files. It is very commonly frowned upon to publish a binary/compiled (consider the PDF a compiled version of the eagle schematic) file into the actual source repository, even in closed source environments. This would be like asking all code developers to publish binaries along side their code at every single revision change.
The accepted practice is to make versioned releases of the compiled/final form of the project (in this case, a PDF) and release them seperately using some other hosting service, not Github. -
Hi, Here is a few things i have found using ChipQuick
You can reuse the ChipQuick blob of solder around 6 to 8 times before it fails to work, reclaim it from the chip & PCB.
You need to 'wash' the ChipQuick under the pads so it mixes properly with the on board solder, the flux helps with this. Use a circular motion as you move along the pins, this will help mix the alloys. I found this method most useful on quad-j type packages.
Dave ! Flux ! ya gotta use flux, it will help the ChipQuick to ball up to the chip pin under surface tension, this also keeps the heat at & in the chip & pins (& gives you more time).
ChipQuick is totally useless on through hole parts
-
Regarding git missing the PDF.
I disagree, I believe your argument only applies to for instance application binary files that can only run natively for specific OS or processor and that is frowned upon because after all that goes against open source. A PDF is not an OS specific file, it is a portable document file format that contains data for display or printing.
This is a common missunderstanding is what git actually is. Git is a source control tool, it is not designed to track binary files. It is very commonly frowned upon to publish a binary/compiled (consider the PDF a compiled version of the eagle schematic) file into the actual source repository, even in closed source environments. This would be like asking all code developers to publish binaries along side their code at every single revision change.
The accepted practice is to make versioned releases of the compiled/final form of the project (in this case, a PDF) and release them seperately using some other hosting service, not Github.
Git is a source revision control source software, it is to help maintain versions of software files. Git does not mean open source. The problem lies with the source being open source and whether the inclusion of PDF breaks the practice of being open source or not. -
I disagree, I believe your argument only applies to for instance application binary files that can only run natively for specific OS or processor.
Java files are platform & hardware independent yet the .jar files are never published into a git repository because they are the compiled form of the source.
Another good example is dogygen documentation, this is a platform independent HTML output generated from comments in code. The doxygen comments are stored in code, but the generated output is never put into the git repository, as it would be pointless to do so. If you want the documentation, either hope that the developers have published it somewhere (just like a PDF schematic), or generate it yourself by running doxygen on the top level directory.frowned upon because after all that goes against open source
Incorrect, binary form does not go at all against open source, what does is when the source to that binary is not available.Git is a source revision control source software, it is to help maintain versions of software files. Git does not mean open source.
I never said it did, I even stated that even in closed source git repositories that pushing binaries into it is generally frowned upon. It is source control, in this case the source is the schematic file from the editing package, not the PDF, the PDF is the result of the source and can not be edited/updated/modified easily, just like a compiled binary.
The main reason why this is not done is to prevent duplication of useless files that anyone with a little patience is able to generate from what is available. Having to maintain both the source file and remember to also update the compiled/generated form on every single change doubles the work and risks inconsistancies.
And finally, what if you have a PCB version V1.0, and the schematic gets revised 100 or so times before it is tagged as v1.1? The developer/manufacturer needs to take a copy of the design files, generate a static PDF or whatever, and package it up outside of git for v1.0 boards manufactured from that exact revision of the schematic. A git repository is a moving target for developers, not end users. -
Has the wafflemaster replaced the 74 logic?
Yeah, I figured he was going to replace all the chips. -
Does anyone else find these half-complete repairs really frustrating?
I need closure!
-
Don't forget the enamelled wire trick!
I guess it's good if you have nothing else and want to preserve the chip... but it looks likes something you should practice a few times too, unless you want to rip pads off...
I'll stick with hot air. -
Instead of use "patented" ChipQuick kit you always may use https://en.wikipedia.org/wiki/Rose%27s_metal
Retail price for this alloy is about $3 for 100g pack (Russia).
Do you have any link where to buy it?
-
The accepted practice is to make versioned releases of the compiled/final form of the project (in this case, a PDF) and release them seperately using some other hosting service, not Github.
Everybody is obviously not sharing this view, and when you see the trouble Dave had in order to get the schematics,
it is clear that posting it with the eagle code will be helpful for many.
As an example, you can find on the adafruit repository
https://github.com/adafruit/DS1307-breakout-board
the png output. ds1307.png
You cannot say that Adafruit is not experimented and makes a wrong use of github.
-
Is there an easy way to do the thumbs up. I find I have to go to youtube to do this.
Also Dave always says he will post the links below but for me they always appear above. Is that normal. -
You cannot say that Adafruit is not experimented and makes a wrong use of github.
I do not say that anyone is wrong in doing so, I am simply explaining why it is more common then not to not retain these files in a git repository. -
In the case of PDF schematics they avoid the problem of the CAD package version changing and not being able to open or mangling the design files. They also allow people who don't have the CAD pacakage installed to view the schematic, with a full version history.
Again, I am not saying that it is a bad idea, I am just educating people as to why these files are often not put into a git repository. -
From the Material Safety Data Sheet for ChipQuick:
Tin 17% - 30%
Bismuth 57% - 56%
Indium 26% - 14%
I can't find an exact match for this as a standard casting alloy, as ChipQuick's formulation has no lead or cadmium, presumably for RoHS reasons.
To give an idea of the price of similar materials here, the extract below from McMaster-Carr's "casting alloy" section
Weights are in pounds and prices in US$.
Melting Temp., °F Approx. Composition Approx. Size Approx.Wt., lbs. Each
Bismuth Alloy
158° 50% Bismuth; 27% Lead; 13% Tin; 10% Cadmium 1/2"×3" Hex 1 1/2 8921K12 $48.78
160° 43% Bismuth; 38% Lead; 11% Tin; 8% Cadmium 5/8"×1 1/2"×4 1/2" Bar 1 1/2 8921K22 45.78
203° 53% Bismuth; 32% Lead; 15% Tin 1"×1"×3" Bar 1 1/2 8921K23 49.89
255° 56% Bismuth; 44% Lead 1/2"×3"×3" Square 1 1/2 8921K14 45.09
281° 40% Bismuth; 60% Tin 1/2"×1 1/2"×4 1/2" Bar 1 1/2 8921K18 64.80
281° 58% Bismuth; 42% Tin 5/8"×1 1/2"×5 1/2" Bar 1 1/2 8921K16 63.84
Indium Alloy
117° 45% Bismuth; 23% Lead; 19% Indium; 8% Tin; 5% Cadmium 1/4"×2"×4 1/2" Bar 1/2 88635K2 101.84
134° 48% Bismuth; 25% Lead; 5% Indium; 13% Tin; 9% Cadmium 1/2"×3 1/2" Disc 1 1/2 88635K26 133.56
136° 49% Bismuth; 18% Lead; 21% Indium; 12% Tin 1/4"×1 7/8"×4 1/2" Bar 1/2 88635K4 111.04
142° 48% Bismuth; 25% Lead; 4% Indium; 13% Tin; 10% Cadmium 1/2"×3 1/2" Disc 1 1/2 88635K28 117.72
-
Just wondering, was there any particular reason why cutting off the pins of that chip was not an option?
Dremel + diamond disc + very gentle force applied will do.
When i don't want a chip, but want the PCB i take a good hold of the chip with tweezers, apply some upwards force to it, then heat it up with a small blowtorch.
Unsolders pretty much anything without damage to the board.
...Is that actually a good idea, or was i lucky all this time?
EDIT: Ah, got to the part of the video where Dave mentions it.
-
Don't forget the enamelled wire trick!
cutting, ripping off, prying with a wire - its all dangerous and retarded, its basically a car mechanic removing screws with a hammer
its you saying 'yep, im incompetent and will risk removing pads/microfractures because DERP'
this is what hotair gun is for, you heat shit up and lift it off, no need for 10 minutes of effing around with some low temp alloy or cutters.
Low tenp is useful when you are constrained by space and have low temp melt-able elements nearby that could get damaged. Downside is waste of time. It might be ok if its your hobby and you are willing to spend tens of minutes on one chip. But then you end up with dave doing rework video of ONE CHIP because he ran out of time
Regarding git missing the PDF.
This is a common missunderstanding is what git actually is.
no its not, this is a religious debate
git is just a tool, you can shove whatever you want in there, its supposed to be useful for you and your usersDoes anyone else find these half-complete repairs really frustrating?
I need closure!
yes
if you really want to have your itch scratched go to
RetroGameModz https://www.youtube.com/channel/UCwkbpzAEUNnAH28UxPsHsxA
or Core https://www.youtube.com/channel/UCzlXwR8hC2cY2r_TbME3ORw
people that repair shit for a living will post full repairs, because they actually repair shit instead of talking about it