EEVblog Electronics Community Forum
EEVblog => EEVblog Specific => Topic started by: EEVblog on November 30, 2014, 09:38:34 pm
-
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 (https://github.com/shackspace/bttf)
http://shackspace.de/?p=4343 (http://shackspace.de/?p=4343)
Video of Dave blowing up the Time Circuits:
https://www.youtube.com/watch?v=ehXw1Amj40E (https://www.youtube.com/watch?v=ehXw1Amj40E)
EEVblog #689 - How To Rework Solder SMD Chips - BTTF Time Circuits Repair! (https://www.youtube.com/watch?v=HGgat5IVfFE#ws)
-
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!
https://www.youtube.com/watch?v=tlSY1uaw0GA (https://www.youtube.com/watch?v=tlSY1uaw0GA)
-
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.
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.
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.
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! :scared:
-
Don't forget the enamelled wire trick!
https://www.youtube.com/watch?v=tlSY1uaw0GA (https://www.youtube.com/watch?v=tlSY1uaw0GA)
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. :palm:
-
Don't forget the enamelled wire trick!
https://www.youtube.com/watch?v=tlSY1uaw0GA (https://www.youtube.com/watch?v=tlSY1uaw0GA)
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 :D
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 users
Does anyone else find these half-complete repairs really frustrating?
I need closure! :scared:
yes :(
if you really want to have your itch scratched go to
RetroGameModz https://www.youtube.com/channel/UCwkbpzAEUNnAH28UxPsHsxA (https://www.youtube.com/channel/UCwkbpzAEUNnAH28UxPsHsxA)
or Core https://www.youtube.com/channel/UCzlXwR8hC2cY2r_TbME3ORw (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 :P
-
As an example, you can find on the adafruit repository
https://github.com/adafruit/DS1307-breakout-board
the png output. ds1307.png
But I see also that the .png last commit is 4 years ago, while every other file (.sch, .brd) is 4 months ago.
I would wonder if the png is still up-to-date and still trustable picture of the current project status.
-
Instead of use "patented" ChipQuick kit you always may use https://en.wikipedia.org/wiki/Rose%27s_metal (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?
I think it should be available from local stores/suppliers. Export from Russia is expensive.
I see several lots on ebay (from Eastern Europe countries), but price not so good.
http://www.ebay.com/itm/Roses-metall-100g-Roses-metal-alloy-Rose-Legierung-bismut-wismut-Zinn-Blei-/221249414876?pt=Labor_Zubeh%C3%B6r&hash=item33837e26dc (http://www.ebay.com/itm/Roses-metall-100g-Roses-metal-alloy-Rose-Legierung-bismut-wismut-Zinn-Blei-/221249414876?pt=Labor_Zubeh%C3%B6r&hash=item33837e26dc)
-
A few things I noticed here:
#1 the footprint on the circuit board was generated from a crappy library - it probably had an english to metric unit conversion done, and the rounding errors add up. Note how the pad for pin 1 is offset a little high, and as you get down to pin 12 (or whatever the bottom pin on the same side is) the pad offsets are now on the low side. I experienced this once using a library that came with Eagle, the 100 pin QFP footprint was 1/2 a pad (.025 at least) off on a .050 pin pitch part. Now I always make my own footprints, and I use Altium. Start in metric or english, whatever the part is natively (usually metric) when making the footprint.
#2 I would first use solder wick on all the pins, maybe with a little extra flux on the solderwick.
Then I would apply the iron to a pad (not pin), and use the tip of an exact-o to pry that pin up.
Repeat til done. With practice, it is pretty fast.
#3 Also using a dremal to chop off the pins will allow desoldering them individually with little chance of pad damage.
Although seeing how easy the chipquick worked, I think I'll be getting some in the future.
-
Although seeing how easy the chipquick worked, I think I'll be getting some in the future.
I think you will find hot air is cheaper in the long run.
-
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.
Did a repair on a Samsung tab3, looked like a 3 year old had tried to insert something pointy into the micro USB, result?
Mashed connector :'(
I had two issues:
looks like the Tab 3 uses a non standard usb connector (found on ebay) and the old connector was a mongrel to remove.
Tried chip quick , did not budge surprising given I use a lot of chip quick and have used it successfully on removing connectors, ICs etc. Used hot air and it still did not move, got the board really hot and was beginning to melt the captan tape that was used to mask some small SMT caps. :wtf: never had this before as the board was very thin so there can't have been much thermal mass to the small board. Did not want wreck the board so....
In the end I used a Dremmel, I was really careful and just ended up cutting the connector to bits and removing by piece meal. Got the new connnector soldered on ok and the tab3 is working again :)
The problem was the the connector casing was soldered to the ground plane in three places, two sides and the Middle not to mention the fine connections for the socket as it was dead easy to rip the pads on that thing. Dip SMT ICs are far easier to shift.
-
Does anyone else find these half-complete repairs really frustrating?
I need closure! :scared:
I dont find it frustating. I saw a soldering method. Done. If the circuit was working after this or not is a matter of interest for the owner :P Not my shit :)
Dave has in the past made long repair videos where the subject is trouble shooting...this time i think it was soldering...so my needs were satisfied :P
-
Tried chip quick , did not budge surprising given I use a lot of chip quick and have used it successfully on removing connectors, ICs etc. Used hot air and it still did not move, got the board really hot and was beginning to melt the captan tape that was used to mask some small SMT caps. :wtf:
WTF indeed considering Kapton melts >250'C, you arent supposed to burn off solder, just melt it
-
I find it much easier and precise to glue the chip down with rosin flux and then solder pins while the chip is fixed on the board. This way is it possible to precisely place the part while the flux is slowly getting viscous as its solvent evaporates. After precise placement just push it on the top a little bit and after that you can't accidentally move it with iron tip, provided, everything is done carefully.
-
Tried chip quick , did not budge surprising given I use a lot of chip quick and have used it successfully on removing connectors, ICs etc. Used hot air and it still did not move, got the board really hot and was beginning to melt the captan tape that was used to mask some small SMT caps. :wtf:
WTF indeed considering Kapton melts >250'C, you arent supposed to burn off solder, just melt it
Totally agree ! The trick is when things are not going smoothly. STOP! You are probably doing it wrong :palm: got there in the end though and the board was not scorched or damaged :phew:
-
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 file into the actual source repository
All true. Problem is that Github has become a project hosting platform, and therefore a place where people do expect to find rendered output / tarballs / whatever as well.
I don't think that's what was intended from the beginning when Github started to offer a place to host opensource repos, but it happened, and we're seeing the collision of two separate ways of doing things here. There's no obvious correct solution. (Other than setting up a "conventional" website as well.)
Edit: I should add that this isn't isn't intended as criticism in any way, just wanted to add something. (You never know how people are going to read something that's clear and unambiguous inside one's own head, as we've seen.)
-
Don't forget the enamelled wire trick!
https://www.youtube.com/watch?v=tlSY1uaw0GA (https://www.youtube.com/watch?v=tlSY1uaw0GA)
That looks like it will rip the last pad off the PCB with incredible ease. Watch again the part of the video where he reaches the last pin, I don't know how the pad stays on the board (maybe it didn't...!)
Compared to chip-quick? No thanks...
-
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.
Don't use a knife use the specialized tools for it, like the Erem Miniature angle cutter for delicate cutting applications, it costs €60 a piece but it is a beauty for cutting smd pins under a microscope, just don't use it for anything hard or above AWG24 or it will be damaged which is how I ended up owing one for free (with a dent in it) :D
Best is, there is no excessive heat to the pcb so zero chance of doing damage to the pads. The chipquick looks nice but the same as mixing up leaded with unleaded solder you need a lot of time (if you want to do it right) to clean it up and make sure nothing is left. I would only use that if I needed the chip itself again alive.
-
I just checked the KiCad board file for the display board located at https://github.com/shackspace/bttf/blob/master/display/Uhr_Modul.brd
and just as I suspected, it has a crappy footprint. using KiCad I measured, and the pads for the part Dave replaces are located on 0.653796mm centers.
(I measured from the top edge of pad 1 to the top edge of Pad 11, and divided by 10. )
Not the 0.65 mm centers I suspect the Data Sheet says.
It doesn't sound like much, but it accumulates, and will be a problem on higher pin count parts (like I said before, quoted below).
It was off by enough for me to see while watching Dave's video.
I tend to not trust libraries that I didn't create. I trust but verify the Altium ones.
A few things I noticed here:
#1 the footprint on the circuit board was generated from a crappy library - it probably had an english to metric unit conversion done, and the rounding errors add up. Note how the pad for pin 1 is offset a little high, and as you get down to pin 12 (or whatever the bottom pin on the same side is) the pad offsets are now on the low side. I experienced this once using a library that came with Eagle, the 100 pin QFP footprint was 1/2 a pad (.025 at least) off on a .050 pin pitch part. Now I always make my own footprints, and I use Altium. Start in metric or english, whatever the part is natively (usually metric) when making the footprint.
-
Does anyone else find these half-complete repairs really frustrating?
I need closure! :scared:
I dont find it frustating. I saw a soldering method. Done. If the circuit was working after this or not is a matter of interest for the owner :P Not my shit :)
Dave has in the past made long repair videos where the subject is trouble shooting...this time i think it was soldering...so my needs were satisfied :P
This video was never meant to be a repair troubleshooting video. It was supposed to be a video about replacing the chip and the technique used.
The unit was fully expected to work after this repair. It didn't for some as yet unknown reason.
I have since shot some troubleshooting video which the original designer has seen and he's scratching his head too.
I still haven't finished that video and it's already 22 minutes long edited. Could take another hour of edited footage to find out what's wrong at this stage.
The original repair video stands on it's own and IMO is very useful. I'm not going to hold off on that and make some super ridiculously long video just because a few people bitch that I didn't get it going in the one video.
Some video bloggers do that, hold off on a video until it's all done and dusted, and then edit it to make it look like it was an easy fix, but I don't do that. If I think a video has enough material to be useful in it's own right, then I release it.
-
It doesn't sound like much, but it accumulates, and will be a problem on higher pin count parts (like I said before, quoted below).
It was off by enough for me to see while watching Dave's video.
Interesting, I got that same vibe. I thought it was just me going a bit wonky...
-
I though of way to easy remove some ic chip with a small clearance. use a magnetic wire. feed wire between two middle pins and lightly pull up while heating the pins on the the ic chip like Dave show in the video.
Can you check if green isn't requiring more voltage/amperage that the other 2 need. All 3 should be at the same brightness, there could be a problem with SS0P24 ic chip on the green one.
Have you check to if you Arduino unit too verify it OK.
Have you though about just take a Breadboard and see if there nothing wrong with how wired.
-
I though of way to easy remove some ic chip with a small clearance. use a magnetic wire. feed wire between two middle pins and lightly pull up while heating the pins on the the ic chip
and pull that pad of the pcb
Amazing, no matter how many times you tell people to NOT USE ANY FORCE, especially together with high temperature, they keep insisting on inventing new ways to break something :( :palm:
muffenme copper tracks are 'glued' to pcb, they are easy to rip off to begin with. When you heat up pads they become even more fragile. There really is no need to innovate here, just heat all the pins at the same time, wait for solder to melt and lift whole part (using vacuum picker or tweezers).
Those 'easier' homemade methods end up more time consuming and potentially destructive, and all because people are afraid of SMD and/or dont have hotair. Its only ~$40 (for the cheapest chinese hotair, the one Dave tested some time ago for example), just buy one already and enter wonderful world of no hassle smd rework.