[weblacky]

Yo All,
 I was looking into a new glitch product for the DS1286 pointed to by user Dylanear, here: http://forums.irixnet.org/thread-2511.html

A new board-top was mentioned and I found a reference here: http://www.vcfed.org/forum/showthread.ph...tchdog-RTC

So...how in the world was he de-potting these for only $10, getting the original DS1284 chip that's inside, then placing a top on it....HOT AIR, it turns out the potting material is two separate layers.  A harder "shell" of plastic and an inner layer of "other".  Both seem hard...until you heat them.  Now the outer plastic is tough stuff and when heating you have to get under the edges and crack it off, puncture it, roll it with miniature pliers to "de-skin".  Once you get to the outer material, you get a pair of pliers (different kinds of flat and long nose will help you) and you hold the IC by the side edge (NOT THE legs) and heat the top epoxy and surrounding outer edge.

I set my hot air for 313c with some air force (50% I guess) and went around the top and edge.  I the inner material will partially liquify and bubble black, a little.  Then you know it's ready.  After you tear into the edges enough, the actual IC will naturally want to separate from the material.  I jammed a pick between them and got a huge chunk off that way.  it's hard to get it totally clean.  But fast heating, then gripping with pliers and twisting to crack material away from the IC works!.  Also the pins are soldered to the IC legs, they will fall off or move, try not to heat them!

In the end, I got it nearly 98% clean.  May coin cell was already removed from a drill attempt (this unit was my drill test unit).  So I barely have the IC legs to attach the battery, and it's unknown if you'd just desolder the battery or have to cut it away when using hot air.  Mine was already half drilled...so it's not quite the same.


OK, so what does this mean, with practice the Dallas chips could be de-potted instead of drilled!
   
   
   
   

Since I barely have all my legs on my Ds1284, I'm going to try to order the top board.  I also just tried my soldering iron's knife tip.  This took care of ALL the leftover epoxy very easily...shaved right off.  So it looks like the secret it likely a combo between a soldering iron and hot air.

It looks like the battery interfaces to the Pin 25 (Vbat) and Pin 21 (GND), I have solder balls on what's left...so with some wire..I think I'll make it!

[soviet]

That was a lot of work, but look great at the end.
Have done this before but only sanding where the battery legs are and hooking the cr2032 there.

[weblacky]

Yeah, I did this on a whim based on what I saw for the end result on glitch’s post. I think they are using some kind of hot knife or such and cut off the majority of the head, then cleaning it up.

I was already missing half my DS1286 top, so this took me about 30 minutes. A lot faster than you’d think.

I went ahead and ordered 3 DS1286 crystal + battery tops @ $10 each last night (found $7 shipping + tax). Since it also includes batteries, I don’t think that’s a bad deal.

I have 2 more intact dead/used DS1286 ICs that are outside of a system and two more unmodded still inside systems. I’m going over in my head how I can improve my technique without jamming a piece of epoxy up my finger nail again (that was not a great feeling when it happened last night, doing the IC above). I only own three Indigo2 systems and two ICs are also side-drilled and working so I have extra to play with or swap around right now.

I’m considering trying to heat and pry off the outer plastic shells on only four sides (perimeter), then heat it up again and place it inspire down in a bench vise, and crush/separate the epoxy top just by applying a lot of force to crush the two longest side of the epoxy while the real IC is just outside the jaws (legs facing upwards).

This may allow me to cause a sudden separation right at the epoxy to IC boundary. Which is what I observed last time.

I’m still worried on the best way to disconnect the battery as I can’t literally just pull on it. I need to cause cracks and some major separation and then whittle around the pairs of IC legs so I can cut/desolder the both the crystal and battery while they are partially trapped inside the epoxy.

I’ll develop it more as I experiment.

[jwhat]

Thanks Weblacky,

this is really helpful and great timing, as I am preparing for some major Dallas DS1742W surgery.

This heat gun trick should really help make this pretty straight forward.


Cheers from Oz,

John

[Irinikus]

Maybe also try "boiling" them! (Use distilled water if possible!)

@Pentium used this method to remove the casing and epoxy covering from a cars ECU.

The heat from the boiling water causes the epoxy to plasticise, allowing for it to be relatively easily and cleanly removed with a pair of tweezers.

It's worth a try! (The Battery may pop due to the thermals, but we won't know until someone tries it!)

[weblacky]

I'm kind of deathly afraid of the constant, long term, heat on the IC. I have no idea how the original potting was done and what temps I should expose the IC to for more than a 40 seconds or so. I just got the tops today in the mail. I've sent a question to the seller to ask about adhesives and maybe what secret they are using to de-pott.

Overall, I'm still of the opinion that you need to de-skin the device first with heat and pliers. Then I think I'll take a knife-tip iron and carefully work the two pairs of pins trapped in the epoxy. Once they are free and desoldered, I'll do large-scale heating and crack the entire epoxy block off in chunks or whatever comes off!

Those legs are important and simple wrenching and pulling puts them at risk. I'll update when I get a response to my inquiry.

Again I will state that having a larger knife-tip for your iron works very well for cleaning up the residue. If you don't have one...get one now...don't do this without it. I can pretty much confirm this IC packing is ceramic and the iron knife-tip just glides over the IC surface without digging in or melting with a knife tip on 650F. The last scraps of the epoxy just clear right off, much better than "sanding" the top surface. It works so well, sanding shouldn't be considered here (really).

[weblacky]

Wow, OK just heard back from the seller. Uh...I guess hey don’t have the same caution. They recommend 5-min epoxy to mount the PCB. While they didn’t say an example, I’m finding Araldite products apparently could fit the bill. Normal stuff isn’t compatible with ceramic to PCB mounting due to solvent leaking during cure, heat, and shrinking. However stuff like Rapid Araldite claims not do this.

And they claim they been using a hacksaw/bandsaw to cut off the tops. They claim they cut through the legs because their board is designed to attach to cutoff legs...uh...no?! It’s not, if you don’t have a full leg, the metal to metal connection doesn’t reach, you be using a solder bridge. The IC legs internally don’t have a long pin-like bottom, they just have the initial IC leg which is very short.

So there is “an” answer. It’s not a huge problem to maybe cut most of the epoxy away with a saw, but right to the IC boundary? Yeah, no. A hot knife iron tip works so very well that those kinds of sacrifices are unnecessary. However it’s likely easier to use the 5min epoxy between left over epoxy on the cut (cut flat) surface than it is on the actual ceramic casing of the IC.

At least we now know, saws.

[weblacky]

OK, I just did this AGAIN with another DS1286  - it went a lot better and I was able to plan it out.  More Pics Here!.

OK first the de-skinning was very much the right answer.  I carefully heated one end and used a pocket knife to get between the plastic sheet the epoxy.  It turns out the "glossy" layer between the epoxy and the plastic shell will soften first allowing you to "pry" the plastic away from the epoxy.  This was my first step.

       
       


I left the top plastic on the longest, and used a soldering iron with a knife tip @650f to gently work into the hidden pins.  I was able to uncover all the pins and discovered something very interesting.  Both the battery and the crystal legs are soldered to the pins but they are originally stamped into "looped" ends that lasso the IC legs and are then soldered in place.  You see this in some of the pictures.  So I had to free the component-end of the part first, in order to either cut or desolder and push, the part off the leg to preserve the IC legs.  But it was doable.


Here's what made the big different from last time....channel lock pliers (small versions). I could grip just the epoxy end with a small set of channel pliers, then heat the opposite end and work on it.  Once I had enough of the legs free I found I couldn't reasonably dig underneath the legs...wasn't working well.

       

So I decided to chance heating and "crushing" and then cracking the epoxy using the channel locks, just on the epoxy region, not contacting the IC.  I was hoping that since I had taken pressure of the back of the legs, that the cracking wouldn't damage them. This turns out to be correct.

So I used a hot air tool @ 330c and heated a section, then configured the channel lock jaws to allow me to grab with some extra travel for some crushing force without being open too wide open (for control).  This allowed me to crack the epoxy in sections.  I only cracked it, I didn't pry it, yet.  Once cracked I examined the area, I found it's prone to cracking around the part or the legs.  If that occured, I switched to the knife tip iron and carefully freed the IC legs or the battery/crystal legs (to cut if needed). Only once the IC legs were safe did I heat & pry, using two channel locks (one for both hands), in a slight twisting motion.  Then it all came apart wonderfully.  For the last top chunk of epoxy...I didn't dare put the channel locks on the IC.  So I used my smaller chanel locks to cruch the last section independently after heating (no second pliers for my other hand).  Then cleared the cracked section with a small pair of pliers while I was holding one end of the IC with my hands.

       

The results speak for themselves.  You don't have this level of fine control with a dremel, thought I wouldn't say it's wrong to get into the shell and the components using a dremel, if you could.

       

I used a small pair of pliers to clean up the edges of the IC legs and excess epoxy along with the knife tip, of course.  Gently working between the legs of IC and also the extreme sides to remove what I could.


While my vise idea would have worked, my vise isn't conveniently near my soldering area, so I didn't use it.  But I'd still think a linear crushing force from a vise in the middle of the epoxy, after de-skinning would likely be faster!

So I'll stick with this for ICs I'm going to use the glitchworks tops on.  It works and I have the control I think I need.  My only concern is...it's very easy to lose reference of which part is the front of the IC.  I'm having to go by the asymmetry of the legs during socket insertion.  I wish the PCB had better indicators on it.  But that's a minor issue.

I did just get the rapid epoxy I ordered as well, So I'm close to assembly...bad news is I don't have setup space for my Indigo2s right now.  So testing will have to wait until another time...yeah I know I'm not thrilled either.

[Dylanear]

Very interesting explorations! Keep sharing your work on the various RTC modifications, restorations. A real service to the community. I look forward to seeing where you go with this and how the Glitchworks boards work out.

[jan-jaap]

This is interesting stuff. I'm not new to Dallas dremeling myself, but I like quality of the mod with Glitchworks board.

Oh, and when tearing the Dallas apart, consider putting it in a matching DIP socket to avoid accidentally bending or breaking the legs maybe?