Monolithic Soldering Critique

[networkpc3000]

did similar case without high-end machine.
http://www.rflashdata.com/8978537007c-3 ... m-a-1214b/

such job popular now. i think the point is had to make it read well after operation.

[HaQue]

Sunny, nice work, you do some interesting cases. Did you scratch away the pcb to get at the traces inside? or another method, and how did you find pinout pints of subterranean traces..

[jeremyb]

just my luck, I get in a uSD with only about 6 vias on the whole PCB.. I need:
1. Money
2. x-ray
3. laser
4. time
5. reality check!

What good is an X-Ray if you don't know the pin-out to the controller or NAND, I still don't know how this is derived. You're taking pictures of wires with no reference, like a map without street names.

You don't need a laser, use fine grit automotive sandpaper and be gentle, create a jig to put the device in so you're not sanding the edges off.
To get a proper laser is a HUGE waste of money. Yes it can do other things, like remove thermoset BUT it will destroy everything inside as well. Maybe 1 out of 100 times there is a blown fuse inside, fuses are in predictable locations, removing the thermoset will destroy the fuse, but you are replacing it anyway so it doesn't matter. I don't think it's worth it. (?Almost?) Nobody uses a laser. Put the money towards a sports car.

In school they would say:
With enough time, money and effort you can build a bridge the moon with nothing but popsicle sticks and glue BUT it will take a lot of time, money and effort.

If you really want to crack monolithics start innovating, test out new ideas, I'm constantly rummaging around my office trying to create a solution out of something else. Don't give up, make it your obsession. If it's not worth your time move on to something else you don't "have" to solve it. Think like the chinese, they make something using stupid simple solutions, it's usually crap but it kind of works. Start small, can you remove a PCB layer without damaging the IC's inside, hone your skills. Anyway enough of my motivational speeches.

[HaQue]

"Make it your obsession" - haha, yes done! I have a number of ideas that either worked or didn't, I usually come up with some weird way to do something, then halfway thru figure out something so simple it hurts.

I have 2 "pet goals" that I am always looking at things around that might help get there, and will share If they every get realised

[networkpc3000]

Find the right location(I have broken it make it don't work in my case cause I don't have high end machine. carefully Operation ) ----connect logic analyzer----make it work again And analyzing it.

[jeremyb]

I've never dug for traces but from experience I'd try this

Get a cnc (cheap one may do, lead screw or ball screw, accuracy numbers are probably fake you need to calibrate it), replace the controller with a planet cnc usb controller, it's easy-ish, many cnc machines that come with this controller have an old open source controller that doesn't have the needed features.

Get a vacuume table and vacuume generator.

Get copper clad fr4 and measure its height.
Do a z-axis calibration / transformation on the copper clad. This will create a virtual overlay accounting for height differences in different locations, then use pcb milling bits to remove the monolithic Fr4 0.05mm at a time.

It would take some trial and error but that is how I'd do it, total setup would probably cost $1000 USD.

Or use an Xacto knife carefully.

[jermy]

It would take some trial and error but that is how I'd do it, total setup would probably cost $1000 USD.

is it worth it ?

[jeremyb]

is it worth it ?

Depends how many monolithics you see like this, maybe a better option would be to send it to a decapping service to safely remove all the epoxy thermoset.

[jermy]

Depends how many monolithics you see like this

OK, lets see, how long time does it take with a machine like this ?

[HaQue]

I think with a bit of thought and ingenuity you could develop something that is pretty quick, and would, IMHO, be worth it. I agree with Mr Brock that getting a laser and xray is not the way, though right at this moment, cost effectiveness would be to do the "normal" jobs and reject these "harder" more involved or whatever ones.

Lets face it, we want to take these jobs on for our own interest, and to post ego pics on here! but if you run a business, you simply cannot afford to spend days on 1 job.

[Amarbir[CDR-Labs]]

Superb Post ,
And HaQue i pat your back mate for learning this is such a small time .I have seem you grow day by day but few last months since i got all HDD Recovery tools i have been busy .Amazing Talent .I am also setting up a Special Section for Flash To Handle Flash Cards , Pen Drives And Mobile/Tablets .

[jeremyb]

Haque that is profound, at the same time don't loose hope, in time you'll find a solution.

As far as the cnc goes once you have everything clamped and calibrated (that can take a time depending on how much accuracy you are willing to accept), you create a jig (you can use fr4) and learn how to use cam software (cambam is my favorite) the whole process takes under a minute.. You'll probably need a few weekends to practice and for the love of God use limit switches, I'd recommend a cnc machine with a spindle that grounds the bit for a z axis limit that way you don't drill straight through the table when your learning. You can use a software z axis limit if you go with a router spindle (these are mainly sold for wood work) but a grounded spindle is ideal. At the end of the day your milling a rectangle at a specific depth, not rocket science.

I was a big cnc fan if you couldn't tell.

Haque if you think expensive tools will help, instead of acquiring all this equipment outsource each step requiring machines. Most company's have 24hr turn times and cost $20-$70 USD per step. I can't speak for AU but if you shop around you can find some good deals in the USA. All these steps for monolithic are child's play to these companies, some will remove the NAND chip and wirebond it to a breakout pcb if you ask.

Last edited by jeremyb on October 9th, 2015, 5:15, edited 2 times in total.

[Amarbir[CDR-Labs]]

Hello ,
If One Needs To Accure a Xray What is The Recommendation .How About a Dental Xray Should Be Enough .Can You Pour in Your Suggestions .

[jeremyb]

Hello ,
If One Needs To Accure a Xray What is The Recommendation .How About a Dental Xray Should Be Enough .Can You Pour in Your Suggestions .

The picture from a dental X-ray is unusable, very low quality

[Amarbir[CDR-Labs]]

Hello ,
If One Needs To Accure a Xray What is The Recommendation .How About a Dental Xray Should Be Enough .Can You Pour in Your Suggestions .

The picture from a dental X-ray is unusable, very low quality

Hello ,
Sir Unusable Why So .What if You Have Good Quality Powerful Xray .What is Your Recommendation As a Senior In The Trade .

[Sasha Sheremetov]

Hello ,
If One Needs To Accure a Xray What is The Recommendation .How About a Dental Xray Should Be Enough .Can You Pour in Your Suggestions .

Being a big fan of Nikon, I can suggest you this one. It's very nice, with nanofocus, 0.1um and 3D.

[arvika]

Dental Xray is to weak to our purpose, quality and resolution is bad.

[Amarbir[CDR-Labs]]

Hello ,
If One Needs To Accure a Xray What is The Recommendation .How About a Dental Xray Should Be Enough .Can You Pour in Your Suggestions .

Being a big fan of Nikon, I can suggest you this one. It's very nice, with nanofocus, 0.1um and 3D.

Man Sasha ,
You Made My Day ,Can i Have a Link .Is This What You Have .Seems So Interesting .Whats The Typical Pricetag You Can Get This For

[HaQue]

Amarbir, the link is in Sasha's post in the word "this"

[jeremyb]

Children's toys, you all disappoint me. If we're talking about "this" let's throw a 3D CT Scanner into the mix. It makes your X-Ray look like a toy car. Picture is of a SSD BGA with Bond Wires.. I'll take that any day over the spaghetti nightmare that you call an X-Ray.