Building a cheap ACF tape bonding machine

[tibimakai]

Re: Building a cheap ACF tape bonding machine

These Romanians and there improvisations.
In almost every forum, I find some smart Romanian doing something, that others didn't.
I'm from there as well, but left in 96.

[Th3_uN1Qu3]

Re: Building a cheap ACF tape bonding machine

Pressure wise things aren't as critical as one might think, the only risk Hitachi mentions in the datasheet is the tape squishing and coming out the sides, which would make for a weak bond in the middle.

The way the tape works is it has microscopic strips of conductive material (gold for the tape destined for glass bonding, nickel for the one intended to bond to the PCB), which are significantly thinner than the spacing between 2 traces on the flex and panel. So shorting between traces is impossible.

[Th3_uN1Qu3]

Re: Building a cheap ACF tape bonding machine

Cracked panels...

[llonen]

Re: Building a cheap ACF tape bonding machine

I might be interested in this and did look into it in quite some detail a while ago, one of the stopping points for me was sourcing replacement COF assemblies.

[Th3_uN1Qu3]

Re: Building a cheap ACF tape bonding machine

The panels may need to be preheated locally only if the bonding heater is weak. Glass has a higher thermal resistance than copper, so if anything, the PCB bond will need more power than the glass one.

As for cleaning, Hitachi says in their datasheet that the stuff will come off with acetone or toluene.

[stj]

Re: Building a cheap ACF tape bonding machine

the panels arent virgin, how do you clean them first? !!

[dick_barton]

Re: Building a cheap ACF tape bonding machine

Very interesting.

1. Does the panel also have to be preheated so as to minimize the stress on the glass when attempting to bond.

2. This one is a bit more difficult to explain.
I believe though may be wrong that the adhesive has conductive particles in suspension within it and when sufficient heat and pressure is applied to the adhesive, a connection is made, top to bottom.

The tracks on the panel have a certain height (thickness of track).
The tracks on the ribbons also have a certain height (thickness of track).
The applied pressure must be enough to make a connection between top and bottom tracks (A) but not between the spaces between each track (B). This is achieved because distance A is less then distance B.
So would it be critical to know the amount of pressure required on the bonding head to complete the task successfully.

[Th3_uN1Qu3]

Re: Building a cheap ACF tape bonding machine

I have a HD webcam i bought for $3.50 at a yard sale. You can find cheap stick-on microscope zoom lenses on ebay. A few high brightness LEDs should do for light source.

[Alastair E]

Re: Building a cheap ACF tape bonding machine

This looks a great project what are you going to use to align the tab bonds back up?

A good lightsource and a simple microscope would be OK for this...

[diif]

Re: Building a cheap ACF tape bonding machine

Yeah, definitely interested, once my daughter's 18th is out of the way i'll have some spare cash to chip in on some tape if needed.

[vinceroger69]

Re: Building a cheap ACF tape bonding machine

This looks a great project what are you going to use to align the tab bonds back up?

[Th3_uN1Qu3]

Now that ACF tape is available to buy and doesn't cost an arm and a leg (it's still pricey, but 50m of tape for $80-something sounds fair to me, considering the size of a tab), i have decided to build my own bonding machine.

The few builds i have seen use a commercial bonding head and that costs $3000 to $5000 - OUCH. I was thinking more like $50.

I have read the datasheet for the Hitachi ACF tape and it occurs to me that the requirements are not very stringent - the tape should reach 160 to 180C for 20 to 30 seconds, at a certain pressure applied by the bonding head. A suggested temperature (305C) for the bonding head is also given in the datasheet, and the only requirement for the head's shape is that it must have an even surface where it presses down on the tab (duh). There are no slopes required as it is with soldering BGAs, temperature just needs to get up there as fast as possible, stick there for 30 seconds then cool down. So the design of the control circuit is also easier.

I have already chosen my heating element (it's a surprise, you'll laugh) and completed the design of the control circuit - all that's left is making the PCB. There is no microcontroller and no firmware involved, it's all analog. The output drives a relay so any kind of heating element can be used, and the input is a normal type K thermocouple. Control is accomplished via a couple pots and switches. The circuit turns the heater completely off once the time set has elapsed, and can also run a fan if desired.

Anyone else interested?