The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Mixed-Signal Design >> Metal fuse and poly fuse
https://designers-guide.org/forum/YaBB.pl?num=1215137612

Message started by ywguo on Jul 3^rd, 2008, 7:13pm

Title: Metal fuse and poly fuse
Post by ywguo on Jul 3^rd, 2008, 7:13pm

Hi Guys,

As well known, there are both metal fuse and poly fuse. It opens a window above the poly fuse. The window is combined with PAD window. Based on my understanding, the remaining glass on the poly fuse is very thin so that it will be burn out when the poly fuse is blown out. Right?

Question 1. Is there the same window above a metal fuse?
Question 2. Does metal fuse need much more current to blown out because of its much lower sheet resistance?

Thank you in advance.
Yawei

Title: Re: Metal fuse and poly fuse
Post by thechopper on Jul 17^th, 2008, 7:21pm

Hi Yawei,

I´m not an expert but I have some experience with poly and metal fuses.
I think you are correct concerning the poly fuse remaining glass being thin.
As for question 1) I'm not totally sure, but certainly it is always better to use the upper metal layer so that heat is dissipated as much as possible during blowing. When blowing metal fuses there is always a trade off involved between amount of energy required and speed. If you go to slow with your pulse you´ll need a lot of current density, while if you go too fast the current density is much smaller but you might end up getting leakage. This trade offs is more noticeable when using metal fuses of lower metal layers.

Question 2) is simpler: the answer is definitively YES.

Hope this helps.
Tosei

Title: Re: Metal fuse and poly fuse
Post by loose-electron on Aug 3^rd, 2008, 3:43pm

Poly fuses with an oxide opening are commonly used when doing a laser cut trim of the fuse. The poly cuts more cleanly with a laser than metal. (splatters)

Both types of fuse get used when doing brute force current opeing of fuses, again, usually with an oxice window opening.

A suggestion on using fuses, DONT let it define a logic state directly. Take the voltage created (with or without blown fuse) and feed that into a comparator to generate a logic state.

The reason for this is that a blown fuse can have a lot of resistance variation from fuse to fuse, or the residue from blowing the fuse can shift and change the state of the "open connection" --- That open connection may actually be 80K (or similar) of resistance.

Using the comparator allows you to define a circuit that can tolerate a lot of variation in the blown fuse and still work properly.

Title: Re: Metal fuse and poly fuse
Post by vivkr on Aug 4^th, 2008, 12:06am

loose-electron wrote on Aug 3^rd, 2008, 3:43pm:

A suggestion on using fuses, DONT let it define a logic state directly. Take the voltage created (with or without blown fuse) and feed that into a comparator to generate a logic state.

The reason for this is that a blown fuse can have a lot of resistance variation from fuse to fuse, or the residue from blowing the fuse can shift and change the state of the "open connection" --- That open connection may actually be 80K (or similar) of resistance.

Using the comparator allows you to define a circuit that can tolerate a lot of variation in the blown fuse and still work properly.

An excellent point from Jerry.

Actually, poly fuses or other fuses are usually built into a module containing comparators and all necessary support blocks for this very purpose. Otherwise, one would have very unreliable operation, particularly over time.

So be careful not to build your own versions, unless you plan to build several of them and characterize them in detail first before you use them in your real system.

Vivek

Title: Re: Metal fuse and poly fuse
Post by Tlaloc on Sep 8^th, 2008, 1:52pm

With some lasered poly fuses that I've seen in the past, they did have PAD opening, however, this did not generate holes in the ILD during fracture. So, when you assume that the oxide is thin, it depends on what you consider thin. With three levels of metal, you would probably have at least 10k angstroms left. I seem to have been told at one time that the passivation glass can have other compounds in them that can affect the laser. I might, though, be getting that explanation confused with EPROM technology which does depend very strongly on the glass used.