[Hbi-svt] GDS file of PA

[Brian Eng]

His answers to my initial questions are below.

I'm going to tell him to reduce the separation between the bias and pads in order to get the 42.000 mm size of the PA.

Unless any other issues are raised I'm going to approve the design sometime this week (I'm hoping Thursday at the latest, which was 1 week from when we received the design).

On Jun 27, 2011, at 12:30 PM, Miguel Ullan wrote:

> 
> On 23/06/2011 20:56, Brian Eng wrote:
>> Hi Miguel,
> Hello Brian,
> 
>> A few quick observations, there should have more detailed questions/comments by the middle of next week after more people have looked over the design.
>> 
>> Is there any reason for the unsymmetric fanout? Most of the designs we've seen so far either fanout along a horizontal line or have a pyramid/arc shape. While all the designs should be nearly identical electrically, having an unusual design is a sure-fire way to have questions asked during reviews, something we'd like to avoid.
> The fanout is unsymmetric because the 128 chip pads are not centered with respect to their equivalent 128 detector pads (this is the 1.25 mm "left" displacement that you mentioned). The shape is due to the optimized construction we have made: We try set a minimum separation between tracks, and then, starting from the edge chip pads (one by one) we draw the metal tracks. When the separation is going to be smaller than the established minimum, then we go vertically the minimum distance needed so that the track can be drawn in the appropriate angle without compromising the minimum separation.
> 
> We have set a minimum separation of 30 microns, which is better for the yield, but if you prefer a more symmetrical design (just for aesthetics), then we can set a smaller minimum distance (20-25 microns). Let me know.
> 
> 
>> 42.450 mm is to wide for the width. The backing structure the PA will be mounted on is only ~42.2 mm. Ideally we'd like it to be the same width as the sensors, 42.000 mm.
> If I maintain the 500 microns separation between pads and bias, plus the 200 thick bias, plus the fiducial, plus the distance between fiducial and edge; and I separate the bias and the fiducial in ~100 microns, I can go to 42070 microns. I wouldn't go below that. Let me know.
> 
>> Also, one of the mechanical engineers wanted to know what the tolerances on the various dimensions are, both the length/width and thickness.
> Tolerances:
> 
> Thickness: +/- 25 µm  (This is the tolerance from the producer of the wafers. We have not characterized the wafer bowing due to the fabrication, but given that the PA are small, I do not expect much differences from that after cutting)
> 
> Length / Width: ~ +/- 25 µm (This is the tolerance expected for the saw cutting)
> 
> 
>> As having the center of the fiducial (the corner ones anyway) located 200 microns from the nominal cut edge is the important part, if their current size is a problem I think it would be better to make them smaller rather than move them in.
> It should be all right, although we cannot "formally" guarantee that all the edges of the metal rectagles of the fiducials will be perfectly defined (the center crosses should be fine).
> 
>> Off hand I don't see any issues with extra metal being placed in unused areas, especially if it will improve the quality.
> Ok
> 
>> Brian
> Miguel.
> 
> 
> 
>> On Jun 23, 2011, at 10:14 AM, Miguel Ullan wrote:
>> 
>>> Dear Brian,
>>> 
>>> Find attached the gds file containing the first iteration on the design of the pitch adapter. Also I attach the layers conversion table in case is of use to you. Please, have a close look at it and let me know if everything is all right or if you see any problem. I have a few comments about it:
>>> 
>>> - On the detector pads I have increased the metal width so that the passivation opening can be 60 microns with a metal overlap of 5 microns
>>> - On the chip side I cannot do this because the room is too tight, so I have reduced the metal overlap with respect to passivation opening to 3 microns. This way the passivation opening has a width of 54 microns.
>>> - The total size of the PA is 42.450 x 4.012 mm2
>>> - the minimum separation of the tracks is 30 microns and the minimum track width is 20 microns, apart from the space between chip pads where the metal track is 13 microns.
>>> - I have made two bias lines of 200 microns width and at a minimum of 500 microns distance of the tracks. I didn't know very well where to put their pads at the chip side, so I placed them at 3.6 mm of the pads by looking at the hybrid design.
>>> - As the fiducial is 220 microns wide, when its center it is placed at 200 microns of the PA edge, the border of the fiducial is at 80 microns of the physical edge. This is a bit too tight and I would prefer to place the fiducials at more distance (300?). We always play safe and never place anything at less than 200 microns from the edge, but in this case, as it is not an electronic device, only a metal layer, I do not think there will be any problem.
>>> - There is a big metal sheet at the center of the PA. This is to facilitate the metal etching and enhance the quality by reducing the amount of metal to be removed. We would place more metal sheets at the other big empty spaces. They are always at more than 300 microns from any track, so there should not be any electric field problem (breakdown), but let me know if you think they can create any other problem to you.
>>> 
>>> Please, let me know if you see anything wrong or different from what you expected.
>>> 
>>> Kind Regards,
>>> 
>>> Miguel.
>>> 
>>> -- 
>>> ------------------------------------------------------
>>> Miguel Ullan
>>> 
>>> Centro Nacional de Microelectrónica (IMB-CNM,CSIC)
>>> Campus Universitario Bellaterra
>>> 08193 Barcelona
>>> SPAIN
>>> Tf: + 34 93 594 7700 ext. 2203
>>> Fax: + 34 93 580 1496
>>> Miguel.Ullan at imb-cnm.csic.es
>>> http://www.cnm.es/
>>> ------------------------------------------------------
>>> 
>>> <PA_Jeff_lab.gds><layers_PA_JL.map>
> 
> -- 
> ------------------------------------------------------
> Miguel Ullan
> 
> Centro Nacional de Microelectrónica (IMB-CNM,CSIC)
> Campus Universitario Bellaterra
> 08193 Barcelona
> SPAIN
> Tf: + 34 93 594 7700 ext. 2203
> Fax: + 34 93 580 1496
> Miguel.Ullan at imb-cnm.csic.es
> http://www.cnm.es/
> ------------------------------------------------------