I think I may have painted myself into a via corner........
Working on a blind via board - I don't do this very often so bear with me - it is an 8 layer board with a few 144 pin 0.5mm pitch BGA's. Directly below these BGA's are a couple high fine pitch 168 pin connectors. Oh and throw in some diff pairs and 50 ohm RGB video signals.
Instructed to use blind vias - no buried.
Stackup:
Layer 1 - signal
Layer 2 - gnd
Layer 3 - signal
Layer 4 - pwr
Layer 5 - pwr
Layer 6 - signal
Layer 7 - gnd
Layer 8 - signal
I worked with the board house to come up with the layer to layer and trace width spacing to meet the diff and impedance requirements.
I have structured the blind vias as follows:
L1 to L2 to pick up gnd
L1 to L3 to get first inner row of BGA to signal layer
L1 to L4 to pick up power
L1 to L5 to pick up other power (board has several powers 1.8, 2.5, 3.3, 5 and 12v)
L1 to L6 to fan out inner BGA pins
Then from the bottom up
L8 to L7
L8 to L6 connector pins to signal
L8 to L5 connector to power
L8 to L4 connector to power
L8 to L3 connector to signal
Then the standard thru via used for many other non-BGA area signals
After reviewing this and thinking about it a bit I believe that my L1 to L3, L1 to L5, L8 to L6 and L8 to L4 vias (which are the most utilized) are not manufacturable since they end up on the top surface of an inner layer. Am I correct in this thinking? I have a call into the board house to ask them but have not received a reply yet.
One option is to convert the L1 to L3 to L1 to L4 and suppress the unused pad on L4. Same with the L8 to L6 - convert it to L8 to L5 and suppress the unused pad on L5.
Am I mixed up or ........... Need some expert input.
Thank you.
Tom
