> There's the BFI (Brute Force and Ignorance) method- just check continuity= at > untented vias or scratch through the solder mask like an animal until yo= u find > the break. Get the PCB artwork up on the screen, divide and conquer. >=20 > If it's a bad via, consider strongly discarding the board and maybe the e= ntire > batch. Bad vias have a way of multiplying (it's usually due to bad contro= ls at > the PCB maker which causes cracks at the edges of the hole- it's easy for > there to be many marginal vias which pass 100% test but open up later or > under temperature). Broken traces are usually due to sloppy handling and > are cause for reconsidering the supplier next order, but not as much > immediate concern. That is one problem source, another is the original laminate manufacturer. I had occasion recently to reject a PCB on incoming inspection due to an in= clusion in the laminate that shouldn't have been there. Representatives fro= m this same PCB manufacturer had previously told us of having problems with= laminate due to the laminate factory essentially being a large shed so as = workers went in and out they carried all sorts of debris into the processin= g area, and the result being that sometimes such debris gets into the epoxy= as the laminate is manufactured. They related a story of how a PCB they ha= d manufactured had been loaded with components passed all its testing event= ually making its way into a fighter plane. After some months in service it = developed a fault eventually traced to this PCB. With everything on the PCB= apparently otherwise checking out alright they eventually delaminated the = PCB to discover some charcoal substance in the inner layers of laminate, I = think between supply planes. This was eventually determined to be the resul= t of a seed th! at had got into the epoxy mix and because of the solid copper planes could= n't be detected by visual inspection. Eventually there had been some breakd= own due to voltage stress across the impurity which eventually carbonised e= nough to cause problems. This was of such concern that ESA now require all planes to be done as cros= s hatches unless there is a DGR (damn Good Reason) otherwise. This is to al= low inspection after etching but before layer build-up of a multilayer PCB.= As this particular PCB was coming from a supplier whose process is certifi= ed by ESA I was most surprised to see this defect in the PCB I was inspecti= ng. I also have a similar story over a minicomputer I used to be involved in se= rvicing. Every so often after a power down for maintenance it would have pr= oblems powering up again, sometimes putting strange voltages on peripheral = grounds. Eventually it totally failed, and every PCB was swapped one at a t= ime into another machine. Only the backplane caused the fault, and I couldn= 't understand how this could be until I was told the story above about the = seed in the laminate, and I suddenly understood what had happened. As to the OP's original problem, the only way to determine just where the b= reak is on an inner track will be a TDR measurement compared to a good PCB,= as mentioned previously in this thread. The problem is you will need a TDR= capable of doing picosecond measurements to get any sensible length measur= ement unless it is a very long track. Maybe time to call on your local (wha= t used to be) HP or Tektronix rep for an instrument demonstration. :) The r= est of us would just put a patch wire on, call it fixed and wait for anothe= r failure if it really is a via breakaway problem. --=20 http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .