BH Rerange revisited

We have two PTP 450i links that are very stable with the exception of the periodic BH Rerange that will disconnect the link.  At first, I I thought I narrowed down the source as being the 24hour key refresh setting -which did significantly reduce the number of rerange occurances.  However, it has not eliminated them and I get one approximately every month at all different times of the day.  

Both links have 2ft parabolic dishes, high RSSI, low noise floor from spectrum scan, great SNR.  Keep in mind that these units replaced ePTP links that had no problems .  I replaced to try to squeeze 30 or 40mhz channels on a tower with the ability to sync and bring these two radios closer together in frequency guard band.  

Things I noticed: On the second link we set up, the link was running for 3 or 4 months on a tower without a collocated radio in it's band.  Once I collocated this link with a spur ePTP link, BH Rerange began happening.  The first 450i link always has occurances of rerange and has always been collocated with ePTP link.  The collocated links are in the same band and have 10ft of vertical seperation and have 30-40mhz of seperation between them and are either 90 or 180 degrees of azimuth apart. Also, I am still only generating sync on these links since their collocated partners (ePTP Force) do not sync.  

Cambium has posted its reasoning for BH Rerange occurnaces and bascially says its not the radio, it's your RF conditions.  Well...If this is indeed a collocated interference issue, why is the PTP 450i the unit having problems while the Force PTP units continue to work fine (and always did work fine before we swapped them for the PTP 450i)?  Isn't it supposed to filter this type of interfence?  I do not believe the rerange is due to RF conditions and other people have complained about this before as well.  I would like Cambium to look into this and provide an answer please.