ZAB CTLR EXPERIENCED LOSS OF SEPARATION WITH ENRTE ACFT AT FL390.

Narrative

I WAS WORKING THE SECTOR ALONE; LIGHT TFC; NO CHOP OR WX. I WATCHED THE 2 ACFT INVOLVED FOR 6-8 MINS; MONITORING THEIR SPDS; AND SEPARATION. THE B747 WAS NWBOUND; ABOUT 20 KTS FASTER; AND IN FRONT. AT 8 MIN; 4 MINS; AND 2 MINS; IT LOOKED TO ME LIKE I WOULD HAVE 6-8 MI SEPARATION. CONFLICT ALERT ACTIVATED AT ABOUT 3 MINS; BUT IT STILL APPEARED I HAD 6-7 MI. I HAD A HALO ON THE B747 AND I ISSUED TFC TO BOTH ACFT. AT ABOUT THE 2 MIN POINT ANOTHER ACFT; SBOUND; AT FL290 CAME INTO CLOSE PROX AND I BELIEVE THE RADAR BECAME CONFUSED. THE SPD ON THE B747 DROPPED APPROX 40 KTS AND THE SPD ON THE B757 INCREASED APPROX 40 KTS; AND THE BACK TARGET JUMPED INTO THE HALO. IT PRINTED AT 4.85. THE SAR TAPE AT OUR FACILITY (ZAB) WAS NOT WORKING. WE HAD NO INTERNAL DATA TO RUN A SATORI OR NTAP; HOWEVER THE ACFT HAD BEEN HANDED OFF TO ZDV; SO OUR FACILITY USED THEIR DATA TO PROCESS THE ERROR. I BELIEVE WHAT CAUSED THE ERROR WAS ERRONEOUS RADAR DATA CAUSED BY OVERLAPPING SORT BOXES; THE CLOSE PROX OF 3 ACFT WHICH CAUSED SPD FLUCTUATIONS; AND THE DIFFERENCE IN ZDV'S DATA AND OUR FACILITY'S DATA WHICH WAS NOT AVAILABLE. I ASKED OUR QA OFFICE TO DO SOME RESEARCH ON THE DIFFERENCE OF ZDV'S DATA AND OURS; BUT TO MY KNOWLEDGE; IT HASN'T BEEN DONE. IF WE DON'T DO MORE ABOUT TARGET JUMPS AND SPD FLUCTUATIONS; MORE ERRORS WILL OCCUR.