For　rotating　machinery　condition　monitoring　and　fault　diagnosis　under　speed　variation　conditions,　order　tracking　has　been　considered　as　a　very　powerful　non-stationary　vibration　signal　analysis　method,　when　compared　with　the　frequency　analysis　methods　based　on　stationary　assumption.　However,　the　conventional　order　tracking　methods　require　additional　hardware　to　provide　a　phase　reference　signal.　This　constraint　limits　the　conventional　methods　in　industrial　applications　significantly.　In　order　to　further　improve　the　applicability　of　the　conventional　order　tracking　methods,　some　tacho-less　order　tracking　methods　have　been　proposed　in　the　past　few　years.　Despite　the　tacho-less　order　tracking　methods　successfully　getting　rid　of　reference　signal,　the　algorithms　are　very　complex　and　the　computational　burden　is　increased　correspondingly.　To　address　the　aforementioned　shortcomings,　a　straightforward　tacho-less　order　tracking　method　based　on　order　spectrogram　visualization　is　proposed　in　this　paper.　In　the　proposed　method,　a　ridge　extraction　approach　is　used　to　estimate　the　instantaneous　frequency　of　a　certain　rotating　frequency　harmonic.　And　then　the　vibration　signal　is　resonance　demodulated　and　the　time-frequency　distribution　of　the　demodulated　signal　is　obtained.　A　subsequent　transform　is　conducted　and　the　frequency　axis　of　the　time-frequency　distribution　is　rescaled　based　on　the　estimated　instantaneous　frequency　of　rotating　shaft.　Then,　an　order　spectrogram　is　constructed　and　thereby　the　non-stationary　interference　introduced　by　rotating　speed　fluctuation　is　suppressed.　Finally,　fault　orders　are　uncovered　and　bearing　fault　type　can　be　identified.　The　effectiveness　of　the　proposed　method　has　been　validated　by　both　simulated　and　experimental　rolling　bearing　vibration　signals.　The　results　illustrate　the　improved　features　regarding　previously　developed　tacho-less　order　tracking　method　in　bearing　diagnosis　under　speed　variation　conditions.　©　2018　Elsevier　Ltd