If you want to use lower g-force, you should increase the time of centrifugation proportionally. So instead of 1 hour at 100,000 g, you would need to spin for 10 hours at 10,000 g. To speed things up, keep the density of the suspending medium as low as possible (for example, avoid a lot of glycerol or sucrose, and add DNase and RNase to degrade the nucleic acids) and keep the particle size as large as possible by avoiding excessive homogenization. Start with a short spin ( 5 or 10 min) to pellet unbroken bacteria and large chunks. Keep everything ice-cold throughout.

100,000g is the accepted force for pelleting membranes. However, if you wish to isolate a membrane protein from membranes, then there are other small-scale ways to do it without having to pellet the membranes such as water lysis. It depends what organism your membrane proteins are located in

Membranes will precipitate at lower speeds (I would try common microtubes at 14 000 rpm, has to be cooled for 1 hr)

It will depend very much on the way you break the cells,

for example, sonication tends to make  smaller, harder to pellet membrane vesicles.

There was a paper for isolation of membranes from small amount of plant material using standard table-top centrifuge with centrifugation for ~3 hours. The trick was to dilute the extract (viscous) with dH2O and split by 200ul per tube. If you want, I can look it up for you.

If you want to use lower g-force, you should increase the time of centrifugation proportionally. So instead of 1 hour at 100,000 g, you would need to spin for 10 hours at 10,000 g. To speed things up, keep the density of the suspending medium as low as possible (for example, avoid a lot of glycerol or sucrose, and add DNase and RNase to degrade the nucleic acids) and keep the particle size as large as possible by avoiding excessive homogenization. Start with a short spin ( 5 or 10 min) to pellet unbroken bacteria and large chunks. Keep everything ice-cold throughout.

Thank you all for  kind suggestions