For starters stacking a zillion short exposures will work. After a while you will certainly long for something better. I lasted about 2 months before I gave up on this as just not providing the results I wanted.

Each time you read out an image you add read noise to the frame. Each frame, if you ignore read noise reduces the total noise in the image by the square root of the number of frames. So going from 4 to 16 frames cuts noise by a factor of 2. Going to 64 cuts it again by 2. You see the problem. Next you need 256 frames then 1024. Obviously there's a point of no return here. All this time read noise is being added to the result so the gain is really far less than 2.

But since read noise is constant each read by reducing the number of reads you reduce the noise level. So one 5 minute shot will have less than one third the read noise of 10 30 second ones.

The time you can go without rotation of the field being a problem depends on where in the sky you image. By taking the object near the meridian you get the longest exposure with the least rotation. By imaging at declinations closer to the pole you again increase the exposure time you can use. How long it will be depends on the image scale of the CCD/telescope combination. You'll have to exeriment and see how long your can go.

Also with longer exposures possible with a wedge you need a way to guide these. Most mounts can't go more than a minute or two without errors creeping in no matter how accurately you are polar aligned unless the image scale is very small. That adds to the cost and likely will delay the switch between methods.

Rick