The balls are fuzzy and intended to have a lot of friction with surfaces like the court and the racket; on the rather shallow incline there is little weight that is transferred/supported from ball to ball (think sin(slope)) but in the funnel, being nearly vertical, nearly all the weight is. This means that when the bottom ball tries to roll it is stopped by friction with the ball(s) above it.

In ordinary material handling for materials that behave the same way a heavy vibration source is used to cause the items or, in the case of powdered materials, particles, to bounce out of contact and eliminate the friction between the items and each other or the pipe/funnel. The term for what may be a problem for the tennis balls is bridging: https://youtu.be/Twp0OY0nn5w?t=12

What could be used is some mechanical arrangement that sweeps the bottom ball from the position under the funnel each time a ball is launched.

I note that many ball launchers use long spirals to compact the ramp of individual balls to avoid this problem. Here is a video of the feeder for one that uses a large hopper: https://youtu.be/ReUYT4M-5v4?t=265

Instead of a funnel I would be tempted to try a large platform with a hole in one corner and a V-slope crease leading to the hole. Make it with a 2 ft X 2ft (or so) sheet of cardboard, creased from corner to corner. Cut one corner just large enough for a ball to drop through and add a raised edge to more easily contain balls dropped onto it. For easier use one could put netting on 2 or 3 sides so that balls tossed at it would be stopped and drop onto the platform.

One thing you might want to do is to look at USPTO.gov (US Patent and Trademark Office) and look for patents on tennis ball feeders to see what has been tried in the past.

One obvious problem is the CARDBOARD, which begs the friction issue noted by 3DDave. A smooth funnel would allow the ball to simply slide down the funnel. Nevertheless, the tray suggested by 3DDave sounds like a simpler way to go; however, I would suggest a smooth slidable surface.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert! https://www.youtube.com/watch?v=BKorP55Aqvg
FAQ731-376: Eng-Tips.com Forum Policies forum1529: Translation Assistance for Engineers Entire Forum list http://www.eng-tips.com/forumlist.cfm

They make teflon coated aluminum foil - some spray contact cement on the aluminum side and on the cardboard would be a quick install.

Spray outside if possible - that overspray will be capturing dust and pollen and making them a permanent part of anything the overspray can get to.

I put a 2 foot margin when spraying a headliner in the driveway once and in a few days the edge of where I covered was clearly delineated by dozens of tiny ants, like a drew a line with a magic marker.

Thanks everyone. The v platform sounds like a winner. I will create one and report back.

Yikes, this has been a major thorn in my side. I made a new prototype that was a disaster again. I do believe I need to adjust my slope from back to front to make it steeper, it just complicates things a bit as I have to find a way to attach it to the machine.

Also, I may need to adjust the funnel side to side slope. I like the idea of having a string or whatever like I did so I can easily increase or decrease it. However, I think I grossly underestimated the friction between the balls. They really seem to be locking themselves up as they touch. You can't see it in the pics but in the white circle there is a little paddle wheel that turns 1/4 rotation I think between each ball to advance them. I'm not sure a simple design without some type of jiggle device or feeder that constantly spins at the bottom to advance things will work.

Vibration would help. Maybe just a small motor with an eccentric mass.

Making it a shallower V should lower the force between the balls and lower the mutual friction.

I will decrease the side to side slope to decrease the ball friction and also increase the front to back slope somehow.

The steeper the slope the more force is on each ball increasing the friction between the balls and discouraging them from rolling. Matching the slope of the feeder on the machine is likely sufficicent.

Seems to me that some sort of shaker mechanism would be more effective that futzing with the slope, although lessening the side slopes is a good move.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert! https://www.youtube.com/watch?v=BKorP55Aqvg
FAQ731-376: Eng-Tips.com Forum Policies forum1529: Translation Assistance for Engineers Entire Forum list http://www.eng-tips.com/forumlist.cfm