I am a long time developer of electronics devices and products, but have been busy writing software for several years while SMT has taken over the world of electronics. Now I want to make a series of new devices and products, but... virtually all components I want to adopt are SMT.
I used to layout my own PCBs with tape (aarg) and assemble my PCBs manually on my workbench, but that no longer flies!
I have successfully switched to software for schematic capture and PCB layout, but the components on my PCBs are simply too small to assemble manually (even under stereo microscopes). The reason isn't as much component size (given the microscope), but instead the inability to place components accurately. In theory it should be possible, but static cling between tweezers or even vacuum tip devices and components means the components shift when they are released and the tip moves away.
So I'm stuck. No way to continue without buying stencil printer, pick-and-place machine, and reflow oven.
After many months of research I settled on the exact same pick and place you purchased... the ddm novastar LE-40V. I also chose their new high-end automatic stencil printer SPR45VA because I've heard solder paste problems are the biggest cause of problems for many people... plus my designs require 0201 discretes, 0.40mm pitch QFN, and 0.50mm pitch BGA (and perhaps 0.30mm or 0.35mm pitch in 1 or 2 years). I also chose their GF12-HT or GF120-HT reflow oven to support these small components and lead-free solder (if necessary).
I am close to deciding to purchase, but recently one fellow who does PCB assembly keeps telling me horror stories about how every PCB will not work due to improperly placed components, bad solder joints, etc. But he is working with a 10 year old machine that he purchased used, so I am a bit skeptical that his claims apply to all brands of SMT lines and new equipment (though I don't doubt his experiences are what he says).
Since you have the same equipment I plan to buy, I want to ask you what experiences I should expect? And if you are willing, I would love to come to your lab, garage or basement and watch you assemble some PCBs so I can experience the whole process for myself. Nothing I design is even remotely competitive with your products, so you need not worry about that!

To help you understand the kind of PCBs I design, I post a couple links below that you can click on to view images of the two PCBs in a "robotics vision" subsystem I designed.
the PCBs
the PCBs (other side)
gerber file for smaller ice_eye PCB (about 75mm square)
gerber file for larger ice_quad PCB (about 150mm square)
BOM for ice_quad (2 pages) and ice_eye (last page)
Note: the vias in the pads under the BGA solder balls have been filled with conductive material, and coated/overlayed with copper (as part of the PCB manufacture process). In other words, there are no vias under the BGA balls any more, so the BGA solder balls will not flow into the vias and cause problems.
Otherwise the PCB designs are fairly conventional. The trace/space is 0.004" minimum, the smallest caps are 0201 (usually between BGA pads, but also elsewhere), etc. The smaller ice_eye PCB is a 6-layer PCB. The larger ice_quad PCB is an 8-layer PCB. The inner layers are mostly ground and power planes, but a small quantity of traces are run on one inner layer (not that this should matter to assembly).
The PCBs do have a few through-hole components too, but I figure those would be installed and soldered manually. For example, the following are through-hole components: large electrolytic capacitor, power-in-jack, microphone-in-jack, headphone-out-jack, RJ45 jack for gigabit ethernet, linear array of LEDs, and quite a few conventional ribbon cable headers (26-pin, 40-pin, etc).
In case you wonder, I tell you what this device is. The "ice_quad" PCB is a "camera controller" that connects to a PC via the RJ45 gigabit ethernet connection. Up to 4 of the "ice_eye" PCBs (in cases of course) connect to the "ice_quad" controller. Each "ice_eye" is a 2592x1944 pixel digital video-or-still camera.
You can probably imagine the obvious applications for a 1 to 4 camera system like this. The most obvious is a security camera with 4 cameras that point in 4 directions and feed video (or still) images to the PC on command. The PC can control each camera separately, so they can run at different frame rates, different resolutions, different sub-window positions and sizes, etc.
Another obvious application is "motion capture". Stick some colored [retroreflective] dots on an object or human, position the 4 cameras at ceiling level in the 4 corners of a room (or area of the room), and point them all towards a central area. Now the object or person can move around and the video images will be sent to the PC, which can collect and record precision motion capture information.
My next projects is a super-high speed/resolution camera (roughly 8K x 5K pixels @ 300 frames per second). That will involve even faster signals (much faster in fact, either 2.50Gbps or 3.125Gbps differential signals). That product needs to be absolutely as compact as possible, so the push towards smaller components will continue to be very strong.
I look forward to your reply, either here or via my email address (which I assume you have via the forum signup process).