I believe in what's been said thusfar on this thread, the principal of sync pops has been lost overall. The whole point of a pop is to ensure sample-accurate sync (and phase coherence) of every track, of every predub, of every stem, of every printmaster to pix, in addition to each other.
This begins with popping every single edit track (but mute the other ones so they don't overload). As such, one DX pop should play, one FX pop should play, and so on. If a pop must be checked further down the road for a specific track, we can simply unmute the pop and check it.
In my opinion what JBerg has said is quite dangerous and misses the principal of pops entirely. The pops must follow the identical signal path of how our edit tracks are routed, or else it calls the pops (and sync) into question altogether. Pops must be checked each and every time they are needed (such as starting a new session and placing them in physically). Period. It is incredibly dangerous to assume every show you receive will have consistent pops with the previous one.
When we become complacent in our popping (such as looking for a template or just "routing in" a pop from a dedicated track), we've done a disservice to what a major role of our job is, and a disservice to the client - ensuring 100% sample-accurate sound sync from the base edit track right through to the final printmaster. The only way to ensure this is to not be complacent in habits, and thus build your pops each and every occasion manually, allowing them to follow the same signal path as all other sounds - that way, if there is a sync issue which arises, it is visible by what the pops are doing.
Personally, I pop all my edit tracks, and my personal mix path of choice is to feed my DX, FX, MX edit tracks and sidechains (noise suppression, reverbs, delays, etc) into respective DX, FX, MX Submasters (Auxs) - these Submasters feed DX, FX, MX Stem print tracks at unity gain but have a very high-threshold brickwall limiter like Maxim or the L2007. The Submasters are treated like master faders and don't receive any plugins, panning, volume automation, or anything except for an EQ and brickwall limiter These stem print tracks then feed the DM&E printmaster at unity gain. By allowing the pops to go through the entyire signal path, one can troubleshoot sync issues which might not be realized until you are ready to print a stem and realize "uh oh, why is this 4 frames out for the other stems". The whole idea to is that these pops on the edit tracks are not automated or altered in any way, they basically feed 1:1 unity gain from the edit track to the stem, which means that all things being equal, the pop is unchanged because it lies outside the program material which does get processed. What you can do, and is something I have done for a long time, is simply automate a bypass on each and every plugin that the pop feeds through - but only bypass them over a 15-20 frame area of where the pop resides. How come? Bypassed plugins still are utilizing the DSP and feeding the signal through, the plugin's processes are just not getting a signal fed through it so what comes out the other end is 1:1 (think of it like a switch track - the original track still operates, but there's no train on it because it was switched over to the other track). This is why feeding the pops the whole way through is critical - and it ensures that when you bus your tracks, everything is coming out properly.
Here's one particular issue with what JBerg's suggestion is:
So lets assume none of your edit tracks are popped. Everything feeds through your Submasters and stems and so forth, which results in a lot of complicated bussing to keep this basic hierarchical organization in the edit, predub, stem, printmaster signal flow. Basically, it's easy to mis-route or not route something at all, even down on the edit track-level or as top-level as a Submaster send to a stem.
You're working along on your mix for days or weeks, everything you're hearing sounds wonderful, it's coming together layered, full, and fantastic. The client loves what they hear and so on. So now, let's print this baby, knowing we bussed everything. Need a pop? No worry, we'll just send one straight to the printmaster. Okay fantastic, now lets lay this print off.
You hit record and kick back to listen to the symphony of the mix as a it prints, hearing everything exactly as you and client intended. Great, it's finished, let's ship it off to the client.
You get a call from the client or QC house next week, and they are crimson red because the print bounced QC because some BGz were missing (or a line of dialogue maybe, pick your poison). Long story short, it's going to cost both you and the client a lot of time and money to fix this, and they are none to happy.
But you ask yourself... how could this be, I know I heard the print as it was happening and everything was there? Well, we did hear it, but that particular track was bussed to the Mains instead of the submaster! So we could "hear it", thinking it was getting printed, but we never caught the routing problem because we didn't pop our edit tracks and just sent a pop to the printmaster - so what actually happened is that track was feeding to the Mains for us to hear, but it actually wasn't gonig anywhere or being "heard" through the submaster/printmaster signal path. The point is, there's no way of knowing or being sure unless the tracks are popped.
Morale of this? Pop your tracks - manually, regularly, and with due diligence. There is no free lunch on this post sound line-item in my opinion. That's my two cents