Fouled plugs don't work well, spark or glow, and after the first use with BP they will be pretty well fouled.
......or the element would break and it would be single use. Kinda pricey proposition if you ask me. I think a delay and BP is the most economical way to go with trying to come up with a "reusable" ejection charge ignition simply an academic exercise to absurdity.
Get a new delay grain and BP charge with each commercial load one buys.
Say this over and over again...... "It's only a hobby. It's only a hobby. It's only a hobby." In that regard, K.I.S.S. ing it (Keep It Simple Stupid) with ejection charges is the way to go. Delay grain and BP or electronics and BP is the simplest way to go.
(Errrr, ummmm, in my humble opinion electronics is the way to go as apogee detection is so much better than a fixed delay grain that one might screw up on the drilling. Get a good, low energy deployment at apogee and avoids zippers with cardboard tubed rockets.)
Yeah, yeah, yeah. Run a ton of simulations in different wind conditions on ones rocket and one get a half way decent delay time to drill a grain but what a PITA. I could spend an hour on a rocket doing a sim. Go to the field and take a wind reading with a hand held anemometer (though winds aloft can be totally different than what they are on the ground) and thank you Kestrel instruments for that. Push the button and get a nominal flight with no zipper.
What a pain. I have to admit though due to the lower mass of modrocs and the rubberband shock cords, zippers are not as much of a problem in that realm.
A subtle "trick" for a low energy deployment is to point the rocket on the pad a few degrees "downwind" as the rocket will have a tendency to weathercock into the wind and have a curving path to apogee with a low energy deployment. Run a sim and will get an idea of this phenomena. Rocksim has the video simulation mode and that's where I got the hang of it.
I vaguely remember the offering of a glow plug ejection charge system back in the day when I first started out with HPR around 2004 or thereafter. Slick website but I thought I was "too" young and stupid to try something like that. Probably a lot of folks felt that way and the site (and product) disappeared.
Oh for the uninitiated, drilling a delay grain is using a drill bit of manufacturers recommended size and drilling a hole of a certain depth to shorten the delay time. I've done that a lot with research motors that fit the small AT cases and pack a kick. I usually have to buy the longest delay grain and drill accordingly to what the simulation suggests.
Kurt Savegnago