Yup, guilty of buying into thread drift by replying to an off target previous reply. Back to the OPs question.
Rounding up the usual suspects - a punch list
1 - design ( or redesign ) the sound system to reduce the likelihood of feedback. Move the loudspeakers closer to the listeners and further from the microphones. Choose loudspeakers that have a narrower dispersion pattern. In bad acoustic spaces good design very often means adding delay speakers further into the listening space. Since voice intelligibility is the goal here the delay speakers can be small and of limited bandwidth. They also don't need to be where you might think initially. If you can't hand them over head on put them around the perimeter I've had good luck in things like aircraft hangers ( don't ask ) but distributing small Hotspot style speakers under the seats. Thing of the little round speakers in the ceiling of the highs school cafeteria but upside down.
2 - Choose the right mic for the job. If this is a lectern mic there are more choices than plain old cardioid mics. An SM58 is a pretty bad mic all around except for that one application of people screaming into them from a centimeter away. Try cyber cardioid or even short shotgun mics.
3 - Tune the sound system to reduce the energy at resonant frequencies ( the long way of saying reduce feedback ). Lots of Youtube vids on this but the short version is increase the gain til it feeds back at one frequency. Find that frequency with the equalizer ( an RTA app on your phone can help ) turn that frequency down 3db and increase the overall gain again. Rinse and repeat for up to six frequencies and a maximum of 9 db of cut at any ONE frequency. Now stop and listen to the sound system. It probably sounds pretty bad so it's time to start restoring some of the cuts you've made til you find the sweet spot between feedback resistance and sound quality.
What? It still feeds back before it's loud enough to meet you needs? Return to step one and take it more seriously this time.
The text in bold is what Dave M is having problems with. There are 2 or more bad practices in play and figuring it all out via text on a forum is... time consuming. I wonder if we have a LABster somewhere close to him that could offer a site visit and offer some advice.
