Does anyone know what height relative to the track the 153IR needs to be in order to be effective? I'd like to lower mine by cutting holes in the foam underlayment, so that the IR window is right at wheel level to make it (slightly) more realistic. But am not sure if it will work. Any advice?
I think it's a your-mileage-may-vary. Before you start excavating your layout, I suggest you take a piece of scrap track and raise it by the amount you think you'd lower the 153IR. In other words, experiment by raising-the-bridge (the track) instead of lowering-the-river (the 153IR). Then run a few cars to see how the 153IR responds...probably setting the delay to minimum so you don't have to wait 10+ seconds every time it trips.
As you say, the Infrared emitter and detector components are centered behind the red window.
The emitter sends out a conical beam and the detector senses likewise - both centered. So sensitivity is maximum at that height above the track.
One issue I see if you lowered the target height to, say, wheel axle height is you might no longer get a reflection in the gap between trucks. I guess you'd have to fuss with the delay setting - probably to maximum. If you're a hard-core DIY'er you can open up the 153IR and using a meter you can find the point in the circuit before the delay timer and hence can in real-time see when the detector is triggering without the confusion of the delay. I took apart a 153IR and detailed how it works in a previous OGR thread which I can dig up if interested.
Separately, I'd think by now that someone with a 3-D printer or whatever would have come up with a DIY trackside equipment box that accommodates the IR sensors by Azatrax discussed above. I have written (and can dig up if there's interest) OGR threads about rolling your own IR occupancy detectors for just a few bucks but it's only been from an electronics perspective, requires soldering, messing with small components, etc. and not the total package ready for prime time so to speak.