That's pretty nice - although I don't consider the PING or most of any of the other ultrasonic sensors expensive; honestly, they are really cheap compared to what was available in the past. You basically had the equivalent of the the SensComp kits:
AKA - Polaroid rangers - which weren't cheap (and as you can see, still available - and still not cheap). Maxbotix also has fairly good sensors, though they aren't as cheap as others, either.
Something I am curious about (and I see this often in homebrew ultrasonic sensors) - why didn't you attempt to use a single sensor for both sending and receiving? Alright - I know it is more difficult to do, and hard to get your head wrapped around; I've always thought to try it myself, based on some designs from Hexamite:
...but using sensors similar to what you are currently using. I'm just curious more than anything; everybody seems to go for the simpler route of the PING style (not too mention, since the raw sensors used are so inexpensive, why not)...
Thanks for those links! I have to admit that building something from scratch does feel good regardless of the material cost.
Getting back to the one sensor question, The main challenge is to design a transmitter that is powerful and then at the same time to design a receiver that is sensitive.
In a one sensor design, the transmitter side usually is not bridged since the transducer needs to be multiplexed between sending and receiving. If the sender needs to be bridged the circuit will become even more complex.
Also a good analog switch chip that can deliver the power needed and has a low noise level is probably going to be expensive. But I think I might experiment with a single transducer design sometime just to see it myself.
As a side note for possible improvement, I recall playing with an exterminators Ultrasonic kit that Polaroid sold in the early 80s. In their receiving circuit they added a AGC (automatic gain control) function to the op-amp stages. Seems the return echo amplitude decreases with distance, so their amp stages would start up with a lower gain and increase gain over time to better detect the received signal and have better signal to noise ratio overall. I'm sure this would increase distance capability but at some added circuitry complexity.