You could just do a serial dilution... Whatever dilution where the heat drops off completely should be equal to the inverse of the scovilles. (so a 1/4000 dilution that is just barely detectable = 4000 scovilles)
Here's an example. Take say a shotglass of 1 ounce of the sauce you want to test, or finely blended chile puree you want to test, then add 9 ounces of water and mix. You now have a 1/10 dilution. Take 1 ounce of that mixture, add it to a separate glass with 9 more ounces of water. Thats a 1/100 dilution. Do again, your third glass will be a 1/1,000 dilution. The next one will be a 1/10,000, the next 1/100,000, and the next 1/1,000,000.
Taste the 1/1,000,000 solution, if you can taste ANY heat at all you will need to do more dilution, unlikely unless you have a REALLY hot sauce. If you cant taste any heat, then your sauce is under 1,000,000 scovilles.
Taste your 1/100,000 dilution. If you can still detect heat, you know your pepper is hotter than 100,000 scovilles, and need to do a new dilution. If not, its lower and you will need to work from your 1/10,000 dilution glass.
Lets pretend it is hotter than 100,000. I would just guess and check with further dilutions from the highest dilution where heat is still present. In this example, take 1 ounce of the 1/100,000 mix, dilute it into 4 ounces of water, giving you a new 1/500,000 dilution glass. Taste. Heat? If yes, do a new dilution of the 1/500,000. If no, then do a smaller dilution from the 1/100,000 glass, like 1 ounce of the 1/100,000 mix, and 1.5 ounces of water, to give you 1/250,000 dilution. Just keep on honing in like this until you get two fairly close dilutions, where you can taste heat in one, and not in the other. Your sauces scovilles can be estimated to be between those numbers.
Thats essentially how a scoville test is done, though I am sure more precisely, with special water mixtures etc, and with tasters who probably notice lower heat levels than most of us