Doug,
Any type of “brake” dynamometer, be it electric, hydraulic, or mechanical requires a considerable amount of instrumentation, servocontrols, software etc. to operate successfully and maintain adequate control over what is happening. While watching a “pull” on an installed commercial dyno seems pretty straightforward, behind the scenes there is a lot of automatic electronic controlling going on to make it work properly. It is doubtful that it could be done manually and produce any useful data or even avoid damaging the engine or dyno.
I would suggest that you investigate using an inertia dyno. These are commonly used for low horsepower engines and are relatively simple to operate. As you may know, the engine drives into an appropriately sized flywheel through a clutch or overrunning coupling. By recording the RPM as a function of time (a fairly simple instrumentation and data acquisition task) and consequently the change in RPM per unit time, the torque and horsepower can be calculated or, you can simply compare the slopes of the curves for the baseline and test engines to see which accelerates harder. It would probably be best to do the calcs and generate the torque and power curves to really see what and where the differences are. This would be easy to do on a spreadsheet using the rpm data.