An investigation was conducted to design instruments to directly measure skin friction along the chamber walls of supersonic combustor models. Measurements were made in a combustor at the General Applied Science Laboratory (GASL) and in the Direct Connect Arcjet Facility (DCAF) supersonic combustor at the NASA AMES Research Center. Flow conditions in the high enthalpy combustor models ranged from total pressures of 275-800 psia (1900-5550 kPa) and total temperatures from 5800-8400 R (3222-4667 K). This gives enthalpies in the range of 1700-3300 BTU/Ib_m (3950-7660 KJ/kg) and simulated flight Mach number from 9 to 13. A direct force measurement device was used to measure the small tangential shear force resulting from the flow passing over a non-intrusive floating element. The floating head is mounted to a stiff cantilever beam arrangement with deflection due to the shear force on the order of 0.0005 in (0.0125 mm). This small deflection allows the balance to be a non-nulling type. Several measurements were conducted in cold supersonic flows to verify the concept and establish accuracy and repeatability. This balance design includes actively controlled cooling of the floating sensor head temperature through an internal cooling system to eliminate nonuniform temperature effects between the head and the surrounding chamber wall. This enabled the device to be suitable for shear force measurement in very hot flows. The key to this device is the use of a quartz tube cantilever with strain gages bonded at orthogonal positions directly on the surface at the base. A symmetric fluid flow was developed inside the quartz tube to provide cooling to the backside of the floating head. Bench tests showed that this did not influence the force measurement. Numerical heat transfer calculations were conducted for design feasibility and analysis, and to determine the effectiveness of the active cooling of the floating head. Analysis of the measurement uncertainty in cold supersonic flow tests show that uncertainty under 8% is achievable, but variations in the balance cooling during a particular test raised uncertainty up to 20% in these very hot flows during the early tests. Improvements to the strain gages and balance cooling reduced uncertainty for the later tests to under 15%.