Here is small modification I did to improve the overall accuracy of my particular HM-15 SWR meter. I did not build this kit but bought it from a local guy many years ago who did a nice job putting it together. I am not sure if he was a Ham. The manual indicated that it would operate up to 6 meters.
I attempted to use this meter to tune a gamma match on a 6 meter homebrew antenna. No matter what I did I was unable to reduce the VSWR below 1.6 to 1. When I connected my Bird Termaline Wattmeter/Load to the meter the VSWR also showed 1.6 to 1 indicating an inaccuracy with the meter at least on 6 meters.
I later came across a schematic drawing of a Radio Shack CB SWR meter using the almost identical circuitry as the Heathkit. The fixed resistor R2 (100 Ohms for a transmission line impedance of 52 Ohms) in the Heathkit meter was replaced with a calibration potentiometer in the Radio Shack unit. I modified the Heathkit meter to include a non inductive calibration pot (250 Ohms). Calibration was accomplished by connecting a precision load to the SWR meter, supplying power from a 2 meter transmitter and adjusting the calibration pot for a meter null with the meter switch in the reverse direction. Using a 2 meter transmitter for this test produced a sharper null than using a 6 meter transmitter for the same test. Applying power from a 6 meter transmitter using the same precision load connected to the SWR meter resulted in a 1 to 1 VSWR. This test was also performed on 20 meters and also resulted in a VSWR of 1 to 1. The problem seemed fixed.
A basic accuracy test was accomplished by connecting 2 precision 50 Ohm loads in parallel resulting in a load impedance of 25 Ohms. Under these conditions the SWR meter should have and did read a VSWR of 2 to 1 on 6 meters.
The manual does not give a minimum power requirement for the meter. Lower power transmitters seem to give inaccurate readings even if the sensitivity can be adjusted to the SET position. For example, a 10 Watt signal appears to read incorrectly when the SWR is high, whereas a 100 Watt signal appears to read OK. This also applies to an unmodified meter.