[G0HZU_JMR]

See below for an excel spreadsheet showing the calibration results and test results for my old RLB. Bear in mind that this was built in an evening probably 30 years or so ago using surplus parts.

The idea was that the diode detector port fed to my DVM and the cal point was set at 1.000V into an open circuit.

If the detector had been linear then the DVM would indicate the reflection coefficient directly but you can see that the column A (true reflection coefficient) does not agree with column D (DVM indicated reflection coefficient). This is because my detector was not linear and this is why I had to use a calibration chart to correct for the detector non-linearity.

One crude rule is that if you want to measure return loss within about 1dB then the directivity of your RLB needs to be 20dB better than the return loss of the test load. If you look at the test results in column G they agree with the column D quite well right down to about 22dB return loss. This is because my RLB has about 40dB directivity.

The errors get worse as you scan down and compare the grey coloured cells in columns D and G and I've colour coded them towards red as the errors mount up. However, this is still a very good return loss bridge in my opinion.

Your RHO device should be a bit better than this. If you can drive its detector into the linear region for most of the range it should be able to indicate the reflection coefficient fairly well down to about 0.1 or 0.15 without needing to calibrate it. You could just use an analogue meter with a scale of 0 to 1.0 FSD. You could then mark it with a secondary scale for return loss using a simple equation to tell you where the scale pips should be.