At HC-500 I showed some VNA plots of the HC-500 matching a 50+j0Ω load at 3.5MHz.
The following commentary is on a single load scenario, a 50+j0Ω load at 3.5MHz, and while the results are not simply extensible to other loads and frequencies, it does provide some interesting insight into the devices.
THP HC-500 (Ultimate Transmatch (McCoy 1970))
Above is the behaviour of the unmodified HC-500 (an Ultimate Transmatch).
Losses at match are 12% of input power. At its rated 500W maximum power, that is 60W (which might seem high but heat tolerant insulation materials are used). On modification to a T match, losses at match were reduced to 8% or 40W at rated maximum power.
MFJ-949E (T Match)
Above is a MFJ-949E in the same scenario.
Losses at match are 12% of input power. At its rated 300W maximum power, that is 36W.
Above is the W9CF T match simulator configured for E inductor setting and Ql calibrated the same lossesas measured, Q=120.
Part of the reason for fairly high loss is that the inductor steps drive a sub-optimal network at some frequencies.
Above is a simulation with the same inductor Q, but a stepless inductor. Note the different component values and match, and that power lost is lower at 7.6%. In reality, Q is not independent of inductance, and actual losses might be a little higher.
The chart above from (Duffy 2009) is a simulation of the MFJ949 on high VSWR loads. The sawtooth effect is due to the discrete steps of inductance available.
Whilst that might give the impression that stepless inductors are naturally better, they tend to have lower Q, and may be better or poorer. Despite the problems of roller inductors (contact resistance, reliability, higher maintenance and lower Q), they remain the preferred choice of most hams and are about the only choice with high end ATUs.
ATR-30 (T Match)
Above is an ATR-30 in the same scenario.
Losses at match are 5% of input power. At its rated 1500W maximum power, that is 75W.
Above is the W9CF T match simulator configured for 500pF capacitors and Ql calibrated the same losses as measured, Q=75.
References
- Duffy, O. Feb 2009. Optimum length of ladder line. https://owenduffy.net/transmissionline/LOLL/optlen.htm.
- McCoy, L. July 1970. The Ultimate Transmatch In QST July 1970 p24-27, 58.