This article is a review of an inexpensive QRP ATU kit sold widely on eBay and possibly other online stores.
Above is the seller’s pic of the kit.
Above is from the sellers listing, it shows the internals of a constructed tuner.
Note that there is no expectation that two kits bought on eBay will be the same, even if bought from the same seller in the same purchase. The details presented here a review of a single ATU kit.
No construction instructions or schematic is provided by the seller.
A Google search reveals a constructor’s article which includes a schematic though some values are not labelled (notably Cin and Cout).
There are three components that are critical to T match performance, Cin, Cout and the main inductor. Their values and losses critically determine T match performance.
The variable capacitors supplied have 2 x 26pF sections, 52pF in total when paralleled. These are unlikely to be large enough to permit matching on the lower bands over even modest impedance ranges. Even if it does match, small capacitances in T match drives high loss on low frequencies.
The impedance of the caps was measured across HF, and on the lower bands Q was about 500. Q is lowish and that will roll up into total ATU loss.
They were tested for voltage withstand at 500Vrms AC (50Hz) for 30s, passing with no apparent increase in leakage current.
The kit is supplied with a toroid of approximately T106 size. It is painted with the Micrometals code for a T106-2. A T106-2 uses material with µ=10, and has Al=13.5nH.
Note the chamfered corners. Micrometals powdered iron cores have radiused corners to reduce the risk of wire insulation failure since the cores are conductive (which is also the reason that they are usually painted).
Refer back to the second pic and you will see radiused corners on the core in the constructed kit.
An 18t winding was applied to measure and validate the core behaviour.
At 100kHz, indicated inductance is 7.16µH, making Al=22nH and µ=16.3. This does not look like a T106-2.
(19-01-2018: I purchased two more of these from another seller, and they also measured Al=22nH and exhibited low Q at 10MHz.)
Looking at 7MHz, Z=30.5+j334, giving Q=11.0. I would expect Q around 200 from a Micrometals T106-2, this is a very lossy core material and experience suggests probably unsuited to the application.
Because the core is judged unlikely to deliver reasonable performance, a set of models based on the measurement data was used to calculated likely loss at 7MHz. The load used was simply 50+j0Ω as it was easy to provide for actual measurement should construction proceed.
W9CF’s T match tuner simulator was used to obtain an estimate of the T-match loss with the given load.
The settings above are for the components in use.
Above, the supplied components will match a 50+j0Ω load at 7MHz, though matching range is very limited and it will not match at 3.5MHz.
Importantly, whilst it does match 50+j0Ω load at 7MHz, 96.5% of the input power is converted to heat.
The model above uses 200pF caps, and efficiency is improved, just 38% of the input power is converted to heat.
200pF caps and Q=200 core
Above, caps are 200pF and the inductor has Q=200.
With just 3.2% of input power converted to heat, efficiency is good.
The kit is cheap, but does not have the capability of a reasonably efficient T match tuner on the lower bands. The three components critical to performance all fail miserably, it is Cheap Chinese junk, not because of concept or design, but supply of unsuitable components.
Needless to say, as a result of the measurement and models, the tuner was not constructed.