OwenDuffy.net 


Palstar AT2K

The following documents a study of an AT2K and forms a record for experiments using the AT2K.

Fig 1:
 

Fig 1 shows the front view of the AT2K

Fig 2:

More interesting is the inside, Fig 2 shows the interior of the AT2K.

Variable capacitors

"Input" (really the output)

Fig 3:
 

Fig 3 shows a cubic spline fit of measurement of the capacitor in-situ at 10 equally space settings of the dial on the antenna side variable capacitor, "Input" from the panel label. The capacitor was isolated from the other circuitry as much as possible using the main switch, and measurements were made with the cover removed and using a LCR meter at 100kHz.

The red and green traces show the characteristic as capacitance was increased and decreased respectively. The significant hysteresis evident in the other capacitor is not evident with this one.

Capacitance increases with increased dial reading.

"Output" (really the input)

The first thing to note is that there are apparently several versions of the AT2K. Some have an auxiliary inductor and relay switching, some have an internal balun.

Note also that the labeling of the capacitors as "Input" and "Output" may vary with different versions. On the version documented here, the "Input" capacitor is the one on the antenna side (or output side) of the T match, and the "Output" capacitor is the one on the transmitter side (or input side) of the T match. Confused? Palstar advised that the changed wiring was to improve 6m performance.

The manual doesn't help clear this confusion, it refers to "Input" and "Antenna" capacitors, but then the table of recommended starting points refers to "Input" and "Output" and the circuit diagram doesn't label either capacitor. This might seem trivial, but it is inconsistent and dumbs the user down to suggest there is no difference in operation. Small values of the actual input capacitor warn of high voltage impressed on the inductor and potentially high internal losses... knowing which capacitor is the actual input capacitor is important to those with basic knowledge of how the ATU works!

The AT2K uses a pair of air-spaced capacitors of nominally 400pF.

Fig 4:
 

Fig 4 shows a cubic spline fit of measurement of the capacitor in-situ at 10 equally space settings of the dial on the transmitter side variable capacitor, "Output" from the panel label. The capacitor was isolated from the other circuitry as much as possible using the main switch, and measurements were made with the cover removed and using a LCR meter at 100kHz.

The red and green traces show the characteristic as capacitance was increased and decreased respectively. There is a clear significant hysteresis around dial marking 40. Cursory visual inspection shows an obvious misalignment of the the shaft from the panel mounted reduction drive to the capacitor, probably the cause of the hysteresis issue noted above and a slightly lumpy feel to the control.

Capacitance increases with increased dial reading.

Further investigation

Fig 5:

Fig 5 shows a side view of the misaligned coupling as delivered. Though a little out of focus, it is obvious that the shaft axes are approximately paralleled but misaligned vertically. They are also misaligned horizontally. The flexible coupler is designed to deal with angular displacement of the shaft axes, but not severe offset and hence the cocked coupler bosses in the picture, distorted by the poor alignment.

In fact, on releasing the screws securing the capacitor to the bottom panel, it needs to be 2mm higher to line up with the panel shaft. Observing this poor fitup, the alignment left to right was examined and it was out as much so the total misalignment is more than 2.8mm or 44% of the shaft diameter. The problem was resolved by filing out the holes in the bottom panel a little and packing the capacitor mounts with 2mm of washers to achieve near perfect alignment of the shafts. The binding effect has disappeared and hysteresis noted in Fig 3 has gone.

Fig 6:

Fig 6 shows the packings that were used to correct the alignment problem, two washers under front and back of the capacitor. 

Palstar advised the alignment problems are caused by inaccurate cutting of the capacitor end plates, a known problem! So much for Palstar's brag DESIGNED AND MADE IN THE USA as if it implied superior quality.

Fig 7:

Fig 7 shows the measured capacitance after the works described above.

Inductor

Fig 8:
 

Fig 8 shows a cubic spline fit of measurement of the in-situ at 11 equally space settings of the dial. The inductor was isolated from the other circuitry as much as possible using the main switch, the capacitors were set to minimum capacitance, and measurements were made with the cover removed and using a LCR meter at 100kHz.

Inductance decreases with increased dial reading.

Efficiency

Efficiency was measured when matched to a 50Ω dummy load. Power in and out were measured using an LP100A wattmeter.

Frequency (MHz) Input C
(really output)
Output C
(really input) 
Inductance  Efficiency (%)
1.8  65  70 71 91.3
1.8   90  100 117 92.9 
1.8   83  90 108 92.7 
3.6  94  100 213 96.8
7.0  100 94 244 94.1
7.0 80 78 245 89.4
 

The table above shows the results of the efficiency measurements.

The first row of the table was guided by the AT2K manual. Clearly such low values of C do not achieve best efficiency with this load.

Frequency (MHz) Input C (pF) Inductance (µH)  Efficiency (%) Implied coil Q 
1.8  418 10.2 91.3 94 
3.6  418  3.1 96.8 124 
7.0  392  1.1 94.1 45 

The table above shows the interpolated input C and L, efficiency, and the implied coil Q assuming that almost all loss is in the coil. The performance on 7MHz was surprising, but measurement at some other combinations of C, C and L indicated that the figures reported were the best available efficiency.

Coil Q will vary with frequency, but for the roller inductor coil, Q will vary with the tapping point. The low indicated Q at 7MHz is possibly the result of the dimensions of the active part of the inductor (about 3-4 turns)... these are not proportions for a high Q inductor.

The loss might seem insignificant, but consider that for the load scenario measured at 1.8MHz (a fairly ideal one loss wise) that at 1kW of RTTY input, 90W will be lost as heat in the roller inductor... or about 3/4 of it that is in use. It is not designed to dissipate 90W.

Efficiency on other loads will differ. In particular, efficiency on extreme loads often used on the lower bands may be much worse, less than 50% for some practical scenarios and the risk of damage to ATUs during tuning processes becomes real.

ARRL review

(Hallas 2013) reviewed the AT2K and commented on the smoothness and repeatability of capacitor setting, so his tuner was of better construction quality than the one tested here. The pictures reveal different couplings to the ones pictured above.

 Interestingly Hallas did not comment on the incorrect labeling of the capacitors though his pictures show identical connection and labels to the ATU in this article... but the ARRL is always very kind to their advertisers and these evaluations are not much short of an advertisement.

References / links

Changes

Version Date Description
1.01 11/11/2012 Initial.
1.02    
1.03    
1.04    
1.05    

 


© Copyright: Owen Duffy 1995, 2021. All rights reserved. Disclaimer.