YU1AW
Click on call for e-mail

YU1AW is a prolific designer! Watch this space for additions at any time!


Dragan YU1AW, has kindly granted me permission to present his RF power amplifier designs, even though he already has a web site of his own! His home page is on QSL.net!

Although my purpose is to present information, or links to same, regarding Power Amplifiers which are designed and built around "surplus" Russian tubes, I would be remiss in not stating that YU1AW is an RF engineer. Not only does he design and build Power Amplifiers, he an excellent line of high dynamic range preamplifier designs on his pages, as well! Please check these at his preamplifier web page.


YU1AW QRO:  PAs using Russian Power Triodes

Although thermal drift problems and physical size (too large in terms of electrical wavelength) render it problematic for use at power output levels greater than about 250W on 23cm (links to or drawings of several designs are shown on this site), the GI-7b is well suited for use in multiple-tube amplifiers from HF through 70cm. Several successful designs have been developed to use this inexpensive tube to produce respectable, drift-free power output on these lower frequencies. The excellent YU1AW contributions to the cadré of GI-7b PAs is presented here. Additionally, designs using the GS-31B and GS-35B are aslo shown. The "Easy-Build" ("EB") PAs are Linear Amplifiers designed using HF design techniques. The result is a family of 6m & 2m PAs which use standard inductors and capacitors rather than tuned lines in the anode circuit.

YU1AW designs loaded on this site
(*) indicates most recent additions or updates.
1xGI-7B 2xGI-7B 1xGS-31B 2xGS-31B 1xGS-35B 2xGS-35B

6m

6m & 2m

2m

"EB" 2m

6m

2m

"EB" 2m

135cm

70cm

"EB" 6m (*)

6m & 2m

2m (3 PAs)(*)

"EB" 2m

135cm

70cm (*)

"EB" 2m

135cm

"EB" 6m (*)

6m & 2m

2m (*)

"EB" 2m

135cm

70cm (*)

"EB" 2m

The GS-31B is the "little brother" to the powerhouse GS-35. The primary difference between the two is the lower anode dissipation of 1000W for the GS-31 vice 1500W for the GS-35. This is due to different design of the anode cooling structure (both the physical interface with the detachable radiator, AND the type of radiator used with the tube). Russian tube manuals also say the design life of the GS-31 is not less than 1000 hours of operation, while for the GS-35, 650 hours is claimed. There are no other significant differences; these two tubes can be interchanged, electrically. The GS-31 is a good performer, capable of high-duty-cycle output of 1200W at 2m. A single GS-35 can easily exceed 1500W output on all bands through 432MHz.


DESIGN NOTES:

Input Circuit Q:  There are at least three reasons to design input circuits for a relatively high Q and accept the compromises with respect to a lower Q of 3 or 4:
First, higher input Q allow higher transformation ratio at input circuit. This is very important because many drivers do not exhibit non-reactive 50 ohm output impedance. This impedance, transformed via a length of coaxial cable connecting the driver to the PA, can result in "wild" excursions of impedance far from 50 ohms. Even with input circuit Q = 20, it is possible to exceed the range of acceptable impedances. If in tuning PA input it is found that capacitors Ctk and Cck reach limits of their values (max or min capacity), change length of coax cable between driver and PA to produce different impedance transformation and retune. Repeat until input circuit tuning is successful.
Second, if operating frequency is moved too far from plate circuit resonance, tube damage can result if it is still possible to drive the PA to full power output. It is easy to avoid this situation by making the bandwidth of the input circuit narrower than that of the output circuit. This prevents driving the PA to full power output under off frequency conditions.
Third, driving voltage is rectified at the input to the active PA device (tube or transistor) which acts like a diode. As a result "half of Fo" frequency is produced, which must be suppressed. One of the best way to reject this "half Fo" is narrowing input circuit bandwidth enough to decrease impedance of input circuit on Fo/2 frequency to some low value which effectively shorts Fo/2 voltage to ground. This improves PA stability significantly! Observation: Higher Q produce higher losses, but it is not very important because it only increases drive power requirements by 1dB or so. Input circuit Q has to be low enough to allow operation over an acceptable bandwidth without retuning the input circuit. Input circuits with Q = 20 yield a wider band than it is necessary for normal VHF work but, as above, result in bandwidth narrower than that of the plate circuit!

GI-7B Bias Voltage:  Due to significant differences between GI-7B tubes (new and surplus), and because plate voltage can increase significantly from full load no load condition with some power supplies, it is difficult to predict exactly what bias voltage to use. Start with around 30V and adjust up or down to achieve an idle current of about 30-50 mA per tube. Very precise value is not very important, but it IS important in multi-tube PA that all tubes have same plate current (Ia)! Bias resulting in higher idle current gives higher gain but less efficiency, and vice versa. The most important consideration in multi-tube design is that all tubes have very similar maximum Ia! Measurement of air temperature for each tube gives good information on dissipation and the relative Ia for each tube.

GI-7B Plate Voltage:  The GI-7B can operate with up to 2500VDC on the plate. This voltage can, however, cause the tube to arc in a catastrophic manner (ruin the tube) if the PA is too lightly loaded. It is best to limit plate voltage on this tube to about 2200VDC.


home back QRO home home
go to the top
Last Updated: 20 November 2005
Feedback: Paul S. Goble, III, ND2X
Copyright © 2001 - 2005 by ND2X & YU1AW, all rights reserved