Kit building 101 Part 4 Opportunities for improvement

Art WØBA & Dave W6OAL

Operation of our new transverters uncovered a few minor concerns. The primary concern was higher than expected 5V regulator temperature. Let’s see if there is room for improvement;

Art – W0BA and Dave (W6OAL) have put in a great number of hours in the re-design assistance of the W7BAS 222 transverter. It is a nice little kit project and can be enhanced with additional engineering. I’ll first address the original design. The 222 transverter was designed around the SNA-586 Stanford Microdevices MMIC for the active device in the receive chain and a pair of them in the transmit chain. The data sheet on these devices suggests that they should operate at 5V with 65 ma of drain current or ‘Id’.

In the original unit the designer used a 7805 voltage regulator (5 V) to supply Vdd to the devices. Two things came to mind when doing a circuit analysis; where is the “V” headroom (not that a class “A” operating device really needs any), and the heat generated from an 8.8 V drop from 13.8 to 5 V in that little package. The max current drawn would be 23 ma for the L.O., plus 95 ma in receive and 190 + 23 ma in transmit. The bias resistors in the kit were 1 ohm and were unlikely to provide current control permitting the MMIC’s to operate anywhere near spec. Additionally, the regulator and SNA-586 require at least 5 Ohms of bias resistor between them to avoid possible interaction between the devices and instability or oscillation.

Worst case analysis; The 190 + 23 total current in transmit and dropping 8.8 V resulted in a power loss of ~2W that needed to be dissipated which is not insignificant. In the receive situation the power to be dissipated was a little over a watt. The 1 ohm bias resistors were doing nothing. Do the transverters function or really transvert? Yes, but the MMIC’s were screaming! Dave had multiple thermal failures prior to doing an in depth analysis of the circuitry involved and made some hardware changes.

The spec. sheet from Stanford MicroDevices suggests best performance is to be achieved with a Vdd of 5 V and an Id of 65 ma. We looked at how to achieve this on the existing board. The 5 V regulator posed the largest problem as it was  a D pak that would probably handle 0.5 A. What we really wanted to do was create some headroom and therefore considered a 9 V regulator (7809) but once into the values of bias resistors needed, it was soon abandoned for a 7808, 8 V regulator. This device would run cooler dropping 13.8 to 8 V rather than to 5 V. (A large resistor between the 13.8 input and 5 V regulator was tried to reduce the power dissipated by the regulator but there was still be the voltage headroom issue.)

Our approach was to run a series resistor between the L.O.  which required 5 V and the 8 volt regulator. We had to drop 3 V  at a current draw of 23 ma which would require a 130.4 ohm resistor. Standard resistor values were 120 and 10 which we series-ed to provide us with the 5 V.  The power dissipation amounts to ~0.07 W so 1/8 W resistors will work just fine in this application. If chip resistors are to be used in the sanitized version I suggest paralleling 3) 390’s which is a standard value.

Alternatively, the 5V regulator can be moved to the bottom of the board with a 4.7uF and .01uF capacitor at its output to provide operating voltage to the oscillator.

The target current for the SNA-586’s is 65 ma at a voltage of 5 V. First the receive bias resistor (R9) needs to be replaced with a 47 ohm resistor  from the 1 ohm chip resistor (chip or carbon film will work just fine). The receive MMIC requires 5 V so from the regulator, 3 volts have to be dropped  from the 8 V and at 65 ma  the calculation indicates a 47 ohm resistor (3 / 0.065 = 46.15 ohms). The transmit chain employs a pair of MMIC’s which will entail a total current draw of 130 ma. Again 3 V has to be dropped  so (3 / .130 = 23.08 ohms). The power dissipated would be ~.4 W so a pair of 47 ohm ¼ W resistors paralleled will suffice. Again, in the sanitized version a pair of 47 ohm chip resistors paralleled will provide the resistance and power handling capability required.

In summary, the hardware changes required is a 7808 regulator for VR1 from the 7805 D-pak. A 130 ohm or series combination  to make 130 ohms installed in the Vcc line to the L.O. (X1). A 47 ohm resistor  to replace the 1 ohm chip resistor R9. A pair of 47 ohm resistors paralleled to replace R10. The unit should now run somewhat cooler and still the receive MDS will be on the order of -135 dBm, the transmit output power will still be on the order of +20 dBm.

Tests on the bench and actual operation on the air shows the transverter is working well. The regulator runs warm to the touch but not hot.

This opportunity for improvement was now exploited and MTBF and operation enhanced.

Bruce (W7BAS) is in the process of sending out the circuit update, components, and an instruction set to implement this modification.

Other modifications and additions are in the advanced drawing board stages. Dan (W7RF) is working on an amplifier. Dave is working on a multiband idea. . . And I’m enjoying the elevated interest in VHF/UHF and SHF.