LO/interface board
The orginal design seems to be fairly solid in the RF/IF sections. At the output of the second mixer, the nice design tends to falls apart.

I read some dissussion on the mailing list about the inductance of the coil used in the bandpass filter. The measured inductance doesn't match the design. Bob sort of explained what was going on, but failed to mention that the real problem is the core material is wrong for the frequency range, so likely that the core saturates somewhat causing the inductance to swing.

The transmit circuit uses a complementary pair amplifier and transmission gates to connect the DSP output to the second mixer. Bob points out in the article that the transmission gate on resistance is a problem. Simple solution, don't use them. AC couple the output of the complementary pair amplifier directlyto the mixer. Then apply the 5VDC to the amplifier on transmit only. The tranmit output is effectively gated off.

The transmit signal applied to the second mixer is only -10dbm into 50 ohms. Therefore, the amplifier need only supply about 1.5ma which an LM358 can easily handle. Hence, the complentary pair transistors are not needed.

The receiver part of the second IF is largely duplicated from someone else's design. It isn't absolutely necessay to mute the reciever on transmit. The second half of the opAmp is unused and could have been utilized for additional filtering or in place of the LC bandpass filter.

Improvements
Total redesign of the second IF for the reasons given above. I will replace the LC filter with a four pole active filter. The filter will be followed by a two stage amplifier prviding the 50dB gain. There will be no T/R switching here, the amplifier drives into the output jack to the DSP.

The transmit block will consist of a 2-pole active low pass filter followed by an amplifier stage for gain adjustment. There will be no T/R switch. A 50 ohm resistor in series with the opamp will provide impedance matching to the second mixer. During receive, the opamp output is driven low, providing a 50 ohm termination for the mixer.

Changes
I will use MC14170-2 phase lock loops for both local oscillators. This is a newer version capable of operating to 185Mhz. I will use opamp loop filters with differential output from the PLL. I will use a 10Mhz TCXO for for reference with previsions for external station reference input. The PLLs have a buffered oscillator output, so one will feed the other which in turn will feed the microcontroller.

The VCOs will probably be unchanged except for some possible minor tweeking required from changing the PLL.

The MMICs that buffer the VCO could be changed to a pair of ERA-2's. However, power consumption would increase considerably with some gain in performance.

Control will be provided with a local microcontroller which will probably be from the PIC family. The micro will control frequency setting of the PLL, level control of the attenuator, T/R power switching and status reporting. Communication with the DSP and host will be via RS485 as outlined in the DSP board page.

The crude discrete D/A will be replaced by a IC primarily to reduce parts count and board real-estate.

The power supply will have a 10VDC pre-regulator with six low power 5VDC regulators for audio, micro, VCO and PLL. I will use separate 10VDC regulators with shutdown control for transmit and receive. While this seems like an excessive number of regulators, it minimizes interactions between sections. Besides, multiple regulators are cheaper than filtering one master regulator.

Practical considerations
This board is going to get crammed full of stuff. Six connectors, preferablely board mount, have to get worked in too. Therefore, maximum usage of surface mount components is a must.