Ft 37-43 Bifilar Broadband

Amplifier Input (Fig 4)

Post Amplifier

_) Input/Output s- to First Mixer 0 01 RF Input (fig 3)

Fig 2—Receiver post amplifier and ca ier oscillator circuits.

_) Input/Output s- to First Mixer 0 01 RF Input (fig 3)

Fig 2—Receiver post amplifier and ca

C20—Exact value depends on the crystal used. It might even be omitted if Y4 is appropriately low in frequency compared to Y1.

D4, D11, D12, D13, D16—1N4007 rectifier diode used as a PIN switching diode. More expensive PIN diodes such as the 1N5767 may be used. (Some of these are shown in other figures.)

ier oscillator circuits.

Q6, Q15, Q16, Q18—2N5109 BJT. (Some of these are shown in other figures.) Q7—MPS 918. A 2N5179 or 2N5109

will also work. R32—May not be necessary. Added to prevent amplitude modulation present on one of the oscillators built. T2, T6, T7—Broadband transformer. 10 bifilar turns #28 enamel wire on FT-37-43 toroid core. (Some of these are shown in other figures.)

T3—Broadband transformer. 10:1 turns ratio; 20 turns of #26 or #28 enameled wire on an FT-37-43 core (primary). Tap is 13 turns from the collector. Secondary has 2 turns of #26 or #28 enameled wire over the primary.

Type 6 core, it has a slightly lower temperature coefficient (30 versus 35 ppm/°C). It doesn't hurt to get the most stable parts possible, given that we want a VFO stable to within at least 10 ppm, if not better. You may wish to anneal the core by boiling it in water and letting the water and coil cool slowly, but I've not done measurements to see how effective this is.

One of the critical circuits I almost forgot to include is a receiver incremental tuning (RIT) control, which is the first thing I would have noticed if I used it on the air! RIT is really important if you are attempting to work someone using a DC receiver/transmitter combination that has no frequency offset capability. They can only work stations that can operate split frequency. RIT is implemented by U2, an adjustable three-terminal voltage regulator IC (see Fig 1). The voltage applied to D2, the varactor tuning diode, is controlled by R12 when receiving and by R15 when transmitting. I haven't seen this particular method of obtaining transmit and receive offsets used previously. It allows you to add transmit incremental tuning (XIT) fairly easily, though the biggest problem is likely to be finding the panel space! A compromise may be to use a single control and a DPDT switch to choose either XIT or R!T. Since a dedicated 8- to 10-volt regulator is needed in order to get enough range from a tuning diode, I decided to just switch the resistors that determine the output voltage. This allows the use of cheap NPN switching transistors (Q4 and Q5). There should be no interaction between the two variable resistors, and none has been noted. This method also has the advantage that it is usually possible to set up the RIT with just a voltmeter, although a frequency counter would eliminate any error caused by the transmitter pulling the VFO. This error, while sometimes severe if the transmitter operates on the same frequency as the VFO, is usually negligible in a properly designed heterodyne system. D3 and R10 are suggested by a Motorola application note as temperature compensation for the MV2101 series of varactor diodes, though I've not verified their results. An LM2931 low-dropout regulator was tried at U2 with mixed results: the high output capacitance required for stability hinders its transient response. Thus the circuit I tried doesn't change voltages quickly enough for QSK use. Wasting current with a small-valued resistor load would speed things up, but conserving current drain is usually a design criterion.

The BFO arid CFO

You may be wondering why I don't just use the cheaper method of using the same oscillator for the BFO and carrier oscillator, swinging the VFO to obtain the proper offset. Yes, it's cheaper, and it works cheaper, too, as the offset you get depends on what frequency you are on. That's not too bad it you only cover 20 or 40 kHz of the band, but it certainly can be noticeable if you want the VFO to cover the entire band.

I added 47- and 100-Q swamping resistors to the beat frequency oscillator (Q8 in Fig 5) after I discovered that the circuit had a tendency to take off on the third overtone of the crystal. This happened because of the poor load provided by the mixer, as demonstrated by the fact that such techniques as rolling off the gain of the oscillator with a capacitor from collector to ground had little effect. Not only isn't crystal overtone operation desired in this case, but I've found that using overtones the crystal manufacturer didn't expect you to use can lead to unreliable results, caused by spurious responses close to the desired frequency. As an example, I looked at the 5th overtone of an 18-MHz crystal and found two responses 300 kHz apart showing only about a dB of difference in series attenuation in a 50-Q system. Careful crystal manufacturers make sure that the spurious responses are at least 3 or 4 dB down.

Crystal Filters

Since I know how much people hate to wind transformers, I designed these crystal filters to work well with 50-Q input and output impedances. As a result, the only transformer needed to match the filters to the rest of the circuitry is the one that couples to the high-impedance MC1350 IF amplifier output (T4 of Fig 4). A second crystal filter is used just before the product detector. Since a

7-MHz Band-Pass Filter to Receive Post Amp (Q6) in Receive, 1 Carrier Osc (Q1 7) f in Transmit (both in Fig 2)

7-MHz Band-Pass Filter to Receive Post Amp (Q6) in Receive, 1 Carrier Osc (Q1 7) f in Transmit (both in Fig 2)

VFO Input (from Fig 1)

Band-Pass C31 Filter -^ Input/Output

(to Transmit Amplifier Input & Receive Amplifier Output, Fig 8)

All capacitances are in picofarads. SM = silver mica.

Fig 3—Mixer and band-pass filter.

L2, L3—22 turns of #20 enameled wire on a T-68-6 iron-powder toroid. (2.7 nH, Q=340 at 7.1 MHz)

0 0

Post a comment