18-08-2020, 12:36 PM
AM
The AM IF is ideally outside the LW band and MW band. Below 160 kHz means more difficult image rejection and more pickup from SMPSU. Though we think of 128kHz as 1935 and earlier IF, the Eddystone 830S (1966) only uses Dual Conversion of 1.35 MHz IF on SW, it used 100 kHz as 2nd converted IF, and I think on MW. The low IF means gain, stability is easier and narrow bandwidth is inherent.
The Amateur Bands on SW start at 1.810 MHz. USA MW extends maybe to 1.7 MHz (Sports band?). The latest incarnation of the ZN414 and MK484 is the TA7642, which does work to over 2MHz. Some Racal sets used 1.6 MHz.
So I've chosen 1.75MHz for DIY AM double tuned IFTs. This means less winding. About 120 to 140 turns depending on the former diameter. Litz isn't needed. Just random wind about 7mm wide on a cut up fibre tip pen body. 40 SWG is manageable. I'll use no core, just tuning with a fixed capacitor and variable in parallel. So about 100uH with a 47 pF fixed and a variable cap that can do easily 20 pF. Wind two separate coils and mount apart for desired coupling, which helps set bandwidth.
FM or dual Conversion
Various frequencies of IF for single conversion of 6.5MHz to 44 MHz have been used. 10.7MHz to 1200 MHz for dual or triple conversion. Since the AM is at 1.75 MHz rather than 0.455 to 0.48 MHz, we want to at least double the normal 10.7 MHz, to have the FM and AM IFTs in series. If it's too high, then the gain is an issue for valves and also layout is critical. Capacitances get tiny, The 21.5 MHz has been used as an IF and it's just above the 15m Amateur band, which goes to 21.450 MHz.
Conveniently on a fibre tip pen barrel about 25 turns gives 4uH. The 4uH means a small trimmer and no fixed capacitor will work.
Does it work?
[attachment=19889]
Here is the dual band IF on simple matrix board. It's been tested feeding one end by 47 K from the signal generator and having the other end on a scope using 10:1 probe (still too much capacitive load).
A better test is driving from a valve or bipolar transistor and have secondary feed a valve grid or JFET. Note a JFET or MOSFET is poorer to drive a tuned circuit than a Pentode or BJT because it either looks like a low impedance triode or a resistor depending on device type. The BJT and Pentode can approximate to a constant current drive, i.e. high impedance.
Top from left to right: 1.75MHz Primary, 1.75MHz Secondary.
Next row: 21.5 MHz Primary and Secondary on one former.
The two primaries are in series with the 21.5 MHz driven and 1.75 MHz to the supply.
The two secondaries are in series with the 1.75 MHz at the earthy end and the 21.5 MHz at the IF out connection.
The photo was taken before adding the two fixed caps for the 1.75MHz. Without them the minimum frequency is just under 2MHz.
A single aluminium cover is needed for either the entire board or the part with the coils.
The final FM IF would have a centre tap on the secondary and a smaller coil over wound on the primary with one end to the secondary centre tap. Plenty of space for ratio detector capacitors, resistors and diodes. The tertiary coil looks to be about half the size or a little less of primary on commercial ratio detector final IFTs. I can't find details on that.
The ratio detector has an advantage over the PLL, Quadrature and Foster Seeley detectors that the resistor load and 2uF capacitor (1uF to 10uF depending on design) and bleed resistors vary the damping on the IFT according to signal level, so it's less sensitive to AM without a limiter or enough signal to operate a limiter. The main reason for popularity of the Quadrature detector on ICs is that it is a simple coil or ceramic and it's essentially free to have a limiting amplifier and multiplier to make it work. Later the PLL is used as it can do AFC over a wider range and is less distortion on a mistuned radio. The 7088 is the ultimate PLL detector IC as it uses the AFC on the LO varicap to track so that the FM becomes Phase modulation and thus an Op-Amp (built in) RC filter at 70 kHz is the IF.
There is the gated Beam tube, which is superior to the EQ80 (developed from the EQ40). The EABC80 worked out cheaper in Europe than an EQ90 or 6BN6 for AM/FM sets. Even in the USA, the 6BN6 was used in only a few high end FM Tuners, it was more used in TVs. The 6BN6 is an interesting valve that is a Limiter and Multiplier as a single valve with lower distortion than a conventional quadrature detector.
The AM IF is ideally outside the LW band and MW band. Below 160 kHz means more difficult image rejection and more pickup from SMPSU. Though we think of 128kHz as 1935 and earlier IF, the Eddystone 830S (1966) only uses Dual Conversion of 1.35 MHz IF on SW, it used 100 kHz as 2nd converted IF, and I think on MW. The low IF means gain, stability is easier and narrow bandwidth is inherent.
The Amateur Bands on SW start at 1.810 MHz. USA MW extends maybe to 1.7 MHz (Sports band?). The latest incarnation of the ZN414 and MK484 is the TA7642, which does work to over 2MHz. Some Racal sets used 1.6 MHz.
So I've chosen 1.75MHz for DIY AM double tuned IFTs. This means less winding. About 120 to 140 turns depending on the former diameter. Litz isn't needed. Just random wind about 7mm wide on a cut up fibre tip pen body. 40 SWG is manageable. I'll use no core, just tuning with a fixed capacitor and variable in parallel. So about 100uH with a 47 pF fixed and a variable cap that can do easily 20 pF. Wind two separate coils and mount apart for desired coupling, which helps set bandwidth.
FM or dual Conversion
Various frequencies of IF for single conversion of 6.5MHz to 44 MHz have been used. 10.7MHz to 1200 MHz for dual or triple conversion. Since the AM is at 1.75 MHz rather than 0.455 to 0.48 MHz, we want to at least double the normal 10.7 MHz, to have the FM and AM IFTs in series. If it's too high, then the gain is an issue for valves and also layout is critical. Capacitances get tiny, The 21.5 MHz has been used as an IF and it's just above the 15m Amateur band, which goes to 21.450 MHz.
Conveniently on a fibre tip pen barrel about 25 turns gives 4uH. The 4uH means a small trimmer and no fixed capacitor will work.
Does it work?
[attachment=19889]
Here is the dual band IF on simple matrix board. It's been tested feeding one end by 47 K from the signal generator and having the other end on a scope using 10:1 probe (still too much capacitive load).
A better test is driving from a valve or bipolar transistor and have secondary feed a valve grid or JFET. Note a JFET or MOSFET is poorer to drive a tuned circuit than a Pentode or BJT because it either looks like a low impedance triode or a resistor depending on device type. The BJT and Pentode can approximate to a constant current drive, i.e. high impedance.
Top from left to right: 1.75MHz Primary, 1.75MHz Secondary.
Next row: 21.5 MHz Primary and Secondary on one former.
The two primaries are in series with the 21.5 MHz driven and 1.75 MHz to the supply.
The two secondaries are in series with the 1.75 MHz at the earthy end and the 21.5 MHz at the IF out connection.
The photo was taken before adding the two fixed caps for the 1.75MHz. Without them the minimum frequency is just under 2MHz.
A single aluminium cover is needed for either the entire board or the part with the coils.
The final FM IF would have a centre tap on the secondary and a smaller coil over wound on the primary with one end to the secondary centre tap. Plenty of space for ratio detector capacitors, resistors and diodes. The tertiary coil looks to be about half the size or a little less of primary on commercial ratio detector final IFTs. I can't find details on that.
The ratio detector has an advantage over the PLL, Quadrature and Foster Seeley detectors that the resistor load and 2uF capacitor (1uF to 10uF depending on design) and bleed resistors vary the damping on the IFT according to signal level, so it's less sensitive to AM without a limiter or enough signal to operate a limiter. The main reason for popularity of the Quadrature detector on ICs is that it is a simple coil or ceramic and it's essentially free to have a limiting amplifier and multiplier to make it work. Later the PLL is used as it can do AFC over a wider range and is less distortion on a mistuned radio. The 7088 is the ultimate PLL detector IC as it uses the AFC on the LO varicap to track so that the FM becomes Phase modulation and thus an Op-Amp (built in) RC filter at 70 kHz is the IF.
There is the gated Beam tube, which is superior to the EQ80 (developed from the EQ40). The EABC80 worked out cheaper in Europe than an EQ90 or 6BN6 for AM/FM sets. Even in the USA, the 6BN6 was used in only a few high end FM Tuners, it was more used in TVs. The 6BN6 is an interesting valve that is a Limiter and Multiplier as a single valve with lower distortion than a conventional quadrature detector.