Galaxy GT-550 Cal 25 - Same case, modern circuit
I have several GT-550 radios that I like to tinker with, one of which had the cal-25 crystal calibrator. It hadn't ever worked since I got it, and it was starting to get rather cumbersome not having frequency calibration. I managed to find the manual on mods.dk, here. I believe this was a precursor to the [probably] more functional cal-250, which had a cascade of D-flip-flops (the RTL MC-778P) for dividing down the crystal output from 100kHz to 25kHz. I imagine integrated circuits were very new at the time. The cal-25, however, didn't have any ICs, just transistors. The circuit was an Colpitts crystal oscillator, followed by a common emitter amp, coupled into a discrete part, transistorized Schmitt trigger - refer to schematic below.
Galaxy GT-550 Cal 25 Schematic
I spent quite a bit of time trying to get this circuit to work. I tried replacing various components - it looked like all of the original components were actually fine. The output of the circuit is supposed to be a 25kHz square wave. At first the output was totally dead. In the course of troubleshooting, I put a resistance box in parallel with R6 (the "factory select" resistor), and varied it until I got something at the output. I was able to get various frequencies; 100kHz, 50kHz, 33kHz, 25kHz. I didn't like the output, though, because the duty cycle was very small (probably 5% per cycle). I never tried listening to it. Given the sensitivity of the circuit to this "factor select", and the tendency of carbon composition resistors to drift over time (this is a big problem, and simply replacing all carbon composition resistors with newer types can save a lot of troubleshooting when restoring old tube radios) - I decided to scrap this effort and try a slightly different approach, somewhat similar to what had been done when the engineers upgraded from the cal-25 to the cal-250. I tried looking for something to use for frequency division (D flip-flops) around my parts bins, and stumbled upon the perfect part: the NTE4040B. The pinout is shown below.
NTE 4040B CMOS Binary Counter/Divider
I did a quick test, and found that the output, directly from the crystal (between X1 and C2) could easily drive the high input impedance CMOS NTE4040B; also, the max supply voltage was 18V. There were only 5 pins needed: VSS, VDD, R (reset), C (clock), and Q2 (divide by 4). The connections were as follows (nte4040b connections in bold, others are to the cal-25 schematic):
- Connect Vss to R, and then to the ground rail on pin 4.
- Connect Vdd to the positive (12volt) supply line.
- Connect C directly to the crystal, X1-C2 terminal.
- Lift the connection from C7 to pin 6, and then connect Q2 directly to pin 6.
The final result is shown below:
Cal-25 modified with the NTE4040B
In all, there were only 4 connections to the Cal-25 from the NTE4040B, and all are visible in this photo.
I tested the circuit, and got a roaring 13Vpp (more than 12Vpp; maybe from the crystal oscillator?). I used a few zip ties to secure the NTE4040B perf-board loosely to the Cal-25 board; also, I used an IC socket, so if I ever have to replace the NTE4040B, it will be a quick fix.
When I installed it into the GT-550, it worked great. The output tone was really loud when calibrating the frequency. I am very happy with how it turned out.