A Precision VXO for the WC2XSR/13 Transmitter
by W5JGV
July 22, 2002
Schematic Diagram Revised 17 AUG 2002
Transmitters operating on the LF bands have unusual frequency stability requirements because most operation on LF is done using computers running weak-signal detection software. It is essential that the frequency of the transmitter be accurately known, stable, and adjustable by the operator. This page describes an oven stabilized VXO designed for this service.
It will be helpful to download and print out a copy of the circuit diagram of the VXO. The PDF file is located here.
DESIGN REQUIREMENTS
It was decided that in order to accomplish the goals of accuracy, stability, and adjustability, the most important factors were:
CIRCUIT DESIGN
The Crystal
The first item was to be selected was the crystal. Since physically larger crystal slabs tend to handle more power without excessive aging effects, an HC6-style holder was chosen. A socket was used so that the crystal could be changed should a different frequency be required.
A crystal frequency of 2664.000 KC was decided upon, since that was within the 2 to 3 MHz range required for best stability of the crystal. That frequency is also 16 times the desired carrier frequency of 166.500 KC, thus allowing the use of a simple divide by 16 circuit to generate the carrier frequency. This method of carrier generation also divides any oscillator frequency error or drift by 16 as well, resulting in a very accurate carrier frequency.
The Oscillator
The wiring of the unit is not critical, just avoid any excessive lead lengths. Use high-quality components throughout. Make sure the electrolytic capacitors C6, C7, and C8 are rated for +105 degrees C or higher, since the oven gets HOT! Lower temperature capacitors will cook dry in short order, resulting in a dead or unstable VXO.
Should you wish to order a duplicate crystal to build this VXO, it is available from Sentry Manufacturing Company, 1201 Crystal Park, PO Box 250, Chickasha, OK, 73023, telephone 405-224-6780 or 800-252-6780. Specify stock # 3583, frequency 2.664000 MHz, reference: Spectrotek Services. At the time of this writing, the cost was USD$ 20.98, plus shipping. This is a custom item, so expect a 4-6 week delivery time.
Most of the other components are available from Digi-Key Corporation or other components distributors.
A single transistor, modified Pierce oscillator circuit was designed, followed by a single buffer stage. The first buffer then drives several output amplifier/buffer stages which feed the divide by 16 circuit, as well as providing a 2664 KC test signal output and several spare outputs.
The oscillator is loaded by the input capacity of the first stage of the buffer amplifier IC U2, so it will be necessary to adjust the value of the C4-C5 capacitor combination to put the oscillator spot on frequency. These capacitors should be Silver-Mica units for the best stability. It is fairly easy to vary the crystal frequency +/- 1.5 KC by choosing the appropriate value for C4-C5. If you cannot get the oscillator right on frequency, try another IC for U2. Manufacturing variations of IC's can cause performance problems. I tested ten IC's for U2 and they all resulted in an oscillator frequency within 500 Hz with the values shown for C4-C5. Of course, a single capacitor can be used for C4-C5; I just happened to have these partucilar components handy, so I used them.
The oscillator inductor L1 is one I had on hand. It is a standard two-winding, common mode line filter choke salvaged from a scrapped computer power supply. I placed both windings of the choke in series, and the inductance was just about right. Anything from 1 to 5 millihenries should work fine. You can find these chokes in old TV sets, VCR's, and most any switching power supply.
Varicap VC1 and it's associated parts is optional, but including it allows you to trim the oscillator frequency exactly where you want it. The range of the VXO is about +/- 50 Hz, with the values shown. This gives you a carrier frequency range of +/- 3.125 Hz at 166.5 KC. The tuning range of the varicap may be increased somewhat by applying as much as +20 volts to the tuning pot instead of +10.6. Note that the tuning voltage must be VERY WELL REGULATED or the frequency will shift as the voltage changes.
Injecting an AC or DC voltage at the end of R6 will allow you to FSK the oscillator as desired. This is optional, of course.
The Buffer Amplifier
Yes, I know that IC U2 is a digital IC, but it is a Schmidt Trigger chip and will work with an analog signal if the signal level is sufficient. With the values shown on the diagram, the chip converts the sine wave output from the oscillator to an asymmetrical square wave, which is suitable for driving the 16 X divider.
The Voltage Regulators
Since the +5 Volt regulator IC is located inside the oven enclosure, the heat it dissipates goes toward maintaining the oven temperature, so it's not wasted power in this case.
The DC voltage feeding the regulator should be somewhat regulated, since these 3-terminal regulators are not perfect (but they're cheap!) and there is a slight output voltage change when the input voltage changes.
The Oven Temperature Controller and Heater
The crystal oven temperature is controlled by IC U1, an LM-399H. This is an unusual application for this IC, which originally began life as a precision voltage reference source. In that application, an internal heater maintained the device temperature at approximately +80 degrees C. This served to stabilize the IC's internal voltage reference source quite well. Luckily for us, the manufacturer thoughtfully brought out the heater control voltage on a seperate lead so that it is possible for us to drive an external heater circuit.
As constructed, the circuit will function with voltages from +12 volts to +28 volts. I am using +20 Volts.
Transistor Q1 serves simply as a heater, and will try to draw whatever current it can from the power supply, depending on the base drive it receives through resistors R3 and R4. With the circuit values shown and a +20 volt supply, the heater current will be about 5 amperes at cold turn-on. This is a heater power of 100 watts, far greater than conventional crystal ovens use. Within about 45 seconds, the power level drops to about 20 watts, and tapers off over the next 10 minutes or so, until the oven reaches the set point temperature of +80 C.
IC U1 must be placed very close to the heater transistor Q1 to properly sense the temperature of the system. If there is excessive separation between Q1 and U1, the thermal resistance between the heater transistor and the controller IC will increase excessively. This will cause thermal oscillations, temperature overshoot, and unstable operation of the oven.
Note that as sold, the LM-399H. is shrouded in a thermal plastic cover. This must be removed carefully so that the metal case of the IC may be pressed into the hole in the heater bar. The case is not connected electrically, so no electrical insulation is required between the case of U1 and the heater bar.
If your power supply cannot handle that much power (200 watts), you may increase the resistance of the R3-R4 combination to limit the cold-start current draw to what your power supply will handle. Oven warm up time will be increased, of course. Note that the start-up current draw must be greater then the hot-oven current draw (which represents the thermal loss makeup power required to keep the oven at operating temperature) or the oven will never reach operating temperature.
The oven power supply should be reasonably regulated, since the set point temperature is slightly dependent on the voltage supplied to it. The maximum temperature set point is reached at about +20 volts, decreasing slightly as the supply voltage varies above or below that point. A variation of +/- 5 volts results in a temperature change of about +/- 2 degrees C.
The oven set point temperature CANNOT be changed, as it is set by the internal "guts" of the LM-399H.
The oven temperature meter is optional, but handy. Beware - shorting the meter connections will destroy U1.
OK, 'Nuff Said; lets look at some pictures!!
A view of the completed Precision VXO. The unit has not had its connecting wires attached to the circuitry yet.
The oscillator, buffer stage, and voltage regulators are assembled on a section of Radio Shack perf board. This board is placed inside a length of Aluminum "U" channel stock measuring about 3 inches wide by 2 inches high by 6 inches long. The wall thickness is about 1/4 inch.
To this "U" channel is bolted a section of Aluminum bar stock (the heater bar) measuring about 1 inch wide by 2-3/4 inches long by 3/8 inches thick. Heat sink compound is used between the heater bar and the "U" channel. The heater bar is visible on the right hand side of this picture. This bar holds the oven temperature control IC, the heater transistor and the associated oven components. Heat is transferred from the heater transistor to the bar and then to the "U" channel. This represents a substantial thermal mass.
The temperature control IC, U1, is carefully pressed into a hole drilled into the heater bar. It is placed close to the heater transistor Q1. Since the heater transistor is a PNP transistor with the collector grounded, it does not need a heat sink insulating washer, but is bolted directly to the heater bar. Heat sink compound is used between Q1, U1, and the heater bar.
An end-on view of the "U" channel and the heater bar. Note the heat sink compound oozing from between the heater bar and the "U" channel. Q1 is screwed to the heater bar and is visible just to the right of center at the top of the heater bar. U1 is barely visible just to the left of center at the top of the heater bar. You can see the solder-tipped gold leads of U1 extending from the barely visible gold header plate of U1. This is flush with the top of the heater bar.
This picture shows the VXO circuit board with the connecting wires attached. The heavy copper wire to the right of the picture is the ground wire. The silver wire just to the left of the ground wire is the oven heater power wire. Note that a separate power wire for the oven should be used to prevent any voltage variations caused by the heater current draw from affecting the oscillator. The yellow and green wires leading from the heater circuitry go to the meter to monitor the oven current draw. The bundle of wires to the left side of the picture connect to the oscillator itself. All the spare buffer amplifier outputs have been brought out for future use if needed.
A Styrofoam case for the oven was cut (well, hacked, if you must know!) out of some available packing material I had handy. It was stuck together with 3M "ATG" ® double backed adhesive tape. This is very thin - about the thickness of a sheet of paper - and allows you to make almost air-tight joints. After construction, I wrapped the box in Aluminum tape, both for electrical shielding and for air-tightness. A snug-fitting Styrofoam lid was also fabricated, and a "tape tab" handle was made from some Aluminum tape so that I could get the cover off easily when I needed to get into the oscillator assembly. The box was sized so that it could be snugly squeezed into the transmitter cabinet. The wall thickness is about an inch, and is quite adequate for the task. Feel free to make it thicker, that will reduce your oven power requirements.
Here you see the VXO placed inside the Styrofoam case. When the cover is put in place, it will fit down flush against the edges of the "U" channel. The small air volume inside the case will rapidly heat up as the channel warms up from the heater.
As you can see, the completed VXO in its foam case fits quite nicely in the back of the transmitter cabinet The extra wires from the VXO will be coiled up and placed against the side of the transmitter case, ready for future use. The divide by 16 and CW keying board is visible in this photo. See the page on the transmitter itself for further transmitter construction details.
73, Ralph W5JGV
[Home]
The entire contents of this web site are Copyright © 2002 by Ralph M. Hartwell II, all rights reserved.