I decided to try a CTS OCXO Board from ZL2BKC. It is a small board with a brand new 10 MHz OCXO made to complement the ZLPLL board, and it is available in kit form for just $20: a bargain!
The CTS OCXO board in kit form
The kit comes with a PCB, the OCXO, a SMA connector and a few SMD components. It can be mounted in a few minutes even taking into consideration the supplied sheet with instructions is a bit confusing: some components must not be mounted, the component list has one set of values, and the schematic has another set of values... but not a problem after all.
The finished OCXO
The problem appears when you power up the OCXO. During the warm-up phase, it draws around 600 mA (610 mA for my unit) and this is just too much for the little 78M05 used, so during this phase it gets EXTREMELY HOT. Any 78M05 datasheet will told you the maximum output current is 500 mA, so placing a 78M05 for something that consumes 600mA is not a good idea at all. After the warm-up phase current is reduced to just over 100 mA.
The CTS OCXO board bottom side
Interestingly this seems to be an intended behaviour. From the web page:
A linear regulator on the board drops the input voltage (8 to 14V) down to the 5V required by the OCXO. Since the heater runs at 5V a higher current is required resulting in a power disapation over twice that of an equivalent 12V rated OCXO. By mounting the series regulator under the OCXO the wasted heat is put to use to heat the OCXO module and help reduce the current.That is just wrong. The voltage regulator is under the OCXO, but at the other side of the PCB. Its heat must go through the PCB itself (very bad thermal conductor), the upper layer and a 2-3 mm wide air gap. Only a insignificant portion of regulator's heat will arrive the OCXO.
I know it "works" but that regulator is working totally out of specifications. I know 78M regulators have internally limited dissipation but making it work repeatedly out of specifications can lead to a premature failure. And if the regulator goes bad there are little chances the OCXO will survive.
The regulator is stressed only during warm-up about 30-40 seconds working at room temperature. But if you power it up at very cold temperatures, for example working portable in winter at the top of a mountain, the warm-up phase can last much longer, and if during this time the regulator starts to limit current or dissipation (cross your fingers for this and not for a failure), the voltage applied to the OCXO will be less than 5 volts, and the OCXO could not get thermally stable or will not work at all.
So, what are your options? From worse to better:
1.- Don't do anything and thrust on manufacturer's claims of internally limited values and protections.
2.- Remove dissipation from the 78M05 regulator placing a 3W, 5.6 to 8.2 ohms resistor in series with the regulator input. This resistor will drop the voltage during the warm-up phase reducing heat dissipation in the regulator, but it will be still working out of specifications.
3.- Remove some solder mask around the 78M05 footprint and place a full size 7805 regulator. In this way only the excessive heat will be the problem because of the small heat sink capabilities of the supplied PCB
4.- Apply solution 2 and 3 at the same time: place a 7805 regulator in the PCB and 3W 5.6 - 8.2 ohms voltage drop resistor at its input. In this way most of the heat will be dissipated by the resistor.
5.- Use an external full size 7805 regulator with appropriate heatsink. If you choose this option make sure the cables between the regulator and the OCXO will not pick up any noise or unwanted signals.
I used the 4th solution. Works fine. In this way the regulator gets hot during warm up, but not extremely hot: you can maintain your finger over it without any problem during warm-up.
Miguel A. Vallejo, EA4EOZ