I’ve finished the OCXO upgrade. Read about it here
While browsing on eBay for my new lab at home (I’m getting into electronics again), I came across a Philips PM6665 Frequency Counter, an old but functional unit so I couldn’t resist, I had to buy it. So I downloaded the Users Manual and Service Manual for the frequency counter and started to check the schematic. As usual the unit, when new, could be bought with several options. One of them was a better crystal oscillator, an MTCXO that can be automatically calibrated using a 10 MHz external reference, giving an accuracy of 3*10-8 but my unit doesn’t have that option, unfortunately. This will be fixed though. But it does have a 70 MHz to 1.3 GHz optional input!
So here’s what I plan to do…
First some information on the different oscillators available on that model.
Specifications of the Philips/Fluke PM6665
|Time Base (Crystal oscillator)|
|Per Year||<5 x 10–7 (5Hz)||<1 x 10–7 (1Hz)|
|Per Month||<5 x 10–6 (50Hz)||<5 x 10–7 (5Hz)|
|Temperature Changes 0°C to 50°C||<1 x 10–5 (100Hz)||<2 x 10–7 (2Hz)|
|Line Voltage Changes 10%||<1 x 10–8 (0.1Hz)||<1 x 10–9 (0.01Hz)|
I’m waiting for delivery of a Trimble GPS disciplined OCXO with 10 MHz lab reference output, and the unit has provisions for hooking up an external clock on the back panel via a BNC connector. Now we’re talking about PPB resolution instead!
But why stop there? I’m going to modify mine with an electrically controlled ovenized crystal that I bought cheaply from eBay. There’s a lot of them on eBay for $17-$25 (purchase new? think thousands of dollars!), so instead of ±50 PPM which is the usual run of the mill oscillator, I’m getting an Isotemp 131–100 Oven Controlled Crystal Oscillator where the specs are ±5 PPB!
So how come they’re so cheap? There’s a lot of phone cell towers and the like that need exact time, so there’s a significant surplus of these devices. They’re not hard to find on eBay
So after studying the schematic for the frequency counter, I found the description of the standard crystal oscillator. Unfortunately, it didn’t include the electrically controlled MTCXO oscillator option, so there’s no information on the protocol used when calibrating the MTCXO. There are data and clock pins for communication between the central controller board on the unit, but there’s no description of how it works. It would have been nice to be able to let the apparatus do the calibration automatically, but for now, I’ll do the calibration manually. Maybe I’ll do some hacking and decipher the calibration routine and make a PCB board for it, but for now, this is what I have planned.
The Isotemp 131–100 OCXO
I found this great article after watching his channel on Youtube, where Gerry Sweeney talked about upgrading one of his frequency counters. There’s a blog post with some useful information on how to build a controller board here. I measured (approximately, because I only have the schematic, not the unit or parts yet) and designed a board so I can upgrade the Philips PM6665 with an OCXO. For less than $50 I think it could be worth it, and anyhow, who want’s to miss out on hacking an old frequency counter?
I’ve added the possibility to select to use the internal +5V rail provided, but I don’t think it can handle the power requirements for the OCXO, so I have also added mounting holes for external +5V. I also put an SMC connector on the board if I want to use the card for something else. If it works, I’ll upload the Gerber files here for anyone interested.