LTX Corporation 10Mhz Clock Reference (partial?)

[mdubinko]

At surplus I picked up a "LTX Corporation OEM 10Mhz Clock Reference", but it looks like the main component (U1) isn't present, along with some capacitors and an inductor down in the corner. Doesn't look desoldered--it looks never-populated. This was supposedly pulled from working equipment, so I'm hoping someone here has a better idea of what exactly this board is, and what it would take to coax a nice accurate 10Mhz reference signal out of it.

A few notes: the AD5PS-1 is a signal splitter (fancy. $30 on Digi-Key in small quantities). Input on pin 1; output on pins 4-8, each of which routes to one of the SMA jacks J1-J5. The 6th SMA J6 is possibly an input. On the picture below, I drew in a trace that is otherwise hidden by the black plastic bezel.

This is definitely a 4-layer board, making it hard to trace everything out.

F1 and F2 are filters, also pretty fancy, Murata BNX002-01. TH1 is maybe a thermistor? but I can't find that part number. The electrolytics are 50V.

U2 is barely marked (perpendicular to J5 and L3) but it looks like one of those RF amplifiers, taking DC power through L3.

Everything here looks reasonably well-made, though without anything to actually produce a 10Mhz signal. How do I go about figuring out what goes in the U1 footprint? OCXO I presume? What model? The corner pins on that footprint form a square of about 33.5mm on a side. Or is this designed to take an input signal through J6 from an external source like a rubidium driver or something exotic?

Any hope for this as a working frequency reference?

[mawyatt]

JABIL is a local (massive) assembly house here in Tampa Bay area, Florida. The large missing rectangle is likely the 10MHz OCXO.

Best,

[ZigmundRat]

U1 is definitely an OCXO footprint, so no 10MHz as it is now. J6 looks to be connected to the OCXO output. There’s not much  active circuitry on the board it seems, so a plain OCXO with no EFC is probably what you want. You’ll need to be sure of the voltage, pin out, and pin spacing before selecting a candidate. I don’t see any regulators etc so I’m thinking the power input is simply filtered before reaching the OCXO. I’m not sure the function of Q4, but it could be a buffer or a squaring circuit for the other outputs.

[mdubinko]

Here's what I can find with a matching footprint
https://www.vectron.com/products/ocxo/ox-042.pdf
and also
https://www.vectron.com/products/ocxo/ox-045.pdf
and
https://www.vectron.com/products/ocxo/co714718s.htm

These even have an adjustment screw in the same spot as the little notch on the footprint.
50mm x 50mm (or 2in x 2in) but the pin spacing is father from the edges than all other such sized OCXOs I can find online, other than the above.

Finding one for sale is another matter. I suppose I could jumper across to something of a more common size.

[mdubinko]

And here's the device with the 04 marking code. Not easy to find.
A 0-1Ghz monolithic amplifier, MAR-4SM+ : https://www.minicircuits.com/pdfs/MAR-4SM+.pdf
Neat little device. It has an 8db gain, which about offsets the expected -7db loss just from splitting the signal 5 ways.

But more importantly, the datasheet has a chart of V_bias resistors vs VCC. This board has 3 series 64R9 resistors, which almost exactly matches the 15V recommendation. So now I know what to power this board with...

[edpalmer42]

15V is unusual for powering an OCXO.  Is there a circuit between the 15V and the OCXO power pin that would drop the voltage to 12V?

Ed

[hpw]

Yeah, the beef is missing 

May search on ebay.com about "OXCO 10MHZ" or shop "Queen*s_land" plenty of junk while you do not know the real PN (Phase Noise) and look about 12V brands.

So prepare what you will get. Also the given Chinese clock distribution comes often with a different OXCO brand

Also missing on your part, the reference power and trimming for fine adjustment. Your OXCO should be an sine based, but often seen a square wave OXCO and than a sine filter 

[mdubinko]

Thanks all for your comments. This looks a bit less like a good deal, and more like a new project. 

@edpalmer42 all data align on it being one of these: https://www.vectron.com/products/ocxo/co714718s.htm and these are listed as taking 15V standard (except for part numbers with the N suffix). The V+ comes in on pin 4 of the big connector, and goes only through TH1 and both filters before popping up on the chain of components starting with L2.

@hpw and these 714/717 units have a physical adjustment screw, even marked on the silkscreen. But yes, any kind of voltage regulation seems to happen off-board.

Anyway, I also identified component TH1 -- it's a thermal resettable fuse @ 0.5A right on the positive V+ input. And D2 looks like a TVS, which makes sense to protect those open-to-the-world SMA ports.

-m

[edpalmer42]

Thanks all for your comments. This looks a bit less like a good deal, and more like a new project. 

@edpalmer42 all data align on it being one of these: https://www.vectron.com/products/ocxo/co714718s.htm and these are listed as taking 15V standard (except for part numbers with the N suffix). The V+ comes in on pin 4 of the big connector, and goes only through TH1 and both filters before popping up on the chain of components starting with L2.

If you run the board at 12V, your options for an OCXO will be much greater whether you go for a used unit from ebay or a new unit.  The MAR-4SM+ might work just fine at 12V but, if not, a quick change to the bias string will be easy.

Anyway, I also identified component TH1 -- it's a thermal resettable fuse @ 0.5A right on the positive V+ input. And D2 looks like a TVS, which makes sense to protect those open-to-the-world SMA ports.

These things are often called PTC (Positive Temperature Coefficient) thermistors or PTC fuses.

Ed

[mdubinko]

For posterity. The full circuit. (This is going several notches down on my TODO list)