The Method to Have the Signal level to be Up to 3V

You need to amplify the signal - and you probably want to doiit with a comparator, since most comparators are designed to give logic level signals at their output.

Unfortunately, you haven't told us the frequency at which your voltage controlled crystal oscillaotr is working, so we can't immediately recommend a particular comparator.

Regular comparators putting out TTL/CMOS logic level signals are often too slow to follow a signal switching at more than a few MHz, and faster parts canbe expensieand difficult to use.

snipped-for-privacy@ieee.org =C0=DB=BC=BA:

Thank you for answer.

The frequency is 40MHz and this VCTCXO has +-0.2ppm(max) frequency stability. So if you may have a proper comparator for me, let me know.

Thanks.

is

Have a look at the Linear Technology LT1711 comparator

It certainly isn't the only part that could do the job - John Larkin can probably suggest cheaper alternatives from the National Semiconductor low voltage digital signalling (LVDS) range of line receivers, like the

but I've not used these parts, and they may well be trickier to se up.

--=20 Bill Sloman, Nijmegen (but in Melbourne at the moment)

Unless you need more power to drive the input as well as higher voltage levels, I'd consider a C-L-C matching network. You can tolerate a lot of damping, so low Q (3 or more is all you need) is no problem. A transformer/tapped coil could also be used, but those are less easy to find in stock. Your oscillator is 'way more signal than should require the gain of a comparator.

Getting a low enough output impedance out of a transformer may be difficult.

A relatively low gain line receiver will certainly provide enough gain to act as a clock buffer, and some logic gates have enough gain to do the job, but the chance of fidning a gotcha is even higher.

If this oscillator signal is running continuously Jay might also consider the white knuckle method: Take a buffer or (buffered) inverter that is fast enough at 3.3V, bias it at VCC/2 and AC-couple the 1V signal into it. But if the oscillator ever stops it'll get quite hot and do weird things. It has to run all the time. I know that this isn't considered kosher but it has been done.

is

P2077,D1848

Agreed - my most recent post in this thread mentions this as a possibility.

My background is in small volume high value products, so I'm very nervous of any solution that might involve a service enginer having to get to the equipment to replace a part, but it would be fine in a lab prototype.

--=20 Bill Sloman, Nijmegen (but still in Melbourne at the moment).

It's been done in regular products. As long as this oscillator is always running and has a somewhat orthodox duty cycle. In most cases, if it stops running, it's a warranty or repair return anyhow.

I had the same problem when I needed to clock a PLL chip like the LMX2353, IIRC what I did was ac couple the output of a crystal oscillator and make a voltage divider with 2 resistors after the cap. worked fine. this was for a frequency synthesizer at 2.3-2.7ghz you didn't say what you were trying to clock, it's alway nice to read the datasheet of the input clock requirement of the device...

this

ve

or

ge

en

,P2077,D1848

High-Q crystal oscillators don't start up instantly - I'd probably want to monitor a bunch of turn-ons with a temperature sensor glued to the top of the logic gate before I'd sign off on such a design.

--=20 Bill Sloman, Nijmegen

Yeah, you'd have to have a long enough BOR and hold that one gate to a corner for the reset period.