High fashion crystal oscillator?

It looks like a fairly tightly specified crystal - one way of getting tight specification parts is to make a lot, test them all, and sort them into progressively narrower bins, and charge enough to stop people emptying the best bins too fast.

The 1N821 to 1N829 reference diode range seems to have worked that way.

The audiophool products are aimed at less discriminating buyers.

Radio Spares always was a slightly shady supplier - you never knew quite whose parts you were going to get - but they didn't actually go after the audiophool customers.

Probably a quote error from RS. Exactly the same manufacturer reference (IQD LFPTXO000295) is priced 40-45Eur by Farnell, Digikey or Mouser. Still not cheap, but that's a 1ppm VTCXO...

Robert

Infamously their 741s never had zero offset voltage, but the Vos was distinctly bimodal and near the limit.

But the "whichever supplier is cheapest" is to be seen in many places, from some PCB suppliers to Farnell's "multicomp".

It's RS, what did you expect?

NT

That's worth maybe $7. We buy one Fox TCVCXO for about $2.25 in modest quantity.

You can get a seriously good OCXO for around $70 these days.

Curiously, John Larkin hasn't posted the exact part number or a link to the specification sheet for the part he can buy for $2.25.

Granting his grasp of electronic design, he may not see the point.

A crystal oscillator in an oven runs hot (or always used to - a Peltier junction could let you stabilise it's temperature close to room temperature).

This means that it eats power, and ages fast.

Since John can't be bothered to quantify what he means by "seriously good" this isn't a particularly useful observation.

Farnell did at least sell stuff that was identified as coming from a particular manufacturer, as well as stuff from "whoever is cheapest today".

Some relatively low frequency crystals have a parabolic frequency vs temperature curve. The crystal is then cut to have zero Tc at the top of parabola at the desired operating temperature. Thus, if higher oven temperatures are used, the crystal must be cut for that temperature.

The power consumption is an issue, if mains power is not always available and you have to use batteries to maintain temperature. I once used an oscillator in a thermos bottle to minimize oven heat losses.

I've seen 'em in cans with fiberglass insulation, and in small thermos bottles, as well as what looks like hermetic sealed cases. The two-ounce Dewar flask is a lovely little thing...

A oven eats power with 100% efficiency, a Peltier is much worse. What we want is maximum thermal impedance between the xtal and its environment, some mass for thermal low pass filtering and the temperature the crystal is cut for.

everything; there are also things like activity dips and mode jumping.

A Peltier has all the attributes we don't want there. It has a very low thermal impedance between plate A and B, so there is not any low pass filtering. It generates COLD on one side that we don't need

the verge of dying.

Hope you had a nice winter solstice. Gerhard.

Some people from the HP corner claim that SC does not stand for stress compensated but for Santa Clara :-)

junction could let you stabilise it's temperature close to room temperatur e).

The point about the Peltier junction is that you can stabilise the temperat ure close to ambient - probably a little above, since the resistive heating in a Peltier junction helps when you are using it to warm something.\.

Because the average temperature difference can be a lot lower, the heat you need to shift is less, and the power you need to expend doing it is also l ess.

So get a crystal cut for a slightly lower minimum temperature coefficient t emperature. 70C was a sensible choice when heating was a lot less trouble t han cooling. It isn't now.

It was the summer solstice here in Australia.

Most adults don't believe in Santa Claus. Saint Clara is just the female ve rsion.

Just so. There is a place for both, especially if the limitations of each option are highly visible.

junction could let you stabilise it's temperature close to room temperature).

Sounds ike a good idea, but not in practice, as the peltier is very power inefficient and a short life when cycled regularly. Considering that you can buy a double oven vcxo in less than a cubic inch, with warmup of a few minutes and long term stability of a few parts in

Chris

I used to build my own OCXOs, and got this crystal from Lap-Tech in Canada. It's vacuum sealed in glass, very clean, low ageing.

The "60" on the bottom is the measured "turning point" temperature, the peak of the parabola.

Nowadays, we just buy OCXOs. They are a lot cheaper and better than they used to be.

Peltiers are very inefficient if the heat you try to extract is the same of magnitude as the Peltier nominal power. If the Peltier nominal power is an order of magnitude greater, the situation is not as bad.

Agreed that these days temperature compensated oscillators are quite good.

If you aren't using much power to maintain a temperature close to ambient, efficiency isn't all that important.

Quite what inputs and outputs you are computing to generate your claim that a Peltier is "very power inefficient" escapes me. If you use a Peltier jun ction to warm an object that's just slightly above room temperature, you ge t more heat out than the power you put in because the ohmic heating in the Peltier ends up in the object, and the Peltier effect is moving heat into t he object from the ambient.

The "short life when cycled regularly" is news to me. You shouldn't cycle a Peltier junction - its a non-linear device and works best if you keep the temperature you are stabilising as stable as possible.

My boss for the 1993 project that I published in 1996 had a horror story ab out some idiot who set up a Peltier for bang-bang control and melted the so lder holding the junction together. I did pulse-width-modulate the current going into that junction, but the current was heavily low-pass filtered bef ore it got into the junction, and the extra square-law resistive heating wa s negligible.

It keeps the crystal at a temperature above ambient, where it ages faster.

I've designed these ones into a few boxes:

--Spehro Pefhany

The "analog temperature compensation" part is good. Some people do the temp comp with digital methods, and get superb frequency stabillity specs and ghastly phase noise. I think someone did a TCXO that just skipped whole cycles now and then to pull the average frequency down.

I assume that the analog temperature comp is laser trimmed.