Searching \ for '[EE:] Measurig Crystal Drive level.' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: techref.massmind.org/techref/index.htm?key=measurig+crystal
Search entire site for: 'Measurig Crystal Drive level.'.

Exact match. Not showing close matches.
PICList Thread
'[EE:] Measurig Crystal Drive level.'
2011\03\14@225224 by Richard Prosser

picon face
Hi All,
I have a bit of a problem that I'm trying to sort out.
Basically, we have had a small number of crystal failures and the
crystal manufacturer claims that this is due to us overdriving the
crystal. So I am trying to repeat their measurements and adjust
component values to meet their, and our, requirements.
The problem is that the maximum specified drive level is 1uW and the
maximum crystal ESR is 3.5k. So the maximum rms crystal current is
about 17uA. In the past (different crystals) I have used a current
probe but that only operates down to about 20mA/division on the scope.
Adding more turns would increase the series inductance and upset
things. I can insert series resistance and use FET probes, but while a
1k series resistance will give me a reading, it is not much above the
noise level (I'm getting about 12mV of noise in my minimum bandwidth
of 20MHz). Also, it's close to the ESR value so is probably upsetting
the readings. If I use passive probes, I get less noise, but the
capacitance of the probes upsets the circuit and gives me false
readings - and is possibly the cause of me damaging my origonal
crystal sample.

The basic crystal spec is 455kHz, 1uW max, ESR = 3.5k max, 1.4k
typical, 4pF loading.

Any ideas? I'm even thinking of using an AM radio receiver with a
455kHz IF to see if I can get some sort of amplification, but will
need a bit more thought as to how to connect and calibrate it.

I have a good  range of test equipment available, including some RF
specific gear such as signal generatrs, spectrum analysers and VNAs
etc although it's mostly aimed at VHF and above frequencies.

Richard

2011\03\14@232826 by Harold Hallikainen

face
flavicon
face
Instead of trying to measure the current, I wonder if you can measure the
differential voltage using a couple scope probes. That seems a lot easier
than trying to measure the current. Then calculate the drive power based
on the ESR of the crystal. Do they give a typical and minimum ESR? Do they
maybe even give a maximum recommended differential voltage across the
crystal?

Harold




-- FCC Rules Updated Daily at http://www.hallikainen.com - Advertising
opportunities available

2011\03\14@234943 by Richard Prosser

picon face
On 15 March 2011 16:28, Harold Hallikainen <spam_OUTharoldTakeThisOuTspamhallikainen.org> wrote:
> Instead of trying to measure the current, I wonder if you can measure the
> differential voltage using a couple scope probes. That seems a lot easier
> than trying to measure the current. Then calculate the drive power based
> on the ESR of the crystal. Do they give a typical and minimum ESR? Do they
> maybe even give a maximum recommended differential voltage across the
> crystal?
>
> Harold
>
>
>
>
> --
Thanks for the input.

Yes, I do have typical and max ESR values (1k4 and 3k5) for the
crystal but have had difficulty in the past getting correlation
between differential voltage measurements and the recommended current
probe measurements. The ESR resistance is in series with the reactive
components of the crystal and so there is a significant phase shift.
It might work if I had an equivalent parallel resistance value but I
don't, and the crystal equivalent circuit lumped component values are
not readily available either.

Also the drive voltage is close to a square wave while the output
voltage is nearly sinewave so I'm not sure a direct rms difference (or
rms conversion of the difference) is appropriate. And the output node
is quite high impedance so is sensitive to any loading at all.

The various manufacturer app notes etc always use either a current
probe or a series resistance to derive crystal current but I'm not
having too much luck this time.

Thanks,

Richard Prosse

2011\03\15@002505 by Bob Blick

face
flavicon
face
On Tue, 15 Mar 2011 16:47 +1300, "Richard Prosser" <.....rhprosserKILLspamspam@spam@gmail.com>
wrote:
> Yes, I do have typical and max ESR values (1k4 and 3k5) for the
> crystal but have had difficulty in the past getting correlation
> between differential voltage measurements and the recommended current
> probe measurements. The ESR resistance is in series with the reactive
> components of the crystal and so there is a significant phase shift.
> It might work if I had an equivalent parallel resistance value but I
> don't, and the crystal equivalent circuit lumped component values are
> not readily available either.
>
> Also the drive voltage is close to a square wave while the output
> voltage is nearly sinewave so I'm not sure a direct rms difference (or
> rms conversion of the difference) is appropriate. And the output node
> is quite high impedance so is sensitive to any loading at all.

I don't have experience at 455 KHz but some at 32. I found it hard to
measure using two standard probes in differential mode because you are
influencing the phase, so even compensating for the amount you calculate
the voltage drop you create you don't get a clear idea of the current
since you can't trust that the phase has shifted on one side enough that
your differential voltage is now different. And since the impedance is
radically different on the two sides it is very frustrating. What I did
to make it work was add a 90 megohm resistor on the tip of each scope
probe and probe across the crystal. I used several small resistors to
make the 90 megohms and the metal guts of a small connector to grip it
onto the probe tip. With a x10 probe this makes it x100. Make sure your
scope is set to AC coupling because even 100 megs of DC to ground will
load some processors (mine was a Sanyo 4-bit that had start-up and run
drivers for the crystal and run was very weak). You wll get some
waveform which you can't completely trust but enough to satisfy you
aren't totally wrong.

> The various manufacturer app notes etc always use either a current
> probe or a series resistance to derive crystal current but I'm not
> having too much luck this time.

Yeah, I eventually figured they either made their own test equipment or
made it all up :(

Cheerful regards,

Bob


-- http://www.fastmail.fm - A fast, anti-spam email service.

2011\03\15@014225 by Kerry Wentworth

flavicon
face
Here are a couple of sites you might find interesting:

www.cliftonlaboratories.com/Documents/Crystal%20Motional%20Parameters.pdf
covers measuring motional parameters

www.oscilent.com/spec_pages/PNDescrpt/ESR_Mode.htm
covers calculating ESR from the motional parameters.

It seems to me that if the question is overdriving the crystals, you would want to know the minimum ESR, in order to calculate max voltage.  At 1.4K, max voltage is a paltry 37mV.  Is your circuit operating at that low of a voltage?

Kerry



Richard Prosser wrote:
{Quote hidden}

-- Internal Virus Database is out-of-date.
Checked by AVG Anti-Virus.
Version: 7.0.289 / Virus Database: 267.11.13 - Release Date: 10/6/05

2011\03\15@014622 by Brent Brown

picon face
On 15 Mar 2011 at 15:52, Richard Prosser wrote:

{Quote hidden}

Sounds like a tricky little problem. Try an Instrumentation Amplifier across your current sensing resistor to buffer the signal? A quick look at the Burr Brown INA118 datasheet (because it was handy) seems like it would be close but not quite, eg. common mode input impedance 10^10 Ohms would be ok, but common mode input capacitance typ 4pF probably not ok. Pushing bandwidth too, at gain of 10 = 500kHz.

Perhaps there are others with better specs?

-- Brent Brown, Electronic Design Solutions
16 English Street, St Andrews,
Hamilton 3200, New Zealand
Ph: +64 7 849 0069
Fax: +64 7 849 0071
Cell: +64 27 433 4069
eMail:  .....brent.brownKILLspamspam.....clear.net.nz

2011\03\15@020144 by AK

flavicon
face
It sounds like you have the specs to attempt to model it in your circuit.


On Tue, Mar 15, 2011 at 1:46 AM, Brent Brown <EraseMEbrent.brownspam_OUTspamTakeThisOuTclear.net.nz> wrote:
{Quote hidden}

>

2011\03\15@085203 by Olin Lathrop

face picon face
Richard Prosser wrote:
> Yes, I do have typical and max ESR values (1k4 and 3k5) for the
> crystal but have had difficulty in the past getting correlation
> between differential voltage measurements and the recommended current
> probe measurements.

That's probably because you didn't take phase into account.  One way to do
this is to measure both ends of a known impedance with a dual trace scope.
Then you can see the magnitude and also the phase relation between the two
signals, and thereby calculate the true voltage accross the element and the
current thru it.

For example, consider the basic one-pole low pass filter.  Sine voltage
source from ground, to resistor, to capacitor, back to ground.  Measure the
AC voltage independently accross the resistor and capacitor and note how it
adds up to more than what the voltage source is putting out.  This effect
will be maximal when the frequency is at the -3dB point of the filter.  At
first glance it appears that Kirkchoff was wrong and you get something for
nothing.  (This is a favorite trick of those peddaling free energy,
mysterious power factor scams, and trying to convince you a car can run thru
the hills of Turkey for 100 miles on a thimble of gasoline.)  However,
things make sense again when you look instantaneous voltages, or the
averages with phase taken into account.

> Also the drive voltage is close to a square wave

Ugh.  There's your problem.  You need more capacitcance on the crystal
input.  If necessary, add some series resistance from whatever the driving
source is.

Another way to do this is to figure out what the crystal needs to resonate.
Keep on dropping the drive level until it doesn't work anymore.  Make the
actual drive level maybe 4x that to make sure it works accross part
variations and the range of conditions.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000

2011\03\15@160040 by Richard Prosser

picon face
On 16 March 2011 01:52, Olin Lathrop <@spam@olin_piclistKILLspamspamembedinc.com> wrote:
{Quote hidden}

Thanks All.

Yes, Once I get a handle on how to measure tit  the plan is to adjust
the series resistance to reduce the power to the correct level.
Unfortunately the origonal design notes are lost in the "mists of
time" somewhere between San Fransico & Christchurch & we don't know
how the origonal values were derived. (The design is at least 5 years
old & has not previously been a problem).

I have discovered an Epson app note that gives a bit more info on how
to measure very low level currents for tuning fork crystals and it
looks like I may be able to get a good reading by connecting my
current probe to the spectrum analyser & reducing the bandwidth.
Otherwise I'll be adapting an AM radio & measuring the AGC voltage or
something.

The Epson app note suggests that for a 32kHz tuning fork crystal a
series R of 220k may often be appropriate, so the 1k series R fitted
at the moment appears way too low - even at 455kHz.  One concern is
that the crystal data sheet indicates that in this frequency range the
crystal is operating in overtone mode, but there are no components
added to encourage this. Maybe there's mechanical constraints within
the crystal package?
I know it still oscillates with 33k in series and I may have to go a
lot higher. Then I can test for startup and power level at temperature
& voltage extreames & try & give myself a bit of safety factor etc.

I'll also have a go at modelling the crystal and see if  that sheds
any light on the situation - for example, measuring the output voltage
using  higher resistance probes as previously suggested. by Bob.

We have also now discovered that this is a recent event - earlier
batches of crystal survived, while some later batches have close to
100% failure. We're guessing that the manufacturer has changed
something but hasn't let us know about it - not that that's any excuse
for a design running at 20 x the permitted level.

RP

More... (looser matching)
- Last day of these posts
- In 2011 , 2012 only
- Today
- New search...