Selecting load capacitors for crystal without datasheet

Thread Starter

hrs

Joined Jun 13, 2014
351
Hi,

A while back I purchased a 32768 Hz crystal at a local store. Now that I'm going to use it in conjunction with a CD4060 I realize that I need to know the load capacitance. I've asked them for a datasheet and they say they can't provide one. It's a HC49/U type, whatever that means and it looks like this:


If I stick it in a breadboard with 50 pF or other values that I have, what should I be looking for in a crystal that Is running well? I guess I can't scope a crystal directly but maybe I can buffer it with a jFET opamp?
 

Attachments

MrChips

Joined Oct 2, 2009
24,650
Try 20-33pF capacitors.
Probe the signal at Q4. You will be looking for square wave at 32768/16Hz = 2048Hz.

Edit: Some documents show pin-7 as Q3 while others show it as Q4.
I am not certain if the output at pin-7 is divided by 16 or 32. I will look into it.
 

Thread Starter

hrs

Joined Jun 13, 2014
351
Thanks for the heads up MrChips. The Fairchild CD4060 datasheet shows Q4 at pin 7 but I'll keep this in mind. The main take-away is perhaps that if the ripple counter ripples then the crystal is running fine.
 

DickCappels

Joined Aug 21, 2008
8,000
Be aware that for some 32768 crystals the parasitic capacitance on the board is about right and adding discreet capacitors will kill the oscillation. You will have to experiment. I would start with MrChips' suggestion in post #2.
 

sparky 1

Joined Nov 3, 2018
612
Because you know the frequency adjust a trimmer capacitor observing the best results.
Possibly between 12pF and 24pF the nice thing is when the frequency output is right the capacitance will be close.
If for some reason it does not work you can always go to higher capacitance.
 

Thread Starter

hrs

Joined Jun 13, 2014
351
After trying the above I cannot get the crystal to start.

What tolerance is there on the load capacitance? How sensitive is a crystal to get working? Having tried 2*20p, 2*30p, 20p + 30p and some combinations with 15p, and 2*47p but only with 180k for Rs(see below). No oscillation was observed. I also tried without caps per DickCappels advise.
Since I didn't have 15Meg resistors I used 10Meg for the feedback resistor. From what I read this shouldn't be to critical but I could botch in an extra 5Meg series resistor.

According to this document a 32768 Hz crystal may take between 1 to 5 seconds to start up. I have waited that long.

According to this document typical loads specs for the crystal are 18p or 20p "for a clock design". Assuming 5p input and output capacitance and 2p stray capacitance 30p loading caps should be about perfect. I realize that that doesn't mean the crystal load spec actually is 20p.
It also says that a good initial value for Rs, the current limiting resistor, would be Xc of the loading caps. So 1/2pi*f*C = 162k, then I replaced the 330k with 180k. This is R301 in my schematic. No joy. Bad idea? Change it back?

When I apply a square wave to pin 9 of the CD4060 I see division going on at the various outputs so that works. When powering the circuit as shown I see 5.6V on pin10 and 4.8V on pin11 so I guesstimate there's a 60Meg leakage path from pin11 to ground. I don't know if that is normal. Maybe I damaged the gate by improperly handling the part. Can the current of an Ohm meter destroy a crystal? I'm sure I didn't drop the crystal but maybe the guy that sold it to me did.

Is there something obvious to try before I go out and get a new chip and crystal?
 

Attachments

Thread Starter

hrs

Joined Jun 13, 2014
351
Indeed, as you have said in post #2. So that's what I do, look for a square wave on any of the CD4060 flipflop outputs. They are all latched either high or low. I do not measure the crystal directly.
 

sarahMCML

Joined May 11, 2019
42
Hi,

A while back I purchased a 32768 Hz crystal at a local store. Now that I'm going to use it in conjunction with a CD4060 I realize that I need to know the load capacitance. I've asked them for a datasheet and they say they can't provide one. It's a HC49/U type, whatever that means and it looks like this:


If I stick it in a breadboard with 50 pF or other values that I have, what should I be looking for in a crystal that Is running well? I guess I can't scope a crystal directly but maybe I can buffer it with a jFET opamp?
Funnily enough, I have one running at the moment on a bread board. I'm using a tiny old xtal salvaged from a watch, together with a 10M, a 220K, 10pF for C302 and 33pF for C301. It starts instantly at 5 volts.
The CD4060 is an ancient AE suffix version that I must have had for about 50 years! I have it driving a 2N7000 buffer via pin 9 into an equally ancient CD4031AE shift register. It won't start reliably if that pin is used without the buffer.
So keep at it!
Sarah.
 

Thread Starter

hrs

Joined Jun 13, 2014
351
Funnily enough, I have one running at the moment on a bread board. I'm using a tiny old xtal salvaged from a watch, together with a 10M, a 220K, 10pF for C302 and 33pF for C301. It starts instantly at 5 volts.
The CD4060 is an ancient AE suffix version that I must have had for about 50 years! I have it driving a 2N7000 buffer via pin 9 into an equally ancient CD4031AE shift register. It won't start reliably if that pin is used without the buffer.
So keep at it!
Sarah.
Hi Sarah, is that a tuning fork by any chance? Mine is a HC49/U. In an attempt to find a datasheet, any datasheet of an HC49/U 32.7 kHz crystal I find that I can't. They all start at 1Mhz or higher, also in the sheet that Boba just posted. Mouser and Digikey only list tuning forks for the 32.7 kHz. Apparently 32.768 MHz crystals exist and I suspect that what I have might be a 32.768 MHz crystal ...
 

Papabravo

Joined Feb 24, 2006
17,316
Hi Sarah, is that a tuning fork by any chance? Mine is a HC49/U. In an attempt to find a datasheet, any datasheet of an HC49/U 32.7 kHz crystal I find that I can't. They all start at 1Mhz or higher, also in the sheet that Boba just posted. Mouser and Digikey only list tuning forks for the 32.7 kHz. Apparently 32.768 MHz crystals exist and I suspect that what I have might be a 32.768 MHz crystal ...
Crystals at that frequency do exist, but they usually come in a "grain of rice" package for use in watches. Your inability to get it to work might indicate a host of other problems, not the least of which could be package mislabeling or a counterfeit part. If you had a 'friend' with a VNA, maybe you could sweet talk her into to sweeping it for you. Then you would know all the answers.
 
There's never been a 32.768kHz crystal offered in an HC-49U package. So I also suspect you've got a 32.768MHz part.
AT cut crystals in the MHz freq range fit HC-49U but not low freq tuning fork crystals, which is why those are in a long, skinny tube.

I would short R301 330k ohm resistor to have full gain and see what happens. That resisistor limits drive power to the crystal, typically 0.5uW for a watch type but you can drive it harder for a test. Low drive is for low aging. CD4060 limited to 3.5MHz to maybe 12MHz at 15V so if you have a 32.768MHz crystal it still might not do much.
Any test equipment and probes touching the crystal oscillator's pins upsets the circuit due to loading. This is a freq shift or the oscillator just collapses, but you can see the ripple-counter outputs toggle.
 

sarahMCML

Joined May 11, 2019
42
Hi Sarah, is that a tuning fork by any chance? Mine is a HC49/U. In an attempt to find a datasheet, any datasheet of an HC49/U 32.7 kHz crystal I find that I can't. They all start at 1Mhz or higher, also in the sheet that Boba just posted. Mouser and Digikey only list tuning forks for the 32.7 kHz. Apparently 32.768 MHz crystals exist and I suspect that what I have might be a 32.768 MHz crystal ...
Yes, it's about 2.5mm dia. x 12mm long at a guess. It's running a little slow, but I just grabbed the nearest caps to hand at the time since I wasn't too fussed.
If I get a chance tomorrow I'll see if I can get one of my ancient 1KHz bar type xtals in a B7G glass valve package going, just for fun! That's a big chunk of quartz about 4mm square x 25mm long.
 
Top