P3 with SASM module. Losing 10sec/24hrs.

[Handy]

I have a P3 with a SASM module. The module is very nicely finished and looks very 1970s.
The crystal is a modern one in a cylindrical package.
I was wondering whether adding some capacitance, in series with the crystal, would pull its frequency up a little.
Your thoughts please.
Cheers,
Ian

[rewolf]

10s/24h is too much IMHO, quite likely there's something wrong with the oscillator or crystal.

I don't know the oscillator circuit of the SASM module, but I bet it's a standard "parallel" Pierce oscillator. This type is fine-tuned by changing the load capacitance which is actually in parallel with the crystal, not in series.

I'd ask the manufacturer for help, he's a member and regular visitor of this forum.

[Handy]

Hi rewolf,
Thanks for your reply. I'll try another crystal, when they arrive.
I was thinking of a Colpits oscillator.
Cheers,
Ian

[Old Tom]

10sec in 24Hr is about 1 part in 8650 which is too much to expect adding/removing capacitance to change (the maximum is about 1 in 25000 for small crystals before oscillation becomes unstable). I think you would do better to return the module to Strikes and Spares for repair rather than trying to change the crystal yourself (1. They fix modules day in/day out and 2. They know the actual circuit used backwards and would be able to diagnose the fault accurately & quickly)- I'd lay money that the oscillator is stop/starting randomly rather than running off frequency- the check for this is to use a really accurate frequency meter and see what the actual oscillation freq is. I have a Trafalgar LCD that does this and when you check the freq it is as near to 32768 as can be, but the watch loses 20sec/ 24Hr.

[retroleds]

Rewolf probably has some thoughts on this, being an electrical engineer....
I've been working on a programmable chip module, and the manufacturers stress that the length of the leads on the quartz oscillator can have a noticeable effect on the capacitance. Does an extra 1/8"(2.5mm) on each lead(1/4" or 5mm between the two) really make a difference that would change timing? If so, maybe is best to not think you have tons of room and to keep the leads as short or shorter than what is already there. I'll differ to those with formal electrical education on this topic.

[rewolf]

Rewolf probably has some thoughts on this...

Influence of lead length: yes, 2.5mm extra CAN influence timing, theoretically measureable, but not noticable. We're talking about less than 1pF here, and the standard load capacitance is 12.5pF with say +/-7pF for adjustment.
5mm lead length is no problem at all.

Old Tom made a good point: unstable oscillation.
This can sometimes be cured with a resistor of ~10M in parallel with the crystal.

But before you try the resistor: remove any solder flux residue from the oscillator circuitry. Isopropanol is a good flux removal, or spirit.
Some low power microcontrollers are VERY sensitive in this regard - they won't oscillate reliably with even the tiniest traces of solder flux present. Some years ago I had this problem with a an MSP430F435-based design: on 15 out of 20 prototype boards the oscilaltor sometimes wouldn't start - just cleaned them and all was fine. This was a known problem of that µC, not a board design flaw.

[Handy]

Thanks for your input gents. I'll see what I can do this coming weekend.
Cheers,
Ian

[Handy]

Rewolf probably has some thoughts on this...

Influence of lead length: yes, 2.5mm extra CAN influence timing, theoretically measureable, but not noticable. We're talking about less than 1pF here, and the standard load capacitance is 12.5pF with say +/-7pF for adjustment.
5mm lead length is no problem at all.

Old Tom made a good point: unstable oscillation.
This can sometimes be cured with a resistor of ~10M in parallel with the crystal.

But before you try the resistor: remove any solder flux residue from the oscillator circuitry. Isopropanol is a good flux removal, or spirit.
Some low power microcontrollers are VERY sensitive in this regard - they won't oscillate reliably with even the tiniest traces of solder flux present. Some years ago I had this problem with a an MSP430F435-based design: on 15 out of 20 prototype boards the oscilaltor sometimes wouldn't start - just cleaned them and all was fine. This was a known problem of that µC, not a board design flaw.

I would have thought that long leads would add a tiny amount of inductance, rather than capacitance?
Cheers,
Ian

[rewolf]

I would have thought that long leads would add a tiny amount of inductance, rather than capacitance?

Inductance, capacitance and resistance, to be exact.
But at 32kHz a few nH inductance have no effect at all. Capacitance is by far the dominating effect here, and even this effect is barely measurable.
Wire inductance calculator

[Handy]

Point taken. I'll see if there is any flux residue on the IC.
Cheers,
Ian

[Handy]

Today, I changed the crystal and it's keeping time!
Thanks to all who replied with their suggestions.
Cheers,
Ian

[Handy]

Well,
The module is now dead! It just dispalys "SAS".
Fit for only one thing....the dustbin.

[BENRUS]

I recently had this one and only problem with my SASM-14...Displayed only "'SAS", It just needed a new quartz crystal. I sent it in for repair to Hanno of "Strikes & Spares" and it is now working fine. It lost about 3-4 seconds its first month home, but I reset the time, and am recalculating to see if its any better. It was originally running with its original quartz crystal, and it was keeping perfect time...It now has a barrel style installed, and I don't know if that makes a difference??? Maybe it needs time to break in being that it is new?
Hope this helps

[Handy]

I recently had this one and only problem with my SASM-14...Displayed only "'SAS", It just needed a new quartz crystal. I sent it in for repair to Hanno of "Strikes & Spares" and it is now working fine. It lost about 3-4 seconds its first month home, but I reset the time, and am recalculating to see if its any better. It was originally running with its original quartz crystal, and it was keeping perfect time...It now has a barrel style installed, and I don't know if that makes a difference??? Maybe it needs time to break in being that it is new?
Hope this helps

Hi Benrus,
I have changed the crystal and it lives again but it is erratic.
I'd rather use original modules as they have the light sensor on them.

[BENRUS]

Is this module under the 1 year warranty? Hanno is easy to work with, and repair/replacement parts work may not cost you much even if it is not covered under warranty.
The quartz you just installed might be a bad one still if its erratic?
I had the idea of spot gluing a light sensor in place for the original look through the time-screen, but of course if your going for a complete original module, they are not going to last forever without the need of repair/replacement parts soon enough...right?
In my opinion the light sensor is for aesthetics only.
Good luck on your repair

[Handy]

Is this module under the 1 year warranty? Hanno is easy to work with, and repair/replacement parts work may not cost you much even if it is not covered under warranty.
The quartz you just installed might be a bad one still if its erratic?
I had the idea of spot gluing a light sensor in place for the original look through the time-screen, but of course if your going for a complete original module, they are not going to last forever without the need of repair/replacement parts soon enough...right?
In my opinion the light sensor is for aesthetics only.
Good luck on your repair

I bought it secondhand so I don't know how old it is. It seems to be behaving itself at the moment.
The light sensor extends battery life by applying a lower voltage to the LEDs. It's also nice to see a dimmer display when in a dark environment.