Quartz oscillator replacement in Wittnaeur Polara

[charger105]

Hello Everybody.
I am new to this forum and would like some help with the replacement of the quartz crystal in my Wittnauer Polara LED watch. It is of the type which flush mounts to the PCB. The chip and display are on the other side of the PCB, so access to the solder tabs is prevented from here. How do you remove the existing quartz crystal ? Are there any tricks ? If I hold it up to the light, I can just see the two legs from the crystal in the crack between the bottom of the crystal and the PCB.

On another note, there is a sticker on the quartz crystal saying 'W 9.76'. Does this indicate a manufacture date of September 1976 ?

Any help would be much appreciated.

[bruce wegmann]

It may be necessary to heat the crystal can to solder-melting temperature, and let heat conduction do the rest. Re-soldering a replacement may be difficult; it might be necessary to use conductive epoxy to made the actual connection. Your interpretation of the numbers as a date-code is correct.

[redled]

Yes, 9/76 is the date the module was manufactured. Can you clip the leads of the old crystal and solder a new one onto the wires that remain? Can you post a close-up picture?

[Majestyk]

I got one of these the other day with the dreaded flickering numbers (IE quartz not working or not soldered in properly). In the past I used a system of fixing it in which I won't report because it only had a 20% repair ratio. This time I decided that lifting up the black cover was going to happen and the crystal would be properly soldered in. 5 minutes later I had the watch running. It's really not that hard. You have to peel off MOST of the adhesive around the cover but leave most of it at the top. This way when you lift it up, it will stay in a relatively fixed position and "swing" back closed the way it opened. You do have to use something to hold it open while doing the work. As far as opening it, get a VERY small and flat screw driver and pry it open on the left side where the clear plastic is. This is the area where the fewest wire bonds are, incase your screw driver slips a little. You may find a better tool to do this. Then very SLOWLY and GENTLY lift the cover up and hope that it doesn't pull any wire bonds along with it. As long as you do it slowly, you should be ok. You might also want to move up to a larger tool once it's lose enough.

In my case, there was no connection with one of the quartz crystals leads, since the solder had come loose. I Re-soldered all three of them, gently closed the cover, making sure it came down the way it came up. When it came time to reseal I put some glue around the cover to keep it closed but I wouldn't recommend that unless you use gel glue. It’s best to use the same type of adhesive compound that was originally used.

Jeff

[retroleds]

As soon as you lift that "lid", I would suggest you drip a few drops of clear nail poilish over the entire wire bond area - this will prevent 99% of the wire bonds accidents that can occur while you are messing with the crystal. Also protects the wire bonds for posterity. I'd second it that conductive epoxy is a good choice if it wasn't soldered.

Lifting the lid: generally if you go around it with a screwdriver or knife tip just to remove the adhesive,not going under the lid at all, you will find the lid comes off with a fingernail. The adhesive wasn't ran under the lid, it was only AROUND it. 8)

[Majestyk]

Lifting the lid: generally if you go around it with a screwdriver or knife tip just to remove the adhesive,not going under the lid at all, you will find the lid comes off with a fingernail. The adhesive wasn't ran under the lid, it was only AROUND it. Cool

Doesn't work with me, but if it works for anyone else, go for it. I find even with the adhesive off, it takes quite a bit of force to get it open and too much potential for ones finger to slip inside. It also depends if you actually have fingernails. I cut mine very short. :)

M

[retroleds]

Jeff: I trim mine extremely close too(I play guitar and bass) - that's the beauty of it, there is very little to go under the lid. I've plucked off hundreds of those. Maybe I just have really strong fingers and nails

In any event - caution is always advisable when lifting the mask off any module with open wire-bonds. :wink:

[charger105]

Thank you for your responses everybody. I knew there must have been a trick !
Anyway, I've lifted the "lid" with no mishaps, and found 3 legs on the bottom of the PCB (Jeff said he re-soldered all 3 legs, which puzzled me slightly). I anticipated only 2 legs. I have some 32kHz 2 leg crystals which I was going to use as a replacement.
Excuse my ignorance, but can anyone explain the 3-legged variety please ? Also, a manufacturer/part number or source would be much appreciated.

Rgds,
Andrew.

[Majestyk]

Hughes Aircraft not only used 3 wired (or pinned) quartz crystals but they also do not run at 32,768Hz. They are 786,432Hz. Now the last production Hughes QC's actually had two pins, which are rare but they still run on Hughes exclusive frequency. (Someone correct me if I'm wrong on the exclusive part). So it doesn't really matter, you need a Hughes quartz crystal from another Hughes module. In more likelihood, the QC that is on your module works fine but one or more of the pins needs to be re-soldered. Their quartz crystals were almost bullet proof and the most reliable ones that came of the 70's, IMO.

[redled]

Wow, that's amazing! I had no idea they ran at a different frequency, and so fast! Were they capable of much greater accuracy because of it?

[Majestyk]

Were they capable of much greater accuracy because of it?

Yes indeed! They were, and still are, among the best for accuracy!

[retroleds]

I'll have to differ with you on that one Jeff(Majestyk)- I've never been impressed with the accuracy of Hughes modules. I've often thought they put the biggest and best quality trimmer on their modules because they needed it(those are fantastic to pull off and put on other brands that use square drive trimmers). That higher frequency meant they needed extra circuitry in the divider circuits. And market share(albiet that is not neccesarily a fact/quality based occurance) went to the 32,768Hz. To this day.

And now that odd crystal speed is an albatrose around the neck of those who repair and those who collect. But hey, they were made 30+ years ago...they had no idea how things would shake out. Statek created the tuning fork style crystal oscillators that have ruled the watch market since about 1976....by the time Hughes started making LCD modules, even they had switched to the lower speed and the fork design crystals.

[charger105]

Thanks for the explanation. I'll try re-soldering the 3 legs of the existing crystal, in case there's a dry joint.
Each of the 3 solder pads seem to have conductive epoxy, and a fine gold wire loop from the end of each pin connecting to the track leading to each pad. How do I remove this epoxy, so that a good solder joint can be achieved ? By the way, what would be a safe temperature for this work ?

Rgds,
Andrew.

[Majestyk]

Very strange Ed. The Hughes modules are the only ones that I rarely ever have to trim. I would say out of 100's of them, I have had only a few that were over the +/- 30 sec a month mark and and a few I had to replace. Most of them run at 2 to 10 seconds a month. This is the result of two properly calibrated Zantech quartz timers. Whether it's the QC's themselves or the trimmers they use is irrelevant, in my opinion, since the 786mHz QC's are the only ones that work with the Hughes modules. In fact, I don't remember the last time I trimmed one. They generally only keep bad time when they are just on the verge of not having a 100% connection to the module.

Now on the other end of things, National Semiconductor modules are one of the worst for time keeping. In this case it's the QC's they used. If you ever get one with those flat, rectangular QC's, it's almost assured that it will either keep bad time or eventually fail. I have to replace 50% of the QC's on NSC's.


Jeff

[Majestyk]

One thing I should add is the Hughes QC's do have a slightly lower tollerance to temperature changes. With an extreme cold to warm temperature change, for example, they can drift -/+ a few seconds. But they will drift back the other way with the opposite conditions. It barely affects the timing but I have been noticing this since I got my first watch, with a Hughes module, back in 2000.


Jeff

[retroleds]

Most of them run at 2 to 10 seconds a month. This is the result of two properly calibrated Zantech quartz timers.
Now on the other end of things, National Semiconductor modules are one of the worst for time keeping. In this case it's the QC's they used. If you ever get one with those flat, rectangular QC's, it's almost assured that it will either keep bad time or eventually fail. I have to replace 50% of the QC's on NSC's. Jeff

So I take it you HAVE had to adjust them;otherwise, why mention the Zantech timers? The NSC crystal problem is well known(ironically, it has the same bar style crystal as the Hughes, albiet at the lower, common speed. Mercifully, the crystals are easy to get and it is also the easiest to replace, bar none.

[Majestyk]

No I rarely have to adjust them, which is what I stated in my post. I run all my watches over my timers to see how much the lose or gain. Putting a watch on a timer doesn't mean your adjusting it.

The only thing the "bar style" QC's, on many of the NSC's, have in common with the Hughes are they are rectagular in shape. :) This is very fortunate, otherwise, I would be replacing as many Hughes as I have been replacing NSC's and NSC's are much easier to replace, thank goodness.

There are other rectangular 32mHz shaped QC's from that time period as well. Some are worse than the NSC's, some are better. I have found, aside from the Hughes, that the cylinder 32mHz were the best.

Jeff

[redled]

Does anyone know if a bad or loose crystal can result in a module appearing 'dead', or do they always just have the 'single digit' or 'frozen time' symptoms? I'm wondering if I can revive any of the many dead modules I have with just a crystal change! Thanks,

Steve

[Majestyk]

There is no set pattern. Sometimes it can be nothing displayed, sometimes one or more digits. One way to tell if it's the QC (although this is not set in stone) is to power it on and off rapidly. If you get random digits "dancing" around, it's almost always a bad QC.


Jeff

[retroleds]

No I rarely have to adjust them, which is what I stated in my post. I run all my watches over my timers to see how much the lose or gain. Putting a watch on a timer doesn't mean your adjusting it.

The only thing the "bar style" QC's, on many of the NSC's, have in common with the Hughes are they are rectagular in shape. :) This is very fortunate, otherwise, I would be replacing as many Hughes as I have been replacing NSC's and NSC's are much easier to replace, thank goodness.

Jeff: "Bar style" has to do with the internal construction - nothing to do with the exterior. Those little round can's are tuning fork design inside. Statek originated that construction and would actually tune them with a laser thru the glass top...you and I have discussed working on Litronix modules, Jeff, so I know you've seen their glass topped crystals. Hughes switched over to that internal design and speed eventually - they swtched the Dual LED/LCD module over to a two lead, 32,768 in January of 1978. Their early crystals were very heavy and expensive to make; it is fortunate that they were a good run.

Actually, your post said "the result of" - which sounds like the calibrators were put to use. Maybe you mispoke and meant, "checked on".
~Ed

[Majestyk]

Actually, your post said "the result of" - which sounds like the calibrators were put to use. Maybe you mispoke and meant, "checked on".

Understood, but that's not what I meant.

As for the Hughes and NSC crystals I thought you were associating them because of shape. :) I will maintain, that the crystals are nothing a like in reliability. The ones on the NSC's have a much shorter life span; the majority on the Huges are alive and kicking today. If I had kept all the dead NSC's crystals today, I would have a huge pile of them. I would only have a few bad Hughes QC's. Now there are two types of NSC crystals (possibly more). The worst offenders I'm referring to are the ones that are very skinny and have the very skinny, cheap wires that easily break. There is another one that looks more like the Hughes QC and a little more robust in construction. I have never personally noted which of these two NSC crystals are more reliable but from now on I will for curiosity sake. They used far more of the cheaper variety so the "Hughes style" one was probably on their earlier modules.

I don't recall discussing the workings of Litronix modules with you...I did ask you if you had one to sell, however...Those modules I won't even bother repairing. :)

M

[retroleds]

Yeah, Litronix really suck to work on...could they have used more clips or what HUghes is a good module...I think the NAtional Semi got a bad rap because some bum crystals. I've replaced a lot of those too - and then they are right as rain(cool when you discover you have a dual time zone or a military time one in your hand).

Man, nobodys asked about my new avatar...it was a customers but it was just so darn blingy I couldn't resist. :P