Longines VHP accuracy

I really don't know whats going on here. I was diligently counting to 14 and seem to have ended up with 21.
The French instructions don't mention impulses. They talk about one or more 'contacts'. Only the German and English talk about 'impulses'.

Wasn't your watch over-corrected at one time in the shop?

Are we having the 0.33 seconds wrong?

If I do my sums again, with only 2 days worth of operation to go by, it would make a lot more sense if 1 impulse equated to a 0.5 second correction.

Hans Moleman said:

...Wasn't your watch over-corrected at one time in the shop?

Are we having the 0.33 seconds wrong?

Click to expand...

Yes, it was "overcorrected" in the service centre.

The 0.33 second/month step must be right!

I think the source of our problem is the incorrect measurements of the rate of the movement before and /or after calibration. We might be guilty of rushing things by not giving enough time for proper check of the rate of the movement. Without properly calibrated professional tools we only can rely on our observations. I'd say we should give the movement(s) at least 1 month (2-3 months would be even better) before we attempt to establish its rate. Even if it was obvious within days that the movement was out of specification we would only know the current rate with reliable accuracy on the long term. Any other approach is a bit like "hit and miss". Once in a while, we can be lucky and hit "bullseye" even with a premature attempt but that would be the exemption to the rule, in my opinion.

Hans Moleman said:

I am impressed you've managed to make a correction with the borrowed hands!
There doesn't seem to be enough space, even for my own two hands...

Click to expand...

Which brings us to an other explanation (specially in an akward position like that) for the "overcalibration": the trembling hands effects...

My rate was established over an 80 day period... that said, who knows?

I'm just glad I don't have to send it back somewhere to try again! :-d

Its very early to draw conclusions, I agree.

What I know so far is that my watch used to gain a second every 6.5 days. And that is measured over a few months.

It now loses seconds. Not too sure how many exactly, but its getting closer.
:-!

I never expected this to be the full and final adjustment. I am happy enough to know that I can adjust it!
We'll get there eventually. I won't tighten the back as yet.

Cheers
Hans

As long as the attempts converge, I too will be quite happy!

Hans Moleman said:

...I never expected this to be the full and final adjustment. I am happy enough to know that I can adjust it!
We'll get there eventually...

Click to expand...

I fully agree!:-!

Does your attempt converge Eeeb?
For what I can see I went from +4.6 to -3.5 per month.

That's 10 corrections too many out of the 24. I really should cut down on the coffee!
I'll have another go after a week.
If I can't get the applied corrections accurate, then there's no point in getting the numbers of required corrections accurate either. I might do a few for practice sake.

Even when "inaccurate", the VHP is way too accurate to get an good timing in several days :-d ... I'll have to wait a several weeks to get a good adjustment count.

Eeeb said:

Even when "inaccurate", the VHP is way too accurate to get an good timing in several days :-d ... I'll have to wait a several weeks to get a good adjustment count.

Click to expand...

It's true!:-d

I am now getting good consistent measurements with recording a video of my watch together with a reference clock.

Well, I did some more corrections and I can confirm 'tapping' on the correction patches is the way to go. As indeed is outlined in post 41 by Ppaulusz.

Hold the correction wire too long on the patch, and you run the risk of registering more than one correction impulse. That is for me the only explanation as to why I over-corrected in my first attempt.

I tapped the 'faster' patch 10 times, and the watch is now within the 0.8 sec per month spec. It is going so well, that my video recordings give no early indication of the yearly rate anymore. Apart from that it's a lot better.

I would like to see one more step after step 4. 'Wait for the date ring to complete its turn.'

heqphoria here I come!
b-)

Hans Moleman said:

Well, I did some more corrections and I can confirm 'tapping' on the correction patches is the way to go. As indeed is outlined in post 41 by Ppaulusz.

Hold the correction wire too long on the patch, and you run the risk of registering more than one correction impulse. That is for me the only explanation as to why I over-corrected in my first attempt.

I tapped the 'faster' patch 10 times, and the watch is now within the 0.8 sec per month spec. It is going so well, that my video recordings give no early indication of the yearly rate anymore. Apart from that it's a lot better.

I would like to see one more step after step 4. 'Wait for the date ring to complete its turn.'...

Click to expand...

Great progress, Hans!:-!
I will include that extra info at the end of step 4 in post #41 as an update. Thanks for pointing it out!

Hans Moleman said:

Getting closer...
I figured it could do with another single correction.

After three weeks of measurements, and realizing that I was dealing with some serious measurement noise here, Mr S.Tochastic advised me that the rate was 0.338 seconds per month slow.
One correction impulse of 0.33 sec/month would fit nicely here...

Click to expand...

You are almost there, Hans!|>

Excellent work, Hans! :-! You may achieve the elusive <1 sec/year mark after your next iteration.

This way of collecting and displaying the data may make summer to winter temperature dependencies clear, too, if they exist.

It's just that I can't stand waiting for a month or three to get my one second difference.
And then find out that I've made another mistake somewhere and have to to it all over again.

Would be interesting to see what would be a good method for rate calculation without a laser and two atom clocks. Start a thread for that?

Hans Moleman said:

Would be interesting to see what would be a good method for rate calculation without a laser and two atom clocks. Start a thread for that?

Click to expand...

I say go for it.