Chinese Timegraphers

[Colditz]

JDM

You know it is what I thought too. I have read numerous posts with contradicting information. Generally where there is contradiction there is falsehood and the statement that 1000 and 1900 can support coaxial escapements just doesn't ring true. The post by an inactive member with links that do not work will make me suspicious of a seller of these TGs trying to hijack the site to get us to buy their wares.

If anybody out there can prove us wrong and show it with video footage and setting detail then do so please.

[m1ks]

I have read in a few places that the 1900 will read co axial escapements.

If anyone would care to send me a George Daniels or Omega I'll gladly check that and video it for you all to see. 

[gary17]

Hi

Wife's bought me a weishi timer for birthday. Anyone no we're there is any info on how you use it and what each number means and how to read what result means. 

Cheers

Gary

 

[jdm]

Manual attached. If you take the time to read some books, or even good postings here you will understand the  terms involved: daily rate, amplitude, beat error, lift angle.

6340_T38209.pdf

[TheFixer]

Hope this helps a little.

https://www.watchtime.com/reference-center/glossary/amplitude/

We are hear to help so please ask.

A timegrapher is one of the best tools you can have imho.

[B42Woody]

Has anybody tried to interface one of these Chinese timegraphers with a PC?  I can see value in the ability to record data.  I don't think there's need to control the machine from a PC, though.  Thoughts and opinions?  Links?

Woody

[jdm]

43 minutes ago, B42Woody said:

Has anybody tried to interface one of these Chinese timegraphers with a PC?  I can see value in the ability to record data.  I don't think there's need to control the machine from a PC, though.  Thoughts and opinions?  Links?

The Chinese machine cannot be connected to a computer, even the more expensive models (not even covered in this topic), only support connection to a printer. If you are interested in using a PC there are varisou application available but you will have to provide yourown microphone.

BTW, we have a dedicated section on this forum where is considered polite for new members to introduce themselves.

[toptime810]

I have one and it works perfect

[B42Woody]

9 minutes ago, jdm said:

The Chinese machine cannot be connected to a computer, even the more expensive models (not even covered in this topic), only support connection to a printer. ....

Thanks for the reply, jdm.

I'm not asking about "connecting" one of these devices to a PC - I see the manufacturers don't offer that option.  Rather, I'm asking if anybody has opened the enclosure and interfaced with the electronics to grab data signals.  My colleague is a wiz with electronics but if somebody has already done it and is willing to share the details, that would be great.

Woody

[jdm]

1 minute ago, B42Woody said:

I'm asking if anybody has opened the enclosure and interfaced with the electronics to grab data signals. 

Personally I think that data loggin is of small importance. beside being a regulation/adjustment help, a timegrapher has basically two functions, first a good / no good indication in different positions, and at different times during the expected power reserve. Scond, in case of "no good", it give some cues about what may be the cause.
Anyway, not even a schematics is available. I would be curious to know the answer you get once present the challenge to your talented friend.

[JohnR725]

17 minutes ago, B42Woody said:

Rather, I'm asking if anybody has opened the enclosure and interfaced with the electronics to grab data signals. 

You don't need the signals out of the timing machine you just need the microphone. Then somewhere in the discussion group we've already discussed that as there are software-based timing machines that can make use of that single.

[B42Woody]

1 minute ago, JohnR725 said:

You don't need the signals out of the timing machine you just need the microphone. Then somewhere in the discussion group we've already discussed that as there are software-based timing machines that can make use of that single.

Thanks, John.  That makes a lot more sense ....

Woody

[jdm]

1 hour ago, B42Woody said:

That makes a lot more sense ....

Which is what I also wrote above:

If you are interested in using a PC there are various applications available but you will have to provide your own microphone.

[watchweasol]

Hi  If you are interestend in PC based timing systems There are a few, Delph electronics do some software, There is tg software and Watch-o-scope which I have used  and find very good giving plenty of information, The only snag is you have to build the michrophone and pick up.  Details are provided along with instructions on the operation of the software. there is a lite version to test if you like it, buying the one off licence opens more features last count licence was $40.

[jdm]

16 minutes ago, watchweasol said:

The only snag is you have to build the michrophone and pick up.

If going for a quality tool the practicality of a ready, articulated, amplified microphone is a must. It can be bought for a little less than the price of a model 1000, and the below details the building on an adapter cable to a PC.

 

[spectre6000]

On 2/27/2020 at 10:12 AM, B42Woody said:

Thanks for the reply, jdm.

I'm not asking about "connecting" one of these devices to a PC - I see the manufacturers don't offer that option.  Rather, I'm asking if anybody has opened the enclosure and interfaced with the electronics to grab data signals.  My colleague is a wiz with electronics but if somebody has already done it and is willing to share the details, that would be great.

Woody

It seems like it might be worth tapping off the microphone input for a daughter board to output the actual audio rather than the beep... Depending on how they're dealing with the microphone's signal in the first place, the daughter board might not even be necessary... Just noodling. 

[JohnR725]

16 minutes ago, spectre6000 said:

It seems like it might be worth tapping off the microphone input for a daughter board to output the actual audio rather than the beep... Depending on how they're dealing with the microphone's signal in the first place, the daughter board might not even be necessary

I'm attaching a photograph of the inside of the Chinese 1000 machine. What you would have to do is get an on oscilloscope look around for the signal and probably just a little audio amplifier with volume adjustment. It's another one of those projects I thought about doing some day if I had more free time.

Then I remember seeing on a discussion group or someone took one of the other Chinese machines apart and I thought they did something with the oscilloscope. So doing a search stumble across the first link at the same discussion group and the second plank of the other machine. I thought there was a little more our oscilloscope use.

Then the other machine is made by somebody else in China. Not as desirable to have because it has limited ranges. The 1000 and the 1900 they paid attention to what witschi as far as the frequencies and ranges and sort of copy that. Versus the other company that came first it had much more limited ranges.

https://www.eevblog.com/forum/reviews/weishi-no-1900-multifunction-timegrapher-teardown/

https://www.eevblog.com/forum/reviews/weishi-mtg-3000-_multi-function-timegrapher_/msg124772/#msg124772

[spectre6000]

There's a preamp in the microphone itself. Depending on what it does once it gets to the primary board (none of the photos are high enough resolution or at the right angles to really see what's going on), it may be possible to tap directly off that right at the plug and pull the speaker out to just play what's coming from the microphone (assuming it's not doing some other critical thing). It's all conjecture without having one in hand, but that'll come soon enough.

[Vacherin]

Just a heads up. I ordered one of these watch timers from China via ebay, but it never arrived. The seller kept me waiting almost 2 months then I opened a case a couple of days within the timeframe to complain, and got an immediate refund from ebay. It looks like a reputable seller suddenly went rogue after 30K+ valid sales. The seller offered to post it again, but that was a stall tactic to get me out of the time period. This happens a lot apparently. In case anyone else is waiting for one, forget the warnings that COV-19 might delay your products. If you don't respond within the timeframe, shortly after expected delivery date ebay/paypal won't refund. I would have gone to a UK seller who has them in stock, but I wanted delivery abroad, and there are import implications, etc.

[quitquitquit]

Just a quick question. Maybe you will be able to tell me if my device is faulty or is it normal.
I bought 1900model and tried it on my watches. Very strange results. Basically there is zero consistency between my tests.
Watch is mounted, i am not touching it. I have let say results of +10sec and 280 amp. I run test for let say 5 minutes and it results areroughly the same. I pause device and start it again. Now it is +7 sec and 290 amp. OFF->ON - and now i can see +15sec and 260 amp. Everything else is the same, watch is in same position, all settings are same, gain on mic is same, room is quiet. Done this test literally 100 times on 7 different watches and machine behave still the same - I can't replicate results.
I just wonder if I have faulty device and I should replace this one, or "you get what you pay for" and another one will act the same ?


 

[JohnR725]

7 hours ago, quitquitquit said:

Just a quick question

not a quick question a confusing question?  Plus it would've been nice to give us a picture  just so we can see that it looks right..

7 hours ago, quitquitquit said:

Everything else is the same

then everything is not the same because this is a mechanical watch. The problem with mechanical watches are fluctuations in power through the gear train that causes fluctuations in timekeeping and amplitude. just the mechanical properties of gears cause power fluctuations and if the watch hasn't been serviced that's going to add to the problem. Like a sticky mainspring causes a slow fluctuation power over a long period of time.

The timing machine tends to be an instantaneous device. It's running internal calculations and averaging off of the waveform. so each time you start and stop the timing the internal averaging has to start over again and it's not averaging over the exact same part of the waveform.

then it would help to have a proper procedure for timing. if you look at the tech sheets for watches related to timing their recommendation is wind the watch up weight anywhere from 10 to 60 minutes then you can start timing. Also change the sampling time the machine defaults to two seconds change that to 20 seconds. That should average out the readings a little more smoothly. 

[spectre6000]

On 5/15/2020 at 12:56 PM, Vacherin said:

Just a heads up. I ordered one of these watch timers from China via ebay, but it never arrived. The seller kept me waiting almost 2 months then I opened a case a couple of days within the timeframe to complain, and got an immediate refund from ebay. It looks like a reputable seller suddenly went rogue after 30K+ valid sales. The seller offered to post it again, but that was a stall tactic to get me out of the time period. This happens a lot apparently. In case anyone else is waiting for one, forget the warnings that COV-19 might delay your products. If you don't respond within the timeframe, shortly after expected delivery date ebay/paypal won't refund. I would have gone to a UK seller who has them in stock, but I wanted delivery abroad, and there are import implications, etc.

I just ordered one from eBay last night. What seller? 

[spectre6000]

On 5/22/2020 at 6:01 AM, quitquitquit said:

Just a quick question. Maybe you will be able to tell me if my device is faulty or is it normal.
I bought 1900model and tried it on my watches. Very strange results. Basically there is zero consistency between my tests.
Watch is mounted, i am not touching it. I have let say results of +10sec and 280 amp. I run test for let say 5 minutes and it results areroughly the same. I pause device and start it again. Now it is +7 sec and 290 amp. OFF->ON - and now i can see +15sec and 260 amp. Everything else is the same, watch is in same position, all settings are same, gain on mic is same, room is quiet. Done this test literally 100 times on 7 different watches and machine behave still the same - I can't replicate results.
I just wonder if I have faulty device and I should replace this one, or "you get what you pay for" and another one will act the same ?
 

What if you take the watch off, restart, then remount? There may be a calibration routine running on start that's not obvious, and having a watch mounted isn't controlled for. I can easily imagine they cheaped out on the mic, stand, preamp, etc. such that it's noisy and there's some code running on start to compensate for that. A few lines of code is usually cheaper than a thousand upgraded components.

Edited May 23, 2020 by spectre6000

[JohnR725]

10 hours ago, spectre6000 said:

What if you take the watch off, restart, then remount? There may be a calibration routine running on start that's not obvious, and having a watch mounted isn't controlled for.

taking it on and off would add a new variation of having to wait at least 30 seconds for the watch to stabilize otherwise that's going to introduce a variation. the timing machine has a start/stop button. Each time the button is pushed it cycles between stopping which is freezing the display then starting a timing cycle starts over from the very beginning no need to remove the watch.

10 hours ago, spectre6000 said:

I can easily imagine they cheaped out on the mic, stand, preamp, etc. such that it's noisy and there's some code running on start to compensate for that. A few lines of code is usually cheaper than a thousand upgraded components.

that's an interesting observation which implies we would have to compare it to something? when the Chinese 1000 machines came out one of their selling points was the screen there were very proud of their backlit screen. At that time there comparisons were with the witschi watch expert II as it does not have a backlit screen. if we use their comparison then you're right the Swiss machine does have a nicer microphone considerably more more analog processing circuitry enhanced features plus a minor price difference.  okay perhaps it's not a minor price difference comparing a professional Swiss machine that  is just under $3000 to a Chinese machine that's under $200 is an interesting comparison isn't it?

then how does the Chinese  timing machines compared to the Swiss? I've already done that experiment couple years ago. I compared the witschi watch expert 2 with the Chinese 1000 and 1900.. for the most part they are very very close so close that I have no problems recommending either Chinese machine.. although I like the display better on the 1900 machine..

being curious about the start and stop problem I decided to try an experiment.. The benefit of collecting timing machines..  The witschi watch expert two  and the Chinese 1000 machine.. Basically the 1000 analog and programming for the most part is identical to the 1900 at least close enough. The watch that I was timing is a Hamilton 992B pocket watch. Mainly because it was convenient and it's a lot easier to have two microphones attached to a pocket watch..

basically the exact same set up I had before just a different watch. but there seems be a problem with the 1000 machine not correct technical term but it's not happy?  The LED is flickering  but not in nice rhythmic way.  the machine is having a problem figuring out  the beats per hour it's just taking a long time. Once it figured out the beats per hour the rate was fine but the amplitude took almost a minute and was wrong.. basically the machine is not getting a good clean signal?

this is a  problem more common with wristwatches in their case.. A lot of times of the case is big heavy cases sometimes the movement is plastic a variety of things can cause the signal not to get out.. This is where the recommendation is rotate the watch until the crown  is pressing against the part of the microphone where the sensor is located.. This should give you the best signal.. repositioning the pocket watch and the microphone the LEDs looking better everything is looking nice now. The witschi microphone is now clip to the side of the watch case it is working just fine. So there appears to be some minor signal problem with the pocket watch that the 1000 as having issues with the witschi is not. This is where in our oscilloscope feature is really nice missing from both of these machines.

now that both machines are simultaneously appearing to work let's look at the numbers and see the difference between them. The witschi indicates +2 seconds 226° 0.5 ms. The Chinese 1000 is +1 225° 0.5 ms so other than the initial microphone issue which I still think it's odd both machines are basically tracking very very close to each other.

I do curiosity I thought I would do a test where did I turn the machine off turn it back on and it's roughly 25 seconds to reestablish all the numbers then looking at the witschi machine it looks like everything is real close the amplitude is off by 5°. Some of this I think is because on power up it defaults to an averaging where sample time of four seconds and that's going to change things a little bit.

now for the final insanity part? Were not exactly comparing the same thing in that the witschi is averaging over 20 seconds updating every two seconds. Watching the display I noticing a amplitude variation anywhere between 229 up to 240 something this is a very slow variation over several minutes and it's not exactly rhythmic or consistent. as a guess this is what happens when you have several variations in issues seeing as how I know the watch hasn't been serviced in a while.

So with the Chinese machine I push the stop button right the numbers down the rates are very very consistent the same as the witschi machine so I'm going to ignore the rate for this test only look at the amplitude because that's the only thing that's showing any variation. So what I push the stop button I write down the number wait a moment push the start button once stabilizes right that number down this is why made a reference to insanity. Of and as a reminder the 1000 machine is set default sample time before seconds and we have no idea how the machine actually averages it would definitely be more stable if I would adjusted 20 seconds but the person who originally posted this never change the setting so I'll keep it where the machine would be for him.

So stopped 237° it starts up at 225° compared to the witschi at 237°. Then running the test again 225° starts up at 236 and the witschi's 237.

so the basic conclusion of all of this is the Chinese machines are relatively stable the microphone isn't quite as nice as the witschi the amplifier and so forth aren't quite as nice but it still doesn't really impressive job. For the original person posting we need more information? Like if you stare at the timing machine do the numbers change at all over several minutes? Are you seeing a slow variation. Because if there is a variation and you stop and start the machine up you going to see that where you had a number and it's going to jump to the variation number. If the graphical display is not a nice smooth line like mine is you're going to see a variation in the rate. Then if you change the sampling time to 20 seconds it will help to smooth things out.

 

[spectre6000]

Are you saying the Witschi and the Weishi are producing similarly variable results like Vacherin said he was seeing, but the Weishi might be slightly worse?