[SOLVED] Cheep eBay tachometer mod

[randall78]

High all,
So I work as a mechanical engineer for a manufacturing company here in the UK and have treated myself to a new contact/non-contact tachometer. When it arrived I was pleasantly surprised as it was a nice looking unit, easy to use and worked well. However, when I checked it's accuracy I found it to be reading twice the speed that it should be. The seller is going to send me a new one by the way. But I decided "TAKE IT APART" and have a play with it. After some fiddling around I found that by changing the crystal oscillator to one of a different value I was able to get it to display almost the correct speed, but not quite. There is no calibration settings by the way. I was wondering if I could just bodge a small potentiometer in there somewhere to allow me to fine tune it as there is no calibration setting. The original crystal was 4Mhz. By replacing it with a 6 Mhz crystal it increased its accuracy. If bodging a small potentiometer in there somewhere is even possible, I would not know where to put it as my electrical knowledge is somewhat limited. Just thought it might be cool to modify it instead of just throwing it away. I have included some detailed images of the PCB if anyone has any ideas.
Sorry, just noticed after posting that I have labeled two ic's the wrong way round.

[uncle_bob]

High all,
So I work as a mechanical engineer for a manufacturing company here in the UK and have treated myself to a new contact/non-contact tachometer. When it arrived I was pleasantly surprised as it was a nice looking unit, easy to use and worked well. However, when I checked it's accuracy I found it to be reading twice the speed that it should be. The seller is going to send me a new one by the way. But I decided "TAKE IT APART" and have a play with it. After some fiddling around I found that by changing the crystal oscillator to one of a different value I was able to get it to display almost the correct speed, but not quite. There is no calibration settings by the way. I was wondering if I could just bodge a small potentiometer in there somewhere to allow me to fine tune it as there is no calibration setting. The original crystal was 4Mhz. By replacing it with a 6 Mhz crystal it increased its accuracy. If bodging a small potentiometer in there somewhere is even possible, I would not know where to put it as my electrical knowledge is somewhat limited. Just thought it might be cool to modify it instead of just throwing it away. I have included some detailed images of the PCB if anyone has any ideas.
Sorry, just noticed after posting that I have labeled two ic's the wrong way round.

Hi

A non-contact tach can work a number of ways. The most common is to pick up reflected light. That has some obvious problems since the  assumption is that the gizmo only has one shiny surface and the rest of it is dead black. The same thing applies to a contact approach. If the shaft (or whatever) is not quite round, it may contact more than once per revolution.

Bottom line: It's very normal for these gizmos to read some integer number times the actual RPM.

Bob

[randall78]

Thanks for the reply.
I understand how a tach works. The one I have has a laser diode for non contact and an attachment with a rotating wheel with a tiny reflective segment that screws on to where the laser is.
My point is. I have managed to increase it's accuracy by changing the crystal but need to add a potentiometer to fine tune it. If that is even possible.

[Fraser]

Stop for a minute and think about what has happened here. You have a tachometer that reads wildly wrong. Previous tear downs of cheap Chinese optical tachometers has shown that they incorporate used components in the build. Used laser diodes, used processors and used detectors are all common practice.

It is interesting that changing the crustal has corrected the situation to some extent. Your unit may have had a used crystal in it and they picked one with the wrong frequency ? A crystal can only be trimmed a small amount using a  variable capacitor across it, or in some cases in series with it. Such trimming is really very fine tuning only as the crystal cannot be pulled off frequency very far.

I suspect these units use a fixed divider in the processing so there will be little opportunity to change the calibration. Such low cost devices are unlikely to be individually calibrated. Finding a crystal that provides the most accurate reading may be your best option. Not all crystals are the same, even when marked with the same frequency. The individual crystal type specs will dictate the actual oscillation frequency in your unit.

There is a teardown of a cheap laser tachometer on this forum somewhere.

It would be interesting to see the crystal frequency used in your replacement unit. It is likely they have chosen a commonly available and cheap crystal that is found on the used part marketplace in China.

Fraser

[danadak]

A crystal by definition has very well controlled characteristics, and
cannot be "pulled" by a significant amount. << 1% is all you could
expect. There are programmable clock oscillators you could consider.

But earlier comments, re unit has issues, probably should be addressed
by warranty.

Regards, Dana.

[uncle_bob]

Thanks for the reply.
I understand how a tach works. The one I have has a laser diode for non contact and an attachment with a rotating wheel with a tiny reflective segment that screws on to where the laser is.
My point is. I have managed to increase it's accuracy by changing the crystal but need to add a potentiometer to fine tune it. If that is even possible.

Hi

A really bad crystal from a true bottom of he barrel supplier *might* be off frequency by 0.01%. These days, even that is a maybe. Having on e off frequency by 2:1 .... simply not going to happen. The most likely reason for a 2:1 speed error is a double reflection or some other optical error.

If you need very accurate numbers, replacing the crystal with a TCXO is probably your best bet. Without a lot of hassle you should be able to find one that is good to less than 1 ppm. Some may have as much as a 10 ppm adjustment range.

Bob

[randall78]

[SOLVED] Thanks for all the replies. I have managed to sort it. When I discovered that by changing the crystal the accuracy was improved I only had a limited number available. 4, 6 and 7.3728MHZ were the only ones I had. Using the 7.3728MHZ got it fairly close. I have since changed it for an 8Mhz and it's now bang on accurate. In conclusion, either the wrong crystal was used to start with, or there was in error in the math used to program it. I haven't received the replacement yet but will be interesting to see if there's any difference from the original unit.