A tale of two meters

[jitter]

Now you're assuming that a revision change only incorporates one component, but often there are multiple. I have also seen that not all revision changes actually make it into production. And sometimes a change in firmare is all it takes to change revision, so unless you keep track of all the changes, there's no way to know for sure what changed when (and why).

[joeqsmith]

Now you're assuming that a revision change only incorporates one component, but often there are multiple. I have also seen that not all revision changes actually make it into production. And sometimes a change in firmare is all it takes to change revision, so unless you keep track of all the changes, there's no way to know for sure what changed when (and why).

Assuming?  I have the two products sitting side by side and am not concerned with taking my chances with the self tappers.    Your comment of "bodges" is why I directed my comment to the diode.   I guess the axial leaded trim resistors are somewhat of a bodge but both meters use this technique and the board does not appear to be revised to improve their bodgeability.   

Obviously changes can require revisions to the board.      I would assume any changes they made were for profit.  Then again we do have Dilbert. 

Your first EX540 has revision 5 on the pcb and your second is already up to 9.
To me it seems most likely that the "bodges" that were put on the "5" to solve some issues found their way into the layout of the subsequent versions.

It's back together again, and unfortunately I'm not going to open it again, I don't like my chances with the self tappers of this meter. Of just three total one is already on the way out.

[jitter]

Now you're assuming that a revision change only incorporates one component, but often there are multiple. I have also seen that not all revision changes actually make it into production. And sometimes a change in firmare is all it takes to change revision, so unless you keep track of all the changes, there's no way to know for sure what changed when (and why).

Assuming?  I have the two products sitting side by side and am not concerned with taking my chances with the self tappers.   

Please don't. As you don't have board revisions 6, 7 or 8, you have no way of finding out what modifications/bodges went into the layout and in which revision.

Your comment of "bodges" is why I directed my comment to the diode.

"Bodge" would not be the technical term used by a manufacturer, "fix" or "modification" sounds a bit better... 

I guess the axial leaded trim resistors are somewhat of a bodge but both meters use this technique and the board does not appear to be revised to improve their bodgeability.   

Obviously changes can require revisions to the board.      I would assume any changes they made were for profit.  Then again we do have Dilbert. 

Any bodge that needs to be put in by hand is way more expensive than when a PNP machine can do that in a correct layout. So yes, cost is a big factor in that, but may only be done if the cost of changing the layout outweighs the cost of modifying by hand. In large production runs, that will be the case.
We also see that some manufacturers change a layout whatever the cost. Sometimes they get to deal with customers that tear down an expensive piece of equipment (sound familiar?) and ask questions why there's a bodge on a board. This won't be your typical consumer, but in industrial electronics, it happens, and protecting a good name can be all that's necessary...

[joeqsmith]

Don't?  Don't what? 

You are correct, I have two meters rev 5 & 9.  That's perceptive!     The bodge on the old meter that was corrected (or forgotten) was the tacked on diode.   Indeed I have no need or desire to see the other revisions.  This meter was bad enough. 

I did not choose the wording.  I do not know if fixes drove the revisions.  Some of it may have been part obsolescence but I don't refer to those as bodges or fixes.     

I wonder if the new behavior seen in next post is a result of a bodge.   

Reputation, manual rework, everything boils down to making a profit.  Well, if a company want's to remain profitable.     

Please don't. As you don't have board revisions 6, 7 or 8, you have no way of finding out what modifications/bodges went into the layout and in which revision.

....

"Bodge" would not be the technical term used by a manufacturer, "fix" or "modification" sounds a bit better... 

Any bodge that needs to be put in by hand is way more expensive than when a PNP machine can do that in a correct layout. So yes, cost is a big factor in that, but may only be done if the cost of changing the layout outweighs the cost of modifying by hand. In large production runs, that will be the case.
We also see that some manufacturers change a layout whatever the cost. Sometimes they get to deal with customers that tear down an expensive piece of equipment (sound familiar?) and ask questions why there's a bodge on a board. This won't be your typical consumer, but in industrial electronics, it happens, and protecting a good name can be all that's necessary...

[joeqsmith]

After the new meter was reassembled,  I noticed it behaves strange in the CV mode.  I made a short video showing the difference. 

[jitter]

Don't?  Don't what? 

Don't take them apart again. If the self tappers are as bad as on my EX830, then the next disassembly might ruin what's left of the thread in the plastic.

You are correct, I have two meters rev 5 & 9.  That's perceptive!     The bodge on the old meter that was corrected (or forgotten) was the tacked on diode.   Indeed I have no need or desire to see the other revisions.  This meter was bad enough. 

I did not choose the wording.  I do not know if fixes drove the revisions.  Some of it may have been part obsolescence but I don't refer to those as bodges or fixes.     

Obsolecence seems unlikely for stuff like resistors, capacitors, or generic diodes.

I wonder if the new behavior seen in next post is a result of a bodge.   

Can you check the link? It doesn't work for me.
Edit: I got to it through the link in your signature:

Reputation, manual rework, everything boils down to making a profit.  Well, if a company want's to remain profitable.     

Yes, indeed.

[jitter]

It's not the exact same behaviour, but my recently calibrated Agilent U1232A doesn't go all the way down to 0 V either in the AC range. Even with the input shorted it stays at 1.9 mV.
The EX830 will drop down to 0.6 mV when the input is shorted.

[timofonic]

Soon it will be safer to design your own DMM than buying one.

Those will be great times for Open Hardware

[joeqsmith]

   That was a link to the edited video..  Thanks for adding the correct link.    That meter is all over the place and that 20ish seconds seems timed by software.  Why is samples and stays at zero, remove the 50 ohm and it drift up even higher than the 1mV, then reinstall and it has no effect.     Maybe this is by design.  Almost wish I had another one that was blessed at the factory as a golden unit. 

You know, I have never had much of a problem with self tapping screws.   They are used everywhere.   I just turn them backwards to make sure they drive into the threads the same.   

Hey Timofonic, I will leave that one up to Dave for now.  I am very interested in seeing how the fat boy, which should be the new standard to judge all meters by, compares with the rest of the pack. 

Soon it will be safer to design your own DMM than buying one.

Those will be great times for Open Hardware

[Cristiano]

Let's assume there were 5 revisions, or 4 after my original meter.   Now lets look at what changed with the PCB.   The plating.  There's one rev.   That gives us three revs to fix the tacked on diode bodge.   

I'm sure there were other changes.

Your first EX540 has revision 5 on the pcb and your second is already up to 9.
To me it seems most likely that the "bodges" that were put on the "5" to solve some issues found their way into the layout of the subsequent versions.

Hi Joe;

     My EX540 also have a big drift in Vca and mV!!!! I didn't have time of test very well this multimeter, but when saw your video I checked the my.
   

videos:
Vac - https://www.dropbox.com/s/rc43stuxq34myva/V_20160615_204907.mp4?dl=0

mV - https://www.dropbox.com/s/ck28hlfkq095tms/V_20160615_205940.mp4?dl=0

   I sent email and videos for EXTECH yesterday. I am waiting response....


Thanks

[joeqsmith]

Thanks! So that delay I see prior to the ACV starting to drift happens on multiple meters.   Mine is a A140 SN. 

After watching your videos with the leads removed, I tried one more test and then stumbled onto another difference between the two meters.    This would be a very easy test for you to try as well and may help Extech solve their problem if they decide it is worth going after.   

Hi Joe;

     My EX540 also have a big drift in Vca and mV!!!! I didn't have time of test very well this multimeter, but when saw your video I checked the my.
   

videos:
Vac - https://www.dropbox.com/s/rc43stuxq34myva/V_20160615_204907.mp4?dl=0

mV - https://www.dropbox.com/s/ck28hlfkq095tms/V_20160615_205940.mp4?dl=0

   I sent email and videos for EXTECH yesterday. I am waiting response....

Thanks

[joeqsmith]

Pulled the can to have a look and straightened it out a bit.  Fits nice.  Can't see why it was so mangled to begin with.

IC3 uses the TI logo and is a TL052.  IC9 is an AD6363 but does not appear to be an Analog Devices part.  It has a CJ logo.   Nice how the large device, IC2, is not marked.   

I'll spend a little time and see if I can sort out what the problem is.

 

[jitter]

No markings on an IC, as in "it was never there, not even sanded off"?

Funny, there are two positions called R62. Both look to be hand soldered. Perhaps placed during testing to pull some spec within tolerances?

[joeqsmith]

More interesting is that the Analog Devices data sheet shows a minimum of 5V for single supply operation.  I am seeing 2.738.   Looks like I need to have a look at the working meter.   

[joeqsmith]

Spent a little time with the AD636 datasheet.  I did not realize is used a split supply.  Rechecking and its 5.460.

I removed the can from the old meter.  Noticed C22 polarity was not the same.  Powering up the meters, it appears like it was placed backwards on the old meter but they corrected it  on the new one.   The older meter's RMS converter was a Cyrustek ES636 rather than the CJ part used on the new one.  IC2 was marked on the old meter.  It is a Cyrustek ES51966P.   Datasheet is on-line for it.

There appear to be some changes to the artwork and design in the analog area.   A transistor was added, some pads for a resistor that is not populated....    No schematic and may need to trace it out.

However, looking at the AD636 pin 6 relative to pin 10 with the inputs of the meter shorted, it will begin to drift after some time.    I strapped pin 1 to 10, and it will still drift.   Looking at the old meter between 6 & 10 with the inputs to the meter shorted, it remains steady.    So at least the problem appears to be with the RMS conversion stage.  I may start with this CJ part.  Looks like there are a few differences between the Cyrustek and the Analog Devices part.   

[joeqsmith]

Shown with C22 polarity change between two meters.   Again, the new meter is correct.   

[joeqsmith]

While there are many changes to the design, after tracing out the RMS stage, it appears very similar if not identical.   To be honest, I am not sure what they are doing after reading the data sheet.   I'm to the point I would like to swap out that CJ part but can't find an a source.   AD does not appear to make one in an SOIC.  It really would be a bodge.   

[retiredcaps]

I'm to the point I would like to swap out that CJ part but can't find an a source.

$16.21 from aliexpress and probably 2 months to ship.  My aliexpress orders are taking almost 2 months to arrive to Canada.

http://www.aliexpress.com/item/spot-IC-CJ-AD636-IGTA1JD09EL-special-Electronic/32642668777.html

[joeqsmith]

Thanks but I actually want the ES636 from Cyrustek that is on my original meter.  I may try and find the 16-pin SOIC version. 

I'm to the point I would like to swap out that CJ part but can't find an a source.

$16.21 from aliexpress and probably 2 months to ship.  My aliexpress orders are taking almost 2 months to arrive to Canada.

http://www.aliexpress.com/item/spot-IC-CJ-AD636-IGTA1JD09EL-special-Electronic/32642668777.html

[joeqsmith]

[joeqsmith]

Warning: this topic has not been posted in for at least 120 days.
Unless you're sure you want to reply, please consider starting a new topic.

Yea, I'm sure....

So after all this time, I drop in a new RMS converter into the brand new EX540 and aligned it.  As much as I would like to report that the meter now works as good as my original, sadly it does not.  If you own a later SN and were planning to try the swap, don't waste your time.  There is still one or more other changes that are causing some of the problems I am seeing with it. 
   

[joeqsmith]

Someone was asking about a twin display meter and I mentioned the EX540 having an RF link.

https://www.eevblog.com/forum/testgear/twin-remote-display-multimeter/

I thought, I should try that out and make sure it really works with Windows 10.  I plug in the adapter for the first time and the PC did nothing.   I found drivers for Windows 10, installed them.  Tried the adapter again, nothing.   

Look at the adapter "brand new" purchased from a major distributor and I notice two of the three screws holding the adapter together are missing!   I take the third one out and pop the cover.  Thread marks where the screws had been inserted at one time.  Would be fun to know the history of this meter. 

Looking at the adapter, first someone had really mangled where the wires attached to the PCB.  Someone had twisted things enough to fray much of the wiring.  I cleaned all this up and applied new hot glue.   Looking on the top side, noticed one joint for one of the ICs had not wicked to the lead.   Cleaned up some of the soldering and tried it out. 

The PC now finds the adapter.  Run their software, turn on the meter and enable the link.  Nothing.....

Decide I better just start with my old meter and adapter that I know works.  Sure enough, everything still works.  I then turn on my new meter with the old adapter.  Nothing.    Old meter with new adapter, works.     EX540 comes back apart....

Remember how the new meter was missing the shield on the back cover and I cut a new on out of some foil?  Well there were some leads on the RF board that made contact with the foil because they had not trimmed them.  Is this why they did not install the shield?  Trim the leads, double check the soldering and tried it out.  Both meters now work with both adapters.   

Was the meter sold, the RF did not work and an unskilled person attempted to repair it?  Was it just returned and the distributor attempted to repair it?  CEM knowing shipped it defective with missing screws?   

Crazy how different these two meters have been...

[3nigm4]

Just a question about graph bar, joeqsmith: it is really updating only with the value and not faster?

[joeqsmith]

Just a question about graph bar, joeqsmith: it is really updating only with the value and not faster?

The EX540 has the worse bar graph of any meter I have looked at.  Watch where I compare it with the TPI 194II, the  Brymen BM869s and the UNI-T UT181A.       

https://youtu.be/fkf0V3Xvq_w?t=1307

[3nigm4]

Definitely unusable 
I see an offer for a CEM DT-9939 for 160€ shipped... but I think is better to wait a bit.
Seems that the UT181A has all I need, except price