Author Topic: 100000 counts DMM frontend I.C identification  (Read 890 times)

0 Members and 1 Guest are viewing this topic.

Online shakalnokturn

  • Frequent Contributor
  • **
  • Posts: 606
  • Country: fr
100000 counts DMM frontend I.C identification
« on: January 28, 2019, 03:31:04 pm »
I've been in a MTX3281 graphic handheld 100k counts DMM.
The frontend/ADC has had the reference sanded down, all that's left is an internet URL that is typical of Fortune Semiconductor Taiwan (http://www.fsc.com.tw now https://www.ic-fortune.com/eng/).
I've looked for datasheets on the website and on google but nothing shows-up above 50k points, there aren't that many 100k point DMM's out there, so maybe other manufacturers have used the same component.

Any clue to the actual part number or what other DMM's it may be found in?
 

Offline elecman14

  • Supporter
  • ****
  • Posts: 156
  • Country: us
Re: 100000 counts DMM frontend I.C identification
« Reply #1 on: January 28, 2019, 03:37:38 pm »
Not sure on the part number. You could try putting some isopropyl alcohol on the part to see if the other chip marking come through when the part is "wet". That trick sometimes works for sanded parts. 
 
The following users thanked this post: dzseki

Offline RoGeorge

  • Super Contributor
  • ***
  • Posts: 1717
  • Country: ro
Re: 100000 counts DMM frontend I.C identification
« Reply #2 on: January 28, 2019, 03:51:12 pm »
I have an MTX3283, also 100 000 counts.  If you suspect the ADC is only 50 000 counts, then that might explain why the DMM is relatively slow.  They are probably making four measurements, then average the result in order to gain one extra bit, thus doubling the max 50 000 counts allowed by the ADC.  This is just my own speculation, no idea if it is true.

I have another question:  How did yo manage to open it?  When I bought mine I wanted to take a look inside (it was a second-hand buy), but give up after finding no way to open the flipping half.  I was not sure if the case was glued by the former owner, or I just didn't spot all the screws.  So, how did you get to the PCB?  Is it necessary to remove the half with the display, too?  So far, couldn't find the service manual or the schematic for MTX.

I'd be grateful if you can attach a few more pictures with the dismantled instrument, so I can figure out how to open mine.

Online shakalnokturn

  • Frequent Contributor
  • **
  • Posts: 606
  • Country: fr
Re: 100000 counts DMM frontend I.C identification
« Reply #3 on: January 28, 2019, 06:34:36 pm »
No wetting the I.C. didn't help in reading the part number.

Well... Yes, until I have proof Fortune Semiconductor have actually manufactured a 100k counts I.C., I'm going to have to suspect they are doing some interpolation and averaging.

I don't mind the whole idea of this DMM but some of the design is just deceiving compared to my favourite older Metrix DMM's. The ribbon to the LCD can only be disconnected at the
LCD side, the battery compartment not sealed enough, the whole mechanical assembly of the thing just feels clumsy to me.

Opening the thing isn't that easy.
Remove the battery cover by unclipping it at the hinge caps.
Remove 4 small posidriv screws, start opening from tapered end, I'm not sure if it is easier to disassemble the lower half of the shell by sliding it away from the hinge (2 small notches on the front side of the hinge cover keep it locked there, there are also 4 small plastic pins that interlock on the rear side of the hinge.) or if it is easier to remove the hinge axis first.

Anyway for that part of the disassembly I was inspired by this: http://www.mjlorton.com/forum/index.php?action=dlattach;topic=107.0;attach=266

The main PCB is held in place near the middle by the 2 screws at the bottom of the shielding.

Then the LCD is a real ass to open when you don't know how it's built, 2 small T6 screws hidden on hinge side, then the frame is clipped onto the LCD polycarbonate protection so it needs spudging off from the screen outwards.

Here's my full disassembly gallery, I'm sure it will be of some help.

https://photos.app.goo.gl/uCUmZvgM6NLxzHAJ7
 
The following users thanked this post: RoGeorge

Offline tsman

  • Frequent Contributor
  • **
  • Posts: 599
  • Country: gb
Re: 100000 counts DMM frontend I.C identification
« Reply #4 on: January 28, 2019, 06:42:45 pm »
My guess is that it is a 50K count chip as well. The pinout matches up with the 50K count FS9704B pinout in the datasheet. The other variant of the FS970X listed is only 5K count so it can't be that.
 

Offline RoGeorge

  • Super Contributor
  • ***
  • Posts: 1717
  • Country: ro
Re: 100000 counts DMM frontend I.C identification
« Reply #5 on: January 28, 2019, 09:47:00 pm »
...
http://www.mjlorton.com/forum/index.php?action=dlattach;topic=107.0;attach=266
...
Here's my full disassembly gallery, I'm sure it will be of some help.
https://photos.app.goo.gl/uCUmZvgM6NLxzHAJ7

Wow, very helpful, thank you!   :-DMM

Last time I couldn't open the plastic covering caps of the hinge.  Either those caps are very stiff, or they were simply glued by the former owner(s).  Will give it another try, and upload pics if I manage to open the DMM. 

Online shakalnokturn

  • Frequent Contributor
  • **
  • Posts: 606
  • Country: fr
Re: 100000 counts DMM frontend I.C identification
« Reply #6 on: January 28, 2019, 11:05:54 pm »
Quote
Wow, very helpful, thank you!   :-DMM

Last time I couldn't open the plastic covering caps of the hinge.  Either those caps are very stiff, or they were simply glued by the former owner(s).  Will give it another try, and upload pics if I manage to open the DMM.

If you have access to the battery compartment it can't be glued, it has to swivel to give access. Then all you need to do is lightly bend one "hinge cover" end of that part of the shell (bat. compartment cover) away from the rest of the DMM.

The real annoying stuff is the LCD frame, as it does resist being opened quite a bit, it's tempting to give it some brute force but you still have the main board attached by a ribbon cable waiting to tear as soon as you slip  :-DMM |O
« Last Edit: January 28, 2019, 11:07:54 pm by shakalnokturn »
 

Online coromonadalix

  • Super Contributor
  • ***
  • Posts: 2062
  • Country: ca
Re: 100000 counts DMM frontend I.C identification
« Reply #7 on: January 29, 2019, 12:00:21 am »
Cyrustek or Fortune Semiconductor ??    fortune goes up to 50k

Cant find the list i had, it listed all the ic's vs display count ....

Uness it use the Brymen 857 867 chipsets who do 50k and 100k counts ??


Cyrustek  up to 60k
http://www.cyrustek.com.tw/product-2-5.htm


https://www.ic-fortune.com/eng/new_product3_3.asp


Gossen Metrahit 28-29s use 310k counts
Gossen Metrahit 30s is 1,200.000 counts via a 24 bit dac if i recall
« Last Edit: January 29, 2019, 12:09:43 am by coromonadalix »
 

Online shakalnokturn

  • Frequent Contributor
  • **
  • Posts: 606
  • Country: fr
Re: 100000 counts DMM frontend I.C identification
« Reply #8 on: January 29, 2019, 02:00:38 am »
Supposing that the "fsc" part in the URL left on the I.C. stands for "Fortune SemiConductor" and going by the LQFP64 package, I'd rule-out Cyrustek as they are all at least LQFP100 packed.

Thanks for bringing my attention to the Brymen 867 that supposedly does 100k counts.
Going by this topic: https://www.eevblog.com/forum/testgear/brymen-bm-867-teardown-pictures/
It uses the BTC AD-85-4, with no extra marking in this case.
But in this topic: https://www.eevblog.com/forum/repair/brymen-bm817-dmm-gone-wrong/
The same BTC AD-85-4 is used, only this time it bears the "fsc" URL...

It has been put forward in other topics on the forum that BTC AD-85-4 is no other than the FS9704B that is a 50k counts I.C. (not quite sure why the datasheet states 80k counts at one point though, maybe a typo...)
https://www.ic-fortune.com/upload/Download/FS970X-DS-44_EN.pdf

So I'd vote for this being used in the MTX meters too, I'd be curious to how the extra digit is gained though!

Gossen are a whole world above all this I doubt they use anything similar.
 

Online coromonadalix

  • Super Contributor
  • ***
  • Posts: 2062
  • Country: ca
Re: 100000 counts DMM frontend I.C identification
« Reply #9 on: January 29, 2019, 02:38:07 am »
Gossen use dedicated asic for their meters, but not all of them ...
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 6310
  • Country: de
Re: 100000 counts DMM frontend I.C identification
« Reply #10 on: January 29, 2019, 07:19:23 am »
The fortune web page lists a 18 Bit ADC chip in 64 pin package. So it might be using a separate µC and control.
 

Online shakalnokturn

  • Frequent Contributor
  • **
  • Posts: 606
  • Country: fr
Re: 100000 counts DMM frontend I.C identification
« Reply #11 on: January 29, 2019, 07:41:21 pm »
It does indeed use a separate µC, an M30626 in the case of the MTX3281 DMM.
The 18bit ADC you mention is the FS9704, AFAIK there is only that one tha will do 50k counts.
 

Online coromonadalix

  • Super Contributor
  • ***
  • Posts: 2062
  • Country: ca
Re: 100000 counts DMM frontend I.C identification
« Reply #12 on: January 30, 2019, 06:38:53 am »
Victor VC8145 has 80k counts display 
https://lygte-info.dk/review/DMMVici%20VC8145%20UK.html

Fortune semiconductor   FSUP01-003

https://www.ic-fortune.com/upload/Download/FS970X-DS-42_EN.pdf
Chart 1  page 5/47  gives : FS9704B  the 80,000 count chip

And a typical schematic page 10/47 ??
« Last Edit: January 30, 2019, 06:47:53 am by coromonadalix »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf