Author Topic: when do you really need 6 1/2 DMM ? , I mean the res and/or the accuracy  (Read 11540 times)

0 Members and 1 Guest are viewing this topic.

Offline eevblogfan

  • Frequent Contributor
  • **
  • Posts: 569
  • Country: 00
hey

nince I've got 2X HP3478A and 1X HP34401 , I thought , when do you guys really use that kind of res ?  or that kind of accuracy  ?

I had to work with 2uV Chopper amp , so low voltage and good accuracy is an issue , hence 5.5Digit is needed indeed  ' but apart from that , what els ?

please feel free to shear any experiance or experiment you've had with that such need , good night guys  :=\
 

Offline Fsck

  • Super Contributor
  • ***
  • Posts: 1157
  • Country: ca
  • sleep deprived
should let you guarantee that the first 3.5-4.5 digits are 100% accurate.
"This is a one line proof...if we start sufficiently far to the left."
 

Offline c4757p

  • Super Contributor
  • ***
  • Posts: 7805
  • Country: us
  • adieu
should let you guarantee that the first 3.5-4.5 digits are 100% accurate.

That's not how it works. For example, the 3478A specifies 0.035% + 40c on the 30mV range, which means if you are measuring 20mV, the meter can read up to 20.011 mV. 300mV range (more similar to the rest of the ranges) measuring 200mV can read up to 200.019 mV, which is only 3.5 "100% accurate" digits. (You did say "3.5-4.5", but you're pretty unlikely to get 4.5 perfect without a really good meter.)

And to contrast, the Fluke 8050A (only a 4.5 digit meter) still gives you 3.5 "perfectly accurate" digits, with a 1-year accuracy on all ranges of 0.03% + 2c.
« Last Edit: May 16, 2013, 10:45:16 pm by c4757p »
No longer active here - try the IRC channel if you just can't be without me :)
 

Offline Fsck

  • Super Contributor
  • ***
  • Posts: 1157
  • Country: ca
  • sleep deprived
should let you guarantee that the first 3.5-4.5 digits are 100% accurate.

That's not how it works.
maybe not 100%, since the confidence interval is probably 95%, but even if you jump to 7-8 sigma interval, and assuming the calibration is correct, the first couple digits will be guaranteed
"This is a one line proof...if we start sufficiently far to the left."
 

Offline vk6zgo

  • Super Contributor
  • ***
  • Posts: 4851
  • Country: au
It's so you can show it off to the other Geeks! ;D
 

Offline don.r

  • Frequent Contributor
  • **
  • Posts: 657
  • Country: ca
It's so you can show it off to the other Geeks! ;D
Digit envy? Lucky the man who owns a 3458A.
Shippers: for the love of Pete and all that is holy, STOP USING UPS INTERNATIONAL!
 

Offline ftransform

  • Frequent Contributor
  • **
  • Posts: 729
  • Country: 00
6.5 digit multimeter is the bare minimum for digital work. I recommend a 8.5 for analog.
 

Offline ben_r_

  • Frequent Contributor
  • **
  • Posts: 411
  • Country: us
  • A Real Nowhere Man
6.5 digit multimeter is the bare minimum for digital work. I recommend a 8.5 for analog.
Never heard anyone make that statement before. Okay, so WHY?
If at first you don't succeed, redefine success!
 

Offline eevblogfan

  • Frequent Contributor
  • **
  • Posts: 569
  • Country: 00
6.5 digit multimeter is the bare minimum for digital work. I recommend a 8.5 for analog.
Never heard anyone make that statement before. Okay, so WHY?

yea , tell us WHY ?  :wtf: :wtf: :rant:

8 1/2 is the world's bets res ( as far as I know ) , most things on the analogue need no more then 0.05%  ( fluke 87V will do ) , that is not justifying the cost of 6.5 DMM , not tallking about 8.5 !!!?

I wonder if someone needs 10.00000V reference , I might make one for myself ( I might put my hands on kiethely 2001 , so the adjustment of that 10V reference will be done just after adjusting the 2001 ) , i presume fluke 5700 has good enough accuracy , so anything will be referenced to him :P

please PM me if you are interest in such reference :P
 

Offline mrflibble

  • Super Contributor
  • ***
  • Posts: 1945
  • Country: nl
6.5 digit multimeter is the bare minimum for digital work.

Agreed!  :-+


(6.5 binary digits that is)
 

Offline AlfBaz

  • Super Contributor
  • ***
  • Posts: 2007
  • Country: au
6.5 digit multimeter is the bare minimum for digital work. I recommend a 8.5 for analog.
Ahh... That's where I've been going wrong!
Logic probe for digital and test lamps for analogue  :)
 

Offline eevblogfan

  • Frequent Contributor
  • **
  • Posts: 569
  • Country: 00
hey

can you share with me ? 

ca you describe scenario where you was in need for 6.5 DMM within digital project ?

5.00000V ? ( the 34401 can go to 1.2 milion count , no more )
 

Offline robrenz

  • Super Contributor
  • ***
  • Posts: 3035
  • Country: us
  • Real Machinist, Wannabe EE
6.5 digit multimeter is the bare minimum for digital work. I recommend a 8.5 for analog.

Fortunately, for my low level analog work my 6.5 digit meter can read to 100 picovolt resolution in statistics mode so I can get by without buying a 8.5 digit meter ;)

Offline eevblogfan

  • Frequent Contributor
  • **
  • Posts: 569
  • Country: 00
hey robrenz , didn't you mean 100 nanao volts ?  :bullshit:
 

Offline robrenz

  • Super Contributor
  • ***
  • Posts: 3035
  • Country: us
  • Real Machinist, Wannabe EE
No I meant 100 Pico volts or 0.1 nano volt.   An 8846A at 100NPLC and low voltage levels has in internal resolution of 100 pico volts. You have to be in statistics mode to see the values.  Actual readings only go to 100 pico volt but the calculated values like average go to 0.1 pico volts

Offline eevblogfan

  • Frequent Contributor
  • **
  • Posts: 569
  • Country: 00
WOW , that's low 0_0

do you or any one else know any keithley instrument who does better than that ? , I mean , what is the lowest voltage you can measure ?

and as for the subject , I've asked Why do you really need 6.5 res ? can someone point some data ?
 

Offline ben_r_

  • Frequent Contributor
  • **
  • Posts: 411
  • Country: us
  • A Real Nowhere Man
Thankfully I dont work in anything that needs 6.5! I have a brand new 5.5 benchtop Agilent on order and thatll be overkill for me as is. But that OLED display is sooo puuurdy! lol
If at first you don't succeed, redefine success!
 

Offline madshaman

  • Frequent Contributor
  • **
  • Posts: 699
  • Country: ca
  • ego trans insani
when do you really need 6 1/2 DMM ? , I mean the res and/or the accuracy
« Reply #17 on: May 17, 2013, 03:18:58 pm »
Useful for graphing the behaviour of a system over a long period and looking for any interesting behaviour, especially when the change in voltages over time can be extremely small (e.g. A battery bank powering a system with low power consumption).  There even just having the precision is useful (as long as drift is very low and it's monotonic).

It's also nice to have extreme accuracy. I haven't gotten this far, but I have all my bench meters networked to my PC over gpib and although there's existing software, I plan to write generic software where I can characterise circuits and components' DC properties very accurately.

That means eventually I'll need a set of calibrated references to calibrate my meters from.

Off-topic, but does anyone have any good pointers or links to anyone who maintains a calibrated lab at home (potentially even NIST traceable) and what procedure is easiest and what is required?
To be responsible, but never to let fear stop the imagination.
 

Offline ddavidebor

  • Super Contributor
  • ***
  • Posts: 1135
  • Country: it
    • Fermium LABS website
when do you really need 6 1/2 DMM ? , I mean the res and/or the accuracy
« Reply #18 on: May 17, 2013, 03:27:56 pm »
You need a 6.5dmm for calibrate 5.5digit ones, a 5.5 digit for calibrare 4.5digit ones
Davide Bortolami,
Fermium LABS srl
 

Offline KedasProbe

  • Frequent Contributor
  • **
  • Posts: 502
  • Country: be
It may be nice to have if you want to check 16 bit DAC/ADC.
Or notice other small changes.
Not everything that counts can be measured. Not everything that can be measured counts.
[W. Bruce Cameron]
 

Offline ftransform

  • Frequent Contributor
  • **
  • Posts: 729
  • Country: 00
I actually sometimes turn down the digits on my meters because they annoy me and I don't need them. For alot of stuff that I found myself doing I basically need a LED probe for 90% of the work and then 4.5 digits is nice for final details.

Remember that this forum has a slight hard on for metrology.
I think that 6.5 digits is nice when working with an analog front end, like determining coefficients for a gain stage.

My original chip sources 10mA, while the 74HC04 can source 16mA.....WTF
« Last Edit: May 18, 2013, 02:20:03 am by ftransform »
 

Offline ddavidebor

  • Super Contributor
  • ***
  • Posts: 1135
  • Country: it
    • Fermium LABS website
when do you really need 6 1/2 DMM ? , I mean the res and/or the accuracy
« Reply #21 on: May 17, 2013, 07:22:03 pm »
Yeah i love to put in 4.5 digit with a hight plc, so the read are super stable and sufficient slow to read.
Davide Bortolami,
Fermium LABS srl
 

Offline robrenz

  • Super Contributor
  • ***
  • Posts: 3035
  • Country: us
  • Real Machinist, Wannabe EE
Remember that this forum has a slight hard on for metrology.

I never noticed that, thanks for pointing that out. ;D

Offline saturation

  • Super Contributor
  • ***
  • Posts: 4788
  • Country: us
  • Doveryai, no proveryai
    • NIST
You can easily check the health of certain batteries like NiMH or SLA if you check the output voltage at the max digit resolution.  You can see in real time self discharge rates on the LSD, be it uV or 10s of uV.
Best Wishes,

 Saturation
 

Offline free_electron

  • Super Contributor
  • ***
  • Posts: 7218
  • Country: us
    • SiliconValleyGarage
6.5 digit multimeter is the bare minimum for digital work. I recommend a 8.5 for analog.

bullshit. for digital work you only need a 1 digit 7 segment common anode display where segments b and c can be stuck in on position work. the others can be broken. the duality of the binary system guarantees that , if it is 1 it is not zero and vice versa.

segments ba and c are always on. segments a,d,e and f are connected together and go to a wire. touch a logic zero with that wire and segments a,b,c,d,e and f will light showing '0'. leave that wire floating or logic high and only segment b and c will light showing a logic '1'.

there you have it a digital binary logicel level meter using exactly 1 7 segment common anode display and 1 resistor.

Professional Electron Wrangler.
Any comments, or points of view expressed, are my own and not endorsed , induced or compensated by my employer(s).
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf