FLUKE DMM autorange VS Manual mode ... Ohm mode Serius Issue .

[Kiriakos-GR]

Well I just finished  the construction of my resistors box ( Decade ) ....

And started testing it ..

Well everything works sweet ,  but when I get at the 100K increments ,
I have an major by my opinion problem ...

Even if I had use the 100K 0.1% resistors ,  I get mixed numbers by the Fluke 87-5 ..

The good part are that I have  another three Fluke bench meters , and my oldie but trusted Taiwanese DMM,  so   there is no chance  to fall off  about  numbers ..  

The problem are that  the   Fluke 87-5  at the auto-range mode , reads the resistor as  100K .

But all the other meters ( not auto-range ones )   they measure the resistor about 240 Ohm sort .
And the 87-5  at the manual mode , comes close to this number too..

I checked the technical specs of three  old Fluke bench meters , and they have at the specific scale 0.6% accuracy ,  and the  87-5 has 0.5% ...

My true problem are that if i multiply  the 240 Ohms loss , by nine times , I get one large number as off set.    


My true problem .... if I cure the problem , by adding resistors to correct the readings at the manual mode,
the auto-range ones will probably show more total resistor , than the real ..
Basically I am guessing here ..  

Any way , tell me your thoughts ..

[Simon]

sounds like some shit meter !

[Kiriakos-GR]

sounds like some shit meter !

I will agree , because the rest Fluke orchestra at my bench ,  finds the Auto-range one ,
to be out of tune ...   

[Simon]

and if it agrees with the others on manual then it's very out of tune as it's disagreeing with itself (split personality - wait, do meters have personalities ?)

[Kiriakos-GR]

Well  I do not know what personality issues it has ...

But based on the book ,  in order to measure one 100K resistor , I will have to use the 600K range ,
or  the 6.000 MOhm  .

And both looks to have the same accuracy factor, so logically they should measure the same.

But they do not .. !!  

By the way, I use the high Resolution mode ... if something with it its cooked up, then there is the problem.

.

[alm]

In the 600kohm range, a resistance of exactly 100kohm should measure between 100kohm*.994 - .1kohm = 99.3kohm and 100kohm*1.006 + .1kohm = 100.7kohm. In the 6Mohm range, it should be between 100kohm*.994 - 1 = 98.4kohm and 100kohm*1.006 + 1kohm = 101.6kohm, so the tolerance is not the same. Any precision beyond that is not guaranteed or specified, although short term stability is usually significantly better. If you need better accuracy, get more accurate equipment.

The difference between auto-ranging and manual ranging is surprising, although not outside specs. Is it in the same range in both cases? If the reading depends on the ranges, it might just be a calibration issue, since each range has different calibration constants. Is the error consistent when measuring multiple resistors in series (eg. is individually measuring ten resistors and adding them the same as measuring all ten resistors is series), is it linearity or offset? Any type of relative functions active? Any voltage sources or active components connected to the setup?

[EEVblog]

There should be no difference between auto and manually ranging, it's either on the same range or it's not.
Yes, different ranges could give different readings.
I would want to see a video of it to see what is actually happening.
Not a probing issue perhaps?

Dave.

[Kiriakos-GR]

Nop no probing error , I use banana plugs at both ends , and crocks grabbing the resistor.
I was even waiting 30 seconds so to stabilize.

I will post  some pictures later on.

[Kiriakos-GR]

Well  I reported the above issue , mostly because  I was testing my new resistors box ( Decade ),
and at the 100K (9 positions) switch I got mysterious measurements.

It looks that the  problem its not actually related with the meter, but with some type of leakage,
caused by (9 positions) switch .

I am still unable to understand what did go wrong, as when I unsolder the resistors they do measure more stably .

The 100K to 900K , its an large number as resistor, every leakage it will strongly effect the measurement.

So I will do one step back , about my primary "false alarm" about the meter  it self .

Even so, there is an deference at the measurements between manual ranges.
But the 87-5 looks to stay in specs.   
     
   

[Simon]

well even if there is a "leakage" I would expect the manual and autorange to read the same, you have two issues here:

1) measuring error
2) inconsistency in measurements

[rf-loop]

This is not maybe related to this problem but also this is good to know and think in some cases when make things with mechanical relay or all other kind of contacts... special difficult is very low DC voltage with extremely low currents. These cases are sometimes very difficult if also need take special care about accuracy and reliability.

This is example where I have "break my head" in "history". This small example text is from Omron datasheet about one small rotary switch with golden contacts. This small text remeber me about this kind of problem. (some time in history I did not read carefully this position about some other manufacturer schwitc...  I pay lot of about these "lessons")

Quote:
"Circuit Design.

Although the minimum current is 10 ?A (3.5 VDC), contact reliability
may need to be improved in some cases."


In my experience this kind of things may sometimes make big (and sometimes very difficult to find and solve) problems with small signals.

I hope (and believe) you have not this problem.  But with this kind of things this problem is sometimes also behind corner waiting...

----

Btw... some years ago I burn some components. I use other person's multimeter what I did not know. (not remember name of DMM)
In manual mode  some range  in my own DMM use higher voltage and other lower voltage (and current) over DUT. Normally I use my meter and automaticly avoid these higher voltage (ohm) ranges in special places. (no need always think becouse it was like routine).

Then in some special situation I use other peoples DMM for "ohm" measure.  Of course meter was "autorange". What it do... it do some kind of "scan" to find range... and in this case it use both... high voltage and low voltage ranges before it find... so just as I do measurement (inside running system becouse factory macines many times service "hot and running"........after my measurements all stop working... (after this I change some components I burn - destroy some sensitive circuits).

I do not know how your fluke work with autorange and manual.

(one method for extremely small current flow trough machanical contact is "wetting". First "wet" contact with higher current/voltage and after this pulse.... small signal go... )

if someone is interested more then maybe something can find with google using words: contact wetting minimum current

[djsb]

Do you connect all the resistors in series? How many resistors at the most are connected at one time? When you have the box set to a particular resistance have you measured each individual resistors resistance as well as the total resistance? Do these figures agree with each other? Could the resistor tolerances (and errors) be cumulative when connected in series? Just some things that occur to me anyway.

David.

[Kiriakos-GR]

Hello people, thanks for your interest on my problem ..

I have just finish the battle with my demons , and I had won ..  

The solution was not simple ...  

About the part that has to do with 87-5 :
The  named as issue was at the "High Resolution" mode .
The DMM at the 600K range  measure my 100K 0.1% resistor as  100K ,
by entering at the "High Resolution" mode , the resistor was measuring 99.95 .
This deference made me think about an possible issue with the meter.
Thanks God , after the 87-5  I got and the bench meter Fluke 8050A  20.000 counts,
that it did confirm the 99.95 K  reading.

Something that helped my confusion to grow, was that at the "High Resolution" mode the 87-5 , it needed more than 12 seconds to settle ,  it started measuring the resistor as  100K and then it started slowly to drop like  99.99 99.98 99.97 99.96  some play between 99.96 & 99.95 so to stay there.
The 8050A  was  spot on from the first second .  99.60 thats it .. this is the measurement .
No play , no waiting .. spot on at fast.  

About the part that has to do with 9 positions switch
I had bust my head for three days dealing with all this complex issue ,
and come to the point to totally dismantle it ...
Found inside excessive amount of  Flux , so much that all the contacts was having an bath in it.
I used solder with Silver , it was hard to melt , and I was using lots of flux so to help it .
The switch are almost sealed , so the fumes of the flux was unable to escape , and they was remaining trapped in it.
When I was heating the external contacts so to solder or de-solder , the flux inside was becoming wet again,  causing all sort of problems in my measurements .   ( leakage )
I took the specific apart , wash anything inside with alcohol , in depth cleaning,  and then I assembled it again.

About the part that has to do with the resistors

I trusted the crop of the cream  " Dale " 100K 0.1% , but even so, not all resistors was correct .
I had to check them , one by one .
I did not do that at the beginning , because of blind trust !!  I payed an large price for them,
so to not need to test them ...    Testing garbage its logical ... testing the top dog ? ( looked as illogical to me )

Any way since the time , that I started to isolate and remove my problems above,
DMM readings & wet switch full of Flux.....

I started to calibrate the last stage of my decade the 100K range,
got many film metal resistors  from 10 Ohm  to 85 Ohm ,
and started to adjust the six from the nine resistors , until to see the 100.01 K  in all .
Now I am dead tired but an clear winner .  

The 8050A helped me allot , with out it , with out the second confirmation , I would become nuts.
Any way, the decade are ready , and the other stages like 10K and lower , they did not effected by the extra Flux in them.
I know that they are full of it , like the last one.


Few words about the Auto-range ,  the Fluke 87-5 its my first Auto-range meter ,
I have to admit , that its a very fast meter .
@rf-loop had speak about some possible Auto-range issues,
well I think that the 87-5 its so damn fast , that its unable to cause damage,
due the speed that it haves.
An very slow DMM , It could be called as danger for sensitive components, but not the 87-5.

The happy end, about my seven stages decade, was the interconnections lowest resistance.
That is at 00,04 ohms .   ( 20 ohms range)  
I did verify that number , by the 87-5 , and the 8012A bench meter that I had fall in love with it,
because it has the  "Low Range Ohm "  extension that found very handy in this case,
and the most amazing part , are that the 87-5 it did agree with the ancient 8012A, at 100% .
So all my meters are finally agree , that I had succeed