-
#625 Reply
Posted by
JxR
on 30 Jul, 2019 19:30
-
OK, let's not argue. JxR gets the credit of finding this flaw after a lot of people used the meter for a year.
I certainly don't think about it this way. If I'm going to be honest I was a tad frustrated that no one was willing to do the test a ~month ago when I first said something. Although, that is on me for not being more vocal about it, providing more data, and test scenarios. I guess I felt that I was still new to the forum and didn't want to be a prick about it. Keithley refusing to talk to me about it also kind of messed with mind.
I'm a career IT and have only been playing at EE for a few years now. Most of you are way more knowledgeable than I am and I'm just glad your here to help me out.
-
-
I just tested Keysight 34470 in this way. It has an -81 mV offset for 1 A current on 3 A terminal and -4mV offset on 10 A terminal. But these numbers don't change on any voltage ranges.
-
#627 Reply
Posted by
KedasProbe
on 30 Jul, 2019 19:50
-
I guess the next question is what should we do about it. All start requesting warranty service from Keithley for our meters?
Simple we push until we get a good answer.
You will see I added the link of the tek forum in the description of the youtube video.
It's their support forum they should use it wisely.
Could be one of the reasons why their latest firmware update is delayed.
-
#628 Reply
Posted by
JxR
on 30 Jul, 2019 19:59
-
Simple we push until we get a good answer.
You will see I added the link of the tek forum in the description of the youtube video.
It's their support forum they should use it wisely.
Thanks. I put in a reply to the topic as well.
Could be one of the reasons why their latest firmware update is delayed.
I admit I have wondered about this myself.
-
#629 Reply
Posted by
Kleinstein
on 30 Jul, 2019 20:38
-
There probably is nothing that can be done about it - it kind of a HW point. It is quite normal to have the actual central ground point at the "negative" side of the shunt chain. So there is a drop to the common terminal and this is measured. The unusual point is having the divider connected the way to get the extra signal. Anyway, when there is a current flowing, the voltage reading is no longer valid. This could be annoying when using the dual measurement mode. Anyway a shared terminal should also be suspicious.
-
#630 Reply
Posted by
JxR
on 30 Jul, 2019 20:42
-
I just tested Keysight 34470 in this way. It has an -81 mV offset for 1 A current on 3 A terminal and -4mV offset on 10 A terminal. But these numbers don't change on any voltage ranges.
The DMM7510 was -35mV on the 3A terminal in all voltage ranges, and -3.5mV in all voltage ranges on the 10A terminal with a 1A source.
-
#631 Reply
Posted by
JxR
on 30 Jul, 2019 20:55
-
Anyway, when there is a current flowing, the voltage reading is no longer valid. This could be annoying when using the dual measurement mode. Anyway a shared terminal should also be suspicious.
I would certainly not argue against your expertise in the matter, but why should we expect the meter not to perform as specified when doing dual measurements? It is a feature of this meter and documented in the manual. The only warning in the manual concerning dual measurements has to do with wear on the relays.
-
-
I just tested Keysight 34470 in this way. It has an -81 mV offset for 1 A current on 3 A terminal and -4mV offset on 10 A terminal. But these numbers don't change on any voltage ranges.
The DMM7510 was -35mV on the 3A terminal in all voltage ranges, and -3.5mV in all voltage ranges on the 10A terminal with a 1A source.
Maybe somebody can test an HP3458 in this way for reference.
There probably is nothing that can be done about it - it kind of a HW point. It is quite normal to have the actual central ground point at the "negative" side of the shunt chain. So there is a drop to the common terminal and this is measured.
This is certainly the easiest way to do it, but on a meter that can do 4-wire Ohms measurement it should be possible to move the ground reference of the ADC closer to the input terminal.
-
#633 Reply
Posted by
lukier
on 30 Jul, 2019 22:27
-
Maybe somebody can test an HP3458 in this way for reference.
It is getting late here today, but maybe tomorrow I can try to do this test on my DMM6500 and two 3458As, as well as previous generation of Keithley (2015 and 2001).
-
#634 Reply
Posted by
iMo
on 30 Jul, 2019 23:36
-
Possible explanation of the issue with the small 9V battery and the 10ohm resistor:
1. the first picture at 10V range shows
PRIM: 9V -> battery voltage when NOT LOADED by 10ohm
SEC: 0.87A -> the current when 10ohm connected to the battery
2. the second picture at 100V shows
PRIM: 6.7V -> battery voltage when LOADED by the 10ohm
SEC: 0.87A -> the current when 10ohm connected to the battery
My current understanding is the 10V range and 100V range differ in the way "when or how" they switch the internals in the measurement sequence.
Thus it is not a drop on the shunt or a 100V range defect, but the 6.7V is the voltage of the battery when loaded by the 10ohm. You may try to connect the 10ohm directly to the battery and measure the voltage at the battery.
-
#635 Reply
Posted by
cozdas
on 31 Jul, 2019 00:23
-
Thus it is not a drop on the shunt or a 100V range defect, but the 6.7V is the voltage of the battery when loaded by the 10ohm. You may try to connect the 10ohm directly to the battery and measure the voltage at the battery.
All of those possibilities are eliminated in the posts above. DMM reads -2.7V
with the input Hi and Lo shorted when 1A passes through the amps terminal.
-
-
Most of these meters have DC Ratio function that measures separately voltage on sense terminals. So, one could setup an I-V function, where the voltage is measured by sense terminals and current is measured using the common and current terminals. This would eliminate the internal voltage drop problems.
However, since the maximum voltage range of sense terminals is 10V, it wouldn't allow higher voltage measurements that are particularly bad on the DMM6500
-
#637 Reply
Posted by
Hugoneus
on 31 Jul, 2019 02:59
-
I tried two other meters:
Keithley 2015 THD: 14.5mV offset is observed on all scales. It is constant.
Keysight 34470A: 75mV offset is observed on all scales. It is constant.
I can try this on several other models (Keysight, Rigol, Siglent, etc.) if you guys really want to see.
-
#638 Reply
Posted by
2N3055
on 31 Jul, 2019 05:55
-
I just tried on Rigol DM3068. 1A into Amp connector, set to volts, it is cca -15mV , all Volt ranges .
DM3068 doesn't have simultaneous I/V function.
MTX3293 shows 5mV, all ranges, it does have VA function.
Regards,
-
#639 Reply
Posted by
cozdas
on 31 Jul, 2019 07:42
-
Fluke 8842A: -35mV for 1A in all ranges.
-
-
I tried two other meters:
Keithley 2015 THD: 14.5mV offset is observed on all scales. It is constant.
Keysight 34470A: 75mV offset is observed on all scales. It is constant.
I can try this on several other models (Keysight, Rigol, Siglent, etc.) if you guys really want to see.
Perhaps you can do a video where you hunt down the cause of this offset and explain the reason why it is there.
Tell about proper measurement techniques etc.
On the keithley it would be interesting to see where that jump in the 100 Volt range is coming from.
-
#641 Reply
Posted by
KedasProbe
on 31 Jul, 2019 15:32
-
Hi,
I made another script to make things easier:
When you have a sensor output 0-10V or 4-20mA or ... you usually want to see the right values live.
You can enter m and b of Y=m.X+b in the menu but usually your sensor documentation isn't m and b but 2 points of that line, so you can take your calculator... but the DMM6500 is a calculator so why not let it do it for you, hence this script enter two points and it will update m and b for you.
It also allows to set a unit to change 'X' I don't think you can do that in the menu, unfortunately limited to 2 char.
(this avoids confusion)
No sound video example:
Script:
https://github.com/Ken26M/keithley/blob/dev/Instrument_Examples/DMM6500/LinearConversion.tspedit: If you just want to do a current measurement and convert your resistor voltage drop to current, just enter
X1=0 and Y1=0 (no current if no voltage is present)
X2 = the known resistor value (volt but think ohm)
Y2=1 (1A if voltage = resistor value)
-
#642 Reply
Posted by
Kleinstein
on 31 Jul, 2019 15:44
-
The offset comes from the resistance from the banana plug at the common input to the internal zero point, that is usually at the lowest shunt. This includes the double plugs, so it's not even all from the meter. It is a pronciple problem to use the same terminal for a high current and precision voltage - these don't go together. So the cross coupling is not a surprise at all.
A solution for the problem could be using the Ohms sense terminal at least for the common side of the voltage measurement. Depending on how the 4 wire ohms is implemented this could be relatively easy and more like just a software thing. As the meter supports ratio measurements via the Ohms sense terminals it should be possible to use them, from the HW side - though one may still need an external link. The difference between the voltage negative side and current terminal is limited, but should be good enough for the offset.
The higher shift for the 100/1000 V ranges is likely due to the point where the divider is connected. If the shift adds after the divider the number gets high.
-
#643 Reply
Posted by
lukier
on 31 Jul, 2019 21:44
-
I've tested most of my meters. Sourced 1A into the current terminals, confirmed by switching to the current function and then switched to DCV and went through the ranges.
Only DMM6500 has this problem, above 10V range 25mV becomes 2.5V.
Other DMMs (2 x 3458A, K2001, K2015, 34401A) usually on lower ranges behave as nothing is connected (collecting charge) and on higher ranges are close to zero, both are to be expected.
-
-
Did you short out (or put some resistor) across the voltage inputs of the meters? Otherwise, they do read zero but its not an accurate representation of what happens if some voltage source is connected.
-
#645 Reply
Posted by
lukier
on 31 Jul, 2019 22:47
-
Did you short out (or put some resistor) across the voltage inputs of the meters? Otherwise, they do read zero but its not an accurate representation of what happens if some voltage source is connected.
No, I forgot to short them, hence why on low ranges they were collecting charge (GOhms). But they behaved exactly the same when nothing is connected (also to current input) so I assume it is fine. DMM6500 was also tested the same way and yet it showed 2.5V out of nowhere, on 10MOhm ranges that don't accumulate that much charge.
-
#646 Reply
Posted by
Mike G
on 01 Aug, 2019 04:47
-
I don't know if I am missing something but Brad O states clearly in reply 271 (either page 16 or 17 ish) that the 6500 cannot measure volts and current together, maybe the answer is as simple as that?
-
#647 Reply
Posted by
Mike G
on 01 Aug, 2019 04:48
-
Sorry I meant page 11 or 12 ish
-
-
He said "MegaVolt is right, the DMM6500 can't measure Voltage and Current simultaneously" - which is right, none of DMMs can measure two things really simultaneously (one needs a simultaneous-sampling DAQ for that).
But it should be possible to measure voltage while current is still flowing through the terminals. In a setup where a DMM is used to measure current it is impractical to keep disconnecting the current for measuring voltage, and it also changes the circuit because of burden voltage change.
-
#649 Reply
Posted by
Kleinstein
on 01 Aug, 2019 05:31
-
It looks like that with the current software, they can not measure voltage accurately with significant current flowing. This can't work using the same terminal: there would be at least some resistance in and between the plugs in the common terminal. The error may be acceptable at low currents, e.g. in the mA range if the voltage is not that low (10 V range). The higher voltage ranges seem to have an additional problem with the DMM6500 from the odd connection of the divider.
In theory they should be able to add this feature with a modified software, if the sense terminal (likely Ohms sens low) is used for the "negative" side of the voltage reading, at least for low voltages 0.1 ..10 V ranges. The 100 V range won't work well with the divider directly to common terminal. To some degree the sense terminals could be seen as secondary inputs and they are used like this in the ratio mode.