New Keithley DMM6500 and now DAQ6510

[Kleinstein]

The sense inputs should be high impedance, similar to the normal voltage input in high impedance mode (so > 10 G) range. There may be however slightly higher bias current and maybe slightly higher noise from the switching and protection part that is often lower grade than with the main input.  So there is no real need for an extra buffer.

[samofab]

Only downside is the already known transformer hum.
My unit is about as loud as my fridge (with the compressor running.) which might not be horrible, but it still feels excessive.

I also got my DMM6500 from welectron and it was quite loud. After some investigation, the unit was set to 220V. After I switched it to 240V, the hum is almost inaudible. Hope that helps.

[HendriXML]

The sense inputs should be high impedance, similar to the normal voltage input in high impedance mode (so > 10 G) range. There may be however slightly higher bias current and maybe slightly higher noise from the switching and protection part that is often lower grade than with the main input.  So there is no real need for an extra buffer.

Because the tests will be (semi) automated, and I only have one scpi enabled dmm, there's also the issue of the scope measuring Vbe. I think its impedance is about 10M with 10x attenuation. That measurement might require a buffer after all.

[HendriXML]

Only downside is the already known transformer hum.
My unit is about as loud as my fridge (with the compressor running.) which might not be horrible, but it still feels excessive.

I also got my DMM6500 from welectron and it was quite loud. After some investigation, the unit was set to 220V. After I switched it to 240V, the hum is almost inaudible. Hope that helps.

Mentioned it before: all my devices (dmm6500 as well) are switched using 2 powers trips. Bolted down at each side (left, right) of the bench. No hum and very practical.

[HendriXML]

I want to use the DC ratio function to measure the current amplification (at 180 μA) and measure Vbe using a scope to be able to match pnp transistors on those parameters using the posted test circuit. (The transistors will be used in a current mirror configuration.)

When calculating values in the test circuit,  I didn't take into account that the sense inputs are referenced to negative input and cannot exceed 10V to that point. (I think the manual mentions 12V).

[HendriXML]

Not the most interesting circuit in the world, but now it should be correct  . I added a voltage follower, just so scope measurement cannot interfere with the low base current.
The idea is get to a precise collector current, fine tuned by VCC using a programmable PSU based on the dmm's secondary measurement.
Then the dmm's DC ratio measurement needs to be scaled to get the current amplification.
Also Vbe will then be measured. I will do that for 60 x 2N3906's and run them through a matching script, which I hope will deliver 2 x matched pairs for the 4 transitor current mirror.

[KaneTW]

Only downside is the already known transformer hum.
My unit is about as loud as my fridge (with the compressor running.) which might not be horrible, but it still feels excessive.

I also got my DMM6500 from welectron and it was quite loud. After some investigation, the unit was set to 220V. After I switched it to 240V, the hum is almost inaudible. Hope that helps.

Oh, wow. I think this might have been my issue as well (Datatec, not welectron, but voltage is closer to 240 here anyways)

[Kleinstein]

Not the most interesting circuit in the world, but now it should be correct  . I added a voltage follower, just so scope measurement cannot interfere with the low base current.
The idea is get to a precise collector current, fine tuned by VCC using a programmable PSU based on the dmm's secondary measurement.
Then the dmm's DC ratio measurement needs to be scaled to get the current amplification.
Also Vbe will then be measured. I will do that for 60 x 2N3904's and run them through a matching script, which I hope will deliver 2 x matched pairs for the 4 transitor current mirror.

The circuit does not work this way, with the FB from the collector to the inverting input. Taking the signal from the collector inverts the signal - so wrong polarity. In addition the collector usually adds gain so just swaping the OP inputs may still not work but oscillate.

I would consider a slightly simpler circuit: just an emitter follower with a rather large resistor at the base. for a fixed voltage before the resistor one would than measure the voltage at the base an emitter. The collector current would not be absolute constant, but one could still measure the base and emitter current from the voltage drops. Matching pairs would also have a current very close.

Ideally the DMM6500 could also measuren the votlage from one sense input to com. Not sure if this is supported by the software.  The hardware should allow this and it would be a nice feature, not just for this measurement.

[HendriXML]

Not the most interesting circuit in the world, but now it should be correct  . I added a voltage follower, just so scope measurement cannot interfere with the low base current.
The idea is get to a precise collector current, fine tuned by VCC using a programmable PSU based on the dmm's secondary measurement.
Then the dmm's DC ratio measurement needs to be scaled to get the current amplification.
Also Vbe will then be measured. I will do that for 60 x 2N3904's and run them through a matching script, which I hope will deliver 2 x matched pairs for the 4 transitor current mirror.

The circuit does not work this way, with the FB from the collector to the inverting input. Taking the signal from the collector inverts the signal - so wrong polarity. In addition the collector usually adds gain so just swaping the OP inputs may still not work but oscillate.

I would consider a slightly simpler circuit: just an emitter follower with a rather large resistor at the base. for a fixed voltage before the resistor one would than measure the voltage at the base an emitter. The collector current would not be absolute constant, but one could still measure the base and emitter current from the voltage drops. Matching pairs would also have a current very close.

Ideally the DMM6500 could also measuren the votlage from one sense input to com. Not sure if this is supported by the software.  The hardware should allow this and it would be a nice feature, not just for this measurement.

Thanks for mentioning the opamp input issue. Should have seen that one myself..
I'll try swapping the inputs first and see how it behaves. It would be something I would do in a normal circuit as well  . Will do some research on it why it might be tricky. It's probably due to phases, a concept I yet need to familiarize with.
I had thought of a setup which used the dc-signal of an awg at the base (+resistor) to control the collector current. But that would take more time to tune for each transistor, so not ideal.

From what I know, there's not a supported way of acquiring the individual sense voltages programmatically. Only primary and secondary measurements. Measurements are done via buffers, with only 2 flavors: with or without secondary measurement. Having more optional properties would be nice. For instance each measurement could also contain the range (when auto ranging) that was used for that measurement. So one could also determine the accuracy of it.

[HendriXML]

From my understanding when a ratio measurement is done it consists of 3 distinct sequential measurements of the potential between negative input vs sens-, sens+, input+.
If one would like to minimize noise due to measurement wires of about 0.5 m, what would be a good approach.
My little test circuit would use dedicated wires with banana plugs soldered to them.
Using one twisted pair to the sense lines would not make sense, unless it were 2 twisted pairs, both using negative input. But that would result in 3 wires carrying negative input. Don't think noise get canceled out then..
The other option would maybe using shielded cables, but should the shield be connected to earth and on which ends?
I'm thinking about just using individual wires, but if there's an easy way to improve the setup that would be nice. I only have some coax cable lying around to use as shielded cable. But I think some low noise signal cable should be bought for in the toolbox. Any recommendations for comparable measurements?

[leighcorrigall]

As a follow-up, the Keithley DMM6500 was found to have a defective board and it was replaced under warranty. It took a bit of whining on my end to convince them that the instrument was faulty, but the factory fixed it. They even calibrated the unit free of charge, so the instrument is good for another year. This was not a firmware issue like the technical support thought it to be.

What bad luck. I feel like I jinxed something yesterday because the same day I posted this message the instrument failed again. 

New symptoms: 
- DCI is displaying a constant overcurrent ( > 3 A).
- ACI initially starts at high amps and quickly decays to a noisy uA reading.

I tried updating the firmware (1.7.3 -> 1.7.5), checking the fuses through the APP, reinitializing it, switching to rear mode. I am going to ask for a complete replacement of this instrument and then possibly sell it.

The Keithley DMM6500 is absolute garbage. The instrument has failed on me twice in 6 months, even after a factory repair. For the price, buy something less complicated and more reliable. I bought the instrument believing that the name had the same quality as the older Keithley models like the 200X, 648X, or the 23X. It's a great paperweight.

[HendriXML]

I've cut a firewire cable which is shielded and also contains 2 shielded twisted pairs, besides 2 unshielded wires.
That cable might be useful because of the shielding but it had been more suited if it had 3 twisted pairs. (If twisted pairs are even beneficial in this case.)
It is mildly flexible so that's a plus.

[HendriXML]

I implemented the circuit and tested it for oscillation. And it did as Kleinstein predicted and as shown in the scope screenshot (across R104).
So I added - a quick fix - a feedback cap C101, that makes it almost stable, but now and then the Vbe voltage (across vcc-ch1) rises, as shown in the 2nd screenshot.
But this happens only each second or so.

Part from that, I do like the circuit. The collector current can be measured and regulated well. The ratio gets measured ok. The base voltage can be measured.
And all those measurements don't fluctuate much. Touching the transitor and heating it only a bit has an effect though. So swapping needs to be done with pliers, and without much temp fluctuations in the room while doing the complete batch.

For the wires to the dmm I took the firewire cable and use the pairs as one combined wire.

While doing this experiment I had to laugh because of the worry about noise I had while looking at large oscillations on screen. 

Advice to stabilize it further is welcome,  otherwise I'll just add more capacitance  :

[Kleinstein]

Even a little oscillation is not good for a good DC measurement. Especially the VBE measurement can be off in this case.

I would try another additional cap at the transistors from base to collector, to slow down the transistor. To capacitor already added is good and needs be there also.  This should also reduce the sensitivity to external radio signal pic-up.

[HendriXML]

Will do that 
If it gets to be very rare, I can discard the measurement because of the scope measuring Vbe and it will then be "out of bounds" for a lot of samples.
The scope measurement should then be overlapping the dmm's. Good enough for the Netherlands 

[HendriXML]

The issue might be more interesting..
I implemented a new circuit, but nothing really changed. However I found out it happens very regulary. About every 832 ms.
But even more strange only when sense+ is connected to the dmm 
That line isn't even an input of something and also not really high output impedance.
I haven't given it much thought yet, but I suspect the dmm measurering interferes with the circuit in someway.

After writing this: I should check whether it also happens when sense- is disconnected and sense+ is connected and other combinations as well.

[HendriXML]

Might have remebered the wrong connector (or it doesn't replicate), it is input+that needs to be connected to get the interference.
Which also makes more sense, because it is connected to the opamps input.
Also the ratio function should be on, if it's off then no issue, it then measures voltage at the same terminals just fine.

I think I'll add another voltage follower and everything is Huncky Dory!

But this issue kind of restricts the usage of the ratio measurement function to a degree, I've got the freedom to adapt the circuit for it, but in other cases?

[Kleinstein]

The sense inputs are measured not all the time, but the are switched to the main amplifier from time to time. This switching can cause current spikes at the inputs and these may interfere with the circuit. Another possible problem is from capacitive loading - the TL072, like most other OPs does not like capacitive load at the output. With more than a few 100 pF (this may be just 1-2 m of cable) it can start oscillating. So the ouputs from the TL072 should have some series resistors like 100 Ohm to prevet this.


The capacitor just from the base to ground does not help that much, the more effective point is for the collector and especially the collector to base version.

Capacitance at the collector could be also an issue:  it can help with stabilit and if the DMM switches to measure at the sense input the capacitance can become smaller.

Anyway things drift a bit off topic -- this thread is about the DMM650 , not about the circuit to measure transistors.

[HendriXML]

The capacitor just from the base to ground does not help that much, the more effective point is for the collector and especially the collector to base version.

My wrong, I've misread your advise as put a capacitor at the base, not specifically between collector and base. (Will think about the difference.)
Thanks for helping out, eventhough off topic! (I didn't want to bring in a circuit with issues, just the "test setup")

The sense inputs are measured not all the time, but the are switched to the main amplifier from time to time. This switching can cause current spikes at the inputs and these may interfere with the circuit.

Probably the remaining issue.

[KedasProbe]

I also got my DMM6500 from welectron and it was quite loud. After some investigation, the unit was set to 220V. After I switched it to 240V, the hum is almost inaudible. Hope that helps.

Europe does have 230V and we have 220V and 240V to choose from, I left it on 220V but maybe we should set it to 240V?
If it reduces the HUM and works equally as well.... (haven't tested it)

That's something for Keithley to answer if 240V is also ok or even preferred on an 230V net?

Edit: I changed it to 240V since I think we are always above the allowed bottom limit of 240V net
It gives the impression that there is less hum, I could only measure less 400Hz hum, 100Hz and 200Hz are the same.

[E-Design]

I also got my DMM6500 from welectron and it was quite loud. After some investigation, the unit was set to 220V. After I switched it to 240V, the hum is almost inaudible. Hope that helps.

Europe does have 230V and we have 220V and 240V to choose from, I left it on 220V but maybe we should set it to 240V?
If it reduces the HUM and works equally as well.... (haven't tested it)

That's something for Keithley to answer if 240V is also ok or even preferred on an 230V net?

Edit: I changed it to 240V since I think we are always above the allowed bottom limit of 240V net
It gives the impression that there is less hum, I could only measure less 400Hz hum, 100Hz and 200Hz are the same.

With mains at 230V, setting to 240V should be ok. If the internal voltage is a problem (too low in this case), the unit has some ability to detect it and provide an error on startup.

[Mike G]

Hi E-Design, glad to see you are still on this forum.
A while ago I was describing ongoing problems I have had with my DMM6500.  These have now worsened to the point where it is almost unuseable.
The continuous sampling just stops and starts completely at random with messages in the reading table, while in the non sampling mood, which mean nothing to me. The annoying thing is that sometimes the meter will run faultlessly for several days and then will suddenly start misbehaving. It is in a temperature stable workshop, in a permanent position in an equipment stack with good ventilation. HELP!! PLEASE it is driving me mad
I am attaching a screen shot, which clearly shows the periods when no sampling is occuring, I also have the csv file from the reading table which shows the messages in the table while it is not sampling, maybe they could give you a clue? Unfortunately this is too large to attach to this post.
Any help would be appreciated since this problem has plagued me since new, although it is now becoming so serious that the meter cannot be relied on.
Thank you and regards Mike
AAAGH!  it's just stopped again after only being on for 20 minutes, screen shot and csv attached
Have had to change csv extension to txt to upload attachment

[Kleinstein]

The times marked look really odd: the data look like they are not even in the right sequence - like 2 streams of data coming in with some 5 seconds off.

The CVS file shows 2 data streems : one from the main ADC and one from the digitizer. So there is something odd with the settings. This should not happen and the data from the 2 source should go to separate buffers.

[Mike G]

Yes I noticed Digitizer in the csv but I am not using the digitizer, this just appears at random. While it is in this fault condition the meter sort of locks up, some hard buttons do nothing, some perform the wrong function, the screen is non responsive.  If I prod butons at random, I have not been able to repeat a sequence that helps, it will suddenly wake up and carry on as if nothing happened but leaves a gap in the readings.  It used to stop until I intervened but now if I leave it alone it does, eventually, restart by itself, this can be seen on the home screen image which shows several hours of data. It is a shame because when it works correctly this meter is a superb piece of equipment. When I initially bought it other people were reporting blue screens, freezing etc so I assumed it was a firmware problem and Keithley would sort it out, unfortunately I am now suspecting a possible hardware issue    
Let's see if E Design can throw any light on it
Thank you for your input, regards Mike

[The Bootloader]

As a follow-up, the Keithley DMM6500 was found to have a defective board and it was replaced under warranty. It took a bit of whining on my end to convince them that the instrument was faulty, but the factory fixed it. They even calibrated the unit free of charge, so the instrument is good for another year. This was not a firmware issue like the technical support thought it to be.

What bad luck. I feel like I jinxed something yesterday because the same day I posted this message the instrument failed again. 

New symptoms: 
- DCI is displaying a constant overcurrent ( > 3 A).
- ACI initially starts at high amps and quickly decays to a noisy uA reading.

I tried updating the firmware (1.7.3 -> 1.7.5), checking the fuses through the APP, reinitializing it, switching to rear mode. I am going to ask for a complete replacement of this instrument and then possibly sell it.

The Keithley DMM6500 is absolute garbage. The instrument has failed on me twice in 6 months, even after a factory repair. For the price, buy something less complicated and more reliable. I bought the instrument believing that the name had the same quality as the older Keithley models like the 200X, 648X, or the 23X. It's a great paperweight.

Have you contacted Keithley technical support about it ? They have been pretty responsive to me when I reported bugs to them
Just curious as to what they would have to say for the issues you are reporting