All about Keithley DMM7510. Bugs and features, recipes, advice, notes.

[aronake]

It seems it got removed in the updated calibration manual not to mislead people that these are the only things needed for all calibration.

Yes, indeed, this was mentioned in the old manual.

The new manual leaves some questions open.

For example, Step 2 of DC calibration mentions 100V and 1000V full range. But it does not say whether the calibration is done automatically or an external source is needed.

The manual also suggests verifying the 100V and 1000V ranges, but does not say what to do if they are outside the tolerance? Is this a malfunction?

Unfortunately, I do not know if it is possible to do partial calibration I am also interested in this question.

I have now tired to do the artefact (cal.adjust.dc) part only. And can report that it worked very well. Before adjustment my reisitance measurement was pretty far of, but max 10 PPM off for 10 to 10K range. Only things needed is a 4 wire short, 10V and 10K ohm.

I have not checked linearity of voltage yet, but will do later.

Step 2 in DC is very easy signal to generate, just leave the terminals open

And for 100V and 1000V ranges being outside specifications after an artefact calibration. There isn't really anything to do about that (maybe try to calibrate again, i doubt that makes a difference, get the DMM7510 repaired or buy a new.... or just live with the error)

I was also wrong on 7 amp needed for AC adjustment. Apparently its possible to use 2 amp (2 amp to 10 amp) instead. I should be able to manage to arrange that. Generation would not be accurate enough on its own, so will try to use 2 3458a in parallel to measure actual current.... Not sure how that will turn out though. And not sure if i feel i need super accurate AC current measurements in the 3 to 10 Amp range.

[HighVoltage]

Before adjustment my reisitance measurement was pretty far of, but max 10 PPM off for 10 to 10K range. Only things needed is a 4 wire short, 10V and 10K ohm.

I have not checked linearity of voltage yet, but will do later.

What do you mean by far of?
How far?
In what ranges?

I have two DMM7510 and I am always surprised how stable they have been over the years.
I do not see any need for adjustments at all.
Maybe I got lucky with my units.

[MegaVolt]

I have now tired to do the artefact (cal.adjust.dc) part only. And can report that it worked very well.

This is good news. Did you save the partial calibration results?

[aronake]

I have now tired to do the artefact (cal.adjust.dc) part only. And can report that it worked very well.

This is good news. Did you save the partial calibration results?

Yes! So old calibration data is gone. The DMM7510 was calibrated in a supposed to be quite OK calibration lab before it was delivered to me by the seller. I got a quite nice looking calibration certificate with measurement points. But I have always been very suspicious about the quality of the calibration. And first got my standard resistors calibrated by the national metrology lab last week, so have not know which of my 10 multimeters was more right than any other. Ill hand in my 732b this week to the national lab. Then i know more about the voltage situation once i get that back.

Even with my quick and dirty calibration the dmm7510 is very well within 24 h specification on all resistance ranges. So will do a more proper calibration and then analyse the results more in detail.

[aronake]

I was just reading though the calibration manual of DMM6500 which I also have and may want to calibrate. As is quite common for a 6.5 digit DMM each range need to be calibrated for each function. All signals must be set exactly (except for resistance). It also requires a fixed no compromise 7A 1khz AC current signal which can be difficult to generate. So a comparable nightmare to DMM7510's artefact calibration

Here for instance are the steps to calibrate open front connectors. I guess all these steps may come with some added flexibility. And would of course be automated in some script. But still.... There are equally many for short front connectors.

--Use front-terminal 4-wire open circuit setup cal.adjust.step.setup("cal_DCI_3A_zero_front")  cal.adjust.step.execute("cal_DCI_3A_zero_front")
--Use front-terminal 4-wire open circuit setup cal.adjust.step.setup("cal_DCI_1A_zero_front")  cal.adjust.step.execute("cal_DCI_1A_zero_front")
--Use front-terminal 4-wire open circuit setup cal.adjust.step.setup("cal_DCI_100mA_zero_front")  cal.adjust.step.execute("cal_DCI_100mA_zero_front")
--Use front-terminal 4-wire open circuit setup cal.adjust.step.setup("cal_DCI_10mA_zero_front")  cal.adjust.step.execute("cal_DCI_10mA_zero_front")
--Use front-terminal 4-wire open circuit setup cal.adjust.step.setup("cal_DCI_1mA_zero_front")  cal.adjust.step.execute("cal_DCI_1mA_zero_front")
--Use front-terminal 4-wire open circuit setup cal.adjust.step.setup("cal_DCI_100uA_zero_front")  cal.adjust.step.execute("cal_DCI_100uA_zero_front")
--Use front-terminal 4-wire open circuit setup cal.adjust.step.setup("cal_DCI_10uA_zero_front")  cal.adjust.step.execute("cal_DCI_10uA_zero_front")
--Use front-terminal 4-wire open circuit setup cal.adjust.step.setup("cal_DCI_DIGI_3A_zero")  cal.adjust.step.execute("cal_DCI_DIGI_3A_zero")
--Use front-terminal 4-wire open circuit setup cal.adjust.step.setup("cal_DCI_DIGI_1A_zero")  cal.adjust.step.execute("cal_DCI_DIGI_1A_zero")
--Use front-terminal 4-wire open circuit setup cal.adjust.step.setup("cal_DCI_DIGI_100mA_zero")  cal.adjust.step.execute("cal_DCI_DIGI_100mA_zero")
--Use front-terminal 4-wire open circuit setup cal.adjust.step.setup("cal_DCI_DIGI_10mA_zero")  cal.adjust.step.execute("cal_DCI_DIGI_10mA_zero")
--Use front-terminal 4-wire open circuit setup cal.adjust.step.setup("cal_DCI_DIGI_1mA_zero")  cal.adjust.step.execute("cal_DCI_DIGI_1mA_zero")
--Use front-terminal 4-wire open circuit setup cal.adjust.step.setup("cal_DCI_DIGI_100uA_zero")  cal.adjust.step.execute("cal_DCI_DIGI_100uA_zero")
--Use front-terminal 4-wire open circuit setup cal.adjust.step.setup("cal_ACI_1A_zero")  cal.adjust.step.execute("cal_ACI_1A_zero")
--Use front-terminal 4-wire open circuit setup cal.adjust.step.setup("cal_ACI_3A_zero")  cal.adjust.step.execute("cal_ACI_3A_zero")
--Use front-terminal 4-wire open circuit setup cal.adjust.step.setup("cal_4W_HiOhm_7Vref_open")  cal.adjust.step.execute("cal_4W_HiOhm_7Vref_open")
--Use front-terminal 4-wire open circuit setup cal.adjust.step.setup("cal_CAP_zero")  cal.adjust.step.execute("cal_CAP_zero")

[kj7e]

Model DMM7510 Firmware Revision 1.7.16
https://www.tek.com/en/support/software/firmware/model-dmm7510-firmware-revision-1716-and-release-notes

[rusoaie]

Hello, everyone!

Due to for its speed, graphing and remote TSP scripting capabilities the Keithley DMM7510 has become my daily driver and I could not stand the fan anymore... so...

Just like to share my [partial] success in replacing the annoying fan inside the DMM7510 (see caveats below).

The main contender was the Noctua NF-A6x15 FLX. It is slower [3500rpm] than the original MB6010X12B1 [5300rpm], however it catches up at airflow due to its 15mm thickness (compared to 10mm of MB6010X12B1) reaching maximum 23.4 m³/h (compared to 31.8 m³/h max of MB6010X12B1).

A few mounting notes:
- I used only the extension cable supplied with the fan (and none of the attenuation/LNA/ULNA cables supplied in the box)
- had to use other screws so that they reach the board and not exceed it
- had to ditch the metal protection grille, as the Noctua is way fatter than the original fan (proceed at your own risk!)
- added rubber o-rings between the fan and the PCB for extra vibration protection (again compromising a bit on airflow)
- The DMM has 2 connectors for fan: 3 and 4 pins; I used the 4 pin connector (although my fan was the non PWM-version), as the 3 pin one starts with the voltage set at ~8.5V by default and the Noctua fan is waayyy to slow at 8.5V to properly cool the instrument
- if someone tries the NF-A6x15 PWM version on the 4-pin connector maybe they can report back in this thread?
- multimeter detects the rotation to be a tad slower in the beginning, but catches up after max 5 minutes of startup (see screenshot)
- for precission measurements, or if you keep your DMM always on, probably replacing the fan with the original MB6010X12B1 is a better choice

Needless to say that you can still hear the A6x15, but it is orders of magnitude quieter than the original one and you can work next to it extended hours.

See attached some mounting details.

Happy ppm's to everyone on DMM7510's 10th birthday!

[aronake]

Model DMM7510 Firmware Revision 1.7.16
https://www.tek.com/en/support/software/firmware/model-dmm7510-firmware-revision-1716-and-release-notes

Its pretty competitive in the filed of lame firmware updates, but thisone is really toward the top! New format of serial numbers and and a new end user license agreement.... Or do Keithely the habit of doing fixes in firmwares they don't disclose? That would be quite strange if that was the case for test equipment.

[KungFuJosh]

Its pretty competitive in the filed of lame firmware updates, but thisone is really toward the top! New format of serial numbers and and a new end user license agreement.... Or do Keithely the habit of doing fixes in firmwares they don't disclose? That would be quite strange if that was the case for test equipment.

I think it's fairly common. Often only larger bugs are listed. I think it's less common to list every single thing in a firmware update, unless there really isn't much to report.

[exe]

- if someone tries the NF-A6x15 PWM version on the 4-pin connector maybe they can report back in this thread?

I bought one, it will arrive on Wednesday, but it will take me some time to get to the replacement. The rubber o-rings and new screws, are they included with the fan?

New format of serial numbers and and a new end user license agreement....

I was also skeptical about this, but now it makes more sense. As a software engineer, I can tell it does make sense to keep all clients at the same software version, even if there are no functional changes. This makes support easier. Imagine they don't release the new version, and only make available with new units. This will piss-off existing customers, and there will be speculations. For transparency, I'm glad they released it.

As a client, you have a choice to upgrade or not to upgrade. That's what changelog is for. Just skip it.

[rusoaie]

- if someone tries the NF-A6x15 PWM version on the 4-pin connector maybe they can report back in this thread?

I bought one, it will arrive on Wednesday, but it will take me some time to get to the replacement. The rubber o-rings and new screws, are they included with the fan?

Nice!

Inside the box you will have some anti-vibration mounting accessories which you cannot use because the Keithley DMM7510 case slides too close to the mounting PCB for the fan (if you manage to use those somehow, please DO share how you did it!).

As far as I remember, screws I used are M4 (or M3?) with a buried head and the appropriate length to accommodate the extra 5mm from the thicker fan + the O-ring.

I got both the screws and the o-rings from my local hardware store (Hornbach, you also have it in the Land from Down Under) - don't remember the exact part number; measure first and then go to the store and ask for it.

Here are some Hornbach references, these are NOT the actual parts you need to buy, just useful to start the search: O-rings category, Flush-mounted screw

[exe]

@rusoaie I'm trying to use rubber mounts supplied with the fan. So far so good.

However, I cannot plug the fan into the board: the connector doesn't match socket on the board. I see in your photo you kinda connected it in a reverse position. Did you apply much force to plug it in? Because mine doesn't want to go in. I'm crimping a different connector.

[rusoaie]

@rusoaie I'm trying to use rubber mounts supplied with the fan. So far so good.

However, I cannot plug the fan into the board: the connector doesn't match socket on the board. I see in your photo you kinda connected it in a reverse position. Did you apply much force to plug it in? Because mine doesn't want to go in. I'm crimping a different connector.

Hello!

Yes, I used The Force. Yes, better crimp a connector.

Did you try to slide the cover? Mine was getting stuck in the anti-vibration thingies.

Best of luck!

[exe]

Did you try to slide the cover? Mine was getting stuck in the anti-vibration thingies.

Best of luck!

Thanks! I didn't slide the cover, but I see what you mean, very little space for rubber thingies.

Meanwhile, with PWM cooler, both 3-pin and 4-pin connectors give the warning that "fan is spinning to slowly (56% of minimum)". It eventually reports 66%, but I think it will never catch up. I'll try with non-pwm cooler.

[exe]

PS forgot to say, I measured waveform on power line for both old and new fans, connected to external 12V linear power supply. Noctua seems miles quieter (in all senses) than the original one. The original one gives very nasty spikes. Noctua is below noise floor.

[rusoaie]

Did you try to slide the cover? Mine was getting stuck in the anti-vibration thingies.

Best of luck!

Thanks! I didn't slide the cover, but I see what you mean, very little space for rubber thingies.

Meanwhile, with PWM cooler, both 3-pin and 4-pin connectors give the warning that "fan is spinning to slowly (56% of minimum)". It eventually reports 66%, but I think it will never catch up. I'll try with non-pwm cooler.

You can try to ignore the PWM pins (or connect them to ground - see datasheet or AI), that should make your fancy PWM cooler act as a normal fan (at full power).
Then you can attenuate the speed using the attenuators in the kit. Oh, wait. The PWM version comes without attenuators? 
You can also check the voltages on the second connector (the 3 pin one) - I'd say the instrument is designed to support a variety of coolers.
There must be a way to hack the PWM one to work

[exe]

So, non-pwm version reached 99% of speed, and I still get warnings)

As of mounting solution, please see the photo. It works well, imo, and there is even a little bit of extra space, so no need to trim standoffs that low. BTW, the stock fan was due to change. It was loud and making unhealthy noise.

Now, the impact on the instrument. The stock cooling solution was looking quite bad to me. With my mount, I was expecting no change at all. Also because the new fan has 72% of performance of the old fan. I was wrong.

Before replacing the fan, the voltage from reference was 6.923_984V. Now the value is 6.923_990V, which is closer to actual reference voltage of 6.923_992V (measured by a calibrated DMM7510).

Resistors also measure somewhat higher. And ACAL shows 6.4C difference. So, yes, the fan does affect parameters.

ACAL shows 6.4C difference (and 1.7C difference to calibration temperature), while room temperature didn't change that much. So, it seems fan has an impact. I think that's because of the gap between top pcb and fan due to standoffs, hot air stays inside the unit for longer.

[exe]

Folks, one thing I noticed is that after replacing fan, the 10V and -10V "ranges" now much closer to each other. Before replacing fan I had like 5uV difference. I thought it's "okay", just some mis-calibration or offset drift. But now the difference is ~1uV, may be even within noise floor of my setup.

Anyway, I changed mount of the fan a little bit, removing the gap between fan and pcb. This increased acoustic noise, but lowered the temperature by 2C. So, with my first way of mounting the internal temp was 37.9C, now it's 35.9. The voltage readings slightly raised (though fan is working better), and now my DMM7510 withing 1uV to what's written on my ADR1399 reference.

And I still get fan warnings(. I consider installing NF-A6x25 FLX on rubber feet. It's a bit risky because it's bigger and has even lower RPM, but more airflow. A complicated solution would be an adapter board that would multiple RPM readings to make it within what DMM7510 expects.

[Scottjd]

Folks, one thing I noticed is that after replacing fan, the 10V and -10V "ranges" now much closer to each other. Before replacing fan I had like 5uV difference. I thought it's "okay", just some mis-calibration or offset drift. But now the difference is ~1uV, may be even within noise floor of my setup.

Anyway, I changed mount of the fan a little bit, removing the gap between fan and pcb. This increased acoustic noise, but lowered the temperature by 2C. So, with my first way of mounting the internal temp was 37.9C, now it's 35.9. The voltage readings slightly raised (though fan is working better), and now my DMM7510 withing 1uV to what's written on my ADR1399 reference.

And I still get fan warnings(. I consider installing NF-A6x25 FLX on rubber feet. It's a bit risky because it's bigger and has even lower RPM, but more airflow. A complicated solution would be an adapter board that would multiple RPM readings to make it within what DMM7510 expects.

I’ve been following this, but not at that point of trying any changes until I know everything is good with the one I bought.
I just had some thoughts to toss in on the subject. I think the fan closer to the PCB would increase the function on how the ventilation should flow. Maybe if you have a 3D printer you can print a thin TPU flexible part to fit in between the fan and the board. This would hopefully still calm the acoustic vibration and give it the seal it needs for air flow without taking up much space. Maybe even a thin foam cutout like a gasket would work?

As for the changes in the voltage readings, wouldn’t autocal be able to correct for those minor changes in temperature by using its onboard temp sensor, 10Ω resistor and on board references?
I know mine was last calibrated and adjusted at a temp thats showing -1.6C compared to my ambient temp in the house. But once I had it for a week and gave it 4 hours warm up I ran the auto calibration and it did a good job of correcting that slight offset in the readings to how it was adjusted last time. Just a thought, not sure if you tried that yet?

I was looking at other fans on DigiKey and mouser but the other ones are rated for the same airflow, RPM’s and noise decibels so I dont think they would be any different. Hmmm, maybe that TPU gasket or cutting out the foam gasket used with the stock fan might be enough to dampen the vibration acoustics and still report the correct RPM’s?
Just some random thoughts. 

[exe]

Maybe if you have a 3D printer you can print a thin TPU flexible part to fit in between the fan and the board. This would hopefully still calm the acoustic vibration and give it the seal it needs for air flow without taking up much space. Maybe even a thin foam cutout like a gasket would work?

As for the changes in the voltage readings, wouldn’t autocal be able to correct for those minor changes in temperature by using its onboard temp sensor, 10Ω resistor and on board references?

Yeah, I'm looking into printing a TPU gasket. I consider varioshore TPU (foamy), but it's kinda expensive. About autocal, it never changed anything on my unit. Which is kinda expected for a stable device.

BTW, I was looking into setting fan settings, and I found :SYSTem:FAN:LEVel QUIET or fan.level = fan.LEVEL_QUIET. It seems it doesn't do anything on my unit with replaced fan, but I think it lowers expected RPM, so now in log I see "The fan is operating properly".

So, my plan now:
- put an even bigger fan, and with lower RPM
- add a gasket and rubber mounts

BTW, could you please report your ambient temperature and output of print(localnode.internaltemp) after 90min warmup?

[exe]

I bought varioshore TPU and designed a gasket. The TPU and fan should arrive tomorrow, so next week I'll make "the most silent dmm7510 in existence".

BTW, do you guys know if TPU is a reliable filament? I don't want it to decompose inside my precious dmm, but I also cannot find any info on the subject other than "there are two types of TPU, and one of them doesn't like humidity". I guess I need to contact the manufacturer. The technical datasheet is here: https://downloads.colorfabb.com/index.php/s/rtfDDRCa723Xdor?dir=/Technical%20Data%20Sheets/TPU/colorFabb%20varioShore%20TPU&editing=false&openfile=true

[Scottjd]

I bought varioshore TPU and designed a gasket. The TPU and fan should arrive tomorrow, so next week I'll make "the most silent dmm7510 in existence".

BTW, do you guys know if TPU is a reliable filament? I don't want it to decompose inside my precious dmm, but I also cannot find any info on the subject other than "there are two types of TPU, and one of them doesn't like humidity". I guess I need to contact the manufacturer. The technical datasheet is here: https://downloads.colorfabb.com/index.php/s/rtfDDRCa723Xdor?dir=/Technical%20Data%20Sheets/TPU/colorFabb%20varioShore%20TPU&editing=false&openfile=true

I think thats more of a user print sheet, not a data sheet for the TPU. But in short the usual two types of based TPU’s are either polyester-based TPU, or polyether-based TPU.
The polyether-based can be better if you’re using the print in high humidity “wet” conditions. But I don’t think you will have your meter outside in the rain?

Like most 3D print filaments they can absorb moisture over time and are hygroscopic (like your car break fluid, change it!) and TPU is no different. But this can be reversed by drying the filament, or dealing with the stringing while printing it. Once its printed it can be affected by hydrolysis, different from hygroscopic as you can’t dry it or reverse the break down of the material. This is where the polyether-based one is better in high humidity wet applications.

But honestly, I don’t think either will matter for something that is being used indoors (I hope) and not directly being exposed to water (I hope not)….

Side Note:
I looked at one of my Noctua fans last night and realized the rubber corners already protrude further out than the casting of the fan. So doesn’t that already absorb some of the acoustic transfer?
BUT I also noticed it’s a soft compressible rubber, so I guess if the screws are tightened those corners are not making much of a difference? After all you don't want those screws to be loose as the vibration could work them free over time and they will fall on to the main board.

[exe]

I looked at one of my Noctua fans last night and realized the rubber corners already protrude further out than the casting of the fan. So doesn’t that already absorb some of the acoustic transfer?
BUT I also noticed it’s a soft compressible rubber, so I guess if the screws are tightened those corners are not making much of a difference? After all you don't want those screws to be loose as the vibration could work them free over time and they will fall on to the main board.

Yeah, rubber corners don't do much if fan is mounted by screws. For the best result, the fan should not be firmly attached to the pcb with screws, but through some shock absorption.

I tried to do noise measurements with different types of mount. But when the fan was "floating" on rubber standoffs, the noise was comparable to the noise from outside (even with all windows closed), so I couldn't get any reliable measurements.

Alternative solution to gasket could be mounting on rubber standoffs with gaps, and cover the gaps with kapton tape. But gasket will look fancier)

[Hydron]

BTW, I was looking into setting fan settings, and I found :SYSTem:FAN:LEVel QUIET or fan.level = fan.LEVEL_QUIET. It seems it doesn't do anything on my unit with replaced fan, but I think it lowers expected RPM, so now in log I see "The fan is operating properly".

Heads up that the DMM6500 defaults to quiet, and normal (NORM) is MUCH louder, so it does something at least on that meter (the command is unlisted in the 6500 reference manual, but still works).

[KungFuJosh]

Heads up that the DMM6500 defaults to quiet, and normal (NORM) is MUCH louder, so it does something at least on that meter (the command is unlisted in the 6500 reference manual, but still works).

Same on the DAQ6510:
:SYSTem:FAN:LEVel?
;; QUIET