Keysight's new 34465A (6.5 digit) and 34470A (7.5 digit) bench multimeters

[Keyview]

Hi,

the data logging for the 34465A is puzzeling me.
Keysight specifies 5000 readings/s.
The user manual talks about 1000 readings/s in data logging mode.
I read that data logging mode allows zooming into data.

I would understand that reading rate is reduced from 5k/s to 1k/s when saving the data to an external file (e.g. via USB).
But, following the manual, it is  reduced to 1k/s when saving to internal memory, too.
Is this really the case?

Thank you for your feedback.

Btw, one of the prime reasons for my interest in the 34465A has not been
mentioned in this thread, yet: The guaranteed accuracy of this device for VAC in the low end down to 1 mV
is only outperformed by the high-end Agilent Keysight 3458A and Fluke 8508A as far as I can see.
Even the 34461A is hard to beat in the range 1-5 mV AC. 

[HighVoltage]

the data logging for the 34465A is puzzeling me.
Keysight specifies 5000 readings/s.
The user manual talks about 1000 readings/s in data logging mode.

Welcome to the forum.
The 34465A and the 34470A both offer 5,000 rdgs/s in standard mode and 50,000 rdgs/s with the fast digitizer option
The 34461A only does 1,000 rdgs/s and has no option for faster readings

The storage memory on the 34465A and the 34470A 50,000 rdgs in standard mode and 2,000,000 rdgs with the memory option
The 34461A only does 10,000 rdgs and has no options for more memory

[Keyview]

Welcome to the forum.
The 34465A and the 34470A both offer 5,000 rdgs/s in standard mode and 50,000 rdgs/s with the fast digitizer option

Thank you for the welcome and your feedback.
Yes, that is why I was wondering about the maximum of 1,000 rdgs/s in the "data log mode" for the 34465A and 70A.
(source: user manual, pages 82 and 182, www.keysight.com/main/redirector.jspx?action=ref&cname=EDITORIAL&ckey=2345839&ml=eng)
Does the 5,000 rdgs/s for the 34465A apply only to "storing data" (page 99, above PDF) which does
not allow zooming into data? (The "digitize mode" on the other hand has 50,000 rdgs/s.)
Have a nice day!

[Sparky]

And the fast DIG option also arrived and works really well.
I will post some test results of the DIG option soon.

Any updates on the DIG option?
It seems the DIG option is only really useful if you get the memory upgrade as well, 50k readings a second fills up fast when you have a 50k readings memory limit.

I'm also looking forward to hearing more about the DIG option...hopefully HighVoltagewill get a chance to write about it soon...

I was looking forward to someone posting about the DIG option and illustrating what 50k readings/sec is like of in terms of high enough sample rate and resolution to capture micro- and milli-amp magnitude currents (in battery powered embedded circuits, which is my application).  That never happened and in the meantime I bought the DIG + MEM options myself.  I'm posting some results to show what can be achieved for anyone else who might be curious about this.

Test Setup
Keysight 34465A with the DIG and MEM options; valid calibration.  Current range set to 1A, and digitize option used to acquire readings at maximum 50k readings/sec for 40 seconds (total 2M readings).  The readings were saved to USB 2.0 drive as CSV file and later plot in MATLAB.

Device Under Test
Custom Texas Instruments CC3200 (MCU + WiFi) board.  Current measured at output of Li-Ion battery before input to LDO regulator. 

I attached 4 .png files which are MATLAB plots of Current [mA] vs Time [sec].  The images are of the same data set "zoomed" in on the x- and y-axis to highlight successively finer details. 

Image 01 10 sec duration showing system starting up; lots of current spikes while connecting to WiFi.
Image 02 1 sec duration of system in "idle" state.  Possible to identify 3 different periodic tasks: 1) MCU wakes at 4Hz (~20mA "blocks"); 2) WiFi applications processor responding to maintenance-type packets from wireless router at 2 Hz (~350mA "spikes"); and 3) A sensor running asynchronously from MCU at ~8Hz (~250mA "spikes").
Image 03 0.1 sec duration zoom-in to the MCU wake; can see various levels of activity as the system comes out of sleep and then returns to sleep.
Image 04 0.01 sec duration zoom-in showing an unknown periodic (~1.56kHz) current waveform.  I really wonder what part of TI CC3200 MCU is causing this, and why it's not staying at the lowest current draw!

Overall I have found the DIG and MEM options extremely useful in validating correct power management and sleep functions.  The ability to "zoom in" at many levels and see finer details is really remarkable.  I did not really expect this when using the 1A range.  It has enabled me to identify various sources of current draw and implement software changes leading to lower average current draw.  It is a bit of a challenge -- at each stage that you identify and then eliminate the biggest current draw you can see, you can then target the next one! (It's a tale of diminishing returns, but you get the idea!) 

Would still love to hear from HighVoltage and anyone else who has used the DMM for these kinds of applications --- it's very interesting seeing exactly how your system is operating in terms of the current waveform!

Have fun!
Sparky

[SharpEars]

A new FW for the +3ppm failure will be released in the next days / weeks, it's currently undergoing their tests.

Frank

It is January 2016 now and it doesn't look like the firmware that fixes this has been released yet. Am I correct in my assessment that the 34465A units are still not functioning properly as you described using current firmware 2.09? I am holding off on my purchase until Keysight fixes this issue.

[Dr. Frank]

A new FW for the +3ppm failure will be released in the next days / weeks, it's currently undergoing their tests.

Frank

It is January 2016 now and it doesn't look like the firmware that fixes this has been released yet. Am I correct in my assessment that the 34465A units are still not functioning properly as you described using current firmware 2.09? I am holding off on my purchase until Keysight fixes this issue.

I'm surprised also, that KS has not yet published 2.10 officially.

Maybe they want to solve other problems also, like the continuity function. Or maybe they discovered, that the 10ppm shift by ACAL really is not working correctly.

Anyhow, if you need this instrument, you may buy it despite these bugs.. it's still performing well inside all spec limits.

Frank

[rosbuitre]

New version of firmware v2.11

http://www.keysight.com/main/software.jspx?ckey=2367633&lc=eng&cc=US&nid=-536902435.1058513&id=2367633

Regards

[HighVoltage]

Fixes from the release note:

[blackdog]

Hi,


Loaded the new firmware on two 34461A, continuity check is stil a mess, morons....

Kind regarts,
Blackdog

[sotos]

Hi,


Loaded the new firmware on two 34461A, continuity check is stil a mess, morons....

Kind regarts,
Blackdog

There’s nothing in the new release notes mentioning the beeper problem fixed.

Maybe it’s difficult for them.

[blackdog]

Hi Sotos, :-)

I know there is no mention in the release notes about this, still i think morons! 

Kind regarts

[sotos]

Hi Sotos, :-)

I know there is no mention in the release notes about this, still i think morons! 

Kind regarts

         

[KE5FX]

Hi,
Loaded the new firmware on two 34461A, continuity check is stil a mess, morons....

Agreed, this is a bit disappointing. 

Come on, guys, this has got to be a pretty easy one to fix.  The continuity tone was better before. 

[sotos]

About the continuity tone, I also posted the problem at their forum here.

http://www.keysight.com/owc_discussions/forum.jspa?forumID=123

Maybe they are made in China and we have to inform them in Chinese’s.

[TheSteve]

I've posted in the Keysight thread - I hope others do the same.

[diyaudio]

I've posted in the Keysight thread - I hope others do the same.

This is very dishearten news. I have a 34465A 34461A  arriving next week, hopefully these firmware issues will be fixed soon... I still need to learn the functions on the instrument. 

 

[HighVoltage]

The continuity beep problem seems to be limited to the 34461A
My 34470A does not show this issue at all.
I would suspect the 34465A also to be OK, may be someone can confirm this.

[diyaudio]

Sorry, I keep thinking 34465A, im getting a 34461A.

[sotos]

With Firmware, AGILENT V 1.10 its working fine, so I don’t think it’s a hardware problem. It gets messed up with Keysight updates.

[Sparky]

@HighVoltage

I believe you have the DIG option, if so, and you use BenchVue also can you test the maximum sample rate when using BenchVue "Digitize" mode?  The reason I ask is that when using firmware 2.09 on my 34465A I could set 50kHz (max) from within the BenchVue Digitize app, but now (using firmware 2.11) the maximum appears to be 25kHz!  It is still possible to set 50kHz from the front panel of the meter for "local" control, but when using BenchVue for remote control I get an error that "50kHz is greater than the maximum of 25kHz".  (I have both DIG and MEM options).

Could you --- or anyone else --- check and confirm?

[HighVoltage]

Hello Sparky
Yes, I can test this on my 34470A but I am gone from the lab for a couple days, will test this early next week.
I did update the FW to 2.11 already on my 34470A meter.

[HighVoltage]

@HighVoltage

I believe you have the DIG option, if so, and you use BenchVue also can you test the maximum sample rate when using BenchVue "Digitize" mode?  The reason I ask is that when using firmware 2.09 on my 34465A I could set 50kHz (max) from within the BenchVue Digitize app, but now (using firmware 2.11) the maximum appears to be 25kHz!  It is still possible to set 50kHz from the front panel of the meter for "local" control, but when using BenchVue for remote control I get an error that "50kHz is greater than the maximum of 25kHz".  (I have both DIG and MEM options).

Could you --- or anyone else --- check and confirm?

I just tested this and come to the same results that you have.
I also have the DIG and MEM option installed in my 34470A.
Since I have already updated the FW to v2.11, I no longer can say, if this worked in v2.09
Hopefully Keysight will fix this in the FW or in BenchVue.

[Sparky]

@HighVoltage

I believe you have the DIG option, if so, and you use BenchVue also can you test the maximum sample rate when using BenchVue "Digitize" mode?  The reason I ask is that when using firmware 2.09 on my 34465A I could set 50kHz (max) from within the BenchVue Digitize app, but now (using firmware 2.11) the maximum appears to be 25kHz!  It is still possible to set 50kHz from the front panel of the meter for "local" control, but when using BenchVue for remote control I get an error that "50kHz is greater than the maximum of 25kHz".  (I have both DIG and MEM options).

Could you --- or anyone else --- check and confirm?

I just tested this and come to the same results that you have.
I also have the DIG and MEM option installed in my 34470A.
Since I have already updated the FW to v2.11, I no longer can say, if this worked in v2.09
Hopefully Keysight will fix this in the FW or in BenchVue.

Thanks HighVoltage for following-up and confirming the findings -- we can say this is an issue for both 34465A and 34470A. 

I posted in the BenchVue thread about this issue and one of the Keysight folks confirmed the issue yesterday on 34470A; he said he will look into it.  Interestingly they said the 25kHz maximum was the case for both v2.09 and v2.11 on a 34470A.

[Dr. Frank]

@HighVoltage
34465A I could set 50kHz (max) from within the BenchVue Digitize app, but now (using firmware 2.11) the maximum appears to be 25kHz!  It is still possible to set 50kHz from the front panel of the meter for "local" control, but when using BenchVue for remote control I get an error that "50kHz is greater than the maximum of 25kHz".  (I have both DIG and MEM options).

Could you --- or anyone else --- check and confirm?

I have a 465A, with DIG option, which is identical to the 470A in this instance.
I already upgraded to FW 2.10 (yep, I got this inofficial version), when I later also received the DMM PRO version of BenchVue.
This is the prerequisite for 50kHz digitizing over the bus. Also, I got no MEM option.

So I tested this 50kHz Digitizing with a few 100k of sample points.. over GPIB .. and had problems.
The sampling rate was set back to 25kHz in most cases, but that completely transferred all the data directly to BV, so that the internal 50k sample memory is no limitation.

By some odd tries with BV, depending on the pre-configuration on the 465A, which I can't recapitulate at that moment, I really managed to get 50kHz sampling rate directly to BenchVue.
But this was really stuttering.. display of BV freezing, and so on.
But it seemed to have transferred all data in the background.

Anyhow, I'm under the impression that the whole digitizing state machine of the 34465A/470A is too slow, and also that the bus transfer might be too slow.
Maybe BV itself has too much overhead to be optimal for fast digitizing.. although it's been promoted to do the job.

I have to point out, that I have some - successful - experience on the 3458A, digitizing 100kHz @ 16bit / 50kHz @ 18bit..
by using 2 Byte / 200kHz or 4 byte / 200kHz integer format transfers over GPIB.

Also, the former 34411A is optimized for such fast data-rates and - transfers, but this instrument provides a 4byte integer transfer, afaik, whereas the 465A / 470A only have an 8 byte real format.
As the 465A and 470A were promoted as successors to the 411A, I really wonder, if an improvement at that point is necessary, either an optimization of internal data crunching, but also formatting to 2Byte / 4 Byte integer.
A brief look into the different data transfer rate specifications of these instruments is also very instructive..

I already planned to write my own 50kHz digitizing program, w/o the overhead of BenchVue, to improve speed & buffer depth.
Got no time yet, but will report here, as soon as I'm successful.

Frank

[Sparky]

@HighVoltage
34465A I could set 50kHz (max) from within the BenchVue Digitize app, but now (using firmware 2.11) the maximum appears to be 25kHz!  It is still possible to set 50kHz from the front panel of the meter for "local" control, but when using BenchVue for remote control I get an error that "50kHz is greater than the maximum of 25kHz".  (I have both DIG and MEM options).

Could you --- or anyone else --- check and confirm?

I have a 465A, with DIG option, which is identical to the 470A in this instance.
I already upgraded to FW 2.10 (yep, I got this inofficial version), when I later also received the DMM PRO version of BenchVue.
This is the prerequisite for 50kHz digitizing over the bus. Also, I got no MEM option.

So I tested this 50kHz Digitizing with a few 100k of sample points.. over GPIB .. and had problems.
The sampling rate was set back to 25kHz in most cases, but that completely transferred all the data directly to BV, so that the internal 50k sample memory is no limitation.

By some odd tries with BV, depending on the pre-configuration on the 465A, which I can't recapitulate at that moment, I really managed to get 50kHz sampling rate directly to BenchVue.
But this was really stuttering.. display of BV freezing, and so on.
But it seemed to have transferred all data in the background.

Anyhow, I'm under the impression that the whole digitizing state machine of the 34465A/470A is too slow, and also that the bus transfer might be too slow.
Maybe BV itself has too much overhead to be optimal for fast digitizing.. although it's been promoted to do the job.

I have to point out, that I have some - successful - experience on the 3458A, digitizing 100kHz @ 16bit / 50kHz @ 18bit..
by using 2 Byte / 200kHz or 4 byte / 200kHz integer format transfers over GPIB.

Also, the former 34411A is optimized for such fast data-rates and - transfers, but this instrument provides a 4byte integer transfer, afaik, whereas the 465A / 470A only have an 8 byte real format.
As the 465A and 470A were promoted as successors to the 411A, I really wonder, if an improvement at that point is necessary, either an optimization of internal data crunching, but also formatting to 2Byte / 4 Byte integer.
A brief look into the different data transfer rate specifications of these instruments is also very instructive..

I already planned to write my own 50kHz digitizing program, w/o the overhead of BenchVue, to improve speed & buffer depth.
Got no time yet, but will report here, as soon as I'm successful.

Frank

Very interesting to read your experience, Dr. Frank. 

I really hope that 34465A with the DIG option (equivalently 34470A) will allow a continuous 50kHz data streaming to PC over USB or LAN.  If BenchVue could do it that would be great, otherwise a custom program with less overhead.  I will experiment with MATLAB sometime and see what I can achieve, and keep an eye out for your report