GPIB practical sampling speed Qn

[3roomlab]

i am curious to know @ 10 NPLC or 100 NPLC. how fast a sampling rate can you fetch readings (specifically NPLC 10 / 100, on GPIB only), could it do so effortlessly at 5Hz or more?

[bingo600]

Do the math

100 NPLC@50Hz = 100 * 1/50 = 100 * 20ms = 2 sec

100 NPLC@60Hz = 100 * 1/60 = 100 * 16.667ms = 1,667 sec

It would be hard to match 5Hz at that rate.

@10NPLC it is "just" possible (50Hz) , and GPIB certainly isn't the limitation.

/Bingo

Ref:
http://cp.literature.agilent.com/litweb/pdf/5989-4879EN.pdf

[3roomlab]

ok i forgot NPLC is based on AC cycle as base "clock" thanks (me getting very very forgetful  )
i was curious because, on RS232, max i could do on NPLC 10 is about 1.5Hz. i guess thats about it.

[bingo600]

ok i forgot NPLC is based on AC cycle as base "clock" thanks (me getting very very forgetful  )
i was curious because, on RS232, max i could do on NPLC 10 is about 1.5Hz. i guess thats about it.

Are you having a filter/avg on ?
Could that slow down things ?

I'd check the triggering , and maybe consider reading samples to "Mem" , and offload from mem via rs-232 or GPIB.

@9600 bps , you'd get ~900 chars/s , and if a reading takes up 10..15 chars , that would be a lot more than 1.5Hz.

For relally fast xfers , most meters have binary readouts.
Saves cycles in converting to ascii (float) , and saves byes as they're usually shortern than the ascii representation.

/Bingo

[3roomlab]

nope, no filter, no averaging

when i first succeeded in using python - RS232 - DMM, i thought it was bitrate, it did help in 1NPLC. i am using 19.2kbps btw.

but in 10 NPLC, when remote set it up, the internal delay is very slow, to get readings out it seems. so i was wondering if other K2000/2015 users see its reaction similarly.

hmmm i didnt try binary/mem mode. maybe i will study it.

in RS232/python, i find that longer NPLC seems to require more delay for the MCU to compute readings. so i was wondering if other keithley users are also experiencing it the same way 

[Kleinstein]

In many DMMs there is also a auto zero cycle. This can be as long as the reading itself and add an extra delay.

Usually GPIB is rather fast (near 10 MBit/s if I remenber right), in the old days there where even hard disks with GPIB interface. Some Interface cards (especially via UAB) may be slower, but this schould not be a problem even at 1 NPLC.