Critical Bug: Use of %f instead of %e for values truncates values

[maor1993]

Hey, I've been working with this library for a while using the DSOX1102G I noticed that once I use a value smaller than 1e-6, the scope does not receive it correctly and uses the smallest value possible.

This issue occurs as the python default %f max decimal points is 6:

The more concise solution would be to use %e which is always supported by SCPI devices:

Looking at the repo there are about 19 devices use %f and not %e

This is uncaught by testing as it was never tested for.

[BenediktBurger]

Thanks for raising this issue.

If your device is able to understand exponential notation, you can change it there.

I'm not sure, that every device understands exponential notation (not even all SCPI ones). Some might truncate at "e" producing undesired results, others might raise an error.
Therefore it is not that easy changing all occurrences.

Another idea: %g produces normal output (%f) if the exponent is between 5 and -4 and exponential output (%e) if outside this range.

Maybe we should put a corresponding note into the documentation as well.

[bilderbuchi]

Thanks for raising this! I agree that %g is probably the more sensible default choice.

For all instruments where you want to move away from %f, you'll have to check (with a device if available, or in the manual) if the resulting numbers format is accepted by the device.

[maor1993]

Thanks for the quick reply.
the list is as follows:

I think I might have a moment to check the programming manuals, will update.

[maor1993]

Okay done:

device	link	supports exp?	Notes
AWG401x	?	?	Can't find programming manual
agillent33220A	http://ecelabs.njit.edu/student_resources/33220_user_guide.pdf	YES
agilent33500	https://www.testequipmentdepot.com/usedequipment/pdf/33500b-33600a-series_manual.pdf	YES
agilent33521A	https://www.testequipmentdepot.com/usedequipment/pdf/33500b-33600a-series_manual.pdf	YES
anritsuMS9710C	https://dl.cdn-anritsu.com/en-au/test-measurement/files/Manuals/Operation-Manual/MS9710C_W1580AE_remote_opm_e_4_0.pdf	YES
danfysik8500	https://www.danfysik.dk/media/1182/software-manual-sys8500.pdf	NO?
keithley24xx	https://download.tek.com/manual/2400S-900-01_K-Sep2011_User.pdf	YES
keithley6517b	https://download.tek.com/manual/6517B-901-01D_Feb_2016.pdf	YES*	some fields do and some don't
keysightDSOXxx	https://www.keysight.com/il/en/assets/9018-07554/programming-guides/9018-07554.pdf	YES
keysightN7776C	https://www.keysight.com/il/en/assets/9018-70058/programming-guides/9018-70058.pdf	YES
ips120_10	https://muquip.triumf.ca/sites/default/files//instruction_manuals/mnl_spec_bNMR_oxford_powerSupply_IPS-120-10.PDF	NO	Uses decimal point only
itc503	http://lmu.web.psi.ch/docu/manuals/bulk_manuals/OxfordInstruments/ITC503.pdf	NO	Uses decimal point only
cnt91	https://www.manualslib.com/manual/1375130/Pendulum-Cnt-90.html	YES
sr570	https://www.thinksrs.com/downloads/pdfs/manuals/SR570m.pdf	NO
sr830	https://www.thinksrs.com/downloads/pdfs/manuals/SR830m.pdf	NO
sr860	https://www.thinksrs.com/downloads/pdfs/manuals/SR860m.pdf	NO
ibeamsmart	?	?	Can't find programming manual

[LongnoseRob]

@maor1993 AGW401x - is probably this product.
But I could not find a public available programming manual on manufacturers website

[BenediktBurger]

Let's put a hint in #779 regarding small numbers.

[BenediktBurger]

Thanks for looking up the manuals.

Would you mind opening a PR with the changes, @maor1993 ?

You do not have to change all instruments possible.
I'd guess, that it does not occur that often, that you want to set a value to below 1e-5. And if it occurs, we can still fix that device (with a person being able to test the change).

It is more important, that devices continue to work (down to 1e-5) than to break at any input. If you're in doubt, whether it will work for a given instrument, leave as is.