-
or what is your opinion?
Ah, I see what you're saying now. Personally, I rarely use trend charts, and I don't know if I would ever use one for the total count. It certainly doesn't look useful in its current implementation for Total. That said, it seems logical in that it only advances when a new count is received.
How would you prefer it works for Total counts? Would you want it to time out after a set period, but keep some record of what came before? At that point, it doesn't sound like a trend chart anymore. -
Maybe my expectation from a "free running counter" is false,
but have a look at the screenshot of a FM modulated sinewave.
(1kHz, +/- 100Hz)
I started the AWG, let it run for 30sec, stopped it for 10sec and startet it again for 15sec (so 55sec total)
Afterwards I would expect, that I can use the trend chart to analyse the signal change over time.
But the green trace of CH4 CNT-FRQ is showing a linear decline of the freqency to zero with a duration of 10sec,
that is not what actually happened.
(by the way, the period trace is showing the same as an incline, but I have clipped the visible range)
I added a red line how the signal should look like in my opinion.
If all the traces would be updated regardless of the trigger event,
the problem should be fixxed.
(and there is allready an interval configuration for the counter, so why not use this)
-
And there is also a different "problem" with the current approach
(check the screenshot)
Input signal is still 1kHz sine (FM modulated) +/-100Hz
Enable/disable of the AWG at some random moments.
If you time the trigger events just right, the first measured frequency value after the AWG is enabled again is not "0 Hz",
so the trace don't drop to "0 Hz" for this at all, as it should (marked in red).
I would call this a bad implementation.
-
First of all, thanks a lot for your posts!
All the points you mentioned actually come down to the same root cause: when the input signal is interrupted, no counter events are generated. As a result, the counter trend charts don’t receive any new data during that time. Once the signal comes back, new measurements are recorded again and added to the trend charts. The charts connect the last valid value with the next one by a straight line - which is usually desired, but in intervals with missing measurements it leads to those misleading transitions.
Your suggestion to keep the counter trend charts updating continuously (for example by holding the totalizer value or reporting a frequency of 0 Hz while no signal is present) makes a lot of sense and should solve all the cases you described.
We agree with you and will revisit this behavior to improve how the counter trend charts handle missing input signals. -
We agree with you and will revisit this behavior to improve how the counter trend charts handle missing input signals.
you realize it's a long weekend, right
maybe the current implementation is usefull for some cases and there could be an option to choose between the two variants,
but that is your job to figure out....thanks a lot for the quick response anyway.
and I wan't to to point out, that I like the new labeling of the x- and y-axis a lot better then the version before....much cleaner. -
Mathematically correct way to deal with missing data (signal was not present) would be not to show 0 Hz but to show gaps in graph.
There was no samples, so it is not 0 Hz it is empty for that timestamp. -
Mathematically correct way to deal with missing data (signal was not present) would be not to show 0 Hz but to show gaps in graph.
There was no samples, so it is not 0 Hz it is empty for that timestamp.
I agree; gaps make more sense. -
I see another strange behavior regarding the new DVM.
The min and max values are not "absolut" values...
with the meaning, that max is always increasing and min always decreasing
See a screenshot from a very slow (20mHz) step funktion (16 steps up, 16 steps down)
left hand side is the trend chart of a "normal" average value of CH4
right hand side is the trend chart of the DVM (250ms interval)
(both charts with enabled min,max,current)
Is it supposed to do that?
(maybe a dedicated "reset button" would be a better option)
And another thing:
at the moment there is no button to stop the DVM,
if you enable it, it runs forever...
in a trend chart not ideal, if you try to take a screenshot :-)
That issue will be prominent also for the counters,
if you change the current implementation as discussed earlier.
(so, maybe we need a STOP button for all the trend charts)
-
Do you mean the min/max value changes in an unexpected direction? If that's the case, then I would assume it's based on the set average number, not the total of the entire test run.
Besides that, tap on the measurement box and you get 3 icons: settings, reset, and trash. You can disable the DVM by tapping the trash icon.
ETA: A popup menu icon in each trend chart is a good idea with similar ideas.
-
I'm a little confused by your post. What else would you expect min/max values to do if the min/max values keep changing?
Have you compared both trend charts?
They are getting the same input data, so why is the output result different?
(maybe it's a feature of the DVM, I only want to know) -
You can disable the DVM by tapping the trash icon.
yes, I know, but I wanted to take a screenshot of the trend charts
and befor I got to the right button, the charts have changed....so I had to start over again....
not a big deal, but anoying. -
I'm a little confused by your post. What else would you expect min/max values to do if the min/max values keep changing?
Have you compared both trend charts?
They are getting the same input data, so why is the output result different?
(maybe it's a feature of the DVM, I only want to know)
I edited my response. I think you're talking about the trend chart showing the complete data, versus the left column showing data based on the set average? Or am I reading it wrong?
I definitely agree that the trend charts should get some buttons too.
Thanks,
Josh -
sorry, it's late I try a different explanation:
If I want to use a DVM to measure the maximum voltage of a signal (or the highest voltage of a bunch of signals)
I enable the DVM, probe all the signals one by one...and at the end, after some time, I read the stored "max value" from the display.
I don't want the DVM to reduce the "max value", automaticaly after some time, back to "0V" only because I don't apply an input voltage anymore.
If I want to start a new measurement series, I press a "clear" button and set max/min manualy back to "0V".
I hope that makes my point clearer. -
sorry, it's late I try a different explanation:
If I want to use a DVM to measure the maximum voltage of a signal (or the highest voltage of a bunch of signals)
I enable the DVM, probe all the signals one by one...and at the end, after some time, I read the stored "max value" from the display.
I don't want the DVM to reduce the "max value", automaticaly after some time, back to "0V" only because I don't apply an input voltage anymore.
If I want to start a new measurement series, I press a "clear" button and set max/min manualy back to "0V".
I hope that makes my point clearer.
Ok, so I think I understand what you're saying. For example, if you're measuring 10V, max says 9.9999, and eventually it changes in the wrong direction, and then max says 9.9889, is that correct?
The DVM min/max values on the left measurement column by default are referencing a fixed Average N value set, which default is 128.
Change the setting to Running average instead of Fixed. Then you should see the behavior you want.
-
Ok, so I think I understand what you're saying.
I think not
another screenshot,
very easy...I measured with CH1 the voltage of an external DC battery (1.5V)
and I want to know the absolute peak voltage.
I enabled "normal" CH-1 Vtop measurement and a DVM for the same channel.
(In the trend chart, I enabled for both the max and current trace)
then the measurement:
first I waited 20sec, then I measured for 15sec the battery voltage, then I waited again for 55sec.
Result with a "normal" CH-1 Vtop measurement = CH-1 (max) = 1.499 V
Result with DVM measurement = CH-1 (max) = 13.596 mV
And you can see in the trend chart of the DVM, that the stored max value,
is reset zo "0V" after ~30sec with no input signal.
So this is not a "max value" but a "highest value from the last 30sec".
And by the way @Batronix, there is another bug with the setting of the DVM vertical unit grid.
To get 1V/div you have to choose 100mV/div.
-
Regarding the trend charts, I think it would be a good idea to be able to export the data. That would also provide a data logger. I think it should not be too complicated to implement this - and yes, the data would be somehow limited.
As for "stopping" the DVM, I think it would be even more useful if the frequency counter didn't display "nothing" for the current value when the signal is absent, but instead showed the last measured value, for example, in red. This way, you can see that the counter isn't currently measuring anything, but you can still see the last value.
Furthermore, a setting to optionally stop not only the DVM but also the frequency counter when the scope is in "stop" mode (and then display the measured values in red) would be helpful. Imagine running four counters in parallel, each with, say, 15-digit resolution; you simply wouldn't be able to read and record the measured values quickly enough. In that case, the "stop" function (with the display) would be very useful.
Regards Thomas -
Ok, so I think I understand what you're saying.
I think not
another screenshot,
I think so.
Try what I suggested and change to a Running Average and that should give you the results you want.
From your screenshot, Count shows 128 = Fixed Average. That means it will disregard the data prior to the current sample set of 128 counts.
Thanks,
Josh
-
Regarding the trend charts, I think it would be a good idea to be able to export the data. That would also provide a data logger. I think it should not be too complicated to implement this - and yes, the data would be somehow limited.
As for "stopping" the DVM, I think it would be even more useful if the frequency counter didn't display "nothing" for the current value when the signal is absent, but instead showed the last measured value, for example, in red. This way, you can see that the counter isn't currently measuring anything, but you can still see the last value.
Furthermore, a setting to optionally stop not only the DVM but also the frequency counter when the scope is in "stop" mode (and then display the measured values in red) would be helpful. Imagine running four counters in parallel, each with, say, 15-digit resolution; you simply wouldn't be able to read and record the measured values quickly enough. In that case, the "stop" function (with the display) would be very useful.
Regards Thomas
I dunno anything about how difficult it might be to implement, but I think having History mode include the counters/measurement data would make that easier from a user standpoint. Maybe an optional mode to include that. -
Here are some examples with the measurements set to Running Average over a slightly longer time period.
The first screenshot is DC only, then the others are all from the same test with the internal AWG using an arbitrary wave with different DC offsets.
As can be seen, there is no time limit for the measurement data. The only restriction previously was using a fixed average.
Thanks,
Josh
-
Try what I suggested and change to a Running Average and that should give you the results you want.
thanks a lot....this makes the difference for the max value.
I thought "average type" is only affecting how the "average" is calculated,
sliding window or continuously.
just learned something new. -
I dunno anything about how difficult it might be to implement, but I think having History mode include the counters/measurement data would make that easier from a user standpoint.
something like that? (check the right hand side of the trend-chart)
(but that's not "history mode"....right?)
-
No, that's not history mode, but the trend chart is history I suppose.
In this attachment, the waveform is showing history but the measurement column is still showing live data.
Thanks,
Josh
-
Sorry, Can't follow you 😢 I don't see any history in this screenshot 🥸. The waveform display is live as is the measure data?! If in "history mode" this would be indicated by "stopped" (instead of "triggered") and the history menu at the bottom - wouldn't it?
Regards Thomas -
That was the wrong image. I'll blame the forum for that one.
Image corrected.
Thanks,
Josh -
now I need to delete my post
