Analog channels with v4.645 and non-A units

Analog channels with v4.645 and non-A units, a forum discussion on Cleverscope. Join us for more discussions on Analog channels with v4.645 and non-A units on our Questions forum.

Back to Forum Index : Back to Questions   RSS

14 Sep 2009
Posts: 22

Your latest releases until v4.645 were a great enrichment, escpecially
for us non-A users. I didn't except more features or further improvement,
like the persistence of signals, for our ""old"" units ;) Great!

Good news first:
This time, v5335 has working trigger 2 functionality ;)

The bad one:
Using v4.645 makes the analog channels appear with an amplitude of
~1/1000 of the actual waveform. The signals are _very_ noisy too.

Downgrading to v4.644-exe immediately solves this issue.


Axel
[editor:html]  

14 Sep 2009
Posts: 399

Hello Axel, Thanks for checking this out so quickly. The answer is that to use 5335 and 4.645 you need to do a 'Once Yearly' calibration followed by a 'Standard' calibration. We have made a change to the internal calibration table that makes 4.645 incompatible with previous versions. We have put much thought into this, but are still not totally convinced. This change allows us to increase accuracy in future iterations, and means we get sub mV DC offset now. Do you think we should be making users do these calibration steps? The Once Yearly calibration requires a 2.048 (or thereabouts) reference. If you do not have one, you will need to open the Cleverscope lid, and probe the built in reference (the Once Yearly dialog shows you a picture of where). Any comments (from anyone) welcome. We put a note on the front page about this, and also the update preferences file includes a 'Note' saying this. Unfortunately the updated preferences file is only used in XP. In vista we need an extra step, which we are implementing now.

15 Sep 2009
Posts: 22

I only had some precision current sources available so I looked out for the internal reference.
Well, the picture does not match my scope and I did not find the reference via...

After a little probing, I found another via beside the AD5318 DAC
which offered a sufficient voltage.

Okay, here we go...

At a glance, I have plausible values for both analog channels.
At least, far better than the three decade error before doing a calibration...

I repeated the calibration several times, used same probes on same channels
and measured the reference with a, not too bad, Fluke process meter.

Can you rate or comment these offset errors (rounded to 10mV):

A/B grounded; 1V/div; averaging on; 1x
Same scale but different offsets (scope window) applied:

scale mid -11 to 11V: A: -260mV B: +35mV
scale up 0 to -22V: A -1110mV B: -1120mV
scale down: 22 to 0: A: +580mV B: +1170mV

I could swear it was better in the past...
Note that the waveform itself (e.g. pk-pk voltage) is still valid but
e.g., a 2Vp-p waveform is shifted by 50%.

Additionally, the signal generator refused being calibrated.
CS claimed that there is no connection (and tells something about ~ -4000 digits).
(Fortunately, the output is plausible until now...)


Axel
[editor:html]  

16 Sep 2009
Posts: 399

Hello Axel,
Ok, we did make it hard for you. We have made a new version, 4646, which now includes a picture like it should have to show you where to probe, and hopefully, better instructions. We would really like you to try it. If this all turns to custard, we will go back to the old calibration table.

You will need to load firmware 5335 into the unit. Wait for a few minutes to warm up, and then do 'Once Yearly' calibration, followed by Standard. You can also do a Signal Generator calibration (sorry, we loaded the phase value wrong in one place).

Best offset results are achieved with the 20 Mhz filter on (note that you now have to set this up explicitly in the Settings/Acquisition Settings dialog, before setting the Filter On).

We achieved, for the CS328
Filter On Filter Off
+/-1 V, mid scale 400uV 1.6mV
+/-5V mid scale -9mV -13.8mV
+/12 V mid scale -3 mV -20 mV
0 -22V up scale -20 mV -47 mV
22 - 0V down scale -9 mV -24 mV

A and B are similar. We still have work to do here, but we think it is better than before.
Please let us know how it goes.

Bart

17 Sep 2009
Posts: 22

Ok, I upgraded to v4.646 but skipped the calibration because of the following:

I already noticed the missing function of the control panel ""Filter""
(was labeled 20[MHz] in the past).

The offsets reduce dramatically and immediatly only by activating
""Pre-Filter X"" in the settings menu (which was new to me).
Offsets now are within 2 digit value for full offset shifts (and within 1
for mid scale).

Well, what is the difference to the ""front panel"" button ""Filter""?
Your (2/2009) manual only mentions the good old ""20 MHz"" button ;)
Ok, I could test this out but I guess you can describe it much quicker
than I can measure...


But anyway, I am missing something again:
Display method does not work anymore.
No matter what you select, the display defaults to ""peak captured"",
except for ""Waveform avg"", which does absolutely nothing...


Axel
[editor:html]  

17 Sep 2009
Posts: 399

Axel, I really am not understanding this. The calibration table has changed. It will not be written to the Cleverscope acquisition unit in the correct format unless you do a Once Yearly, followed by a Standard calibration. So if you do not do a calibration it will never be right. Maybe we should give up on this.

The CS328A now has a moving average filter option as well as a FIR 20 Mhz filter. You can turn on none, one or both. The Filter button now activates whatever you have set up. If you set up nothing, there will be no filtering when you push the Filter button (maybe we should grey it).

I just tried all the Display methods and they all work (except Filtered which is a dummy we are going to make work soon). The manual still needs to be updated.

Please, please do the calibrations.

Thanks, Bart

17 Sep 2009
Posts: 22

Hi Bart,

possibly I did not express this clearly...
I indeed made a calibration before and, except for the offsets, everything
is working fine.

But these large offsets are _only_ there, if ""Pre-Filter X"" is _not_ set to ""20 MHz"".
No matter if you enable or disable this feature during calibration.

Note:
The waveform does not shift off immediatly. The large drift has a time constant
of about 0.5s.

Everything is working fine now, just by setting the pre filtering in the settings menu.


Except for the ""display method"". This really does not work on my system
(And it did indeed, work before. I used this quite often in the past)...


I am slightly confused (but still ready for operation =)



Axel
[editor:html]  

18 Sep 2009
Posts: 399

Hello Axel,
Thanks for this. OK I finally realize.
In Settings/Acquisition Setting, the Prefilter must be set to 20 Mhz, even if the Control Panel Filter setting is OFF to avoid the offsets. Bizarre. We will look into it and fix it.

If the Pre Filter is set to 20 Mhz, then the Filter Button does work correctly On/Off - right?

On the Display Method, it works fine here. So we have find out what is wrong.
Just to check - your app reports 4.646 (which supports sig gen cal for the CS328)
your firmware version is 5335?

Could you email a .apc file of your setup so we can try and recreate it here?

Thanks for taking the time.

Bart

18 Sep 2009
Posts: 22

Hello Bart,

yes, if ""Pre Filter"" is once set, the front panel button works as excpected.
And yes, I am running v4.646 and v5335.

I found out more about the display method ""bug"":
The protocol analyzer was active all the time.
I turned it off and - bingo - at least ""Sampled"" and ""Repetive"" were back again
immediately...

""Waveform avg"" still did not work but I also managed getting this one run:
I just updated the averaging values in the ""Settings/Averaging..."" menu.
I guess the upgrade left some invalid settings in the scope?

Well, everything is fine now!


Thank you very much
Axel
[editor:html]  

18 Sep 2009
Posts: 399

Hello Axel,
Fantastic. You have left us with very good clues for the next upgrade. The protocol analyser always uses peak-captured acquisition to ensure it does not lose clock/data transitions. We should grey the display method while doing this.
Thanks again.
bart
Back to Forum Index : Back to Questions   RSS
You must be logged in to post a reply



You need to Register or Log In before posting on these forums.

×

Your shopping cart is empty.