Wittig(welec) DSO W20xxA Open Source Firmware (Teil4)

>Wenn du dich anmeldest, dürfte das wegen der Seitenaufteilung locker
>eine Faktor 5 und mehr sparen.

Ach ja, stimmt, hatte ich verdrängt... aber Zeit für Teil 5 ist es aber 
wohl trotzdem.

Viele Grüße,

egberto

Hayo W. schrieb:
>> Und warum ist "External Trig" jetzt erstmal gesperrt?
> Hallo Jürgen,
>
> ich wollte das gerade fixen, stelle aber fest das bei mir alles richtig
> ist und der Eintrag nicht gesperrt ist. Ist das bei Dir noch aktuell?
>
> Gruß Hayo
>


Das gibt es doch nicht! Jetzt ist es bei Edge freigegeben (W2024A). Ich 
bin mir ziemlich sicher, daß es gestern beim F2-Button war, d.h. bei 
Edge.
Sorry, vielleicht doch schon das Alter :-(
Jedenfalls funktioniert es jetzt.

Gruß Jürgen

Hallo Hayo,
na also, es scheint bei mir doch noch alles i.O. :-O

Nach Default Setup ist bei Edge External gesperrt. Fragt nicht, wie man 
das dann freigeben kann...

Hayo W. schrieb:
> Bei Pulsewidth lasse ich es aber gesperrt, oder?
>
> Gruß Hayo

Da fehlt meiner Ansicht nach noch ein Stück Programm. Ich kann es auch 
schlecht testen, wenn ich übers Menü nicht dorthin komme...
Vielleicht kannst Du es erstmal freigeben und wir sehen dann weiter?

Gruß Jürgen

Sieht doch gut aus? Es sollen doch immer möglichst viele
Daten angezeigt werden, deshalb machen das eigentlich alle
Speicheroszilloskope so.

Grüße, Guido

Hallo,

kennt Ihr die Anschlusseinstellungen ? Ich bekomme keine Verbindung mit
dem Updateprogramm von Welec.

Es ist allerdings ein RS232-USB-Adapter vor dem Notebook eingesteckt.
Verschiedene Baudraten und Einstellungen brachten keinen Erfolg, bleibt 
wohl nur noch am PC mit RS232-Schnittstelle auszuprobieren ?
Danke

Gruss

Rolf

Ich vergass: es ist Windows7 auf dem Notebook.

Hi Hayo, Jürgen, and guys all!

I am back after a long time.
I am in a hurry now, but I want to say some things.
Necessarily I will be short. ;-)
First I have to thanks all You and especially Hayo for the great work 
You are doing on Welec's oscilloscopes: no words, RESPECT!
I installed the latest firmware release 1.2.BF.4.8 and by the way I 
noticed some things who I report follow here in order to improve them: 
mine are not critical to the work done but only for knowledge.
Again for knowledge, I specify that all screenshots were made ​​with a 
W2012A.

@Hayo
When QM is on, changing Time Base sometimes get corruption on the screen 
who can be removed by setting to default or performing reset sequence or 
again by switch between measure's screen to the informations screen: 
please look at the screenhot CORRUPTED_1.jpg, CORRUPTED_2.jpg, and 
CORRUPTED_3.jpg.

I know the FFT have to be finished in order to improve it, but seems to 
me the division's value for either df, div and fN are wrong: please look 
at the screenshot DIMENSION div fN.jpg, DIMENSION fN div.jpg, DIMENSION 
div_fN.jpg and DIMENSION fN_div.jpg.

Again for the FFT when You switch to Display in order to change 
graticule's dimension, then Grid up to 100% without intensity change and 
turning the knob the grid's intensity increase but never is possible to 
set it down, it is stuck until perform default setting: please look at 
screenshot FFT Grid Level.jpg.

About the Quick Measure I noticed when some values are below certain 
levels then zero is showed: is not possible to increase recognized level 
in order to show a more low level?
For some example please look at the screenshot ZERO in QM.jpg and ZERO 
in  QM_.jpg.

About the Hardware ADC Setup, I tried some settings because the behavior 
of both mine either W2022A and W2012A prevents to use Hig Freq's 
setting.
So right choice is a bit difficult to make for me.
Please refer to the screenshot Factory.jpg, High Freq.jpg, Test 1.jpg, 
Test 2.jpg, Test 3.jpg, Test 4.jpg and Test 5.jpg.
Is not there a precise way to make a choice?
How can I know if an unknown signal actually appears on screen as in 
real life or only for the wrong setting of the ADC Setup?
Can I get trust or I must to hope it is a right setting?

@Hayo, Jürgen
As I understand You are trying to further improve either the internal 
trigger and external, I repeat some things I already wrote.
Seems when the trigger's threshold is low the waveform is moving and the 
signal's amplitude is a bit more low compared to when the trigger's 
treshold itself is high and so the signal's amplitude becomes it itself 
more high and the waveform is more steady on the screen.
Please refer to my former post and here screenshot Trig Lev High.jpg and 
Trg Lev Low.jpg.

@All
Is there somebody noted that signal amplitude seems to depend on the 
trigger treshold?
Is there somebody noted that when the trigger's threshold is low 
amplitude waveforms on the screen are dancing while when it is much 
higher the waveforms are much more steady on the screen?

@Jürgen
Can You kindly explain further about the C37 matter?
What are the benefits?
And contraindications?
Seems to me the modification need for a software support, is not so?
Thanks in advance Jürgen.
Perhaps the C37 modification together with the change of line and 
termination resistance for the MAX121 are the minimun to do on the 
original hardware in order to improve the performances if the Piggyback 
board is not installed.
Of course firmware is much important too.

Hayo W. schrieb:
>- Last but not least - die Rechenleistung unserer kleinen CPU ist mit
>drei gleichzeitigen Messungen schon so vollauf beschäftigt, dass eine
>weitere Messung den Ablauf fast ganz zum Stillstand bringen würde.
I connect here in order to make some considerations.
Of course what You say is right, fast response it is fundamental 
otherwise the device becomes unusable and useless.
In the present and in the past all yours efforts were aimed at improving 
the device without make it unusable.
It is can say the goal was reached, infact now the Welec DSO are much 
more similar to Agilent and other blasonate DSO's brands and I do not 
only refering at the features but to the performances too.
So there was, there is and there will, I hope, a great improve on either 
estetic and features: I like very much new configurable screen setting 
as well some killer applications as Overlay, single shot configuration, 
new mathematical functions and many others.
Said this, I think perhaps
> Ich hätte da mal einen Vorschlag:
> im Quickmeasure sowie im Curserbereich, stehen fett und breit, 3x
> Messbereiche zur Verfügung!
> Wäre es möglich, diese Anzeigen um ein Drittel (Breite)zu verkleinern
> und einen vierten Messbereich dazu zzu nehmen? Genug Platz ist ja
> voranden.
it do not kill ours DSO, is not so?
I know, this and other things are already in the wish list, but since 
there was also talk of aesthetic improvements I wanted to remember also 
this.
About the wish list:
What about the sin(X/X) interpolation implementation?
And centre and span features for the FFT?

Now I must to run away, so thank You to all You.
Vielen Dank!!!

Gruß Luc

Luc schrieb:

Hi Luc,

> @Hayo, Jürgen
> As I understand You are trying to further improve either the internal
> trigger and external, I repeat some things I already wrote.
> Seems when the trigger's threshold is low the waveform is moving and the
> signal's amplitude is a bit more low compared to when the trigger's
> treshold itself is high and so the signal's amplitude becomes it itself
> more high and the waveform is more steady on the screen.

Yes, this is the same on internal and external trigger. When the 
trigger's treshold is high the waveform is more steady on the screen. On 
the internal trigger the trigger logic sees more noise and this is why 
you have more trigger events on a zero trigger level. The waveform is 
moving on the screen.

>@Jürgen
>Can You kindly explain further about the C37 matter?
>What are the benefits?
>And contraindications?
>Seems to me the modification need for a software support, is not so?
>Thanks in advance Jürgen.

On external trigger this new capacitor to C37 reduce the amplitude of 
the reference voltage of the external trigger comparator U6, because 
this reference voltage is build by a PWM signal and a low pass filter. 
So there is with this additional C a more stable timing and the 
waveforms are not so much "dancing".
There are no con's. You can use this hardware modifcation also with the 
original firmware 1.4.

Regards, Jürgen

Hi Hayo and guys all!

Hayo W. schrieb:
>Yes, I know, You are allways short ;-)
Oh man, as usual You are right! :-)
You certainly know your chickens! ;-)

Hayo W. schrieb:
>> When QM is on, changing Time Base sometimes get corruption on the screen
>Yes I noticed this too and I'm searching for the reason. I hope I can
>fix it soon.
For sure you will succeed, I will bet on it!

Hayo W. schrieb:
>> I know the FFT have to be finished in order to improve it, but seems to
>> me the division's value for either df, div and fN are wrong:
>I guess the values are correct. For example:
>Sample Rate 1GSa/s -> fN (span) = Sample Rate/2 = 500MHz
>Points = 512 -> Spectrum = 512/2 = 256. Grid width is 512 -> one
>frequency line is 2 pixels wide.
>div = fN/10 = 50MHz = 50 Pixel wide.
>df = div/25 = 2MHz (one line is 2 pixel wide at 512 point FFT)
>am I right?
Yes of course I think You are right, but I was referring to numerical 
values ​​in the lateral table and not to the physical size of the 
graticule.
Looking at screenshot DIMENSION_div_fN.jpg with 1Gsa/s and 512pts it is:
df=1.95MHz
div=50.02MHz
fN=500.24MHz
And further looking at screenshot DIMENSION_div_fN.jpg with 500Ksa/s and 
512pts it is:
div=977.5Hz
div=25.01KHz
fN=250.12MHz
Seems to me that the accounts do not come back, is not so?
Maybe I am wrong, otherwise sorry, I do not understand.

Hayo W. schrieb:
>> Again for the FFT when You switch to Display in order to change
>> graticule's dimension...
>Yes You are right, there is a bug. I working on it. Thanks.
Thank You a lot Hayo!

Hayo W. schrieb:
>> About the Quick Measure I noticed when some values are below certain
>> levels then zero is showed:
>Hmmm, difficult. I think it is because of the interpolation. I will
>check if it can be improved.
You knew the job was difficult when you took it, Hayo! ;-)
Ok, no kidding anymore, I know it is not simple because in this case it 
is to push the measure under the nS value, pratically in the pS's field: 
really not easy!

Hayo W. schrieb:
>> About the Hardware ADC Setup, I tried some settings because the behavior
>> of both mine either W2022A and W2012A prevents to use Hig Freq's...
>Yes this is a Problem. The adjustments with the best HF-behaviour have a
>problem with spikes which seem to be generated by wrong FPGA timing. So
>I called them Test 1 and Test 2 because they seem to be more
>experimental, but may be helpful at very high frequencies. On normal
>purposes the factory setting should work ok.
Yes, thank You for the explanation, Hayo!
The real problem is some setting are ok with waveforms as for example 
the inner probe compensation's generator but are worst with other kind 
of waveforms as for example pulses.
Not easy set it correctly because could happen with a not suitable 
waveform.
And yes, Factory setting work well apart some spikes problem in some 
circumstance, but upon a time with a former firmware release High Freq 
setting was the better choise for all the situations.
I am sure one of the setting it is the right answer also this time, I 
will pursue it!

Hayo W. schrieb:
>> Seems when the trigger's threshold is low the waveform is moving
>correct, trigger is more stable when the level is not zero but a little
>bit higher. The wave form differs because of the zoom effect of some
>timebases. at 50ns/div the zoomfactor is 1 and it shouldn't occur.
Thank You  very much also for this explanation, Hayo!
I guess this even affects the amplitude.

Hayo W. schrieb:
>> About the wish list:
>> What about the sin(X/X) interpolation implementation?
>I investigated in this theme but it is not trivial. Unfortunately I
>didn't find a good implemention example (as code), but only theoretic
>discriptions. I'm working on it...
Hayo, this is really an amazing news, thanks!
I guess this implementation can improve a lot the shape of the waveforms 
shown on the screen.

Hayo W. schrieb:
>> And centre and span features for the FFT?
>Yes this is also on my list and also a 2048 point FFT if possible with
>the CPU performance.
Oh boys, another great amazing news, thank You Hayo, You are really the 
best one: RESPECT!

Hayo W. schrieb:
>Thanks for Your short report,
Ok, ok, actually it is not so short! ;-)
Instead, thanks to You Hayo for the great patience and kindness and free 
time You provided generously to us!!!
RESPECT!!!!!!!
Vielen Dank!!!!!!!

Mit freundlichen Grüßen,

Luc

Hi Jürgen and guys all!

Jürgen schrieb:
>> Seems when the trigger's threshold is low the waveform is moving and the
>> signal's amplitude is a bit more low compared to when the trigger's
>> treshold itself is high and so the signal's amplitude becomes it itself
>> more high and the waveform is more steady on the screen.
>
>Yes, this is the same on internal and external trigger. When the
>trigger's treshold is high the waveform is more steady on the screen. On
>the internal trigger the trigger logic sees more noise and this is why
>you have more trigger events on a zero trigger level. The waveform is
>moving on the screen.
Thank You very much for the explanation, Jürgen: RESPECT!

Jürgen schrieb:
>>Can You kindly explain further about the C37 matter?
>
>On external trigger this new capacitor to C37 reduce the amplitude of
>the reference voltage of the external trigger comparator U6, because
>this reference voltage is build by a PWM signal and a low pass filter.
>So there is with this additional C a more stable timing and the
>waveforms are not so much "dancing".
>There are no con's. You can use this hardware modifcation also with the
>original firmware 1.4.
Very well Jürgen, I thank You a lot for these useful informations: 
RESPECT!!!!!!!
For me, and I guess for somebody else also, this is really a very 
interesting thing.
Going ahead, in relation to what You wrote this means there are other 
capacitors can be modified putting in parallel of them more capacitors 
in order to get a better behavior of the inner trigger, due C37 improve 
only the external one if I am no wrong.
So, which are the capacitors and what is the final capacitance's value?

Jürgen, thank You very much for the kind reply: RESPECT!
Vielen Dank!!!!!!!

Mit freundlichen Grüßen,

Luc

moin moin Hajo & Co.

wenn du die  Y pos auf die 2. unterste Linie stellst
und dann utility/calib. offset ausfuerst und dann mit
Y pos. fast ganz nach oben wanderst verdoppelt (ca.)
sich das dargestellte Rauschen ;(
ungekehrt ist der efekt aber ebebso (kalibrate ganz
oben).
vielleicht kann das mal noch jemand testen ob das
generell oder nur bei meinem so ist

vlG
Charly

Hi Hayo,  Rainer W., charly, Michael D., and guys all!

Hayo W. schrieb:
>Solved
Hayo, thank You very much for the fix: RESPECT!

Hayo W. schrieb:
>Solved -> measuring is now possible with pico second resolution.
Oh boys, this is really incredible and make me very happy!
Hayo You are the best one, no words: RESPECT!!!!!!!
I must to say the Welec oscilloscopes are a lot improved and now their 
are can compete with other blasonate DSO's brands on the market.
Today I finally managed to complete some measures with the sole help of 
my W2022A without necessarily having to use other tools.
All this thanks to your great work Hayo and all who are involved in the 
Welec project, so I thank all You again: RESPECT!!!!!!!

Hayo W. schrieb:
>The Test 1 is the old High Frequency setting. I only changed the name!
>The actual High Frequency Setting does not work so good as the old one
>but doesn't has spikes.
Great news, thank You very, very much Hayo!
Also this information aid me very much, now I know what is the better 
choice even without necessarily take a look at all the options.
This is amazing for me also because I have not all the necessary 
strumentation in order to investigate the matter.
Perhaps is not it better return the name?
Anyway thank You again Hayo: RESPECT!

@Rainer W.
Selbst wenn ich zu spät Ich erwidere den schönen Sonntag. :-)
I hope to have been understood. :-)

@charly, Michael D.
Sorry but I do not understand very well.
I even have a W2022A and maybe I can check it, so please can You explain 
me what is the problem?
Thanks!

Vielen Dank!!!!!!!

Mit freundlichen Grüßen,

Luc

Hi Hayo and guys all!

@Hayo
Maybe I found a little bug in a 1.2.BF.4.8C2 firmware.
Setting channel delay in Hardware menu, values are not update by rotary 
knob but only by push the buttons, but still on this by changing the 
other delay's channel then get corruption on the screen's tag.
DSO is W2022A.

Mit freundlichen Grüßen,

Luc

@Hajo

nullinie stimmt schon, nur das rauschen verdoppelt sich,
i werd mal versuchen es mit bilder darzustellen sobald
i a bissel zeit habe (im moment muss i auch softw. f.
einen kunden anpassen)

@Luc
i will post a few pics so fast as possible

vlG
Charly

Hallo Jörg,

deine Probleme interessieren uns nicht die Bohne. :-))

Wir wollen wissen, wie es mit der Huckepackplatine aussieht!

Hast du HF-Messmöglichkeiten? Ich meine ja, sonst müssten wir
noch etwas unternehmen.

Grüße, Guido

@Guido

>deine Probleme interessieren uns nicht die Bohne. :-))
>Wir wollen wissen, wie es mit der Huckepackplatine aussieht!
ich nehme an, dass Du das lustig gemeinst hast was Du da in 'unserem' 
Namen geschrieben hast, diese Art kannst Du aber m.E. -so- ruhig stecken 
lassen.

Robert schrieb:
> ich nehme an, dass Du das lustig gemeinst hast

Na klar, deshalb auch der Smiley hintendran. Ich denke Jörg wird das
schon richtig verstehen. Ich glaube aber nicht, dass ich der einzige
bin, der sehr gespannt erwartet, was Walters und Branadics Werk in
der Praxis bringen. Jörg ist schließlich der Erste, der es zur
Funktion bekommt.

Hallo Jörg,

das mit der Verstärkung haben wir früher schon diskutiert. Es wäre auch
im Original leicht möglich diese besser auf die ADCs anzupassen, aber 
auf
Kosten von Bandbreite (erheblich) und Rauschen.

Grüße, Guido

1.2.BF.4.8C2

>Unfortunately I wasn't able to reproduce this problem. Has anyone an
>idea how to force this error? Otherwise it is a littlebit difficult to
>analyze the problem.

Maybe it's for the same reason as this one:
Press Utility -> Default Setup
wait until screen is redrawn
change Time Base
=> for a short time there is a problem sightable in the first horz. 
Gridline.

Robert

1.2.BF.4.8C2

same problem:

Press -> About Oscilloscope
Change Time Base

Robert

Hi Hayo,  Rainer W., Robert and guys all!

Hayo W. schrieb:
>> Maybe I found a little bug in a 1.2.BF.4.8C2 firmware.
>Oh man, how did You make it? I tried hard but all is ok, except the gray
>turn arrows which appear on the empty two places after pushing the
>button 1 or 2.
I simply try to adjust the delay between the two channels of my W2022A 
but sometime by the rotary knob does not work, it works only by buttons.
While this happens if I try to change the delay of one of the two 
channels, tags on the screen show itself corrupt.
Hayo,please  take a look at the screenshot.

@Rainer W.
Rainer, apologize me.
For this time I beg You to bear me, the screenshot is only one. ;-)

Hayo W. schrieb:
>>> When QM is on, changing Time Base sometimes get corruption on the screen
>>Yes I noticed this too and I'm searching for the reason. I hope I can
>>fix it soon.
>
>Unfortunately I wasn't able to reproduce this problem. Has anyone an
>idea how to force this error? Otherwise it is a littlebit difficult to
>analyze the problem.

As I already wrote, for me the bug seemed to gone by installing the last 
1.2.BF.4.8C2 firmware release, but unfortunately it is not: it is like 
Robert says.
Anyway the trouble manifested itself simply by changing the Time Base 
and/or the vertical amplitude when QickMeasure is on, it show itself 
like in my previous screenshots.

Mit freundlichen Grüßen,

Luc

@Luc

As promised I checked the better settings for dso frequency response.
The accuracy I can guarantee is 0,1 dB up to 8GHz,using generators and 
power meters with internal alc and selftest, low loss cable and 
termination at 50 ohm.These are HP,Tektronix,Wiltron..test instruments 
is my job.
Anyway I tested W2022A, FW 1.2BF 4.8C2   HW with no modifications, up to 
100 MHz.
To have a good, flat frequency responce, better is:
ADC setup=factory  Pregain=1.25  Noise filter= IIR stage1
Trigger now works better than version 2.6 , very nice the last FW 
modifications, let’s say firmware improve the hardware limit of our 
DSO,proud to have the best of its class.
Only a hint ,is useful to have a marker peak on FFT,may be Hayo like to 
have it.

Thanks and regards

Sandro

Hi Hayo,
enclosed you'll find the image,pushing one menu key a dot goes to the 
maximum peak on the screen,then display frequency and level.

Regards,
Sandro

>@all -> has anyone the same problem?
W2022A 1.2.BF.4.8C2

Two Gray arrows: Same here

> Setting channel delay in Hardware menu, values are not update by rotary
> knob but only by push the buttons,

Well this is a little bit different here Luc/Hayo

Utility -> More -> CH1 Delay
Press button to next value -> no Update in the below line.
Value still the same as before. You may press thru all values, no 
changes in the bottom line.
Using the rotary knop -> all ok then, values are changing.

This is the same for CH2 Delay.

Scope turned off and on, same problem with CH-Delay as before.

Robert

1.2.BF.4.9C2

die von mir o. a. 'Probleme' sind (bei mir) damit behoben. Danke Hayo!

Robert

Hi Hayo, Robert, Sandro(Alex) and guys all!

Hayo W. schrieb:
>You where right, I found it and fixed it. ;-)
As always I thank You very much Hayo for the great job and free time You 
provided generously to us!!!
Great work, no words... RESPECT!!!!!!!

Hayo W. schrieb:
>But I couldn't reproduce the problem with the delay popup. Is it also
>present on the DSO of Your friend?
Hayo do not worry because seems to me by installing latest 1.2.BF.4.9C2 
version problem is gone!
I have tried to replicate the problem but with the 1.2.BF.4.9C2 I am 
unable to get it, for a very short time can happen there is corruption 
in the tags, but they settle themselves quickly when screen is 
refreshed, hard to see now if You deliberately do not pay attention!
Really an amazing and great job: RESPECT!!!!!!!
Anyway, going to answer to your question, yes the problem was also with 
my friend's W2012A.
The two oscilloscopes are like follow:
W2022A   8C7.0C
W2012A   8C7.0B

@Hayo
Hayo thank You so much also for the rewriting of the ADC's Setup menu, 
for me very very useful: RESPECT!!!!!!!

@Robert
Press thru all values but no changes in the bottom line sometimes 
happens to me too.
Robert, thanks for reporting!

Sandro(Alex) schrieb:
>Only a hint ,is useful to have a marker peak on FFT,may be Hayo like to
>have it.
Sandro, I agree too.
It is roughly the same I meant when I asked for the centre and span 
features for the FFT.

@Sandro(Alex)
Very impressive instrumentations, congratulation Sandro!
I have a unmodified W2022A too, but unfortunately I have not a calibrate 
signal's I prefer to use ADC Setup= High Frequency 2, Pregain=1.25 and 
no filters.
The W2022A is visibly better than the W2012A.
Because I have not a calibrate signal's generator I usually compare 
between the two by measuring a very fast 300pS pulse.generator.
Seems to  me the unmodified W2022A can easy reach up 150MHz but maybe I 
am wrong.
Sure the linearity is not the best but I believe the 150MHz perhaps are 
workable.
If I am not wrong in the Hardware Forum Hayo and others put some 
screenshots where the W2022A were good at 140MHz, seems to me it were a 
resistors modified W2022A.
I am enough interested about the Welec DSO's bandwidth.
I read You use Noise filter= IIR stage1 when You make your measurements, 
but does not this introduce a bandwidth reduction? (IIR 1 Stage ca. 70 - 
80Mhz IIR-filter with 1 stage and coefficient 0.5)
Seem to me it bit introduce a offset level, too.
Is not Smooth filter better for the purpose? (Smooth ca. 80 - 90Mhz -> 
in all other TB cutoff frequ. is half sample rate that means lossless!)
However I believe no filter is better choice when bandwidth is under 
test.

Vielen Dank!!!!!!!

Mit freundlichen Grüßen,

Luc

@Luc
With the ADC high freq setting you have an overshoot in frequency 
response,this the reason because you can go over 150 MHz
with an input stage that is hardware bandwidth limited,you may see the 
datasheet.
High freq.-> the flatness isn't the best in the whole frequency range. 
IIR1 limits the bandwidth,but compensate the Pre gain 1.25
Accurate measure up to 100 MHz with this setting
ADC setup=factory  Pregain=1.25  Noise filter= IIR stage1 /smooth filter
With no filter e high freq setting ADC goes in saturation with your 
settings,try to change the vertical sensitivity to see it,not suggested 
for sine wave.
Let's wait for the improvement in the other features of the firmware, 
than we may do together a DSO specifications datasheet.

@Hayo
Thank you,I am available for further details

@ all
Have a nice day

Sandro

Hi Sandro(Alex) and guys all!

Sandro(Alex) schrieb:
>With the ADC high freq setting you have an overshoot in frequency
>response,this the reason because you can go over 150 MHz
Sandro, what You write is of course correct.
Due the fact I read about it, I know there is a linearity lack on the 
Welec's DSO when I performed the measures.
Without the Piggyback board the only cure is to change some line and 
termination's resistors for the MAX121.
Due this issue I subsequently opted for a pulse generator to evaluate 
the bandwidth, because linearity lack make difficult the measures.

Sandro(Alex) schrieb:
>with an input stage that is hardware bandwidth limited,you may see the
>datasheet.
Yes, of course I agree but You mean the filter bandwidth, not the 
oscilloscope one.
Specifications says about 70-80Mhz bandwidth for IIR 1 Stage and about 
80-90Mhz bandwidth for Smooth filter.
I wanted to try evaluate real DSO's bandwidth.
I know the W2022A is better of the W2012A, if I use the same filter then 
the performances are identical on both the DSOs and this make no sense 
for my purpose.
Anyway, if even there is an overshoot, I can observe the differences 
between the two DSO.
I know it is no much and it is not formally correct, but this is better 
than nothing.
However, during the firsts attemps I tried with Factory's ADS Setup even 
if this no change much the things.

Sandro(Alex) schrieb:
>High freq.-> the flatness isn't the best in the whole frequency range.
Due of the overt linearity lack caused by wrong gain setting of the 
MAX121.

Sandro(Alex) schrieb:
>IIR1 limits the bandwidth,but compensate the Pre gain 1.25
I did not know this, thanks for saying it to me!

Sandro(Alex) schrieb:
>Accurate measure up to 100 MHz with this setting
>ADC setup=factory  Pregain=1.25  Noise filter= IIR stage1 /smooth filter
Anyway, I think the Smooth filter is better than the IIR stage 1, also 
under the aspect of the bandwidth's response.

Sandro(Alex) schrieb:
>With no filter e high freq setting ADC goes in saturation with your
>settings,try to change the vertical sensitivity to see it,not suggested
>for sine wave.
I understand the overshoot matter, but not about the vertical 
sensitivity issue.
Can You kindly explain to me?
I used a no professional sine wave's generator up to 170MHz, it has no a 
calibrated level output but both unmodified W2022A and W2012A can show 
the waveforms on their screen.
Also, unfortunately the output of the my RF wave generator is not 
adjustable.

Glad to read You again, Sandro.
Thank You very much for your report!

Vielen Dank!!!!!!!

Mit freundlichen Grüßen,

Luc

Rainer W. schrieb:
> Korrektur:
> Es geht doch, allerdings mit mir völlig unverständlichen Zeitlimits.
> Solange die Summe der beiden Werte für die Länge des
> Pulstriggerfensters in dem gezeigten Fall gerade noch unterhalb von
> 287µs liegt, kriegt die Triggerfunktion den genau genommen 192µs langen
> Puls zu fassen. Es ist egal, ob man z.B. (<100, >187), (<140, >147) oder
> gar (>187, <100) einstellt.
> Einen komischen "Hystereseeffekt" gibt es, wenn man den ">"-Wert in der
> Nähe der Pulsdauer variiert, d.h. z.B. <50µs fest eingestellt hat und
> dann

Hallo Rainer,

ich konnte dies eben noch mal mit dem Probe Comp Signal nachvollziehen. 
Das schau ich mir nochmal an! Da läßt sich sicher etwas machen.

Gruß Jürgen

Hi Hayo and guys all!

@Hayo
Yes, that is what I did and it is because I say the better choise is to 
disable any filter when trying to evaluate the DSO's bandwidth.
In fact it is very simple to make the test, because stopping the DSO 
then it is easy to view how filters act by changing the filters setup.
Hayo, thank You very much for your suggestion!

Vielen Dank!!!!!!!

Mit freundlichen Grüßen,

Luc

Hallo,

unter 
http://welecw2000a.sourceforge.net/docs/Hardware/Welec-Huckepack-Settings-ScaleFactors.xls 
findet sich unterstützend zur Softwareentwicklung der Huckepackplatine 
ein Tabellenkalkulation-Dokument, anhand dessen man die notwendigen 
Einstellungen des LMH6518 und der Eingangsspannungsteiler ablesen und 
die Software-Skalierungsfaktoren entnehmen kann.
Mit der Huckepackplatine stehen in allen bisherigen Vertikalablenkungen 
224 bis 226 LSB zur Verfügung, der ADC ist also deutlich weiter und in 
allen Bereichen bis auf ±1LSB gleich ausgesteuert. Damit wird auch das 
Rauschen in den 1er und 2er Bereichen um Welten besser und in den 5er 
Bereichen auch noch einmal um gute 10% besser.

Zum Vergleich, in den originalen Einstellungen waren es nur 131 LSB in 
den 1er und 2er bzw. 204 LSB in den 5er-Bereichen.
Dies konnte durch die feste Verstärkung von G=1.25 des ersten 
Verstärkers auf 163 LSB in den 1er und 2er-Bereichen verbessert werden, 
was aber immer noch 20% weniger gegenüber den 5er Bereichen ist und 
somit das deutlich größere Rauschen erklärt.
Diese einfache Gegenüberstellung zeigt noch einmal in eindrucksvoller 
Weise was die Huckepackplatine zu leisten vermag.

Um das in allen bisherigen Vertikalablenkungen auch mal mit Zahlen 
nachvollziehen zu können soll das Dokument zur Aufklärung beitragen.

Darüber hinaus besteht mit dem LMH6518 generell die Möglichkeit die 
Vertikalablenkung fließend zu gestalten und nicht auf feste 
Vertikalablenkungen in 1er, 2er, 5er-Teilung zurückgreifen zu müssen.

branadic

1094 Beiträge sind genug, hier geht es weiter:
Beitrag "Wittig(welec) DSO W20xxA Open Source Firmware (Teil5)"

Jürgen schrieb:
> Rainer W. schrieb:
>> Korrektur:
>> Es geht doch, allerdings mit mir völlig unverständlichen Zeitlimits.
>> Solange die Summe der beiden Werte für die Länge des
>> Pulstriggerfensters in dem gezeigten Fall gerade noch unterhalb von
>> 287µs liegt, kriegt die Triggerfunktion den genau genommen 192µs langen
>> Puls zu fassen. Es ist egal, ob man z.B. (<100, >187), (<140, >147) oder
>> gar (>187, <100) einstellt.
>> Einen komischen "Hystereseeffekt" gibt es, wenn man den ">"-Wert in der
>> Nähe der Pulsdauer variiert, d.h. z.B. <50µs fest eingestellt hat und
>> dann
>
> Hallo Rainer,
>
> ich konnte dies eben noch mal mit dem Probe Comp Signal nachvollziehen.
> Das schau ich mir nochmal an! Da läßt sich sicher etwas machen.
>
> Gruß Jürgen

@Rainer

Ich hab mal wieder einen Blick zurück getan und die Lösung gefunden :-)
Als Anhang mal eine Probe für die Pulse Width Problematik. Du kannst ja 
mal den Anhang ausprobieren, ob es so passt. Es fehlen allerdings noch 
die gesamten Prüfungen zu den Eingabebereichen in der Firmware.
Gruß Jürgen