EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: blabla on March 04, 2016, 07:28:44 am
-
Hello,
Long time lurker, first time poster. It's because of this place I got a ds1054z some time back and have just got around to start using it.
I'm trying to measure my 120v, 60hz mains. Yes, I'm aware of the safety risks involved. I'm using a probe with the ground pigtail removed since the scope's grounded. Probe and scope set to 10x
I have all 4 channels set to 50v/div and zero'd to the center of the screen. Vrms displayed below on all channels. With no inputs, 1/2 show ~2.8v and 3/4 show ~1.4v. I understand this is normal... sounds like its picking up 1-2 ADC counts noise. So far so good.
When I apply the probe to channel 1 (with the tip at hot/live wire) The vrms indicates (in order channel 1/2/3/4) 122/63/1.4/1.4.
When I move the probe to channel 2, I see 2.8/122/63/1.4. Channel 3 - 2.8/2.8/122/64. Channel 4 - 2.8/2.8/1.4/122
If I put two different probes in 1 and 2 (3/4 empty) both tied to the same hot, I get 122/138/64/1.4
I will say this - the "phantom" voltage is only shown on the vrms indicator. The waveforms are as they should be, flatline when theres no input, and ~340vpp sine when there is input. Is this some kind of side effect of measuring the same mains I'm powering the scope from, or is this a bug I've stumbled on? I'm leaning towards bug because the waveform doesn't match the vrms indicator.
-
Oh, and I'm running what I believe to be the latest firmware through the rigol link I found on here, 00.04.03.sp2.
-
I'm trying to measure my 120v, 60hz mains.
Don't do that without a special isolation transformer.
Yes, I'm aware of the safety risks involved.
Apparently you aren't.
If you want to measure mains voltage in a safe, reliable and repeatable way, use something like this:
http://www.yuanxing.net/1-2-1.asp (http://www.yuanxing.net/1-2-1.asp)
-
So it would appear that you've got all channels active while you're doing this, right?
Channel crosstalk would be the suspect in my book.
Check the manufacturers specs, there should be a spec for this.
-
When I first noticed this, I was just using channels 3/4. 1/2 were not active. I only activated all at the same time after attempting to narrow things down and moving the probe bnc connector around.
Crosstalk... hmm, that would happen at the analog / input stage right? If so I imagine that would effect both the picture waveform and the numeric vrms indicator, not just the indicator.
-
When I first noticed this, I was just using channels 3/4. 1/2 were not active. I only activated all at the same time after attempting to narrow things down and moving the probe bnc connector around.
What do you mean by this? ^^^ :-//
Crosstalk... hmm, that would happen at the analog / input stage right? If so I imagine that would effect both the picture waveform and the numeric vrms indicator, not just the indicator.
Technically crosstalk can affect any adjacent signal path.
Where it might or might not originate from is really immaterial.
Again, check the specs, or don't Rigol publish them?
-
No need to speculate about crosstalk in this case. It never can be nearly that bad, particularly not at low frequencies.
The obvious bug is the mismatch between waveform display and automatic measurement. maybe you could provide a screenshot or two to illustrate the issue?
What about lower voltages? Is the behavior the same or is it just an issue with the highest attenuator setting (50V/div)?
If a bug like this actually exists in this scope, then I really wonder how it can be that nobody has stumbled across it yet...
-
No need to speculate about crosstalk in this case. It never can be nearly that bad, particularly not at low frequencies.
The obvious bug is the mismatch between waveform display and automatic measurement. maybe you could provide a screenshot or two to illustrate the issue?
What about lower voltages? Is the behavior the same or is it just an issue with the highest attenuator setting (50V/div)?
If a bug like this actually exists in this scope, then I really wonder how it can be that nobody has stumbled across it yet...
Agree, the decribed mismatch between the curve display and the measurement (Vrms) is the real issue here. Very surprising too -- so far, I was convinced that the measurements are derived from the screen memory, and should hence be consistent. I would love to see a screenshot illustrating the issue!
-
Here we are. I moved the zero of each channel so you can see them easier. I tried getting the vrms indicators to show up in order but I couldn't, so just note they are in order of 1/3/4/2 but the channels are in order of 1/2/3/4 top to bottom. This is with the probe plugged into channel 1, 2/3/4 have nothing connected.
-
Here's two more. This time there is a probe in both 1 and 2, measuring the same mains at same point. First image is with channels in same position as inital screenshot. Second is with 1 and 2 at same zero point. Notice the curves overlap as you would expect, but vrms is different.
-
Here we are. I moved the zero of each channel so you can see them easier. I tried getting the vrms indicators to show up in order but I couldn't, so just note they are in order of 1/3/4/2 but the channels are in order of 1/2/3/4 top to bottom. This is with the probe plugged into channel 1, 2/3/4 have nothing connected.
Could you confirm you are not a troll ? ( no offense intended ;) )
Sorry ... but i have to ask, what are you showing seems weird to say the least.
Anyway, for what is worth, have you tried to "short tap" the unused channels ?
-
Lol. No troll here. Although I'm troll-like, short and hairy. Does that count? :P
What does short-tap mean? Tie the probe tip to ground? If so, it leads to the same result. I tied probe1 to mains, shorted probe 2, and screenshot looks same as first.
I also noticed just now.. With only channel 1 on mains, if all the v/div are set to 50, you see what you do in the screenshots. If I set channel 1 to 20, channel 2 reads 231vrms. If I set it to 100, it reads 64vrms.
-
What does short-tap mean? Tie the probe tip to ground?
Yep, but setting them to 1:1 .
Of course you know that if your scope miss earth ground connection with this kind of connections the whole experience could become "unpleasant" ;)
-
Wow. These 64vrms readings on channels that aren't even supposed to be getting a signal -- that's disturbing all right.
I'm unwilling to try to replicate this one on my scope because... well, because of the safety issue for one thing but also because I'm afraid of what I might see.
-
Here we are.
report it to rigol its a software bug. signal shows nothing but measurement says something... that "software crosstalk" is weird.
-
Well... this is as far as I'm willing to go. 60Hz in from my cheapo DDSFG cranked up all the way. Signal in to CH1, CH2, CH3 and CH4 in succession. Channels set to 10x even though I'm making a direct connection from the FG, so the displayed voltages are higher than the actual 15 v p-p being supplied.
So this bug is definitely here on my scope too. The next channel in sequence shows about 1/2 the RMS value of the actual channel receiving the signal, except when signal is in to CH4. And it's not necessary to use the mains as input to reproduce this bug.
With slower timebase settings the value of the "wrong" channel rms voltage goes up (5th screenshot below).
Selecting "GND" for the "wrong" channel input coupling has no effect on the spurious RMS voltage indication, it's still there. (Not shown in my shots but you get the idea.)
Yes, this is definitely one to report to Rigol. Does anyone have the earlier (SP1) software, to check if this bug is present there too?
-
Sweet! Thankyou for your efforts alsetalokin4017. Does Dave read this? I bet if he can get his Rigol contacts to look at it that would get a faster result.
-
You're welcome. I've been having lots of fun discovering and documenting bugs on this great little scope. (It truly is, but the bugs can be annoying...putting it mildly.)
The problem exists at different frequencies, voltage settings, and waveforms, and even when the "wrong" channel is set to GND input coupling.
-
Good to know. It seems whatever algorithm they use to calculate the vrms references (channel-1) variable instead of its own variable. From what I've fooled with it seems to only do it on vrms, not the other measurements.
-
I don't believe I am able to replicate this on my DS1074Z-S.
I am running one rev back from the newest firmware - 00.04.03.01.05
When was the last time you performed a self calibration?
edit - in my original pic I don't have channels 2/3/4 grounded. I redid it with them set to ground and you can see a slightly elevated voltage on ch2 - but it is pretty minor. If I need to try different parameters just let me know.
-
I'm trying to measure my 120v, 60hz mains.
Don't do that without a special isolation transformer.
Yes, I'm aware of the safety risks involved.
Apparently you aren't.
If you want to measure mains voltage in a safe, reliable and repeatable way, use something like this:
http://www.yuanxing.net/1-2-1.asp (http://www.yuanxing.net/1-2-1.asp)
The OP is not connecting the probe ground clip to anything,& is in fact,removing the ground lead entirely,relying upon the Oscilloscope's earthing via its power cord to provide the return path.
This method is quite safe,providing certain sensible precautions are taken,
(1) Clip the probe onto the Active connection with the power at the socket off,step away from the connection,then turn the power on.
In countries which don't have switches,read this as plugging the power cord in----in any case,it is a good idea to leave the power lead unplugged,until you are ready.
(2)Following your test,switch off, the Mains supply to the point your probe is clipped to,before touching anything.
Experienced Techs working on linear power supplies in equipment,often "have a look" at the incoming Mains to the device,just to check it is there.
One side of the transformer primary should show 120v/230v w.r.t earth,the other side.maybe a few volts.
If both sides show 120/230v,the Neutral connection is broken,if they both say "a few volts" the Active is broken.
If both say "zero" both sides of the Mains are missing.
-
Hi,
I had a try on mine, and cannot reproduce this (still running of SP1). I connected my Rigol DG1032Z to the DS1054Z, one time grounded (with bnc cable, see screenshot 1), and one time floating (with red clip only of a bnc to clip cable, see screenshot 3).
-
Wim, do you still have the sp1 firmware file? If so, please post. I'd like to try older version to see if bug follows it.
TheSteve - does the -S series use the same firmware as the non-S series? If not, then it may not be effected.
-
Unfortunately Rigol prevents "downgrading" the firmware to an older version. There has been a thread about this on the forum. It seemed to be possible to downgrade with a very early bootloader version in early models of the DS1074Z, but not with any of the more recent bootloader versions.
-
Thanks for the heads up. I'd still like to have the file to fiddle with.
-
Wim, do you still have the sp1 firmware file? If so, please post. I'd like to try older version to see if bug follows it.
TheSteve - does the -S series use the same firmware as the non-S series? If not, then it may not be effected.
Yes, same firmware, same hardware with a little add on board for the AWG.
-
So 2 so far that can't replicate with older firmware. Please post if you have it still.
-
I'll load the newest firmware and test again, stand by.
EDIT:
OK, loaded the newest firmware and guess what we see, yes it looks terrible. So it is either the newest firmware or it is because I need to run a self calibration(recommended after new firmware is loaded).
EDIT #2:
Went back to the previous firmware and it is all good again.
-
So it is possible to go back. Can you attach the firmware file to a post?
-
So it is possible to go back. Can you attach the firmware file to a post?
It is only possible to downgrade firmware on very early versions with a really old bootloader rev. I believe the bootloader was updated when the DS1054Z was released to prevent it. The DS1074Z came out a little earlier. They are the same hardware of course. The file it too large to attach to a post and it is doubtful it would work for you anyway.
You can find my post regarding downgrading here:
https://www.eevblog.com/forum/testgear/rigol-ds1000z-series-firmware-downgrade- (https://www.eevblog.com/forum/testgear/rigol-ds1000z-series-firmware-downgrade-)*is*-possible-and-here-is-how/
-
Wim, do you still have the sp1 firmware file? If so, please post. I'd like to try older version to see if bug follows it.
TheSteve - does the -S series use the same firmware as the non-S series? If not, then it may not be effected.
I think it is this one (I have more version on my harddrive, but the rigol names for the files is not clear). First unzip both files, then you get 4 parts and unzip again (this was the only option to attach this file to this forum I could find)
-
It recognizes the file is older but still lets me press ok to attempt the flash. It goes through reading gel, then analyzing, then fails. There's bound to be some way to bypass the check.
I found the Agilent version firmware that rigol makes. That went through the same steps except it never gave a failure message. Seemed like it stalled at the analyzing step and never moved after several minutes. Didn't brick though, powercycle and seems fine afterwords.
-
The darn bug is present also in mine DS1074Z, last FW version installed :
(https://www.eevblog.com/forum/testgear/ds1054z-issue/?action=dlattach;attach=206157;image)
(https://www.eevblog.com/forum/testgear/ds1054z-issue/?action=dlattach;attach=206159;image)
(https://www.eevblog.com/forum/testgear/ds1054z-issue/?action=dlattach;attach=206161;image)
Even if it has an earlier boot looader version, it does not allow the downgrade.
So the Rigol's coder messed up some memory pointer with RMS function, one reason more to issue a new decent FW release.
Software crosstalk in DSO, a new frontier ;D
I wonder if someone get screwed with it in some serious job ...
-
It recognizes the file is older but still lets me press ok to attempt the flash. It goes through reading gel, then analyzing, then fails. There's bound to be some way to bypass the check.
I found the Agilent version firmware that rigol makes. That went through the same steps except it never gave a failure message. Seemed like it stalled at the analyzing step and never moved after several minutes. Didn't brick though, powercycle and seems fine afterwords.
You have to try the procedure that implies to push three times the help button right after the first beep during the scope boot sequence, search the forum for that 3ad.
-
Mark I tried that with the file Wim posted, didn't work. Eventually it just flashed all the LEDs on the unit. May do that with the Agilent firmware.
I also found a article where someone opened the firmware in a hex editor. The version number only appears at the beginning of the file one time. I tried changing the older firmware to have the newer number but that didn't work either. There must be some other hash or crc check.
I guess this isn't a huge deal if you only use one channel at a time. If you use multiple channels though the error adds up as my second post shows. Applying 120vrms to channel 1 and 2, vrms indicates 130something on 2
-
Eventually it just flashed all the LEDs on the unit.
Very same thing with mine scope, so i gave up.
-
Even if it has an earlier boot looader version, it does not allow the downgrade.
It most certainly works if you have bootloader versions 0.0.0.11 or 0.0.0.13.
-
Even if it has an earlier boot looader version, it does not allow the downgrade.
It most certainly works if you have bootloader versions 0.0.0.11 or 0.0.0.13.
System info reports 0.0.1.1
-
Even if it has an earlier boot looader version, it does not allow the downgrade.
It most certainly works if you have bootloader versions 0.0.0.11 or 0.0.0.13.
System info reports 0.0.1.1
That is a higher rev bootloader - so no downgrading is possible.
-
Is board version the same as bootloader version? In system info my ds1054z only shows manufacturer, model, sn, software ver, and board ver. Or is there another menu I don't know about
-
Is board version the same as bootloader version? In system info my ds1054z only shows manufacturer, model, sn, software ver, and board ver. Or is there another menu I don't know about
You can see the detailed system information by pressing "menu menu force menu" quickly, then pressing "utility","system", "system info".
-
That worked. Interesting... software version and firmware version are listed separately and are different.
-
I Just have to ask. I skimmed the posts and saw Steve mention a recalibration but then he back tracked firmware and all was good again. Some firmware update may not compensate for the calibration changes made in the firmware. So the new firmware still read the older calibration that was done for the setting on the old firmware. And I know alsetalokin4017 has a lot of experience with finding/tracking bugs.
But no were did I see a conformation that someone updated to the latest firmware, let the scope reach full warm up, run a self calibration and double check the probes calibration on each channel to confirm the probes are still reading in spec. or adjust them if needed.
Has anyone followed the above procedure to rule this out. I saw one instance a while ago were one probe saw the signal correctly but channel 2 didn't see the same as on channel 1. Same signal, even swapped the probes on the channels and saw the same irregularity. Then I did a recall and checked the probes then when I tested the same signal both channels saw the same thing identically. So from past experience this has fixed irregularities with mine.
So that's why I ask for conformation before running with calling it a bug, and then trying to figure out why some scopes see this and other don't. Maybe going forward we make the warm up and recallabration a required step when trying to confirm bugs or testing new firmware. It might rule out some of these inconstancies.
I've followed several bug finds and never see people mention these steps and it makes me wonder?
-
Someone with the newest firmware and the problem should certainly run the self cal with the scope warmed up to verify 100% for certain the problem is still present.
-
I ran a calibration when someone previously had asked if I did or not and got the same result. Sorry, Thought I stated it but guess I didn't. I wasn't sure how long the scope was on but it wasn't just a few minutes. My scopes now been on a few hours now and I'm running another and will report back shortly.
This is definitely bug territory though. Changing the volts/div on the channel with the probe scales the error on the channel without the probe. Changing the volts/div on the channel without the probe also changes the phantom vrms. Another poster said it showed the error on the channel with no probe while the coupling was set to ground. That should remove any doubt.
Edit: As predicted and expected, same after another calibration.
Edit 2: Another interesting tidbit - if probe is in channel 1, but the only measurement displayed is vrms2, it reads close to 0 as it should. The moment you turn on vrms1 variable, vrms2 jumps up.
-
Edit: As predicted and expected, same after another calibration.
Edit 2: Another interesting tidbit - if probe is in channel 1, but the only measurement displayed is vrms2, it reads close to 0 as it should. The moment you turn on vrms1 variable, vrms2 jumps up.
Current RMS readings are actually nonsense, it's a so cumbersome bug that it shouldn't be a challange to fix for the coders.
Is there any ways to officially commit this situation to Rigol ?
-
I used the contact us form on their site and pointed them to this thread yesterday. Hopefully Dave gets wind of this thread and sends it to his contacts.
-
I seriously doubt Dave spends his time combing through the threads looking for bugs to report to his "contacts". If the bug was preventing him from actually using the scope, like when he's making a review video, then he might take the time to let them know about a bug, but he's not going to send along every bug that gets discussed in the forums. I'm not sure what the proper channels are to submit a bug to Rigol because I've never had the need to try, but the contact tech support on the Rigol site seems to make sense.
-
I seriously doubt Dave spends his time combing through the threads looking for bugs to report to his "contacts".
Huh??? Lol I never insinuated that... IF he stumbles upon this thread it'd be nice for him to pass it along. No more, No less.
-
Is there any ways to officially commit this situation to Rigol ?
since you discovered it, its your responsibility... http://www.rigolna.com/tech-support/ (http://www.rigolna.com/tech-support/)
-
here is mine, though I'm not sure it's a test using similar conditions to be of any use.
I'm not going to put 240VAC into the scope, so instead it's..
DS1054Z with hacked options, h/w 0.1.1, s/w 00.04.03.SP1
input 20V p/p 50 Hz sine high Z output from fun-gen into ch1, all 4 channels "10x", normal aquistion, 5ms/div
ch2-ch4 left unconnected, DC coupling
-
Thanks for another data point. Others previously tried it with low voltages and were able to reproduce. It seems this was something introduced in the current sp2 firmware, others in this thread with sp1 were not affected.
-
Hello eevblog and Dave. First sorry for my bad english and thanks for this great forum. I have just uplodaded a youtube video about this problem.Please watch my video and tell me what do you think. Thanks for your time. https://youtu.be/xpgEC9ghXgE
-
Rigol evolution. :palm:
From good to worst...great video buddy :-+
-
That's an excellent demonstration of the RMS bug discussed in this thread.
Let Rigol know about it! Maybe they will get the message and do something about this. Being able to roll back to earlier firmware would help a lot.
-
Thanks alot for your replies. That's mean alot for me. I am working mostly at home, because my mother have multiple sclerosis i must take care of her and my father passed away last year. About me i am 39 years old, uninsured, self educated, experienced at electronics, computers and mechanics over 28 years. I like exploring thinks, i like reading about new innovations and scientific knowledge. I like try to build prototypes and study on them my understandings on that matters. In my opinion humanity must advanced for better life. But in our little life times there is some bumps like basic support. In example my mother's health insurance does not care what condition she is in it. Or rigol does not care any of us. World become mostly care about money, no life, no earth, no people. I think this is main issue here... Anyway rigol knows about my problem about 1 and half months. And they ask me 4 weeks ago "Could you explain the problem with RMS measurement and FFT more in detail? Regarding the voltage level. Has he done a self-cal?".
And i think i am self calibrated :-DMM. Have a nice life thanks again your precious times. And sorry for my self educated crappy english.
-
Meka77, I like the video; but I have an issue with the probe settings you keep changing. The probe setting just scales the measured value for display, it does NOT change the measured value acquired. You are essentially using 1X probes with your coax cable. The probe setting should be set to 1X and left there for your test video. If you compare Ch1 and Ch2, you will see that the measured values are pretty close if the probe settings for both channels was set the same. The video shows a factor of 10 difference because Ch1 thinks it's a 1X probe and Ch2 thinks it's a 10X probe.
This is not to say there isn't a problem. It just adds an extra confusion factor.