EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: matko on February 17, 2022, 06:27:50 am
-
Hi,
I updated the firmware of my DS1054Z to the latest version: 00.04.05.SP2 (no hack, board version 0.1.1 if that helps)
According to the instructions, a self-calibration is recommended after a firmware update.
I've never run the self-calibration since I bought it years ago, so this was new to me.
I found this manual and followed the instructions: https://beyondmeasure.rigoltech.com/acton/attachment/1579/f-0564/1/-/-/-/-/DS1000Z%20MSO1000Z%20Performance%20Verification%20Guide.pdf?sid=TV2:xBDzVJlWK
The first time I ran the self-calibration, it took more than 1 hour to complete which is way more than the expected 30 minutes but it completed.
Note that I had the oscilloscope running for more than 30 minutes before starting the self-calibration as per the instructions.
The same instructions I followed also explained how to test the self-calibration:
- Set all probes attenuation ratio to 1x
- Acquire>Mode>Average
- Averages=16
- Set vertical scale to 2.00mV per division
I then checked each channels to see if any had over 0.5 division offset.
I found the 3rd channel to have (for a lack of better words) a lot of "noise" compared to the other ones. No probe were connected during the self-calibration or this test.
I wonder if my channel is defective or if it's "normal". This is the only channel with that behavior so I'm a bit worried. I even ran the self-calibration 4 more times with the same result.
I never took the time to perform any self-calibration up until now. And to be honest, never used my oscilloscope up until now.
So I don't know if the channel had this behavior when I bought it years ago, if it changed over time or after the firmware upgrade.
Attached is the output I found. The purple one is from the 3rd channel. Blue one is from 2nd channel. You can see the relative difference between the 2 also.
Please let me know if you think this is defective, a known issue, a setting I forgot to change or something I should just live with.
Thanks for your help!
Edit: I found there is a "Electronics ยป Repair" forum after I posted. Feel free to move it if that's a better location, I can't do it myself.
-
One hour is too long. Unless your lab is very close to a radio station, or you have very strong EMI sources nearby.
The scope looks defective to me. Move the scope somewhere away of any other PC, monitors, LED bulbs or LED light-strings, etc, and try calibrating again, away from noises.
Here is mine, first on 500uV/div, and second at 2mV/div (at 2 different timebases, to get the same 60pts as in your photos). I've set an offset to each channel, so they can be easily seen all at once. All probes are unplugged, the PC and monitor were on, at about 1 meter away.
Later edit:
========
First 2 were with a delay offset, sorry, attached 2 more with delay zero.
The 5th one is with CH2 and CH3 on, CH3/magenta only barely visible behind CH2.
-
Thanks for your reply.
I reset the oscilloscope, reflashed the latest firmware and redid a calibration away from any pcs, monitors or LED lights. Calibrations are taking less than 30 minutes.
It seems to be subjectively less bad but it might just be wishful thinking.
I think the channel is faulty. In fact, it seems to change over time or doesn't seem stable.
Instructions suggested using Average acquire mode. I reverted to Normal and the channel is obviously noisier than the others.
It still seems to kind of work but some glitches are showing randomly.
I guess my last resort is for someone to come up with a "known issue/fix".
But after searching on this forum and Google, it doesn't seem anyone had that specific issue before. Or I'm not looking for the right symptom keywords.
With your reply, at least I know something is definitely wrong with my 3rd channel. And at least, the 3 other channels aren't impacted.
-
My scope has about 0.5 div noise (1mV) on CH3, no jumps and all channels have about the same noise, yours has about 3 div (6mV), jumps around and it's not like the other channels.
I'll say your oscilloscope is defective. Send it back to be replaced or fixed.
It looks like a hardware problem, but you were saying about a firmware update. If the scope was working OK before the firmware update, you can try to downgrade to the older known working firmware, and see if it's just the software or something hardware failed in between, and the firmware update was just a coincidence.
To downgrade the firmware, you can follow the instructions and use the disk attached at the end of this post (you need to download and add your own desired firmware to that disk, by default the Rigol_sign_disk.zip has no firmware update on it, it only contains the hidden Rigol signature needed to allow service operations like a downgrade):
https://www.eevblog.com/forum/testgear/new-rigol-ds1054z-oscilloscope/msg3613778/#msg3613778 (https://www.eevblog.com/forum/testgear/new-rigol-ds1054z-oscilloscope/msg3613778/#msg3613778)
-
Hi RoGeorge,
Thanks for the follow up and guidance about the firmware downgrade.
Something weird happened. I started the oscilloscope and Channel 3 was fine, no more noise at 2mV scale.
I played with the other channels (enabled/disabled) but then Channel 3 reverted back to its noisy state.
I tried reverting the firmware to as far as 00.04.00, same result.
Even with 2-5V scale, the noise is considerable. It definitely is a hardware problem.
Now I need to see if it's economical to get it repaired out of warranty or not.
Thanks for all the help.
-
I started the oscilloscope and Channel 3 was fine, no more noise at 2mV scale.
I played with the other channels (enabled/disabled) but then Channel 3 reverted back to its noisy state.
That's a good sign, the damage is not permanent. Maybe it's just a cold solder joint somewhere, or a bad connector. Could be a fractured PCB trace or a fractured SMD capacitor somewhere on the ch3 analog side.
Since the scope is not under warranty any more, you can open and closely examine the PCB under a big magnifier or with an inspection microscope. There are some videos about opening this model of scopes, Dave has a couple of videos, too:
https://www.youtube.com/watch?v=kb9P1Am9aFU (https://www.youtube.com/watch?v=kb9P1Am9aFU)
https://www.youtube.com/watch?v=lJVrTV_BeGg (https://www.youtube.com/watch?v=lJVrTV_BeGg)
There is also a relay on each channel, you can hear it clicking at a certain range while you change the V/div. Check if the noise appears as a function of the relay state. Maybe it's just a bad relay.
Either way, it might worth to look inside, maybe the defect can be spotted.
-
I disassembled the scope and look at the channel frontend. I couldn't find any obvious fractures, cold solder joints or blown capacitors.
My experience is very limited so unless it's very physically obvious, I won't be able to tell if there is a defect or not.
I did however inspect every joints with a microscope and found nothing obvious.
Once reassembled, the noise was gone. But it reappeared after some time.
When I boot the scope, most of the time, it's fine but then signal degrades after some time, as soon as 5-10 minutes.
If only ch3 is enabled, I see around 1 div of noise. If I enable a second, it looks like noise is halved. And reduced more with 3-4 enabled channels. I supposed this is related to the Sa/s being reduced as more channels as enabled.
At some point, there was 3-4 div noise at 2V.
I also found that there is a weird ~1.2 div offset at 50mV/div and no amount of vertical position can change it.
I ran a self-calibration and it took a couple hours to complete. I think the process jammed or was taking much more time when it came to calibrate the 3rd channel.
I guess it's time to just ignore that 3rd channel or buy a new scope. :-BROKE
Thanks for all the help.
-
I ran a self-calibration and it took a couple hours to complete. I think the process jammed or was taking much more time when it came to calibrate the 3rd channel.
Yes. If a channel is leaping around then it doesn't calibrate well.
I started the oscilloscope and Channel 3 was fine, no more noise at 2mV scale.
I played with the other channels (enabled/disabled) but then Channel 3 reverted back to its noisy state.
That's a good sign, the damage is not permanent. Maybe it's just a cold solder joint somewhere, or a bad connector. Could be a fractured PCB trace or a fractured SMD capacitor somewhere on the ch3 analog side.
I agree.
If you can't see the bad joint (they can be very hard to see) then one thing to try is to heat up that whole area with a hot air gun. Resolder every component.
I guess it's time to just ignore that 3rd channel or buy a new scope. :-BROKE
Thanks for all the help.
Does it still show signals on that channel? If so then a bit of extra noise in the millivolt range isn't a deal breaker.
A 3.5 channel 'scope is still useful. :-)
-
The noise getting less when you enable another channel is indeed due to the sample rate reduction and the resulting change of low-pass filter cut-off frequency in the analog front-end.
To me that looks like there's an intermittent fault in the analog front end of channel 3, something in the compensation network, maybe. A DC offset that jumps around might be due to biasing resistors cracked or a knackered via to GND. It is hard to tell, but I suggest you try to see if it's actually temperature related. You just need some freeze spray or canned air too cool stuff down.
A cracked capacitor might not be visually damaged. A marginal via might respond to mechanical load as well, so you could just press your finger onto components and check if it makes any difference.
-
just press your finger onto components and check if it makes any difference.
This, too. Although it might be difficult to do this while it's running.
-
Also pres on the ADC chip while the oscilloscope is running, the intermittent defect might be anywhere on the CH3 signal's path, including somewhere on the way to the ADC, outside of each channel's analog stages in the metal cans.
If you make it working without the case on, then you may use the other good channels to identify where in the signal path the CH3 signal becomes bad. For example feed both the CH3 and 4 with the same input signal (and with the same channel settings for both), then use the remaining CH2 and CH3 probes to look at the signal in different points, and compare the waveforms observed between the CH3 and 4.
If you do that, while probing with CH1 or 2 use DC coupling to observe any DC offsets, and use only the middle pin of the probes in CH1 and 2 when probing the CH3 and 4 DO NOT CONNECT THE GROUND ALLIGATORS anywhere inside the oscilloscope, or else you may short circuit and burn something. The ground from all probes are already connected internally between all the channels.