Rigol MSO 5000 hardware/software revisions

[NoisyBoy]

02.00.03 is the latest, see the EU or int firmware download site. 

If you have a signal only on Ch 1, and all the others open, and press Auto, does your channel 3 also come on with just noise?

[Gandalf_Sr]

I got my budget LA probe system working, only tested channels 0-7 up to 10 MHz so far.  Check out my thread for details (updated pics are in the second post).

[NoisyBoy]

Nice!  Love to see more update as you move along in your project.  Perhaps that would allow more people to use the LA feature in the scope, as I don’t recall seeing many posts from actual Rigol LA owners.

[sb42]

02.00.03 is the latest, see the EU or int firmware download site. 

If you have a signal only on Ch 1, and all the others open, and press Auto, does your channel 3 also come on with just noise?

Yes if there's an open probe connected to CH3. No if the probe is disconnected or shorted, or if the retractable hook tip is detached. So YMMV I think.

This is on 00.01.02.00.02; not entirely convinced that 01.01.01.04.08 was different.

Edit: forgot to mention - MSO5074, hw rev 01.01.000.

[Cerebus]

02.00.03 is the latest, see the EU or int firmware download site. 

If you have a signal only on Ch 1, and all the others open, and press Auto, does your channel 3 also come on with just noise?

Yes if there's an open probe connected to CH3. No if the probe is disconnected or shorted, or if the retractable hook tip is detached. So YMMV I think.

This is on 00.01.02.00.02; not entirely convinced that 01.01.01.04.08 was different.

Just tried this on mine - 01.01.01.04.08.

Attach probes to 1 & 3. 1 connected to cal out, 3 floating, all channels off. Hit Auto, chans 1 & 3 come on.

Same setup, probe 3 floating but hook tip removed (i.e. all but the last ~3mm of tip shielded) only chan 1 comes on.

So yeah, not a new bug in 00.01.02.00.02, but just the auto setup being very sensitive to the presence of even a small signal on a channel. Repeated the same experiment with chans 1 & 2 - exactly the same result.

[NoisyBoy]

Hey guys, thank you so much for the feedback and datapoint.  I did not involve any probes in my tests, channel 3 was floating with nothing attached, channel 1 was connected to a HP 50 ohm pass through resistor to an AWG.  Never had this problem before in 04.04 and 04.08. 

Let me run a few more tests on it with the Rigol probe attached to channel 3 to see what happened.  I can also test it with the 500 MHz passive probes I got from the Keysight and LeCroy scopes to see if that makes any difference.

[skander36]

Is not about probes .
In 00.01.02.00.02 the channel 1&3 are activated without any probes .
In 01.01.01.04.08 this is not happened .

[NoisyBoy]

You are right, probes made no difference, all my probes do that, from the $47 Rigol probe to the $400+ Keysight probe.  My channel 3 opens when I push Auto with or without the hook tip, and even with probe shorted to ground. 

It is a random event, sometimes channel 3 opens, sometimes it doesn't, but it opens most of the time.  It may not be firmware related, especially given what Cerebus has shown, but my scope definitely have never done that before the upgrade. 

I will run a Calibration tomorrow to see if it makes a difference.

[skander36]

You are right, probes made no difference, all my probes do that, from the $47 Rigol probe to the $400+ Keysight probe.  My channel 3 opens when I push Auto with or without the hook tip, and even with probe shorted to ground. 

It is a random event, sometimes channel 3 opens, sometimes it doesn't, but it opens most of the time.  It may not be firmware related, especially given what Cerebus has shown, but my scope definitely have never done that before the upgrade. 

I will run a Calibration tomorrow to see if it makes a difference.

My fault,  reply was for Cerebus .
The callibration makes no difference . I was doing three times.
But when you downgrade to 00.01.01.04.08 things are back to normal. Upgrade to 00.01.02.00.02 or 00.01.02.00.03 again and channel3 arise uncalled .
So the sollution is back to 00.01.01.04.08 waiting for next update .

[Gandalf_Sr]

I've stayed at 00.01.01.04.08 but what were the supposed advantages of 00.01.02.00.0x?

[skander36]

v00.01.02.00.03 2020/02/27
     - It solves the error of SCPI instruction reading La channel memory data, and solves the error of SCPI instruction reading La channel memory data

v00.01.02.00.02  2020/02/25
       
     - Optimized the connection HDMI start problem optimized the connection HDMI start problem
     - Optimize the vertical gear, channel zero elimination error
     - Optimization of the inconsistency between SPI CLK and SDA names
     - Zoom mode square wave display in optimized 2S time base
     - Added command to get pass / fail times
     - Delete the default email account and password
     - Problems in remote instructions are optimized
     - Optimized 1K storage depth, waveform recording
     - The problem of too many stuck events in optimized decoding

[Cerebus]

v00.01.02.00.03 2020/02/27
     - It solves the error of SCPI instruction reading La channel memory data, and solves the error of SCPI instruction reading La channel memory data

v00.01.02.00.02  2020/02/25
       
     - Optimized the connection HDMI start problem optimized the connection HDMI start problem
     - Optimize the vertical gear, channel zero elimination error
     - Optimization of the inconsistency between SPI CLK and SDA names
     - Zoom mode square wave display in optimized 2S time base
     - Added command to get pass / fail times
     - Delete the default email account and password
     - Problems in remote instructions are optimized
     - Optimized 1K storage depth, waveform recording
     - The problem of too many stuck events in optimized decoding

If we read the Chinglish "Optimize the vertical gear, channel zero elimination error" as something like "Optimised channel gain and offset calibration/compensation" we perhaps have an explanation for the issue with channel 3 coming on in 'auto' from residual noise/offset alone. If they muffed something up in automated gain/offset compensation it would explain this as a regression (in the sense of a bug appearing/re-appearing).

Lordy, I do wish the Chinese would spend a few shekels on translation. Do we have any native Chinese speakers who can confirm my interpretation of the release notes from the original Chinese release notes?

[sb42]

As noted earlier, I'm on firmware 00.01.02.00.02 and cannot reproduce the auto problem.  Because NoisyBoy says it's a random event, I wrote a little script to test this repeatedly:

Code: [Select]

#!/bin/bash

addr="192.168.8.32"
check="2 3 4"
wait="3"
repeat=5000

for ((i = 0; i < repeat; i++)); do
    lxi scpi -a $addr ":autoscale"
    sleep $wait
    for c in $check; do
        res=$(lxi scpi -a $addr ":channel${c}:display?")
        if [[ "$res" == "1" ]]; then
            echo "failed at iteration $i: channel $c is on" >&2
            exit 1
        fi
    done
done

So, activate AUTO, wait 3s, check if channels 2-4 are active. With the scope rebooted and set to defaults, the script has been running for an hour with no spurious channel activations so far.

Looking at the screen however, I notice that activating AUTO will occasionally (say 5-10% of the time) display a (flat) waveform for channel 2 for just a moment, which then disappears when the display settles. All channels other than CH1 stay off and their buttons are not lit. If I toggle channel 3 on and off before running the script, then it's that channel's waveform that flickers on and off instead; likewise for channel 4.

TL;DR - I still can't reproduce the problem but there's now some other thing that I don't know what to make of

[mabl]

v00.01.02.00.03 2020/02/27
     - It solves the error of SCPI instruction reading La channel memory data, and solves the error of SCPI instruction reading La channel memory data

Just as an additional data point: Compared to the larger changes of 00.01.02.00.02, the .03 release really only touches appEntry and the FPGA.

Code: [Select]

        modified:   firmware/fw4linux.sh
        modified:   firmware/fw4uboot.sh
        modified:   firmware/rootfs/rigol/K160M_TOP.bit
        modified:   firmware/rootfs/rigol/appEntry


[NoisyBoy]

What is your hardware version?

I tried calibration last night, but skander36 is right, it made no difference.  Channel 3 came on with nothing attached.

I wonder if Rigol lowered the turn-on threshold with Auto across 5000/7000/8000, and the noisier front end on some 5000 triggers this problem. 

Regardless, the channel should not come on when there’s nothing attached.

As noted earlier, I'm on firmware 00.01.02.00.02 and cannot reproduce the auto problem.  Because NoisyBoy says it's a random event, I wrote a little script to test this repeatedly:

Code: [Select]

#!/bin/bash

addr="192.168.8.32"
check="2 3 4"
wait="3"
repeat=5000

for ((i = 0; i < repeat; i++)); do
    lxi scpi -a $addr ":autoscale"
    sleep $wait
    for c in $check; do
        res=$(lxi scpi -a $addr ":channel${c}:display?")
        if [[ "$res" == "1" ]]; then
            echo "failed at iteration $i: channel $c is on" >&2
            exit 1
        fi
    done
done

So, activate AUTO, wait 3s, check if channels 2-4 are active. With the scope rebooted and set to defaults, the script has been running for an hour with no spurious channel activations so far.

Looking at the screen however, I notice that activating AUTO will occasionally (say 5-10% of the time) display a (flat) waveform for channel 2 for just a moment, which then disappears when the display settles. All channels other than CH1 stay off and their buttons are not lit. If I toggle channel 3 on and off before running the script, then it's that channel's waveform that flickers on and off instead; likewise for channel 4.

TL;DR - I still can't reproduce the problem but there's now some other thing that I don't know what to make of

[sb42]

What is your hardware version?

01.01.000, MSO5074.

I tried calibration last night, but skander36 is right, it made no difference.  Channel 3 came on with nothing attached.

I wonder if Rigol lowered the turn-on threshold with Auto across 5000/7000/8000, and the noisier front end on some 5000 triggers this problem. 

Regardless, the channel should not come on when there’s nothing attached.

I'll see if I can post some noise measurements.

[skander36]

As noted earlier, I'm on firmware 00.01.02.00.02 and cannot reproduce the auto problem.  Because NoisyBoy says it's a random event, I wrote a little script to test this repeatedly:

Code: [Select]

#!/bin/bash

addr="192.168.8.32"
check="2 3 4"
wait="3"
repeat=5000

for ((i = 0; i < repeat; i++)); do
    lxi scpi -a $addr ":autoscale"
    sleep $wait
    for c in $check; do
        res=$(lxi scpi -a $addr ":channel${c}:display?")
        if [[ "$res" == "1" ]]; then
            echo "failed at iteration $i: channel $c is on" >&2
            exit 1
        fi
    done
done

So, activate AUTO, wait 3s, check if channels 2-4 are active. With the scope rebooted and set to defaults, the script has been running for an hour with no spurious channel activations so far.

Looking at the screen however, I notice that activating AUTO will occasionally (say 5-10% of the time) display a (flat) waveform for channel 2 for just a moment, which then disappears when the display settles. All channels other than CH1 stay off and their buttons are not lit. If I toggle channel 3 on and off before running the script, then it's that channel's waveform that flickers on and off instead; likewise for channel 4.

TL;DR - I still can't reproduce the problem but there's now some other thing that I don't know what to make of

Try this :
Set Default -> Activate G1(default settings) connect a cable from G1out to CH1 input->Press AUTO

Let us know what's happened.
Maybe some scopes are defective .

[sb42]

Try this :
Set Default -> Activate G1(default settings) connect a cable from G1out to CH1 input->Press AUTO

Let us know what's happened.
Maybe some scopes are defective .

That displays the sine wave on CH1, the other three channels stay off.

[sb42]

Here's a quick measurement from a rev 01.01.000 MSO5074 running a modified 00.01.02.00.02:



Not too impressed with those noise levels, Rigol

It looks like CH3 is "special", so maybe that's a clue. I tried AC-coupling it to no avail.

Anyone else want to try this or tell me where I goofed?

[NoisyBoy]

Interesting!  I just ran the same test, VRMS on the 1.00 hardware is around 302uV for channel 1/3/4, channel 2 is about 323uv.  It may very well be a manufacturing variance.

[Cerebus]

Not too impressed with those noise levels, Rigol

16-17 nV/√Hz - not exemplary, but not awful either.

[NoisyBoy]

Hi all, an interesting finding to share with you.

If you follow the readme file in the latest 02.03 firmware Rigol released quickly after 02.02, probably there is little to no reason for doing another upgrade for most users.  However, I found the 02.03 update fixes a couple issues that are not documented in the readme file:

1. At least in my case, it fixes the problem with Auto randomly turning on channel that have nothing attached.  I have seen channel 2 (rarely), and channel 3 (almost always) turning on and display traces of noise.  Self-cal does not help, but after applying the fix, the problem is gone.  I have not been able to recreate it, when I press Auto, channels without input no longer comes on.

2. As sb42 had shared, when you press Auto, you sometimes get an intermediate image of other channels flashing on the screen for a moment, before it settles on the final screen.  This problem is also gone, there is no momentary flashing of random channels.

These are important fixes for the new bug introduced in 02.02, but unfortunately in Rigol fashion, the fix documentation did not fully describe what it fixes. 

I hope you will find this tip useful, and perhaps more owners can test it out so we have more than a sample of 1.

And a big thank you for mabl for pointing out the change in the FPGA and appEntry, his effort prompted me to test out the 02.03 firmware even though Rigol did not mention these hidden fixes.

Good luck and stay healthy.

[Gandalf_Sr]

So, are you saying that 02.03 is stable and fixes many issues?

[NoisyBoy]

Seem to be stable so far, at least not having the obvious bug in 02.02 that was in-your-face 2 min after upgrade.  Remember it is a Rigol, stability is all relative. 

[kehall]

Have they done anything with XY mode and the huge square block to represent a point in the signal in this 02.03 firmware?

Not sure it's really 'noise' as such - as even on 100V/div on 10:1 will show a square around 40% of a graticule division (20V noise surely not right?!)

Doesn't seem to be any way to get a reasonable display - colour grading or intensity change doesn't do anything either..

Talking of intensity/grading on waveforms in general YT mode, it'd be nice to be able to adjust intensity for a graded option to fade out the less frequent 'noisy' components to crisp up the waveform, perhaps as a variation on the colour grading instead of changing colour, change brightness of the pixels.