New Rigol HDO1000 12-bit DSO - BUGs

[TK]

I am starting a new topic where members can share their experience with the new Rigol HDO1000 and report / discuss BUGs in the hardware/software.

As first post, I am going to share my experience trying to decode UART and SPI protocols with the HDO1074 scope.  Of course this is the first version of the software and calibration sheet shows manufacturing date October 2022.

The scope froze and rebooted several times during the simple testing.  I cannot make it decode a simple UART signal @ 115200, 8N1.  The signal is from an Arduino Uno board and it is the sketch that prints the ASCII table.  It starts with the message "ASCII TABLE...".  The scope fails to decode the signal.  I tried on my MSO5074 and it decodes correctly the letter 'A' I am showing in the screenshot.

This is the screenshot:

UART Decoder also only allows timebase up to 20ms.  Any timebase 50ms and up the scope disables decoding functionality.  So you cannot capture many packets and zoom to see the frames.  Actually it completely disables the decoding option with timebase higher than 20ms.  Probably software cannot handle decoding large number of frames and Rigol decided to disable it to not affect other areas of the scope.

Decoder divides the screen in 2 sections if you open the event table.  You can put it side-by-side or top and down.  I cannot make the windows detach and move, resize.

Let's start the BUGs discussion from here

[JPortici]

UART Decoder also only allows timebase up to 20ms.  Any timebase 50ms and up the scope disables decoding functionality.

Wow really? A true riglol. Hope they eventually manage to get decoding running at any timebase (as most scope do)

[Martin72]

AFAIK this is 1054Z behaviour..MSO5000didn´t do this, AFAIK too.

[Fungus]

What's the sample rate at 50mS/div? It might not be high enough to do any useful decoding.

[TK]

What's the sample rate at 50mS/div? It might not be high enough to do any useful decoding.

sample rate at 50mS/div is 2GSa/s, sample memory set to 50Mpts

[JPortici]

And one should define useful decoding. 50ms/div is not unusual at all for me:
-LIN, with the scheduler scheduling frames spaced 10s of milliseconds apart
-plain UART, but at 1200bps
-SENT slow channel frame (1-2 bit per sent frame, 14-16 frames per message at about 1ms per frame)
And i don't have to think too hard about test setups involving such long timebases in which i monitor analog sensors with their digital counter parts, or events and communication correlation.
And in all of this i tend to limit the memory to 1MS per acquisition to have some acquisition history and to keep the scope responsive

But i get it if it's an imposed limitation to separate the product lines, makes sense, they have to justify the good scopes somehow

[Fungus]

And one should define useful decoding. 50ms/div is not unusual at all for me:

Jeez.

All I asked was what the sample rate was when people reported it stopped decoding.

[TK]

While trying to decode i2c, I discovered why UART was not decoding correctly.  I had to set the threshold to above 3V for a 5V UART signal.  The same with i2c, I have to set the threshold to 4.4V for the decoder to decode correctly (The i2c signal on the Arduino has a slow rise time due to internal resistor value).  But still many issues, like the decode values being out of position when the signal is moved and stops updating, decode value getting too small when timebase is changed and then staying tiny when timebase is moved back to a setting where decode value should be very visible, constant freezing (RIGOL SCOPE has stopped... Open app again)...

But what is it with being so picky with the threshold value, when at around 50% signal still is the same as at higher values?  Is it trying to use digital logic high and low values?  But the menu does not have any option to select the digital signal (TTL, CMOS, etc).

AND THE MOST ANNOYING BUG... IT RESETS THE COUPLING TO AC AND 20M BW LIMIT ON CH1 EVERY TIME THE APP RESTARTS!!!!!!

[markone]

While trying to decode i2c, I discovered why UART was not decoding correctly.  I had to set the threshold to above 3V for a 5V UART signal.  The same with i2c, I have to set the threshold to 4.4V for the decoder to decode correctly (The i2c signal on the Arduino has a slow rise time due to internal resistor value). 

Wait, pulse distorsion (mismatch in R/F edges slope) destroys any kind of UART transmission, what you see it's not up to Rigol decoder.

Is the probe set to 10:1 ?

[2N3055]

Yeah, they released it too soon.
They were probably under enormous pressure to show something after huge investment and effort in making what is now 2nd chipset generation...

What I see so far is not very good.
We will have to wait until they fix at least basic bugs to make judgment.

[Fungus]

Yeah, they released it too soon.
They were probably under enormous pressure to show something after huge investment and effort in making what is now 2nd chipset generation...

Christmas sales season is coming. The bean counters probably prioritized getting the hardware out there instead of waiting another year.

(and it's not as if Rigol is the only company that does that...)

[2N3055]

Yeah, they released it too soon.
They were probably under enormous pressure to show something after huge investment and effort in making what is now 2nd chipset generation...

Christmas sales season is coming. The bean counters probably prioritized getting the hardware out there instead of waiting another year.

(and it's not as if Rigol is the only company that does that...)

Well apparently R&S does it too... 

Yes, companies do that.. that is right. But there should be a threshold of how low do you go.. It damages reputation, especially if you are in the process trying to build one ....

[Fungus]

Yes, companies do that.. that is right.

Yep. I remember Siglent offering to post little bags of capacitors to to their "early adopters" so they solder them to the PCB and finally be able to compensate their probes properly.

Don't even get me started on second-rate companies like Apple telling people they were holding their new iPhones wrong, or the reason the phones were bending was that people were putting them in their pockets.

[2N3055]

Yes, companies do that.. that is right.

Yep. I remember Siglent offering to post little bags of capacitors to to their "early adopters" so they solder them to the PCB and finally be able to compensate their probes.

That Siglent issue was not the same. It was a production batch problem.
Not saying it isn't a problem, but wasn't the same as releasing something that simply doesn't even work out of the box.
Errors when you have "we didn't anticipate this problem, and we had to learn it the hard way" is not the same as "we only developed 60% of product but we need to start selling it despite we know it doesn't even work. Who cares, we'll fix it later".

And it is not even advocating for Siglent here ( and they in fact learned from that and pay more attention to maturity before release..) but not even Rigol used to do that... Rigol used to be much more solid company in that regard years ago than right now. Let's hope they realize a mistake and make it right.

Siglent and Rigol are only ones keeping western brands in check. Otherwise it would be back to old cartel behaviour and blackmailing us with prices.. We need Siglent and Rigol to do well, so everybody benefits from it. Staunch Keysight users that would never buy "those chinese toys" never had Keysight at current prices. They should write a thank you note to their Chinese friends..

[skander36]

Maybe it is too much said "... simply doesn't even work out of the box..."
It works but it having some quirks  . For sure they will be solved in time as happened with Rigol 5K.
MXO4 has also some quirks, and is in top makers.
About Siglent capacitor problems, they at least, talked about the problem, and offered complete board exchange for free. Rigol does not say a word ... 
Personally I am interested in buying HDO 1000, despite early problems. It is 12 bit scope at the price of a DMM ( Keysight 1282A) or less than Fluke 289 ...   

[Fungus]

Rigol does not say a word ... 

Have you contacted them?

[skander36]

Rigol does not say a word ... 

Have you contacted them?

Don't say you don't understand what I mean...

[Fungus]

It's a bit early to condemn this model, Rigol isn't even selling them yet in many parts of the world.

I'm sure there will be a firmware update soon.

[2N3055]

Maybe it is too much said "... simply doesn't even work out of the box..."
It works but it having some quirks  . For sure they will be solved in time as happened with Rigol 5K.
MXO4 has also some quirks, and is in top makers.
About Siglent capacitor problems, they at least, talked about the problem, and offered complete board exchange for free. Rigol does not say a word ... 
Personally I am interested in buying HDO 1000, despite early problems. It is 12 bit scope at the price of a DMM ( Keysight 1282A) or less than Fluke 289 ...

Well it literally does not ... Look at the TK bug thread...
It sure does not meant it should be "condemn" but at this moment it is not something I would buy.
And given track record  I'm not sure of anything anymore...
And yes, I have been equally critical of R&S..
They all should know better.. regardless of the brand.

[skander36]

"Look at the TK bug thread..."

TK found the right threshold values for deconding to work. I have the same problem for other scopes. At least R&S has a button "Find Threshold" on trigger menu, which help.
Let see if Rigol step up with a fix. Even if they don't talk I'm pretty sure that they listen.
In December a  stock of HDO1000 should arrive.

[Kleinstein]

It makes perfectly sense to already offer the hardare before the software is fine tuned with all details. The pricing of the HDO1000 sugests that they can earn good money from the HDO4000 - there are quite some changes with the higher resolution Scops the last years .  From what Daves videos show, the interface looks OK and workable, at least for the main features. Chances are there are enough customers eager to get one even with some more minor features not (yet) working. Currently it is not so much about would buy, but could buy. 

The HDO1000 is not not so much made for mixed signal tasks - so the seriell decoders are likely not high priority for such a scope. Though in some cases it may still be the only scope for some - so it would be nice to have the seriel decode over a larger range of setttings.  It should be reasonable esay to add / fix, but with the more B brands one never knows when and if things are fixed in the software.
It is still very early - and the limited time scale is more like a missing feature / anoying point that may be fixed in the future.  If this would be Fluke or R&S it may come up a pay extra option though.
So early there is a good chance to get at least some FW updates - though likely not just for this feature.

A not fully working seriel decoder as the first "bug" to come up is a good sign.

[tv84]

A not fully working seriel decoder as the first "bug" to come up is a good sign.

The "first" bug was what they did with the licensing scheme...

[Fungus]

That new Tek 'scope that appeared a couple of months ago didn't have the serial decoders enabled on launch because they weren't finished yet.

If the serial decoders on this one work above 50mS/div then that's something. I'm sure most people would rather have that today than a promise of "maybe next year sometime", like the Tek.

To the people that have one: What's the decoding like? Does it work well when it decides to do it?

[nctnico]

That new Tek 'scope that appeared a couple of months ago didn't have the serial decoders enabled on launch because they weren't finished yet.

If the serial decoders on this one work above 50mS/div then that's something. I'm sure most people would rather have that today than a promise of "maybe next year sometime", like the Tek.

No. Because with the Tektronix it is specified the decoders don't work -yet- so you can base your choice on that. It is all about unspecified surprises (= bugs) that make buying the Rigol an adventure. If you need working decoders today, you don't buy the Tektronix (unless decoders have been added in the meantime) but something else that you know works. MicSig did the same when they released the TO1000 series; the decoders where added later and MicSig was totally open about that. In the end nobody benefits from a feature that is half baked at the release.

[Martin72]

"...Do you know what we call it internally when it is clear at launch that the product is not quite ready yet ?
Banana version...banana version ??
Yes, banana version - matures with the customer."

(V.Pispers, satirist)