Hacking the Rigol DHO800/900 Scope

[lucian]

Thanks, Im still leaning to upgrading it to a 914 or 924 for the  200 μV/div, does the aliasing issue exist on the real DHO900s or only on the hacked 800s

[Antonio90]

The aliasing isue is related to the sample rate.
The absolute minimum sample rate has to be at least double the measured frequency.
With ~300MS/s (4 channels active, or the logic analyzer and 2 channels) you can only measure frequency content up to 150MHz maximum. In practice, more like 120MHz.
The scope doesn't limit the bandwith nor tells you about this.

I get that more is better, but with the 100MHz key you are actually getting close to 200MHz bandwith. I'd leave it at 814, you can always change the vendor.bin later on if needed.

Edit: so, yes, the original 914 will have the same problem, or even worse with the logic channels active.

[Mechatrommer]

The aliasing isue is related to the sample rate.
The absolute minimum sample rate has to be at least double the measured frequency.
With ~300MS/s (4 channels active, or the logic analyzer and 2 channels) you can only measure frequency content up to 150MHz maximum. In practice, more like 120MHz.
The scope doesn't limit the bandwith nor tells you about this.

I get that more is better, but with the 100MHz key you are actually getting close to 200MHz bandwith. I'd leave it at 814, you can always change the vendor.bin later on if needed.
Edit: so, yes, the original 914 will have the same problem, or even worse with the logic channels active.

problem with siglent fanboys is they expect the machine to be "automatic". even 804 70MHz you can put into aliasing if too large timebase. so you have to keep the bolded line above in mind during probing because rigol dho is more or less a "manual transmission" car. meaning if you want to probe 250MHz circuit, make sure you dont activate all channels. 1 or 2 channels maximum. siglent machine will be unable to probe this high frequency content if you forgot to deactivate all channels too. its like spitting bullshit that a manual car cannot automatically change gear similar to automatic car. manual car has its benefit. whats that? i leave that to your imagination. i already mentioned that implicitly.

[ebastler]

siglent machine will be unable to probe this high frequency content if you forgot to deactivate all channels too.

In Siglent's defense, they do adhere to the Nyquist limit in their scopes' design and specs. The entry level models have 2 GSa/s (in contrast to the DHO800/900's 1.25 GSa/s), and they go up to 200 MHz input bandwidth. So even in 4-channel mode, you get 500 MSa/s or 2.5 x the analog bandwidth.

Rigol is pushing their specs beyond that limit. And specifically with the DHO 900 series, where they cut the sampling rate in half when activating the digital channels (and without telling anywhere in the manual or spec sheet!), they are pushing way too far in my opinion. 125 156 MSa/s is sad even in an entry-level scope, and a bit absurd in a scope with a "250 MHz bandwidth!" banner spec.

[Mechatrommer]

siglent machine will be unable to probe this high frequency content if you forgot to deactivate all channels too.

In Siglent's defense, they do adhere to the Nyquist limit in their scopes' design and specs.

yes in effect the machine is crippling itself. in order to probe high freq, we have to remember switching off the other channels too. my point is, we dont operate our machine mindlessly, something need to keep in mind.

The entry level models have 2 GSa/s (in contrast to the DHO800/900's 1.25 GSa/s), and they go up to 200 MHz input bandwidth. So even in 4-channel mode, you get 500 MSa/s or 2.5 x the analog bandwidth.

extrapolating from rigol situation, someone will try to hack the 2GSps machine to 1GHz BW, because sometime they need maximum BW esp rf nuts. while saving money from buying rf grade dso. imho this is why rigol hack is so hot around here. while there are hack threads for siglent too, but siglent is double or more the price... keyword is bang per buck... as i said, if siglent can get SDS800X price close enough to rigol, maybe it could be a game changer, cheers.

[S2084]

This helps solve the problem with time zones.     https://4pda.to/forum/index.php?showtopic=798101&view=findpost&p=57379222

[AndyBig]

This helps solve the problem with time zones.     https://4pda.to/forum/index.php?showtopic=798101&view=findpost&p=57379222

Thanks, the program works great. After the reboot, the time zone remained the one I configured.

[Antonio90]

The aliasing isue is related to the sample rate.
The absolute minimum sample rate has to be at least double the measured frequency.
With ~300MS/s (4 channels active, or the logic analyzer and 2 channels) you can only measure frequency content up to 150MHz maximum. In practice, more like 120MHz.
The scope doesn't limit the bandwith nor tells you about this.

I get that more is better, but with the 100MHz key you are actually getting close to 200MHz bandwith. I'd leave it at 814, you can always change the vendor.bin later on if needed.
Edit: so, yes, the original 914 will have the same problem, or even worse with the logic channels active.

problem with siglent fanboys is they expect the machine to be "automatic". even 804 70MHz you can put into aliasing if too large timebase. so you have to keep the bolded line above in mind during probing because rigol dho is more or less a "manual transmission" car. meaning if you want to probe 250MHz circuit, make sure you dont activate all channels. 1 or 2 channels maximum. siglent machine will be unable to probe this high frequency content if you forgot to deactivate all channels too. its like spitting bullshit that a manual car cannot automatically change gear similar to automatic car. manual car has its benefit. whats that? i leave that to your imagination. i already mentioned that implicitly.

I don't think it is much of a problem, really, but I thought it was worth mentioning, because is an easy trap to fall on for beginners.
I understand the advantages of manual cars, and don't use automatics.
Still, manual or not, the 900 series should state the sampling limitation with LA, and it should be quite visible. There are a lot of beginners using this scopes, I think this is a fair thing to say.

Edit: Although you are right in that it is better to have 250MHz max. that cannot be used with all channels, than crippling bandwith to never have aliasing problems.

[Feuerbard]

i try hack files on 802 model   - all work ok , 100Mhz and 50mpts achieved 

[Silver_Is_Money]

i try hack files on 802 model   - all work ok , 100Mhz and 50mpts achieved

Details please.

[hpmaxim]

I watched both of Retrochannel's videos, and first updated to the 01.01 firmware, and applied the bandwidth/sample depth hack by connecting the scope directly to my router via Ethernet.  Everything seemed to work fine.  I then ordered a USB WiFi adapter and tried to set that up and noticed a lot of... flaky behavior.

When I first tried, the touch screen didn't work with the keyboard plugged in.  Since then, it hasn't been an issue.  However, on about 1 in 5 bootups with the WiFi dongle installed, I will get "LAN Disconnected" and everything works fine.  The other 4 out of 5 times, it seems to just ignore the dongle.  I don't think I ever saw an instance of disconnecting and reconnecting the dongle fixing the issue.  If you reboot enough, it will eventually work, but it seems pretty random.  When it's not working, attempts to turn on WiFi don't work.  It'll go on for a split second and then disable.  I'm thinking I just got lucky the first time when I went to go set everything up.

Anyone got any ideas?  I apologize if this was discussed earlier and I missed it.

[ebastler]

  wever, on about 1 in 5 bootups with the WiFi dongle installed, I will get "LAN Disconnected" and everything works fine.  The other 4 out of 5 times, it seems to just ignore the dongle.  I don't think I ever saw an instance of disconnecting and reconnecting the dongle fixing the issue.  If you reboot enough, it will eventually work, but it seems pretty random.  When it's not working, attempts to turn on WiFi don't work.  It'll go on for a split second and then disable.  I'm thinking I just got lucky the first time when I went to go set everything up.

I take it that you have confirmed that you have the right WiFi adapter version (TL-WN725N, version 2.0 or 3.0)? Do you still have the USB hub connected, from the initial setup via the keyboard? If so, it's worth a try to remove it and plug the WiFi dongle into the scope directly.

[hpmaxim]

I take it that you have confirmed that you have the right WiFi adapter version (TL-WN725N, version 2.0 or 3.0)? Do you still have the USB hub connected, from the initial setup via the keyboard? If so, it's worth a try to remove it and plug the WiFi dongle into the scope directly.

I have tried with and without the hub.  In order to reduce the chances of a problem, I bought it using the Amazon link provided.  Bottom of the package identifies it as: TL-WN725N(US) Ver 3.6, UPC Code: 845973050719

For shits and giggles, I plugged it into my Windows Laptop, the driver appeared to identify as a Realtek 8188EU.  After playing with it more, I'd say it's working less than 1 in 5 times.  I haven't been able to get it to work in at least 10 reboots.

Few other observations: The physical keyboard seems to work consistently.  I tried the USB WiFi in my computer twice, and both times the little green light in the device turned on.

So, it's possible that the USB connection in the Rigol is faulty, but this seems very unlikely.  It's possible the WiFi dongle is faulty (seems unlikely) or it's the wrong chipset (also seems unlikely).  If it is the dongle, I might be able to exchange it at Amazon.  The other possibility is that installing the bandwidth/memory updates somehow hosed everything up.  Seems unlikely too, but it's got to be something, right?  Other people are presumably not having this problem.

[cncjerry]

I seem to want one of these scope for some reason. I have an 1104X-E, upgraded to a 1204X-E with the signal generator (the AWG, full function, not the adapter).  I'm intrigued by the DHO814 and read most of the pages here and they seem to be about all the hacks.  I understand all the advantages of 12b vs 8bit, etc. 

But I'm wondering of people really like the 814?  I realize I won't be getting much out of it, but the user interface is consistent with all the other modern devices and the update rates on the FFT look snappier than my 1104X-E. On the 1104, I seem to have to hit a lot of buttons to make changes.  Even compared to my ancient Tek 3054C that has a vertical knob column for acquisition.  I didn't notice if the battery came with the 814?  I also have an Owon 7102 with battery so no gain there.  Was also wondering if Siglent might try to leapfrog Rigol with their own low-end 12Bit.

thanks

Jerry

[Fungus]

But I'm wondering of people really like the 814?

I like mine a lot! The touch screen is a massive upgrade and the UI is really fast to navigate. You really won't ever go back to twisty-knob models after using one.


Siglent is responding (they have to!):

https://www.eevblog.com/forum/testgear/siglent-new-sds800x-hd-first-bactch-unboxing-and-noise-compare-with-dho800/msg5238453/#msg5238453

[Mechatrommer]

Siglent is responding (they have to!):

why you have to keep relating to siglent even if when not asked?

You really won't ever go back to twisty-knob models after using one.

try vertical offset change using screen when all 4 channels active. if you can get rid of rotary knob, tell me... screen interface on that matter still need improvement, its automatic changing channel to what it thinks closest to where you touch, and it sucks tbh. i will still use rotary knob for that

On the 1104, I seem to have to hit a lot of buttons to make changes.

same thing happened to my rigol DS1054Z, i think the internal need cleaning.

I didn't notice if the battery came with the 814?

no battery, too bad. if switch off dso, date is lost. but well, we dont buy dso as timekeeper, esp the cheap one what do you expect? ymmv.

[Fungus]

Siglent is responding (they have to!):

why you have to keep relating to siglent even if when not asked?

Damned if I do, damned if I don't... 

[Proxy64]

Update to 01.02 or not ?
I don't like to upgrade if I don't see any improvement for my usage, but But maybe they didn't notice all the changes in the readme.txt !
So have you seen any improvements or bug fixed ? (other than the offset error for those who modify the vendor file).
I'm suspicious with this upgrade because it's not even available on their website for download, weird behavior from Rigol !

[tv84]

The touch screen is a massive upgrade and the UI is really fast to navigate. You really won't ever go back to twisty-knob models after using one.

  Weren't you happy with your all-in-wonder Micsig? Or, afterall, their screen/UI was not so good??

[ebastler]

The touch screen is a massive upgrade and the UI is really fast to navigate. You really won't ever go back to twisty-knob models after using one.

  Weren't you happy with your all-in-wonder Micsig? Or, afterall, their screen/UI was not so good??

Could we relegate that kind of squabbling to one of the general unboxing, impressions etc. threads please? This thread is meant to be focused on hacking approaches, and is getting a bit unwieldy anyway. Thanks.

[UK]

since its look like brighter hope, i already had a go at doing this and already cut the front panel of my scope and made youtube video to guide others if interested, its irrevesible, and since its not final yet (the LA probe probably wont work as expected), i dont think to publish the link yet so far.

Wow. You appear to have made significant progress here. I saw your video about cutting the LA port slot, and I want to offer another way... perhaps it may be easier to do it is using a 3D-printed template that will rest on the outer wall of the first BNC deepening and the left USB port. I believe they have more than enough position accuracy for placing the template. With the template you you will have more control over the cutting process. Then, I suggest using a manual jigsaw instead of Dremel.

And since I currently have 914s on hand, I can do all of the measuring, modeling, and several test prints to ensure that everything is properly positioned because everyone has only one try)).

Also, I thought it would be nice after cutting the LA port slot to 3d print the inner walls of the slot to hide the gap between the front walls of the MSO and its pcb, as is done on the DHO900... it may help it to look not so DIY. 

[Mechatrommer]

Also, I thought it would be nice after cutting the LA port slot to 3d print the inner walls of the slot to hide the gap between the front walls of the MSO and its pcb, as is done on the DHO900... it may help it to look not so DIY.

thanks for your offer and idea. another way of doing it is cut the slot however nasty shape it will be, even if slightly bigger not a problem. and then 3d print rectangle or however fancy, sci-fi techy shape covering on outer wall with protrusion to the inside wall touching to metal enclosure inside not a problem, to make it more look neat. but this requires the owner to have 3d printer. and imho its a bit wasteful and time consuming just to make a "one-time use disposable" cutting template, just learn to cut properly with whatever tools you have, cheers.

[UK]

...just learn to cut properly with whatever tools you have, cheers.

I completely agree, however not everyone has jeweler's hands, therefore I believe it took some work to make this slot as beautiful as possible without messing it up... especially if you want to use the MSO frequently and don't want to look at an ugly hole on its neat face.

Anyway, when you finish your project I would be happy to help with this template.

[Mechatrommer]

Anyway, when you finish your project I would be happy to help with this template.

you asked about the dimension of the slot hole and depth, and someone else answered.
https://www.eevblog.com/forum/testgear/rigols-new-dho800-oscilloscope-unbox-teardown/msg5159226/#msg5159226
https://www.eevblog.com/forum/testgear/rigols-new-dho800-oscilloscope-unbox-teardown/msg5166510/#msg5166510

but i think whats missing are distances of the slot to the top, bottom and side wall of the front panel relatively, so we can measure and draw guideline at the correct position, correctly aligned to the inside metal slot, not the slight mistake i've made in the video you saw. i hope you understand i dont have time to draw where they are. if you can measure and post for us, then it probably helpful one day, but not for me, its too late for me

[Randy222]

since its look like brighter hope, i already had a go at doing this and already cut the front panel of my scope and made youtube video to guide others if interested, its irrevesible, and since its not final yet (the LA probe probably wont work as expected), i dont think to publish the link yet so far.

Wow. You appear to have made significant progress here. I saw your video about cutting the LA port slot, and I want to offer another way... perhaps it may be easier to do it is using a 3D-printed template that will rest on the outer wall of the first BNC deepening and the left USB port. I believe they have more than enough position accuracy for placing the template. With the template you you will have more control over the cutting process. Then, I suggest using a manual jigsaw instead of Dremel.

And since I currently have 914s on hand, I can do all of the measuring, modeling, and several test prints to ensure that everything is properly positioned because everyone has only one try)).
Also, I thought it would be nice after cutting the LA port slot to 3d print the inner walls of the slot to hide the gap between the front walls of the MSO and its pcb, as is done on the DHO900... it may help it to look not so DIY.

If 3d printer is available, perhaps just easier to cut the scope bezel in rough dimensions to fit the connector needed, then 3d print a bezel that fits over the connector snugly and covers the scope cut. Result should be somewhat professional finish.