Hacking the DSO2X1X

[DavidAlfa]

edit: @davealpha, I watch David on eevblog YouTube so much I read your posts in a high pitched Australian accent!

We're different Daves!
Run backup builder first, then 2d15 conversion mod.

[Eula]

Hello Aldo,
Not at all, all questions make sense and are appreciated, sometimes the problem is an elephant in front of our noses but we are not able to see it.
And as you rightly say, more after the USB cable (by the way, I have tried with other cables, out of curiosity, and it gave me the same problem with 50% of them. could it be a shielding conductor problem, who knows....)

[Eula]

David, Thank you very much for your reply,

Yes, that's my guess, that there is something wrong adjusted inside the oscilloscope.
The seller has "passed me on" to Hantek's technical service. I have sent them an email before touching anything internal, in case they give me the possibility to return it in exchange for another one.
I will look at that schematic, in case I finally have to do it myself.

Best regards

Edit:
David, I think you have "hit the nail on the head" again.
Although the electrical noise is very high and hardly visible at 20mV/div, the problem disappears when you change the range to 200mV/div and reappears at 500mV/div and above.

This is something I will be able to tell the Hantek technicians when they respond (if they do).
Thank you very much again "maestro" 

[OLderDan]

edit: @davealpha, I watch David on eevblog YouTube so much I read your posts in a high pitched Australian accent!

We're different Daves!
Run backup builder first, then 2d15 conversion mod.

Lol, yes I know you are not him, but even that reply was in his voice to me! I have read through both fnirsi041d and now  hantek 2d10 threads (yes, hundreds of posts!) and have watched your rise from enthusiast beginning to beaten down but still so informative and helpful and was afraid to make a poor first impression.
Thank you for all the effort you have put in here.

[supphol]

Has anyone tried this? Any meaningfull difference? Thank you

My scope is in pieces until the new encoders arrive, so I can't test anything.
If anyone wants to try, I'm attaching modified DTBs increasing the max SPI speed for the FPGA.

Download the platform-tools package, install the drivers, paste the contents of the attached file inside.
Connect the scope in FEL mode before running the script or it won't work!
Run dtb_test, choose the dtb and check the results.

There's no risk, if it stops working properly just flash the original dtb.

[DavidAlfa]

Of course I tried. No difference.

[Eula]

Did you compensate the probe?

Doesn't make sense when any probe works nice in CH1. Something seems wrong in CH2.
Try in different ranges, around 20mV/div (x1) or 200mV/div (x10) you'll hear the relay clicking.

Something might be wrong in the analog frontend, a component missing, or CX1 badly asjusted.
https://github.com/pecostm32/Hantek_DSO2000/blob/main/Schematics/Mainboard/Analog_Input_2.png

Thank you very much David for putting me "on track".
I have finally decided to adjust the channel myself, as Hantek has not responded to me in two days and for such a simple adjustment I doubt that they will offer me a change.
In order to adjust the variable capacitor Cx1, I had to file a few tenths of a screwdriver that comes with the oscilloscope because I didn't have any non-conductive material at hand, which would have been more suitable.
I put here some pictures in case someone who has a similar problem might find it helpful.
The result, as you can see in the screenshots, has been quite satisfactory compared to where it came from.

Best regards!

P.D.: It is not necessary to remove the shielding, but I was intrigued by the hiding place of the relay that is being listened to..... in fact, for the adjustment, this protection must not be removed!

[jovan_kineas]

more photos of my mod.

[BillyO]

Since I am flush with oscilloscopes of all kinds I'm asking this question for a friend.

He is dead set on getting one of these and just sent me an email wanting to know what I thought.  Well, having zero experience with these I can't really give him much advice.  The specs on the DSO2D15 look pretty good and he says he's loves the look of it and he can currently get it for $279 shipped.  I told him "looks" are not an important factor in choosing a scope, but I digress.

So, my questions here goes out to the experienced (David..)

Have Hantek addressed any of the issues?  If he buy's this thing will he get a useful, working interment right out of the box?  Or should he run in the direction of Siglent as fast as he can?

[Aldo22]

@BillyO:

That is far too expensive and the DS02D15 is not worth it. The DS02C10 is (mostly) technically the same, except for the probe.
If you want a DS02C10, you shouldn't pay more than ~$160 for it in my opinion.
E.g. here https://www.aliexpress.com/item/1005006350502458.html
(I don't know this store, I just want to show that you can get one for this price. I paid $130 for mine).

If you want a really cheap scope that can do quite a lot and of course has some weak points (according to the price), you can buy it if you can get it cheap.
I'm a little hesitant to "recommend" it, It's all about the price.

Of course Rigol or Siglent is much better, but I'm still happy with my Hantek, because I like cheap stuff and it's good enough for me.

Have Hantek addressed any of the issues? 

There are still FW updates. The last one (unofficial) in February:
https://www.eevblog.com/forum/testgear/new-hantek-dso2x1x-models/msg5366381/#msg5366381

If he buy's this thing will he get a useful, working interment right out of the box? 

As a non-native speaker, I'm not sure what an "interment" is. The translator says "funeral". 
Do you mean instrument?

Yes, it works out of the box. However, a backup is recommended (follow DavidAlfa's FAQ)

[DavidAlfa]

Only the dso2c10 (Hacked later) for $150-170 does worth it.
For $280 no way! Spend $100 more and get a rigol DHO800, pretty new so still some bugs but nowhere close as Hantek.

[Aldo22]

The specs on the DSO2D15 look pretty good and he says he's loves the look of it and he can currently get it for $279 shipped. 

I would like to express a general view on this.
The problem with the Hantek DSO2000 is that prices vary greatly.
Firstly, it is always the same hardware.
From the DSO2C10 to the DSO2D15, it's always the same.
What's different is the probe, whether the signal generator is unlocked and the label.

It's basically a, say, $150 scope.
And for that it's brilliant!

But if you pay $250 or even $400 for it and expect it to be a "Riglent" for less money, you can only be disappointed.
It's still a $150 scope.
It's certainly not a professional tool and even the ambitious amateur will want something better.

But for $150 it offers an incredible amount of features and as your first scope, it will keep you busy for quite a while and introduce you to almost all aspects of a modern DSO.
Compared to a Fnirsi 1014D, for example, I would even call it a serious tool.
You may not even want more for quite some time.

Just don't pay more than +-$160 and you'll be fine.

[BillyO]

Thanks for your input guys.

He's not the kind of guy that will do the hack himself and I don't want to do it either as I don't have the time or inclination.  I'm going to loan him a scope until he gets his income tax return payment then he's going to buy a Siglent SDS804.  Quite a lot more, but it will work out better in the long run and the "hack" for it, if he wants to do it, is just entering some license keys.

[Aldo22]

He's not the kind of guy that will do the hack himself and I don't want to do it either as I don't have the time or inclination.  I'm going to loan him a scope until he gets his income tax return payment then he's going to buy a Siglent SDS804.  Quite a lot more, but it will work out better in the long run and the "hack" for it, if he wants to do it, is just entering some license keys.

Don't let the word "hack" put you off.
All you have to do is copy a file from DavidAlfa to the USB stick and then select this file as a system update in Hantek.
This is done in 2 minutes and is absolutely painless.

But well, the Siglent is certainly better if the price is right for you.
Good luck!

[dirtmover]

Thanks for your input guys.

He's not the kind of guy that will do the hack himself and I don't want to do it either as I don't have the time or inclination.  I'm going to loan him a scope until he gets his income tax return payment then he's going to buy a Siglent SDS804.  Quite a lot more, but it will work out better in the long run and the "hack" for it, if he wants to do it, is just entering some license keys.

Strange that you're asking this question in the "Hacking the DSO2X1X" thread for a friend that is "not the kind of guy that will do the hack himself and I don't want to do it either".

Anyway, for $150 and 10 minutes of your time you can end up with a piece of test equipment that offers a great value for money proposition if your expectations are set accordingly. Once you account for the lack of support, shitty SW, noise issues, poor triggering, memory depth limitations, measurement limitations, barely functioning protocol decoding etc it's still a reasonably functional and useful piece of test equipment for occasional hobby use. Mine has served me well for the 2 years I've owned it and has adequately handled just about everything I've thrown at it.

I wouldn't touch it at $279 however. In that case he'd be better putting his money towards something else.

[BillyO]

Strange that you're asking this question in the "Hacking the DSO2X1X" thread for a friend that is "not the kind of guy that will do the hack himself and I don't want to do it either".

Don't be silly.  If he or I don't want to hack something that is up to us.  However, this is the perfect place to ask if the thing is serviceable out of the box because the people here have the most technical experience with it.

Once you account for the lack of support, shitty SW, noise issues, poor triggering, memory depth limitations, measurement limitations, barely functioning protocol decoding

And this would seem to indicate it is not really a very serviceable item.

He's a beginner, at least to the scope world, so having something that just works properly is the way to go.  He can use my SDS1104X-E for a few weeks and he'll already have the UI down pat when his scope arrives.

[Aldo22]

Once you account for the lack of support, shitty SW, noise issues, poor triggering, memory depth limitations, measurement limitations, barely functioning protocol decoding

I don't quite agree with this rather unspecific list, but what I have never understood are the "noise issues".
Could it be that this affects older versions only?
I don't see any excessive noise here.
On the contrary, I am very surprised that it can still measure a frequency in the 500µVpp range (half a millivolt, 200µVrms ). That is very good for this price imo.

What do I have to do to see excessive "noise issues"?

[OLderDan]

still confused about the overclock, do i take the number 30 in the sh file and lower it till it crashes around 16? "/usr/bin/oc_manager 30 &" or do I modify the file with no suffix with separate cpu and mem sections?

[Aldo22]

still confused about the overclock, do i take the number 30 in the sh file and lower it till it crashes around 16? "/usr/bin/oc_manager 30 &" or do I modify the file with no suffix with separate cpu and mem sections?

You don't need to change the .sh file, just make the changes in the oc_manager file.
Mine looks like this (relevant part):

# ------ USER ADJUSTABLE ------
# This example: (24*(28+1)*(0+1))/(0+1) = 648MHz
CPU_OC_PLL_N=28
CPU_OC_PLL_K=0
CPU_OC_PLL_M=0

# This example: (24*(13+1)*(0+1))/(0+1) = 336MHz (168MHz DDR)
MEM_OC_PLL_N=13
MEM_OC_PLL_K=0
MEM_OC_PLL_M=0

[OLderDan]

Is a smaller number faster, Dave mentions 16 as the sweet spot.

EDIT: uploaded the latest firmware to the shared folder.

[Aldo22]

Is a smaller number faster, Dave mentions 16 as the sweet spot.

No, it's all in the info.txt

The system will execute the file overclock/oc_manager from the usb drive. If there's no drive, overclock won't be applied.
Now you can try rising the CPU_OC_PLL_N multiplier in that file, rebooting each time to apply the new settings.
When your system becomes unstable, lower the multiplier and try again.
After finding the stable CPU overclock, proceed with the memory overclock, increasing MEM_OC_PLL_N value.
This value have little adjustment, the limit is usually near 16.

I left MEM_OC_PLL_N at 13 because it doesn't help much and it can make it unstable.

[OLderDan]

You know I read that file a million times and this is the first time i noticed 16 was referring to mem! Thanks mate, so far I have backed up, patched from d10 to d15 and the small fonts. Now just want to get rid of the menu\button lag!
EDIT: YouTube clip of still not smooth 2d15 https://youtu.be/p-ngOAyKXiM?si=Z9k8l_1KVBx3a7Bt

[Aldo22]

EDIT: YouTube clip of still not smooth 2d15 https://youtu.be/p-ngOAyKXiM?si=Z9k8l_1KVBx3a7Bt

Try a smaller frequency difference. At your "speed" you can hardly see what is happening.
It looks smoother on my scope, but of course it has a limited update rate. What do you expect for $150?

[OLderDan]

It seems like using different sample rates has an effect on the trace thickness much more pronounced with the latest patch which makes me feel they reduced the smoothing algorithms for less load with the side benefit of increased accuracy or at least the same accuracy with less filtering.
Could someone else confirm or deny this for me please.

[Aldo22]

It seems like using different sample rates has an effect on the trace thickness much more pronounced with the latest patch

Can you describe the test setup in more detail?
The user cannot change the sampling rate directly afaik. It is set automatically by memory depth and s/Div.