I think you may have been running into a limitation of your scope for at least the sine wave past 30MHz or so.
Possible but I hope not
![Sad :(](https://www.eevblog.com/forum/Smileys/default/sad.gif)
I have a MSO5062B hacked to 200Mhz and made Tinhead's frontend hack on channel 1... and the results are the same on both channels.
@Tinhead: what do you think?
On the other hand, if it doesn't come from my MSO then it would be very interresting to compare our boards versions.
Here is a link to 2 high res pictures of my board. Could be interesting to compare with yours...
http://dl.free.fr/nEDtUM0lMYour pictures are nice. I like your DSO interface
![Smiley :)](https://www.eevblog.com/forum/Smileys/default/smiley.gif)
What model do you have?
Have a look at the 2V Pulse @ 20Mhz and compare to the figures of yours... impressive
![Wink ;)](https://www.eevblog.com/forum/Smileys/default/wink.gif)
You can increase voltage accuracy (1mv in DC Arb function) by calibrating the HDG but don't do it without backing up before the htg.cal file in /etc and when following the calibration procedure, stop, save and exit the procedure after entering the 12th value. If something goes wrong, just copy back your old htg.cal file.
@fremen67: Thanks, it's a "upgraded" DS2072A
![Cool 8)](https://www.eevblog.com/forum/Smileys/default/cool.gif)
I will have a look at the hi-res board pics you posted and get back to you. I would be a little surprised that Hantek rolled an new revision of the main board that quickly (unless you bought your unit several months ago).
I think you may have been running into a limitation of your scope for at least the sine wave past 30MHz or so.
Possible but I hope not ![Sad :(](https://www.eevblog.com/forum/Smileys/default/sad.gif)
I have a MSO5062B hacked to 200Mhz and made Tinhead's frontend hack on channel 1... and the results are the same on both channels.
@Tinhead: what do you think?
something is wrong on your pictures, and it is probably MSO. It is always good idea to hack/change both channels, with only one you running into calibration trap.
something is wrong on your pictures, and it is probably MSO. It is always good idea to hack/change both channels, with only one you running into calibration trap.
OK I will finish with the other channel but I already noticed it when the HDG arrived. At that time both channels were still from factory. Bad sign for my MSO
btw, the SCPI implementation is using Agilent VID/PID, so one can use Agilent I/O Suite drivers instead of Hantek driver.
Sure, the Hantek app will then not work anymore (oh well, it is not working for me), but instead of that one can use any VISA/SCPI application. Interessting is the name displayed, htg1102a (instead of HTG2xxxx what so ever). Attached picture is made of HTG2xxx dump running on S3C2416 dev board (i wish i could play with real HTG, but for now it is enought to do things on dev board).
One thing of note is the "noisy" sine as one approaches 100MHz.
aha, and how it looks like near 100MHz? let say 90MHz, 95MHz, 98MHz, etc.
One thing of note is the "noisy" sine as one approaches 100MHz.
aha, and how it looks like near 100MHz? let say 90MHz, 95MHz, 98MHz, etc.
I would have to run another capture, but I do remember seeing the "glitches" as low as 30MHz, but they were more infrequent.
btw, the SCPI implementation is using Agilent VID/PID, so one can use Agilent I/O Suite drivers instead of Hantek driver.
Sure, the Hantek app will then not work anymore (oh well, it is not working for me), but instead of that one can use any VISA/SCPI application. Interessting is the name displayed, htg1102a (instead of HTG2xxxx what so ever). Attached picture is made of HTG2xxx dump running on S3C2416 dev board (i wish i could play with real HTG, but for now it is enought to do things on dev board).
This is what I see with Agilent Connection Expert when I connect my HDG. Did I miss something?
btw, the SCPI implementation is using Agilent VID/PID, so one can use Agilent I/O Suite drivers instead of Hantek driver.
Sure, the Hantek app will then not work anymore (oh well, it is not working for me), but instead of that one can use any VISA/SCPI application. Interessting is the name displayed, htg1102a (instead of HTG2xxxx what so ever). Attached picture is made of HTG2xxx dump running on S3C2416 dev board (i wish i could play with real HTG, but for now it is enought to do things on dev board).
Yes it is working very well. I just had a driver problem (it's late here
![Tongue :P](https://www.eevblog.com/forum/Smileys/default/tongue.gif)
). I tried basic commands like switching ON or OFF the 2 channels ...
You have the list of scpi commands in the documentation. It should then be possible to send an arb waveform like TTSource is doing.
@ Tinhead and Fremen67,
You will find some SA screenshots as follows:
- The first one is the output of my HDG2022B ("software only" modified to HDG2102B) running a sinwave at 90MHz with a -10dbm output (400mvpp for the HDG output) already showing limits of the HW design (high DDS spurious and poor noise level)
- The second one is taken at a more reasonable frequency of 50MHz (same output level) still showing DDS spurious and high noise floor.
- The third one comes from a more serious "calibrated" AWG (G5100A) from PICOTEST (same as KEITHLEY 3390) ....no comments !
BUT if we restrict ourselves to a max frequency of 35MHz .... no more spurious and phase noise degradations...
The flatness from 1MHz to 90 MHz is not so bad (in dbm of course) with a +/- 1 db on my sample....
With such a design I cannot belive that HANTEK can release more than the HDG2032B model to his resellers !!
PS: > 15 MHz Square waveforms are also terrific on the display of a spectrum analyzer (when compared with the G5100A)
And the last one for today: the fan replacement. Exactly like the DSO5000, you can replace the original fan with a less noisy Noiseblocker XS1.
As the PSU is different from the DSO5000 (no easy L7812 replacement to lower the fan voltage) I just added a 47R resistor near the fan to decrease voltage from 12V to 9.4V. (I also tried 110R for 8V but quieter enough with 47R).
OK, not a big mod but you really apreciate it when it's late and everybody is sleeping
The flatness from 1MHz to 90 MHz is not so bad (in dbm of course) with a +/- 1 db on my sample....
Thanks a lot for all those informations.
It seems that I have a serious problem with my MSO as I am the only one to see sine amp decreasing after 30Mhz
BUT if we restrict ourselves to a max frequency of 35MHz .... no more spurious and phase noise degradations...
The flatness from 1MHz to 90 MHz is not so bad (in dbm of course) with a +/- 1 db on my sample....
With such a design I cannot belive that HANTEK can release more than the HDG2032B model to his resellers !!
Could it be that the actual HDG2102B has different hardware?
The unlock just allows you to push the 2022 harder but maybe it doesn't perform like the actual 2102.
@RFasic: I was starting to do those same measurements with my SA; I will post some screenshots tomorrow hopefully.
Hmm...don't know what to think about these spectrums.
What could be the source of these higher harmonics?
I believe, that Hantek knows that they have issues with this FG. It's striking that you only find the HDG2002b low end model which, even if HW or FW isn't perfect, stays in it's specifications. In some months maybe we'll also see the "higher" modells with modified hardware revisions.
I'm sure that this FG could really do better...The specs are very good and the DAC seems to be reasonable.
Hmm...don't know what to think about these spectrums.
What could be the source of these higher harmonics?
I believe, that Hantek knows that they have issues with this FG. It's striking that you only find the HDG2002b low end model which, even if HW or FW isn't perfect, stays in it's specifications. In some months maybe we'll also see the "higher" modells with modified hardware revisions.
I'm sure that this FG could really do better...The specs are very good and the DAC seems to be reasonable.
Hantek has pulled back all those FG's. Everyone I try to buy one from is saying it's at least three weeks until Hantek is shipping.
Thanks for that info.
So you're planning to buy one of these in around three weeks?
A comparision would be great, wonder what Hantek changes right now.
Thanks for that info.
So you're planning to buy one of these in around three weeks?
A comparision would be great, wonder what Hantek changes right now.
I'm not sure at this point. I would like to believe they are actually addressing signal integrity issues and not re-engineering the code to defeat hacking attempts. But I don't know for sure and in the mean time I really need a bench generator.
But I don't know for sure and in the mean time I really need a bench generator. ![Confused :-//](https://www.eevblog.com/forum/Smileys/default/confused0024.gif)
I had seen this on fleabay, if you really need a func gen it might be enough for you:
http://www.ebay.com/itm/221440242609
Yeah I also noticed the glitching with fast sines, I wonder what that is about...
Did you also look at the noise mode? From what I have measured it does not look that random at all...
I discovered the source of the glitches at higher frequencies: the external reference seems to be causing it, regardless of how clean or dirty the reference is. I had been using my Rb few standard, but reran a couple of the 90+ MHz tests with the internal reference and no more spurious glitches! Can someone confirm this behaviour with the external reference?
What model is your MSO?
Hantek MSO 5062D hacked to 200Mhz but I only modified channel 1 frontend so it may be the reason.. but I think I already saw the sine amp problem before starting modifying it. So I may have another problem on my MSO
What model is your MSO?
Hantek MSO 5062D hacked to 200Mhz but I only modified channel 1 frontend so it may be the reason.. but I think I already saw the sine amp problem before starting modifying it. So I may have another problem on my MSO ![Sad :(](https://www.eevblog.com/forum/Smileys/default/sad.gif)
Yeah, I was going to say that if you're seeing weirdness @ 30+ MHz, then it might be something else entirely, and not directly the mod on the frontend. You should be able to see reasonably non-attenuated signals up to at least 60MHz.