Author Topic: Hantek - Tekway - DSO hack - get 200MHz bw for free  (Read 1670555 times)

0 Members and 9 Guests are viewing this topic.

Offline tinhead

  • Super Contributor
  • ***
  • Posts: 1905
  • Country: 00
    • If you like my hacks, send me a donation
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #850 on: September 12, 2011, 09:39:20 pm »
had you though about wifi usb dongle plugin in ?
I think the drivers could be uploaded through the console, we only need to know the kernel vers and architecture. I bet there would be toolchans available.

wifi on backside will not work that easy (as you will have to translate host-client and dso specific protocol),
wireless usb from cypress could work (maybe).

wifi on front side is a different story,  some mods in usb plug script mus be done (this is however not the
biggest problem).

The kernel is 2.6.13, toolchains for QQ2440 will be good (arm-linux-gcc-2.95.3) - however this is exactly the biggest
problem - most wifi adapter have drivers working with 2.6.2x  and later kernel.

As i remember RT2870 was a good candidate for even 2.4 kernels.


I don't want to be human! I want to see gamma rays, I want to hear X-rays, and I want to smell dark matter ...
I want to reach out with something other than these prehensile paws and feel the solar wind of a supernova flowing over me.
 

Offline vlindos

  • Contributor
  • Posts: 17
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #851 on: September 12, 2011, 09:50:19 pm »
Drivers could be mod'ed to work with this kern.
Or kernel could be upgraded ?
Or a suitable wifi dongle could be found.

Although I wonder what could be the benefits of all that, It wouldn't make big difference to use the dongle instead of the UART port. Would Hantek ever release the dso.exe source, so we can mod it and make useful things from the network ?
« Last Edit: September 12, 2011, 09:53:43 pm by vlindos »
 

Offline tinhead

  • Super Contributor
  • ***
  • Posts: 1905
  • Country: 00
    • If you like my hacks, send me a donation
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #852 on: September 12, 2011, 10:26:20 pm »
Drivers could be mod'ed to work with this kern.

maybe, depends on many things


Or kernel could be upgraded ?

not a good idea at all


Or a suitable wifi dongle could be found.

this could be easier as the two above. I do have RT2870 stick here, but since i developed LAN addon card for my
hw0 DSO there was no reason to play anymore with WiFi.

However, the current hw revision have different connector, so i will have to re-develop LAN addon board (on my to-do list)
Not that i have issues with WiFi, but cable is always better in lab than a another one "signal source".

Although I wonder what could be the benefits of all that, It wouldn't make big difference to use the dongle instead of the UART port.

you can in principle use whatever technology to made connection between UART and PC, this can be everything
starting from a simple cable and ending with Satellite transmitter and receiver.
Whatever you decide to use ensure that the piece between is not sending anything to DSO while
DSO is booting - this can stop the DSO at bootloader (delayed power supply or just more intelligent things not sending anything
unless activated).

Probably even this crap will work -> ebay 160550090425 (Wireless Bluetooth Transceiver Module RS232)


Would Hantek ever release the dso.exe source, so we can mod it and make useful things from the network ?

no idea how it will cost to buy Hantek, but this will be probably the only way to get sources of dso.exe
It will be probably (haha) cheaper to buy IDA/X-rays ARM decompiler and spend some weeks to decompile
back to C.

Easiest way (to have nework connectivity implemented into firmware) is to wait a bit - currently Hantek/Tekway are
focused on their new Handheld DSO products (look for DSO1000B on their websites).
These handheld DSOs are 100% based (so far i can see from device drivers, PC software, functionality, user manual)
on Hantek DSO5xxxB/Tekway DST1xxxB DSOs.

Why is this important? Well, because when you download the PC software you will already see that there is LAN
connectivity to Handheld implemented, Hantek confirmed also that they will implement similar connectivity
into DSO5000B series. As long their not ready with that there is a big chance that a firmware (actually only
part of the firmware) from Handheld will allow the bench DSOs to connect over LAN.

Sure, this is now to get full functionality over network, for basic things like pure console access
we don't need anything except whatever technology between UART and PC
or addon LAN PCB (which as i said above is already on my to-do list).

The benefit os LAN-Addon board - it will now console, FTP, etc and later to use the PC software to control/get
snapshots from/to DSO, benefit of whatever technology between UART/PC - cheap and sufficient
for shell connectivity.

« Last Edit: September 12, 2011, 10:29:07 pm by tinhead »
I don't want to be human! I want to see gamma rays, I want to hear X-rays, and I want to smell dark matter ...
I want to reach out with something other than these prehensile paws and feel the solar wind of a supernova flowing over me.
 

Offline vlindos

  • Contributor
  • Posts: 17
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #853 on: September 12, 2011, 11:07:31 pm »
Lan would be awesome to be used of course for taking pictures and controlling the UI, but also it might be possible to be used for using the DSO memory for signal patterns checking software.
Lets see what capabilities would have the handheld scopes that would come and we will think for reverse engineering.

'Bout the dso.exe source I think companies that make brave steps are doing big progress, and open sourcing the platform would make even the DYI scope wave bigger than now.
DSO5x02 is awesome ground for learning and hacking. I bet that lots of the buyers are doing the hack after this thread. Extending the software would make this toy even nicer.
 

Offline pgup62

  • Contributor
  • Posts: 33
    • Elec3i Instrumentation et automatisme
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #854 on: September 13, 2011, 05:19:55 pm »
Tinhead wrote:
For those of you who bought Tekway or Hantek in last days and have followed firmware version loaded:

2.06.3 - 110728
2.06.3 - 110801
2.06.3 - 110806
2.06.3 - 110808

All these versions have a ugly bug (in my opinion), the long memory is not available when you
select 8ns/div, 4ns/div and 2ns/div timebase.

Well I tried 3 x DSO5102B and 1 x DSO5062B, all with 2.06.3 - 110808.0 from Hantek this morning, and they all work without this bug ! ;)
In principle such capability it could be removed for reason
by both manufacturers, but since the zoom is not working then (select 1M, 20ns/div, do stop and zoom to 8ns
- and waveform disapears) i think this is a bug and not a feature.

The same as above, none of my DSO show a lack of waveform, all is correct.... So, perhaps they changed something and updated the fw with the same fw number? :)
« Last Edit: September 13, 2011, 05:22:11 pm by pgup62 »
 

Offline tinhead

  • Super Contributor
  • ***
  • Posts: 1905
  • Country: 00
    • If you like my hacks, send me a donation
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #855 on: September 13, 2011, 05:22:05 pm »
heh, once again perfect timing, i posted 110531.1 downgrade firmware and
today Hantek released new public fw version 2.06.3 110909

http://www.hantek.com/download/DSO5000B%20Firmware(2.06.3).rar

(don't get confused, update file say 110909 but after update you will see 110908)
fixing all these issues in firmwares between 110531.1 and 110901.0


Bugfixes in this firmware:
- long memory bug fixed (known from 110728.0 to 110808.0)
- XY bug fixed (known from 110531.1)
- fine timebase bug fixed (on hw0 only, known from 110728 and 110801.0, sometimes on 110806/8)
- contrast settings bug fixed (don't remember when this bug was observer first, i believe after 110531.1?)
- double window zoom scala bug fixed (known from 110728 and 110801)

New features in this firmware:
- timebase allow fine and coarse settings
- FFT improved, is now much more accurate (actually constant 100-120 wfrm/s instead of up to 360 wfrm/s as before)
- tracking cursor implemented
- sysrunparam button available (utility, page 2, F5) -> you can see all settings at once, also the sampling rate in
  current timebase/memory depth setiing.

Known bugs in this firmware:
-- tracking cursor in FFT mode is not catching signal peak nor manually selectable as marker
-> workaround: just ignore tracking cursor in FFT, use as before the two other cursors

-- the sampling rate of FFT displaying 100MS/s instead of real sampling rate when you enable and then disable dual window (F7)
--> workaround: just turn timebase one position up or down and back. No idea where this crap is coming from, probably
just a default value. It seems to be only bug in 800ns/div to 2ns/div, lower timebase ranges are not affected. It is also visible
when you already in dual window mode with FFT enabled and switch euq. sampling and back to real time sampling.


Known bugs from earlier fw versions:
-- sometimes, when you powered off DSO when it was in Display->Cursor setting position, when you power up
and push immediately after DSO booted up (the first wave screen just showed up and the text "Use V0 to ...")
the firmware will freeze

-> workaround:
do once default setup, change all settings to your pref. setup and save it once. This is mostly error coming from
setups saved with different firmware version, or just stop pushing every button right after DSO
booted up
(this bug has been found by marmad, he seems to be very detailed doing reviews)


-- Display Type Dots/Vectors Icon on top left side of screen not changing when switching between modes
-> workaround: none yet, sometimes it works when you enable dots, equ. sampling and back real time sampling,
but atucally this is only cosmetic thing of bug.


-- chinenglich menus, all languages, except chinese are not perfectly translated, even english is not what
you would expect on a DSO
-> workaround: i will post today later english menu fix compatible with this firmware version, when you like it tell this
Hantek/Tekway, when not use the original firmware to update again over my version. Not that this is really important to
have, but i just don't like words like "Unlitmited" instead of Unlimited or just limit off, or "Close" instead of "Off" :)


-- FFT span higher than sampling rate divided by 2 in 800ns/DIV to 2ns/DIV. When you enable full screen FFT
in affected timebase settings the 8 grid divider is already sample rate/2 (for example in 400ns/DIV the
sample rate is 400MS/s, the FFT center 125MHz, 8 divider on grid 200MHz and full span 250MHz).
This means everything above 8 grid is showed with wrong amplitude because of not suffient high enough
sampling rate/ span ratio. Sure, you could say now "who cares, the FFT abouve 200MHz is not impoerant",
well that's partially truth - when you ahve harmonics on let say 220MHz, they will produce unnecessary
alias signals below 200MHz only because FFT is nto sampling fast enough.
-> workaround : i will post today later an update to this firmware version correcting this issue.
Basically what this update is doing is to setup the sampling frequency higher in affected timebases.
How this work in deatil - later.


-- default setting button resetting color settings, sure default might be everything set to defualt, but on the other
side why? Color setting is not something DSO-setup related.
-> workaround: none yet


-- no warning to user when zoom limit reached in single window mode . Description here
https://www.eevblog.com/forum/index.php?topic=1571.msg63195#msg63195
https://www.eevblog.com/forum/index.php?topic=1571.msg63223#msg63223
-> workaround: when you zoom, remember the max values x1000 in dual window, x10000 in single window with dots display,
x20000 in single window with verctor display
.


-- Scrolling 1M and 512k sampled signal at 4ns and 2ns timebase doesn't work. The signal disappears and the time scale is "jumping".
-> wirkaround: none yet. This was working fine in 2.03.x firmwares and of the the 2.05 or 2.06, but it does not work anymore again.


-- Digital Filters not working - the implementation of digital filters is horrible, they doing everythign but not what they should do.
-> workaround: none yet. My proposal - make dotation!!! I will then buy Hantek and fire these engineers!!


Final note:

when you have Tekway please ask Tekway or your fav. dealer to send you latest firmware version.
When you don't wan't to wait, you can still unpack the Hantek firmware and update manually over UART
or use the Hantek firmware directly - note this will replace Tekway logos with Hantek and rename
the DSO to Hantek naming, but it will not break anything
« Last Edit: September 14, 2011, 11:53:44 pm by tinhead »
I don't want to be human! I want to see gamma rays, I want to hear X-rays, and I want to smell dark matter ...
I want to reach out with something other than these prehensile paws and feel the solar wind of a supernova flowing over me.
 

Offline tinhead

  • Super Contributor
  • ***
  • Posts: 1905
  • Country: 00
    • If you like my hacks, send me a donation
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #856 on: September 13, 2011, 05:25:47 pm »
Well I tried 3 x DSO5102B and 1 x DSO5062B, all with 2.06.3 - 110808.0 from Hantek this morning, and they all work without this bug ! ;)

i had here 8 DSOs from Hantek, two other dealers i know (Hantek and Tekway) had at same time also
some new DSOs - all affected by this error.
Hantek got also bug reports and confirmed it to me as a bug. Actually it has been fixed, see above.
I don't want to be human! I want to see gamma rays, I want to hear X-rays, and I want to smell dark matter ...
I want to reach out with something other than these prehensile paws and feel the solar wind of a supernova flowing over me.
 

Offline pgup62

  • Contributor
  • Posts: 33
    • Elec3i Instrumentation et automatisme
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #857 on: September 13, 2011, 05:28:59 pm »
yes, maybe, mine are new from yesterday  :P
Whatever, updating with last version cannot be wrong... 8)
 

Offline tinhead

  • Super Contributor
  • ***
  • Posts: 1905
  • Country: 00
    • If you like my hacks, send me a donation
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #858 on: September 13, 2011, 06:02:20 pm »
yes, maybe, mine are new from yesterday  :P
yeah, perfectly possible, there wa a lot of beta versions in between, at some time
Jonas got pissed off and stopped to answer my mails and i stopped to bother him
to cool down the situation a bit :P You know how "nice" i can be sometimes to other people.

Whatever, updating with last version cannot be wrong... 8)

yeah, it is worth to update. The 110909 (or 110908) is a good firmware, and
for those who don't like it we have a downgrade firmware too, so not a big deal at all.

Pascal, as you there, did you plan to order some of these DSO1000B/S handhelds?
« Last Edit: September 13, 2011, 06:05:06 pm by tinhead »
I don't want to be human! I want to see gamma rays, I want to hear X-rays, and I want to smell dark matter ...
I want to reach out with something other than these prehensile paws and feel the solar wind of a supernova flowing over me.
 

Offline walt

  • Contributor
  • Posts: 42
  • Country: ua
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #859 on: September 13, 2011, 07:28:30 pm »
Known bugs in this firmware:

Hi, tinhead !
Please, tell Hantek don't reset color of screen, when pressed "default configuration button".
It is no sense! 
 

Offline tinhead

  • Super Contributor
  • ***
  • Posts: 1905
  • Country: 00
    • If you like my hacks, send me a donation
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #860 on: September 13, 2011, 07:31:00 pm »
and here as mentioned above my version of the 2.06.3 110909 (110908) firmware

what better :

- chinenglish menu is now english (more or less, any comments?)

- FFT span bug fixed. Now you have in short memory following sample rates,

in single channel:

800ns/DIV - 250MS/s instead of 200MS/s
400ns/DIV - 500MS/s instead of 400MS/s
200ns/DIV - 1GS/s instead of 800MS/s
80ns/DIV - 1GS/s instead of 800MS/s
40ns/DIV - 1GS/s instead of 800MS/s
20ns/DIV - 1GS/s instead of 800MS/s

8ns/DIV, 4ns/DIV, 2ns/DIV - unchanged, 1GSs
40s/DIV to 2us/div - unchanged as they are not bugy.

in dual channel:

800ns/DIV - 2 x 250MS/s instead of 2 x 200MS/s
400ns/DIV - 2 x 500MS/s instead of 2 x 400MS/s
200ns/DIV - 2 x 500MS/s instead of 2 x 400MS/s
80ns/DIV - 2 x 500MS/s instead of 2 x 400MS/s
40ns/DIV - 2 x 500MS/s instead of 2 x 400MS/s
20ns/DIV - 2 x 500MS/s instead of 2 x 400MS/s

8ns/DIV, 4ns/DIV, 2ns/DIV - unchanged, 2 x 500MS/s
40s/DIV to 2us/div - unchanged as they are not bugy.

Long memory mode is unchanged as it can't be used with FFT together. I know i could speed up it to
500MSs instead of 400MSs, but for now i will not do it.

Feedback is of course welcome, post it here, this is inoficiall firmware. Hantek/Tekway has been informed
about this FFT-fullspan bug and asked to fix english menu, so let's hope one of the future fw versions
will have it implemented (well if not i will patch again).

###################################################################

If you wish to know what has been changed, here more deep informations:

The sampling rate will be changed based on timebase id, so when you on 2ns/div the ID is = 0
When you look on that code

FpgaSel800M1G      
      MOV   R12, SP
      STMFD   SP!, {R11,R12,LR,PC}
      SUB   R11, R12, #4
      BL   Fpga_CheckIsShortStore
      CMP   R0, #0
      BEQ   loc_4DAD8
      BL   Dso_GetWindowTBID
      CMP   R0, #2
      BGT   loc_4DAD8
      BL   Fpga_SetOneChipSpsEQ125M
      LDMDB   R11, {R11,SP,PC}
loc_4DAD8         
      BL   Fpga_SetOneChipSpsLE100M
      LDMDB   R11, {R11,SP,PC}
; End of function FpgaSel800M1G

the cmp r0, #2 is checking if the timebase ID is higher or not as 3 (0_1_2 , so 2ns/div, 4ns/div and 8ns/div)
The hack is in principle a simple cmp R0, #8 to allow in the highest 9 timebases to sample with 125MHz (x amount of ADCs).
So as you can see nothing wild, in principle just an extended capability.

###################################################################

How to use:

- warm-up DSO
- download the original Hantek or Tekway 2.06.3 110909 firmware for your model, run an update
- after reboot check if everything ok, if so copy attached update file to empty stick and update the firmware again
- do "default setup"
- run self-calibration
- setup you DSO to you pref. values and save (overwrite) any previous setups.

After reboot you will see the changed menu (only english changed) and the new sample rates in affected timebase modes.

Attached file is splitted due forum restrictions, run the executable to get the dst1kb_9.99.9_cli(111111.0).up file.

For Tekway users - when you don't have the org. firmware from Tekway or your dealer and you don't want to
rebrand your device as mentioned 5 postings above you can still use my custom firmware to enjoy the
updated sample rate and english menus - the org. Tekway (well or Hantek) firmware is only necessary when you
own DSO with firmware version between 110728 and 110808 as it will fix the contrast bug (these versions have bad device driver).
This means when you have 110531.1 on your Tekway you can use my firmware directly.
« Last Edit: September 13, 2011, 08:37:39 pm by tinhead »
I don't want to be human! I want to see gamma rays, I want to hear X-rays, and I want to smell dark matter ...
I want to reach out with something other than these prehensile paws and feel the solar wind of a supernova flowing over me.
 

Offline tinhead

  • Super Contributor
  • ***
  • Posts: 1905
  • Country: 00
    • If you like my hacks, send me a donation
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #861 on: September 13, 2011, 08:41:40 pm »
Please, tell Hantek don't reset color of screen, when pressed "default configuration button".
It is no sense!

Hi walt,

indeed, this didn't really make sense - except it does ? I will try to get that fixed (means will ask them).
I don't want to be human! I want to see gamma rays, I want to hear X-rays, and I want to smell dark matter ...
I want to reach out with something other than these prehensile paws and feel the solar wind of a supernova flowing over me.
 

Offline pgup62

  • Contributor
  • Posts: 33
    • Elec3i Instrumentation et automatisme
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #862 on: September 14, 2011, 06:44:08 am »
Quote
Pascal, as you there, did you plan to order some of these DSO1000B/S handhelds?

Yes, of course, Elec3i will do, We can report some investigations about them if Hantek allow us. Hand-held DSO is a good market target and we will continue to promote them. 8)
« Last Edit: September 14, 2011, 06:46:32 am by pgup62 »
 

Offline rf-loop

  • Super Contributor
  • ***
  • Posts: 3071
  • Country: cn
  • Born with DLL21 in hand
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #863 on: September 14, 2011, 08:55:57 am »
Downloaded from Hantek side FW 110909.0

Buuh..

Again, the sampling speed and memory as well as scaling the screen messed up.

How hell they do this... maybe they think that not even c tests need make in lab. Customers do first c tests, then b tests... never.

I hope they do not implement now ANY new feature before they repair this buggy softwarespaghetti.

Before they publice new FW they need take paper, pen and start in good order search and check... they need one boss who watch every bit what they do and keep they in good order.
I'm not c tester so I do not report any problem to Hantek becouse they are totally lazy small boys wha do work just as spaghetti programmers do without order, without good documentation.

This is only one and small example:
Set display dots
400us/div
1M memory
capture.
push Utility
look SysRunParam  it show 100MS/s, 800000 sample dots, 800000display dots
Stop
Cursors
Turn sec/div 20ns/div  ok you see dots but not 10ns distance, they are 8ns,
ok turn 8ns/div  oh well now dots are 6.4ns and only left half of screen
try horizontal position: can not move
turn 4ns/div all disappears

Ok same agen but 200us/div
oh, can not select 1M, only 512k
push Utility
look SysRunParam  it show 100MS/s, 400000 sample dots, 400000display dots
cursors
Stop
Sec/div to 20ns
measure 10 dot interval  = 107ns
turn 8ns/div measure dot interval 12.8ns
turn 4ns/div now dots only left half of display
measure dits interval = 6.4ns

and lot of more.

Oscilloscope is not for only show nice pictures... it is measuring tool and specially oriented for time domain measurements.
But what can measure if can not trust what scope show. maybe risetime and all measurements are just this sime bullshit?
Who knows. First is that you need trust your test eguipment. How I can trust this before I check all individual settings and do full measurements and then I can know what this oscilloscope show.

What if I have not any other oscilloscope or test gears..  Then I do measurements and go some place and seriously tell that I have measured this and this... here is numbers - test dta. If I tell that test data is coming from Hantek oscilloscope. Do they cry or laygh. So, for hobby playing with "nice ebents on the screen" it is ok but not any kind of serious work where need get trusted data what accuracy is inside published specifications. These kind of serious work nee Agiloent, Tektronix, R&S... ??

I can not accept this kind of careless loosy paygame working. maybe they play more mahjong than check and test what they have done. Maybe it is not important. But trusted name they do not get never with this kind of game.

Maybe we need one  new study book. "Hantek mathematic fundamentals" so we can ma make corrected calculations betwen truth and scope.
Mode dots, 200us/div 1Mpoints,  10MHz signal from Rubidium standard to scope. Capture.. Stop, change to 8ns/div
measure freq on the screen. 11.7MHz so use 0.855 but specs do not tell that need use this constant for multiply.
Of course then I do other measurements and can find that other way give different. IF I do not know anything about my signal and I need only trust my scope. Which one I know is right... 10M or 11.7MHz. Is it really right to guess that becouse Freq counter give 10MHz do not believe captured data or do not believe 11.7 becouse other kind captured waveform with same scope show different. Oh boys where we are? We nee d first know our signal and then look it with scope.. well. Why need then scope if I know my signal. I have several tons of equipments I can check but why I need then Hantek if I need always check with other machines. First thing what need be over tond of features. Scope signal guality and capturing accuracy and data displayed need be enough accurate. This is not movie and play machine it is as called "test equipment" for making measurements... I think sometimes it is forgetted.

Btw, display vectors have not this problem. So how they make vectors if they do not use these sampled dots, or is it so that display dots are not sampled true dots at all? What is meaning if displayed dots are not sampled datapoints?
If they can scale vector display why not these dots? 

Tracking cursors, really nice thing!
Even better if it can select track t or v
« Last Edit: September 14, 2011, 10:31:01 am by rf-loop »
If practice and theory is not equal it tells that used application of theory is wrong or the theory itself is wrong.
-
Harmony OS
 

Offline tinhead

  • Super Contributor
  • ***
  • Posts: 1905
  • Country: 00
    • If you like my hacks, send me a donation
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #864 on: September 14, 2011, 02:38:05 pm »
First of all the 200us/div and 1M was not working from the beginning. Right after i recognized it on a very
first firmware i send it as bug report - i don't know when exactly but one on next firmwares was able
to enable 1M in 200us/div but zoom was not working. So i asked again and got answer
that the way how the DSO is designed they can't enable 200us/DIV and 1M - the next fw version then was
not capable again of doing 200us/div and 1M. I thought, what a bummer, but anyway,
it works below and above 200us/div, or just with 512k while in 200us/div.

Now the answer i got was send by Tekway long before they got bought, feel free to send Jonas
an email and ask him why this is like it is, maybe Hantek engineers can do more that Tekway org.
designer was capable of (i doubt that, but let's try).

This is only one and small example:
Set display dots
400us/div
1M memory
capture.
push Utility
look SysRunParam  it show 100MS/s, 800000 sample dots, 800000display dots
Stop
Cursors
Turn sec/div 20ns/div  ok you see dots but not 10ns distance, they are 8ns,
ok turn 8ns/div  oh well now dots are 6.4ns and only left half of screen
try horizontal position: can not move
turn 4ns/div all disappears

Ok same agen but 200us/div
oh, can not select 1M, only 512k
push Utility
look SysRunParam  it show 100MS/s, 400000 sample dots, 400000display dots
cursors
Stop
Sec/div to 20ns
measure 10 dot interval  = 107ns
turn 8ns/div measure dot interval 12.8ns
turn 4ns/div now dots only left half of display
measure dits interval = 6.4ns


Mode dots, 200us/div 1Mpoints,  10MHz signal from Rubidium standard to scope. Capture.. Stop, change to 8ns/div
measure freq on the screen. 11.7MHz so use 0.855 but specs do not tell that need use this constant for multiply.

all 3 examples are with zoom higher than 10 000, if you stay within x10 000 everything will be fine (for dots, vectors see below)
Out of curiosity, with which version you saw this working? Becasue none of the firmwares i have here is or was ever
capable of doing properly zoom > 10 000/20 000, not even the first pre-production firmware from Dec 2009.

Btw, display vectors have not this problem. So how they make vectors if they do not use these sampled dots, or is it so that display dots are not sampled true dots at all? What is meaning if displayed dots are not sampled datapoints?
If they can scale vector display why not these dots? 

no, vector have same problem, due the better interpolation you can use up to 20 000 zoom.

So for example 400us/div zoomed to 20ns/div (x20 000) will still work, but 200us/div zoomed to 8ns/div
not anymore. when you don' t believe (luckily there is fine timebase now, no need to patch timebase table to see it)
change the timebase to 10ns/div and you will see the sampled 10MHz signal (with 512k, 200us/div, 20000 zoom)
have length of 100ns. Now when you do the sample for dots sampled you will see 10ns/div is not working
anymore - that's because it can only 10 000 zoom.

Of course this is kind of bug, but actually hard to restrict the timebase, so this is one
of these things where the user have to think a bit about. When you select 2ns/div you
will immediatelly not wonder why you see only one or two dots or just a ine between 2 dots -
because you reached already the smpling / screen size limit.

Now in dual window mode it is easy to restrict zoom in firmware - there is 1000 max, in single
window due better resolution 10 000 for dots and 20 000 for verctors ut you have
only one timebase, so the conditioning is bit more complex (when limit zoom, how limit zoom,
why limit zoom, why just not display "warning, zoom out of accuracy").

I don't know, maybe Hantek is smart enough and can fix that, but knowing that they
never did it (and i'm sure at least the org. Tekway engineer was knowing this restriction) i doubt
they can do it. Again, make a snapshots and ask Jonas. I can work with that, restrictions
are there to be followed.


200us/div, zoom to 20ns/div, measure 10 dot interval  = 107ns

sure, you have 10 dots, sampled with some jitter, multiply by 10 the jitter and you will get something between
104ns (103.5ns) and 106.5ns (well, 107ns are next to 106.5ns). You can actually see the jitter, when you exactly measure
the time between each dot you will see sometimes the difference, sometimes positive, sometimes negative,
sometimes exact time. There is nothign wrong with that (well, maybe the 500ps trigger accuracy on
paper but 350ps to 650ps in the reality, sure avg 500ns but it should be always below 500ps to
follow western standards of datasheets).



I actually don't give a f** what ppl think about Hantek and the way how they work.

I do know Tekway was different - sure more complex to communicate, but that was typical for small company
(well, actually departement of a big company working on own costs with own company name).
But the product which i chose, bought, tested and later announced here is not bad, actually one of the
best you can get for money - the firmware is (when you forget here and there some small things)
really good one - imagine Owon with this firmare, there will be no competitor on market for
long time.

On the other side these small things in the firmware which we can't change, they could be
implemented better with no cost. Everybody would accept that there is max zoom 1000/10000/20000,
when the manufacturer will not hide this fact but clearly show to user a small warninng (nothing else,
i still wish to see more than 1000/10 000/20 000 zoom !!! even if the accuracy is gone)
- "zoom limit reached, out of accuracy", there is nothing to shame about,
1000/10 000/20 000 zoom are acceptable values.

What not acceptable that the manufacturer is not clearly talking about, this remains me on these
faked sample rate things on Rigol. Of course most chinese manufacturers hiding informations,
not only because of competitors but also because they don't know that in western countrys
a good manufacturer is telling everything about the product .... ups, actually not, actually
every manufacturer is trying to cheat a bit :)


EDIT: you know what, i add it into the bug list from page 60 even if i don't see it as a real bug,
but it should be not that complex for them to add just a small warning.
« Last Edit: September 14, 2011, 06:04:33 pm by tinhead »
I don't want to be human! I want to see gamma rays, I want to hear X-rays, and I want to smell dark matter ...
I want to reach out with something other than these prehensile paws and feel the solar wind of a supernova flowing over me.
 

Offline tinhead

  • Super Contributor
  • ***
  • Posts: 1905
  • Country: 00
    • If you like my hacks, send me a donation
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #865 on: September 14, 2011, 06:15:06 pm »
and here as mentioned above my version of the 2.06.3 110909 (110908) firmware

wtf .. only 10 download?

Don't worry, this will not break your scope (and if, you will get new one from me), it will make it faster!
I don't want to be human! I want to see gamma rays, I want to hear X-rays, and I want to smell dark matter ...
I want to reach out with something other than these prehensile paws and feel the solar wind of a supernova flowing over me.
 

Offline walt

  • Contributor
  • Posts: 42
  • Country: ua
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #866 on: September 14, 2011, 07:59:31 pm »
wtf .. only 10 download?

I have already :)

A special thank you for the calibration method with averaging. That is really what is missing.
Now is not ideal, but not disgusting.
 

Offline rf-loop

  • Super Contributor
  • ***
  • Posts: 3071
  • Country: cn
  • Born with DLL21 in hand
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #867 on: September 14, 2011, 08:42:06 pm »
Mode dots, 200us/div 1Mpoints,  10MHz signal from Rubidium standard to scope. Capture.. Stop, change to 8ns/div
measure freq on the screen. 11.7MHz so use 0.855 but specs do not tell that need use this constant for multiply.

This is not any kind of jitter origin. It do this exactly ald systematic all times, exactly.
There is not basic mathematics, they use some hantek mathematics for scaling. Some place have some mistake.

And this example is only one example.

Why I need play this game. I do not want. This is basic work what need do inside Hantek. What hell they do if they can not do these things how many pixel is one div etc..  I feel very stupid if I need test and measure every feature in scope. If can not zoom something without loosing math... then this posibility need close. What 20000 zooming. This data is exactly in memory, every single sample point. Only what need do is pick up them to screen so that displayed time is right enough.
10:11.7  ... this can not accept. I understand well that always can not show exactly becouse screen have only limited resolution.  It can not be very difficult higher level math if want know what is every sampled point time becouse this sampler is really simple becouse interval is fixed. If it use random sampling method with equal time mode it is littlebit more difficult but this is constant interval sampling. (yes if they use small distange calculation ands then scale it and if programmer do not care and go to error accumulate trap... wel then is just only "lack of enough experience for this work"

Also what about jitter. This need not take account. In calculus there is some sampling speed and then some time interval between samples. Multiply 2 and if you multiply it ten or ten million it is just as it is... zooming factor do not mean anything in this case. 100MS/s means 10ns sampling interval. If there is 2, 4, 8 20 or what ever time/div  just plot these pxels right place.

And this problem do not occur with connect dots mode. Zoomed or not but it keeps 10MHz as 10MHz not 11.7MHz Fun?

I do not want more accuracy than is normal for these machines. But basic fundamentals need be right.
If practice and theory is not equal it tells that used application of theory is wrong or the theory itself is wrong.
-
Harmony OS
 

Offline drieg

  • Regular Contributor
  • *
  • Posts: 85
  • Country: cz
    • Silcon Electronics
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #868 on: September 14, 2011, 09:30:16 pm »
I can't but agree with rf-loop on many things. It seems to me too that they have no system of fixing bugs. I mean I'm happy too that they added tracking cursors, but there are still many bugs that should have been already fixed long time ago. Slowly I'm becoming tired reporting again and again same bugs, filming videos (because they almost always ask for it) and asking for the answers after many days of silence (now I'm talking especially about Hantek). I have no time for that and I have to agree with rf-loop again, I'm also not a tester, I'm a customer (not takling now about testing new features like tracking cursors, which I'm always happy to do).

Here are some old bugs that bother me most:

1. Filters - the implementation of digital filters is horrible, my 97 year old grandmother would do it better! The Low Pass filter does only attenuation of amplitude based on selected frequency. It doesn't filter out high frequencies as you go down with the cut-off frequency. Try 1kHz square probe compensation signal and you won't see any "corners rounding" or slowing down the rising edges (until the square looks more like a sine) as you would expect. Back to scool, Hantek engineers!  ;)
On the top of that, selecting any filter massively will slow down the  total response/performance.

Not only Low Pass, but also High Pass filter doesn't work correctly and doesn't show the cut-off freqency correctly when you change timebase settings. Try sine wave, find -3dB cut-off frequency and try to change timebase. The cut-off frequency value changes but the filter settings don't. Then try to press F4 and the value changes again. Crazy  :o

Last thing - select Low Pass and change trigger to Alternate mode. The scope freezees up.

2. Scrolling 1M and 512k sampled signal at 4ns and 2ns timebase doesn't work. The signal disappears and the time scale is "jumping".

3. Display Dots doesn't work in 2ns and 4ns timebase

...

Thanks for the FW patch, tinhead, you're an untiring fighter! :)
« Last Edit: September 15, 2011, 01:32:29 pm by drieg »
Bricked Rigol? This thread might be of any help.
 

Offline tinhead

  • Super Contributor
  • ***
  • Posts: 1905
  • Country: 00
    • If you like my hacks, send me a donation
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #869 on: September 14, 2011, 10:18:52 pm »
Mode dots, 200us/div 1Mpoints,  10MHz signal from Rubidium standard to scope. Capture.. Stop, change to 8ns/div
measure freq on the screen. 11.7MHz so use 0.855 but specs do not tell that need use this constant for multiply.

again, when you out of limit you can not measure anything  properly, stay within org. timebase/10000 when dots
or /20000 when vectors and you will measure proper values.

This data is exactly in memory, every single sample point. Only what need do is pick up them to screen so that displayed time is right enough.

didn't matter what in memory, when the zoom is only able to calculate back data from memory to display
by max. factor /20000 then it is like that. When the timebase would stop at /20000 you will even not
see any crap values, thats all about (or well, you will probably say then "why i cant zoom to 4ns/div?)

One possible solution as already is notice user about zoom limit, another possible solution is to change the
10000/20000 values by whatever max. possible.

Sure you will see then only a long line cross the screen or maybe one dots on whole screen, but this is
actually what you suppose to see by max. zoom (except the sampled dots are left/right out of screen
and the interpolation can't dispaly anything because no data, but this can be easily fixed by increasing
interpolation window size)


10:11.7  ... this can not accept. I understand well that always can not show exactly becouse screen have only limited resolution.  It can not be very difficult higher level math if want know what is every sampled point time becouse this sampler is really simple becouse interval is fixed.

and once again, when zoom out of what the zoom code allows, the data displayed on display is not anymore valid


Also what about jitter. This need not take account. In calculus there is some sampling speed and then some time interval between samples. Multiply 2 and if you multiply it ten or ten million it is just as it is... zooming factor do not mean anything in this case. 100MS/s means 10ns sampling interval. If there is 2, 4, 8 20 or what ever time/div  just plot these pxels right place.

you wrong, there is no and never 10ns sampling step. Each ADCs is having own clock, the jiter apply to
EVERY clock, so you have always differences. What EVERY interleaved DSO is doing is to align the data
during postprocessing based on calibration data. Now the calibration data is having exact the same random jitter error
so it make possible to align the ADC data to see waveform, but the jitter remains.
All these pixels are in right place (except those which has been calculated back when zoom what out of limit),
when the skew in clock created by jitter was 200ps you will see that, there io no magic math, code or engineer
who can fix that - it has been sampled 200ps later/earlier and it will displayed exact like that.
Now of course when sampling with 400MSs to long memory the effect is higher than when sampling with
1GS to 4k memory, this is not driven by sampling accuracy and time domain differences.
« Last Edit: September 14, 2011, 11:14:39 pm by tinhead »
I don't want to be human! I want to see gamma rays, I want to hear X-rays, and I want to smell dark matter ...
I want to reach out with something other than these prehensile paws and feel the solar wind of a supernova flowing over me.
 

Offline tinhead

  • Super Contributor
  • ***
  • Posts: 1905
  • Country: 00
    • If you like my hacks, send me a donation
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #870 on: September 14, 2011, 10:57:16 pm »
I can't but agree with rf-loop on many things. It seems to me too that they have no system of fixing bugs. I mean I'm happy too that they added tracking cursors, but there are still many bugs that should have been already fixed long time ago. Slowly I'm becoming tired reporting again and again same bugs, filming videos (because they almost always ask for it) and asking for the answers after many days of silence (now I'm talking especially about Hantek). I have no time for that and I have to agree with rf-loop again, I'm also not a tester, I'm a customer (not takling now about testing new features like tracking cursors, which I'm always happy to do).

full ack, i love to test something when asked for, but i hate to test everythign and ask myself "is Hantek not having
enough money to give one porduction unit to the engineer?"


1. Filters - the implementation of digital filters is horrible

what filters ? i don't see any filter in the firmware :P
But honetsly, the FIR blackman filter implementation can not be really
shown as filter. Right after they implemented it i said to my customers "it works",
well it was working somehow, but not really exact as it should (slower than today, not really exact)

Now months after these filters has been implemented it is getting worse from one fw revision to another
(rf-loop would say try and error).

I do accept that they slow working (filters, not Hantek), but i can't accept the implementation.
I have not mentioned this as bug, because bug is part of soemthing what implemented, these
filters are not there in my opinion (just play with band pass or stop, what's that?)

2. Scrolling 1M and 512k sampled signal at 4ns and 2ns timebase doesn't work. The signal disappears and the time scale is "jumping".

bingo, i saw this working in one of the 2.05 or 2.06 firmwares, no time to check all of them
but it works perfectly in 2.03.x (yes, 2.03.x). What a crap, they fixed it already (i remember i posted this as bug
and deleted from my list as fixed) and again same bullshit.

3. Display Dots doesn't work in 2ns and 4ns timebase
hmm, you mean when you enable dots, go to 4ns/div and stop you can't see any dots when
you go to 2ns/div ? Well, they are there, believe me. Even in 2ns/div when you stop they are there,
the point is, you always see interpolated dots between, when you would be able to zoom more
(like now from 200us/div to 20ns/div) you wil see them, but let me show you that on attached picture .

What on the picture is firmware with 200ps/div timebase, i did stopped acq. while in 4ns/div
and switched over to 200ps - and bingo, you can see all these points there.

When the interpolation would be really swichable on/off you would see only the real sampled dots
and not the interpolated between, but it is always on. How this looks like ?
Well, check another one picture, this is fw with interpolation off (2ns/div and stopped acq.)


Thanks for the FW patch, tinhead, you're an untiring fighter! :)

... and Hantek is not responding to my emails anymore. Sure i'm sometimes a big asshole
and talking clear words, they should really pay me for support and fight with all of you :)

Anyway, i will once again send them the list of bugs, the last time ? Who knows, maybe
« Last Edit: September 14, 2011, 11:52:00 pm by tinhead »
I don't want to be human! I want to see gamma rays, I want to hear X-rays, and I want to smell dark matter ...
I want to reach out with something other than these prehensile paws and feel the solar wind of a supernova flowing over me.
 

Offline walt

  • Contributor
  • Posts: 42
  • Country: ua
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #871 on: September 15, 2011, 05:50:52 am »
tinhead , in new firmware, and/or you patched save to usb flash work properly?

I am not sure, may by my flash drive is not good (apacer ah323), but update firmware was from it.
Save picture- not work, oscilloscope report "save complite", but usb drive is empty.  :(

offtop
Gays, with you, i learn English  :)
 

Offline drieg

  • Regular Contributor
  • *
  • Posts: 85
  • Country: cz
    • Silcon Electronics
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #872 on: September 15, 2011, 06:37:25 am »
offtop
Gays, with you, i learn English  :)
..and in the next lesson we'll learn the difference between "gay" and "guy"  ;D
Or did I miss something? Tinhead, you?  :-*
« Last Edit: September 15, 2011, 06:40:56 am by drieg »
Bricked Rigol? This thread might be of any help.
 

Offline tinhead

  • Super Contributor
  • ***
  • Posts: 1905
  • Country: 00
    • If you like my hacks, send me a donation
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #873 on: September 15, 2011, 07:43:15 am »
tinhead , in new firmware, and/or you patched save to usb flash work properly?

I am not sure, may by my flash drive is not good (apacer ah323), but update firmware was from it.
Save picture- not work, oscilloscope report "save complite", but usb drive is empty.  :(

it is your flash drive.

The firmware work perfectly here to save screenshots, setups and csv export/import changed/unchanged
sampling rates and memory depths.

 
I don't want to be human! I want to see gamma rays, I want to hear X-rays, and I want to smell dark matter ...
I want to reach out with something other than these prehensile paws and feel the solar wind of a supernova flowing over me.
 

Offline tinhead

  • Super Contributor
  • ***
  • Posts: 1905
  • Country: 00
    • If you like my hacks, send me a donation
Re: Hantek - Tekway - DSO hack - get 200MHz bw for free
« Reply #874 on: September 15, 2011, 07:45:11 am »
offtop
Gays, with you, i learn English  :)
..and in the next lesson we'll learn the difference between "gay" and "guy"  ;D
Or did I miss something? Tinhead, you?  :-*

hihi, depends ... now i'm a bit to old for experiments, but years ago it was a good
game to play gay to get a "chick"
I don't want to be human! I want to see gamma rays, I want to hear X-rays, and I want to smell dark matter ...
I want to reach out with something other than these prehensile paws and feel the solar wind of a supernova flowing over me.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf