Hantek - Tekway - DSO hack - get 200MHz bw for free

[clonecrp]

Hey Colin ... Whew coosh !

Thanks for the info as appreciated by all but ... me ... personally ....
I simply want to change my Hantek from 100 to 200  do you have any advice on that ...?

Please advise...'
Thanks !
Doug

[AnotherRFman]

Hi Carlos

When you say you have found a way to use Tekway's TTScope with Windows 7, do you mean Windows 7  32 bit or 64 bit. I have had it working with 32 bit Windows 7 from the outset but like other members, I am sweating on the 64 bit version becoming available so I can run it on my desktop machine too.

[carloscuev]

Sorry, Win 7 32-bit. Mine wasn't working, dunno why.

[tinhead]

Tinhead,You saw that my upload files?Because i didn't have enough power to do that now. Only if i need to  enough to know to  do it .Can you tell me what the difference is between the received files and your previously  files,and reciprocal measures ?(if you can post the better nandflash backup file of dst1602, so much the better.)I only know a little Linux.used all your upgrade applications, always  reported error code . I don't know what Tekway did?

yes i saw and got it, i will check on weekend (quite busy right now)

ccatt

sorry, but due some undone projects there was no time for fun, i will take a look this weekend.

[tinhead]

Tinhead ...

Removed my new Hantek DSO5102B from shipping container here .... applied power .... the display is quite nice at quick look.
... looking around etc ... I notice Acquire Menu - 1M LongMem is not supported by the model. Is this a SW, HW issue or both ?

Very Interesying Features  ... I think I am going to like this

Probably will do Power Supply mods this week ... got the TI PTH04070WAH Chip as sample.

Comments Please ...

Thanks!
Doug

DSO5102B doen support 1M only in one chan mode, so if even this did'nt work then you have an issue with your DSO (if so, tell me fw version)

[tinhead]

Hi Tinhead.

I've been waiting for the Windows 7 64 bit driver for TTScope. Do you know if this has been released.

Thanks for all the work that you do for us. I also wouldn't have a Tekway 200MHz oscilloscope if it wasn't for you.

RFman.

The drivers isn't really that complex, should be not a big deal to develope 64bit version, unfortunately their
to-do list seems to be much longer than free resources

[tinhead]

1:  I think this is not a good way of understanding what is happening.  Any DSO can under-sample.  I don't care who argues, any DSO can alias and show nonsense - I've done it myself and seen it on 1GS/s scopes with 500MHz bandwidth and huge memory that cost $70,000!!  I don't wish to sound horrible, but there can no argument about this. 

sure, that's what these chinese DSOs are doing is like a combination of flat and gausian response. It is far from perfect, but we should not
forget - that low cost DSOs.

It is a fact of the universe, like the speed of light. 

typical humans ... everything beyond horizon is not possible

If you sample in anyway, you can get alias. 

and of course you can see alias on these DSOs, no doubt. One of the good things in Hantek/Tekway DSOs is hardware freq. counter,
for most ppl the freq. diff on both display should be an indicator for "somethign wrong on the display".

There are some compromises with the Hantek/Tekway for long memory, but they don't make it clear when and how those compromises happen, an information screen is absolutely necessary!

yeah, that's again price driven issue. Actually for real 1M on full speed they should add some additional SRAMs, FPGA should be a bit faster too.
But they just can't, nobody would even think about to buy chinese DSO for 3k $ if there is Tek or Agilent for same price, just because of brand name.

Basically, I believe that a DSO should work for the user, not against the user.   Most Chinese models I have seen don't seem to do this.  It is a shame, I have no trouble with cost based limitations to a design.  Honest disclosure shows respect for the customer.  Would you get into an aeroplane that had "aerobatic" written on it without knowing the exact limits for vertical and inverted acceleration??? *laughs*

i saw so many errors or strange menu on brand name DSOs, even such nice things like touch screen are useless (at least for me) - sometimes
is hard to understand what the developer was thinking (or not) in first place.

I think for you the best would be to get something like an Agilent DSOX3000, there is enough memory running with full speed, much faster waveform
refresh and some nice addons.

[tinhead]

Hello, I've been using my DST1102B (my first scope ) to design a microcontrolled buck converter and everything has been pure sweetness. The only things ths bugs me is when you save to USB, why does it save a .bmp and .gif of the picture? I can do just fine with .gif, Is there any way to make it only save .gif file? Or better ! is there any way to get fast oscillograph screenshots in the PC without using the slow TTScope?

I've found a way to use Tekway's TTScope in windows 7, Just copy and rename:
<Hantek's TTScope Installation Dir>\config\DSO5202B_eng.ini
to
<Tekway's TTScope Installation Dir>\config\DST1202B_eng.ini

This way I can use Tekway's TTScope, but it's still slow I suspect the scope sends *.bmp screenshots instead of some type of compressed image. Is there any third party software ?

fro Win 7 32bit the software from Hantek website is working perfect. The DSO5102B is still sending back identification "DST1102B",
which req. of course proper ini file. Normally this file was included within the TTScope setup, probably again someone updated and forgot
to add everything - work around is exact what you did.


Regards the screenshots, the DSO is doing BMP from the captured memory, then bmptoppm and ppmtogif Sounds crazy but that's the out-of-the-box
tools on linux, so of course the developer could not resist to use them. You can rename the bmptoppm and ppmtogif to something, so you will get only bmp -
or you have to patch firmware to remove the bmp after conversion. However, on the to-do list for future firmwares is an option to select what
will be saved during quick action "save to usb" (picture or/and bin/text file), let's hope this will be implemented soon.

Actually there is a way to get the sampled data from the USB port (and a way to do many custom things), there is a someone working
already on custom SDK, many things are already working but as he is doing without any support from Hantek is costs reverse engineering time.
The good news is, it will be soon published, so please be patient.

[ccatt]

I see.Creation of files takes time.Thanks so much for your time.Compared with a layman myself.your efficiency will Be multiplied several times or even more than .

[colinbeeforth]

Hi tinhead,

I said:  "There are some compromises with the Hantek/Tekway for long memory, but they don't make it clear when and how those compromises happen, an information screen is absolutely necessary!"

Than you said:  "yeah, that's again price driven issue. Actually for real 1M on full speed they should add some additional SRAMs, FPGA should be a bit faster too.
But they just can't, nobody would even think about to buy chinese DSO for 3k $ if there is Tek or Agilent for same price, just because of brand name."

I am not complaining for the sake of it, I understand the compromises in hardware.  I understand memory write times, bus settling times, refresh etc.  I have no problem with compromises used to keep things at a cost the market will pay for.  That's just normal engineering.  Oh, and compared to other Chinese DSOs, Hantek/Tekway are paragons of virtue, some other DSOs are really total rubbish.

Anyway, I've found that so long as you are only using 4k, the FFT screen does show a sample rate.  Why can't they put that on the normal acquisition screen, it wouldn't cost anything?  Why would it wreck their sales?  It 's probably also got a lot to do with culture and their manufacturing environment.

I've been doing some tests with the Hantek DSO, using random fast pulses lately, and like you said, it works fairly well.  However, if we all shrug and go on, it won't teach Chinese manufacturers to make better products.  In 20 years time, if they are open to learning from their customers, they will make better products, and give us good stuff at a reasonable price.  But, we need to tell them what is needed.

Hey, if I had more money, I'd buy Agilent in a heartbeat.  But I don't have that option.  My LeCroy 9310 is really clever, really nice, but only 100MS/S.  Maybe I'm spoiled.  Sadly, the LeCroy power supply maker has gone bust, the switched mode supply has a horrible custom thick film controller (I hate vendor lock in! grr!), it can't be repaired easily, and if it fails, I could buy 4 or 5 Hanteks for the cost of one repair...  So that's why I am looking at the cheaper end of the market.

Thanks again for the excellent work you have done in exploring the software environment of the DSOs, I wish I was better at software, but I'm more of a hardware guy.  Shame we don't live in the same city, we could have some interesting discussions I think.  Oh, it might help if I say that in test equipment and calibration, I like to know - know for sure, not just guess, but you probably figured that out. *waves and smiles*

Cheers, Colin

PS: Maybe I'll send some more email to Hantek, it hasn't done much good yet.

[Gall]

In fact, almost all issues of HanTekWay are software issues. An alternative firmware (maybe including FPGA/CPLD as well, maybe not) could solve them. Hope HanTekWay will eventually go open-source.

[alm]

Hope HanTekWay will eventually go open-source.

Why would they do that? What would be the incentive? Most Chinese companies are even more restrictive about documentation and details than companies like Tek and Agilent, and even they have never released the source as far as I know, even when they published full schematics.

The only exception is WELEC, but that's only because they went bankrupt before releasing a good firmware version, and I don't think the open source version is really usable for end users yet. Not sure how much of the firmware they actually released.

[tinhead]

sooner or later Hantek/Tekway will fix these bugs. For some reason the joined dev. team has to start from scratch, so the original developer "rgj" is not supporting them anylonger ( i guess Hantek thought "we can do it by ourself").
Currently they fixing home made errors, which is typical for learning process when you "overtake" development.

[Gall]

I guess that's why they haven't releases the SDK yet - they just don't have all the sources anymore. I saw such a situation once, I was the one who had to develop new software after the previous main developer ... (to be polite) retired and left only binary compiled version to us.

[tinhead]

I guess that's why they haven't releases the SDK yet - they just don't have all the sources anymore. I saw such a situation once, I was the one who had to develop new software after the previous main developer ... (to be polite) retired and left only binary compiled version to us.

it is of course hard to get internal informations, but in principle every company worldwide is doing exact the same -
1.you buy a company,
2.you plan merger,
3.you fire people because you don't need additional staff,
4.you will get blamed by customers/your product is not "working"
5.you hire contractors because you don't know what to do,
6.you hire again old teams to fix issues.

The good news is - Hantek is already between stage 5 and 6 (they did tried to fix some firmware issues and produced new once,
they released interim non-public fw versions (especially for new hardware) with big bugs, so finally they dropped pants and told the truth
to distributors about development issues due insuficient staff. That's fair enough and good starting point).

From what i know they have definitely SoC sources and at least CPLD/FPGA binaries and documentaton. Which is in principle enough to compile new
firmware (SoC) versions and to publish SDK, but not sufficient to change hardware revisions. The plan was to publish SDK end of jaunuary,
however due the internal dev. team issues they worked on bug fixing and new hardware revision (which caused new bugs).
Finally they deleted all firmwares from web page, dropped pants, and focused on the "right way to work". This is again a good example
for chinese companies, they working much faster than wester-world companies. I'm working since years with Fortune 100 companies on big projects,
and to be very honest within 6 months you will not get even trough all meetings, still light years far away from real work.
Hantek managed to buy, merge, produce mistakes, drop pants, hire people within 8 months which is incredible fast!,
this is in my opinion another one reason to trust their products.

This means for us, we have to wait - for official SDK from Hantek/Tekway or for custom partial-SDK which is almost ready.
A user from this forum spend tons of hours decompiling the firmware and documenting things, let's hope he will publish it soon.

[Bored@Work]

I have no sympathy for companies losing their source code. It is simply gross mismanagement when this happens. Firing people is no excuse for losing assets like source code. In fact, there is no excuse. It is not rocket science keeping track of source code, and the principles aren't new. More like known since 30 or 40 years.

[bilko]

It sounds like they don't do version control either else they would have the archives. The implications are that they can't backtrack and fix bugs properly, hence flaky f/w

[tinhead]

I have no sympathy for companies losing their source code. It is simply gross mismanagement when this happens. Firing people is no excuse for losing assets like source code. In fact, there is no excuse. It is not rocket science keeping track of source code, and the principles aren't new. More like known since 30 or 40 years.

i didn't said they losed source code. Middle of the merger process Hantek designed new hardware revision, and for some reason made
some mistakes within interim firmware, errors which can only happens if you have no sources or if you fired developers
(and new team have not yet "assimilated" the source). As they published in same time some public firmwares they must have at
least part of the source - that's why i said "at least" in my org. posting.

On the other side i know they build new joined team, which means normaly fired some ppl - at least contractors.
The org. developer "rgj" seems to be a contractor. In some countries as contractor you don't own the sources, in some you do.
So it might be that Tekway originaly bought only part of the source, but it can be that they own all sources and the new dev. team
just not managed to understand the source in proper way (which can easily happens in such complex design).
The good part of the story is that they finally recognized this "internal issue", bad aspect of this story that they losed time,
so we will have to wait for bug-free firmware, netwrok support within firmware and SDK a bit longer than promised (and planed by HanTekway).

[tinhead]

It sounds like they don't do version control either else they would have the archives. The implications are that they can't backtrack and fix bugs properly, hence flaky f/w

if you disassemble or decompile the firmware you will see many "improvisations" and things that didn't make sense, version control can
only help if you documment your source and of course don't use such "fast bugfix mechanism" for issue resolution.
In the real world, now talking about over 10yrs experience with Fortune 100 companies, something like that didn't exists.
Sure version control exists, but nobody really cares about clean programming - time to fix is important as this costs money!
So why chinese company should be different? No reason at all.

[bilko]

Sure version control exists, but nobody really cares about clean programming - time to fix is important as this costs money!
So why chinese company should be different? No reason at all.

Clean programming = faster time to fix.
If you let developers run all over the code they cause all sorts of problems and make more bugs than they fix. I worked in industrial automation software development for over 30 years and yes clients were all blue chip automotive and aerospace manufacturers with downtime in the region of $20,000 to $50,000 per minute.
Time to fix was of the essence or people would get laid off.

Not running a proper documented source and VCS is bad and expensive for business and should not be encouraged. The only time it is acceptable is when prototyping or producing products for personal use.

Probably another good reason that HanTekway should release the API, they can concentrate on the hardware and let the developer community decide the functionality of the equipment. Both sides win.

[tinhead]

Probably another good reason that HanTekway should release the API, they can concentrate on the hardware and let the developer community decide the functionality of the equipment. Both sides win.

full ack, and afaik they got during last days some proposals, up to them to decide. In my opinion is not a shame to let ppl help you,
it is a shame to ignore it.

[bilko]

full ack, and afaik they got during last days some proposals, up to them to decide. In my opinion is not a shame to let ppl help you,
it is a shame to ignore it.

There is absolutely no way that HanTekWay can accumulate the knowledge of thousands of developers on the Internet. Their product will see applications that they never dreamed of. Investing software time for them is a pure waste of money and resources. How can they charge more because of their software 'enhancements'. Its still the same hardware and still the same company. Releasing the API and a 'scope development kit' would attract customers from education, industry, R&D. There isn't a product on the market as capable within this price band. Lets hope that they see sense and pursue this market.

[drieg]

Hello to all HanTekWayMen!

I've been watching this thread with interest for some time already and I just could not resist any longer from getting one of these units.
Nice work, tinhead!

In fact I got couple of them. They all are Tekway DST1062B model with English front panel and the DST1102B set of resistors (Rinp = 33R/12R, Ropa = 301R/39R, Rtrig = 100R).

There is a new PCB version inside (Ver1.00.5 2010/11/30, 012-00193FH) which seems to be "ready" for ethernet-on-board and also other stuff (audio AC'97 codec, SD Card interface, 2nd USB,...).

If there is anybody looking for a good price and service in EU, just let me know...

Regards,
drieg.

[rf-loop]

Nice to see that there read Hantek.

I have some (100MHz) units with Hantek front panel.

[colinbeeforth]

I think proper CVS saves time, money and big disasters.  If you lose a critical programmer, others can pick up the threads.  My wife works on software for scheduling nation wide transport systems.  Without CVS, the company would be a train wreck!  Like yachtronics says, only excuse for not using CVS is prototyping or personal projects.  tinhead suggests that many "real world" companies don't do this.  I think he may be right, but the reason why it is needed is in his evidence of bugs and inexplicable software.  It just proves how important discipline in programming is.  It sounds like the Hantek/Tekway source is in a pretty bad way.  It would explain why my scope crashes several times a day, when my LeCroy crashes maybe once per year.  For the LeCroy 93XX scope, the previous 94XX series software was ported across with big rebuilds in the display technology and waveform processing in 3 weeks with a team of 7 guys.  I was told there were some bugs, but I had 9310 serial number 013 for one year, demonstrating it to customers, and I never saw a single bug.  It might help that the team of 7 guys were all Swiss, so maybe there was a culture of precision.  However, to port, modify and extend the software in only 3 weeks with so few bugs shows what can be done.

In reality, the Hantek/Tekway DSO must have far less complex software than the LeCroy, as it does only a fraction of the waveform maths and processing that my 9310 can do.  I know there is better hardware support for averaging in the LeCroy, but it can handle 50k acquisitions without limitations.  Curiously, the Hantek/Tekway also has an ARM processor which must run rings around the LeCroy's old 68020 20 meg processor.  I'm not being critical of anyone, I'm sure the developers of the Hantek/Tekway are trying their best with cost limits to hardware and personnel, but I can't help but think that there is much more potential to be released from that machine and it wouldn't cost so much more to do it.  Don't get me wrong, it's far better then many other products, but it could be a work beating product if the software exploited the hardware to best performance.