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

[Mike Warren]

Thanks for the quick answer. That did the trick.  I guess I should have asked here first.

[RobB]

I purchased a Tekway1102 (now a 1202 ) a short time ago to complement the 30+yo Kenwood 20MHz scope I bought while at college.
I'd like to take This opportunity to thank Tinhead for the invaluable work you have put into the documentation of this scope.

[morris6]

The DM9000 chip most probably also not survived the accident because I can trace no network activity. The LnOE net is also routed to pin 1 of the DM9000 and this net was relatively negative 1.7 V. during the accident. Another soldering job; but I'll have to find a new DM9000 chip first.

i will send you new chip

Received DM9000 yesterday, replaced broken one. Today, doublechecking, fitted LA board in place. Now Ethernet is working too.

Finally the hardware of my MSO is complete.

Thanks again Tinhead. A small token of gratitude is underway by 'snailmail'.

While checking my soldering I found some minor differences with the LA board schematic, I made the following handcorrections in my printout:
DM9000 pin 60 SPEED# routes to pin 51 of J20 P2, E-ACT
DM9000 pin 62 LINK_ACT# routes to pin 52 of J20 P2, E-LINK
DM9000 pin 90 DvDD routes to +3.3D
DM9000 pin 91 IO16 is unconnected
A 0 Ohm R marked R27 (topside) in route LA_CLK25, can be used as testpoint for clock signal to DM9000.

If you have your USB terminal connection open during self calibration you will notice a message when it is ready:
"mv: can't rename '/autochk.tmp' : No such file or directory". An artefact of some form of factory calibration?

[mazurov]

It seems that Hantek tries to prevent people from getting console access. With the latest fw (reply #1589), when I press Ctrl-C, Enter in the serial console, I have only 5 seconds before system sends SIGTERM and reboots the scope.

[tinhead]

It seems that Hantek tries to prevent people from getting console access. With the latest fw (reply #1589), when I press Ctrl-C, Enter in the serial console, I have only 5 seconds before system sends SIGTERM and reboots the scope.

this is the watchdog, if /dso.exe is not running the /dso/app/dsod is rebooting the DSO.
Simply right after you enterd the console try to kill the dsod by (until you see no process)

killall dsod

[tinhead]

Received DM9000 yesterday, replaced broken one. Today, doublechecking, fitted LA board in place. Now Ethernet is working too.
Finally the hardware of my MSO is complete.

that's good news

Thanks again Tinhead. A small token of gratitude is underway by 'snailmail'.

you welcome ... "gratitude " ? huh, whatever it is, thank you!

While checking my soldering I found some minor differences with the LA board schematic, I made the following handcorrections in my printout:
DM9000 pin 60 SPEED# routes to pin 51 of J20 P2, E-ACT
DM9000 pin 62 LINK_ACT# routes to pin 52 of J20 P2, E-LINK
DM9000 pin 90 DvDD routes to +3.3D
DM9000 pin 91 IO16 is unconnected
A 0 Ohm R marked R27 (topside) in route LA_CLK25, can be used as testpoint for clock signal to DM9000.

yeah, the 60/62 i've sa them mirrored and 90/91 is clear schematic error. Changed in my version
as well, not changed the opsted one as i will anyway repost with all pages.

If you have your USB terminal connection open during self calibration you will notice a message when it is ready:
"mv: can't rename '/autochk.tmp' : No such file or directory". An artefact of some form of factory calibration?

i know, there are many other hidden things like :

/soft2.mov - use for dithering calculation i think
/linear_adc - whihc have something to do with cal as well
/chk_base_volt - which is necessary as well, but no idea how the values will be created (it is unique for each device)
/soft1.mov - no idea as well,
/phase.mdy - ditto (where i hink it can be used to change allign phase between ADCs?)
/param/adc_off - as well to modify ADC settings, we know some params already but hell, how to use it?
/ubdb.swi - necessary during update, but why?

So yeah, there is a lot of "work" to do, to recognize them all (i will not spend any time on this anymore, but maybe someone else?)
For sure Hantek/Tekway will not tell us anything about them, which of course make sense as we don't need to know internal things.

It is however really nice to see how much effort they spend to integrate testing routines into fw. This was already the case in
their first DSO, the Tekway DST1000(A) series.

[Mike Warren]

Well, mine only lasted one day. I turned it on today and it ran for about 4 hours and then the screen turned white. I wasn't even using it at the time. I switched it off and back on. It booted fine, but then went white again about 20 seconds later. I switched it off for 10 minutes and it started with a blank screen now. 

At that point it was time to leave work for the day, so I'll try it again tomorrow, but I assume it's going back to the supplier.

[tinhead]

this is one of the very unfortunate things like:
- display going white after some time
- keys freezing (but acquisition still seems to work)
- DSO freezing and not rebooting (we know there is WD, so it "must" reboot on sw crash)

They almost always related to the 2 FFC (display and keyboard) and 2 power rail connectors.
From a technical point of view they good enough, but it seems that from time to time the QC is not checking
them properly - or they simply pluged good enough for burnout test but not for shipment.

Before you send back (depends of coure of warranty sticker existence and what so ever) consider trying the
re-plug them. I saw as well some of the FFC display receptacle somehow full of solder flux, this if course sucks
and need to be send back, manufacturer hae to clean such things.

[brainwash]

Mine did the same out of the box: 5 minutes runtime and then white screen. I re-plugged all the PSU and display connectors and then it was fine, but I stopped using it for a month or so. After one month it did the same so I unplugged and plugged back again all the connectors and I haven't had an issue since. Knock on wood.
Considering it took PCE more than one month to deliver it to me I did not want to consider how long it would take for a warranty round-trip.

[Mike Warren]

That was it. One side of the clip on the connector for the FPC to the display was not latched down.

Er, I mean "It just came right by itself. I didn't take it apart."

[back.to.battery]

My DSO5062B arrived yesterday.  Picked it up from Circuit Specialists for 330 USD shipped (price matched to the Owon SDS6062 at Saelig with the 5% student discount).  Fired it up to make sure it wasn't DOA then hooked up my USB-UART and had a go.  Went as smoothly as one could expect.  (It was actually more difficult getting the custom bootsplash to work than it was to perform the hack, but I couldn't help myself). 

It has the typical 'new' 1007 input circuitry (22 ohm and 280 ohm resistors).  I didn't tear it down much at all (added the UART pins from the top with the board in place), so I didn't tear it down enough to get any quality pictures of the board, however I did get a good look at the DAC.  It's mounted on the top of the board, but they've done a good job of scraping off all identifying marks (and then dabbed it with what looks like sharpie for whatever reason. Cleaned that off with isopropyl and some q-tips, only to find the markings were ground away anyway).  It looks identical in package and placement to the LTC2601 tinhead posted here, if that's at all helpful.

This unit has the 83EB revision FPGA, and I'm not sure if I should bother trying to switch to the 83E9.  Thoughts?

It came loaded with 2.06.3(121023.0), but I'm now running the firmware tinhead posted here, and so far everything is fantastic.  The only question I have is regarding the 24 hour clock.  It seems to be only a 23 hour clock - unless I'm failing to understand something fundamental.

I plan to do some more 'invasive' hardware mods once the parts arrive (low jitter oscillator, low noise 3.3v regulator, etc...), and I'll post updates as I go.

If tinhead or anyone else would like me to take more detailed pictures, identify any components, perform any memory/file dumps, or do any specific testing; don't hesitate to ask.

Anyway, thanks for all your hard work tinhead.  I owe you a beer.

[tinhead]

This unit has the 83EB revision FPGA, and I'm not sure if I should bother trying to switch to the 83E9.  Thoughts?

well, you can go to 83E9, the skew (ADC clocks difference) is bit better than 83E8. You actually don't need really to
run factory cal, the selfcal. is good enought (unless you hacking FPGA clock source as well).

The only question I have is regarding the 24 hour clock.  It seems to be only a 23 hour clock - unless I'm failing to understand something fundamental.

heh, another one small bug, thanks for reporting

It seems simply skip 00: while changing value with V0 knob, when you set is manually (via uart/shell or
the DSO tool http://peter.dreisiebner.at/dso-usb-tool/ ) to 00: it seems to work - no crash or what so ever.

[tinhead]

Thanks again Tinhead. A small token of gratitude is underway by 'snailmail'.

you welcome ... "gratitude " ? huh, whatever it is, thank you!

WOW, i got it today. We (I and my wife) really appreciate your support, thank you very much!

[cynikal]

oooh, what did you get?

BTW i am using the new firmware as well, same story as back.to.battery, got my 5062b today, have the 83E8 as well, i got it for $349 also from CircuitSpecialists, but i picked up some other stuff so i paid more (and extra shipping, and man do i hate UPS for taking so long for getting it to me, coulda been here two days earlier but UPS ground decided to not do anything with the package on two separate days wth?).

Anyway another great big thanks to tinhead and all the other folks working hard on making this one heck of a community/oscilloscope user group

is there a FAQ/newbie page i should check out?  For example, i'm wondering if the fw i loaded has the 200mhz capability "enablement" (and how does one verify this anyway?), as well as how do i go about flashing the FPGA.. i'm pretty tech/computer savvy but electronics i'm still learning a lot about.

[Crenwick]

Hi,

The last version of MSO firmware I found on the web is dst1kb_2.6.3_15202d_fact130201.0.up
 
I thought that the MSO was FW 2.7.1 !

Is there somewhere a status about the FW and bugs (like the good one from Tinhead for the DSO)

Nice weekend !

Bernard

[morris6]

I checked the contents of this firmware, available on Hanteks website. Just followed the decrypt scheme Tinhead gives in msg#23984.

'update' script says: [soft version]2.07.2(130201.0), so far, so good.

It does contain a newer dso.exe, dated Feb 1. And newer language files, English and Chinese. Other files however are same or older than those in the MSO-ready.bin flashfile used to upgrade to MSO.
 
The la_top.rbf, configuration for the FPGA on the Logic Analyzer board, is older and I found that it doesn't provide the 25 MHz clock signal for the network chip DM9000. Don't just run this update, it will kill your LAN.

Next is to update only dso.exe to see what is changed.., after saving the old one.

[tinhead]

right, don't just download versions from Hantek website, for MSO firmware plase check

http://www.mikrocontroller.net/topic/289260#new

If there is something new i will post there. It will be always fw comptible with "our" MSO version.
I need to ask Hantek to stop updating each time the rcS and the la_top.rbf, there is no need for that.

[tinhead]

and here the latest DSO firmware (for following DSOs only):

 - Tekway DST1062B, DST1102B, DST1202B
 - Hantek DSO5062B, DSO5102B, DSO5202B
 - Voltcraft DSO3062C, DSO1062D
 - PCE DSOs: PCE-DSO5062B, PCE-DSO5102B, PCE-DSO5202B

http://www.mikrocontroller.net/topic/205820?goto=3094416#3094416

Feedback as always welcome!

[Mike Warren]

Can this just be installed via the update menu without needing to do any UART stuff?

[tinhead]

Can this just be installed via the update menu without needing to do any UART stuff?

yes

[carloscuev]

Thanks again Tinhead. I'm looking to update the FPGA design in my april 2011 bought model (83E8 FPGA of course). Originally Tekway DST1102B, now DST1202B thanks to you. I've been reading your posts about it, and according https://www.eevblog.com/forum/chat/hantek-tekway-dso-hack-get-200mhz-bw-for-free/msg170862/#msg170862 i guess the best *.rfb I could update to is dn_hw1007_83E9_date111122.rbf . Right? or maybe because is an old HW from april 2011,the best would be dn_hw0_83E9_date_091201.rbf?

I found out in my 83E8 scope that in FFT from 200ns/DIV to 20ns/DIV running this latest 130306.0 firmware and checking the Sys Status i get 1GS/s, not 800 MS/s as this table shows https://www.eevblog.com/forum/chat/hantek-tekway-dso-hack-get-200mhz-bw-for-free/msg170740/#msg170740 all the other timings fit well for the 83E8 column. I hope this isn't a sign of possible problems when updating to any of the 2 83E9s available.

About the factory calibration, I don't understand in the first post it it will be needed or not. I don't have a stable 10 MHz source so maybe its better to keep my current version. What would you recommend?

[tinhead]

some of the things i said there are wrong, the firmware - at the time of testing - was cheating about the enabled memory depth,
therefore i got higher values than they really are (none of the FPGA designs allows 400MS/s when both channels and long
memory enabled). Still, it make sense to change to 83E9 design if someone have 83E8, however it does not make sense to
downgrade to 83E9 if someone have already 83EB (or later) version installed.

Regards the 83E9 "subversions", it does not make sense to use these from hw0 or h1005 - even if you have hw0 you
can still use hw1007 83E9 design (but NOT 83EB, 83ED etc). The latest (working on your hardware), in your case the
dn_hw1007_83E9_date111122.rbf, will be the best option.

Regards the factory calibration, in case you don't have stable source then don't do anything on that.
Simply run the self calibration and your fine.

[walt]

Still, it make sense to change to 83E9 design if someone have 83E8,

1 UNRAR files
2 Copy from folder <save_from_DSO> file dst1kb_2.06.3_15102b_fact(130114.0).up to FLASH storage
3 Plug flash disk in DSO, run firmware upgrade
4 After finish procedure, unplug flash storage and reboot DSO.
5 Copy backup file dn_current.rbf from storage to some else secure storage 

6 Copy from folder <load_in_DSO> file dst1kb_2.06.3_15102b_fact(130114.0).up to FLASH storage
7 Copy target *.rbf file to FLASH storage and rename to dn.rbf
8 Do 3 & 4 paragraph this instruction.
9 PROFIT

This archive file contain just 2 cmd files.
for load file
[DST type]dst1000b
[soft version]2.99.9(999999.9)
[type]123
[cmd] cp /mnt/dn.rbf /dn.rbf
 [cmd] sync
[cmd] sync
[cmd] sync
[end]

for save- another direct of copy

[casinada]

Sorry if this is a little bit off topic but I wanted ThinHead's opinion regarding the Portable Oscilloscopes with isolated inputs from Hantek.
On their web site they have models DSO1062S, DSO1122S, and DSO1202S. On ebay or ALiExpress they have the DSO1062S, DSO1122S, and DSO1152S.
They were mentioned on a post in Dec 2011 but never reviewed. The new Tek or Fluke portable scopes are totally out of my price range

[tinhead]

yeah i remember, i spoke about the isolated version but never had a chance to get one to do a review.
However, i had the non isolated version, and it was ok (except the crazy things on first firmwares).

The only thing i don't liked was the enclosure, i would not try to throw it from a second floor
Fluke will survive it, Hantek for sure not.