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

[alm]

Really? But why they sell these calibration fixtures for every different scope input capacitances seoarately.

To be more accurate, it's a fixed cap with a smaller trimmer cap in parallel (18pF // 2-8pF for the 20pF model according to Håkan). There is also a 1Mohm resistor in parallel (so it forms a 1:1 compensated voltage divider if the scope's input capacitance matches the normalizer). They need different ones because you would need a huge adjustment range (=bad resolution) to cover 12pF-47pF or so.

If it is adjusted capacitance how you can calibrate with it becouse you never know if capacitance is right.

Just like every other calibration, by comparing it to a more accurate source (at some point they probably use a capacitance meter). Not sure what the factory calibration procedure was. But I don't see the point about caring about input capacitance, as long as the channels match, everything should be fine. And you can't adjust it on (most) modern scopes anyway.

Also in tektronix service/calibration manuals ask put this "normalizer"  to scope input (and also extremely fast rising pulse generator) but never it tell that adjust this box capacitance.

That would be the service manual for the normalizer, if such a thing exists (it was designed for internal use, so I'm not sure how well it was documented).

This box can use in calibration routines for adjusting device under calibration! If you adjust tools for calibration what you then calibrate?

That's the whole traceability chain of calibration, all the way to the national standards lab. A 3.5/4.5 digit DMM is calibrated by a DMM calibration system, which may be calibrated by a reference DMM, which is calibrated by a voltage and resistor transfer standard, which is calibrated to a Josephson standard, which is a fundamental physical standard which defines the volt. Same for time/frequency, which is traced to a cesium clock. You can derive capacitance from those.

Found the calibration procedure in the TekScopes archive:

After searching through some data this is the best info you will find.  Straight from the horses mouth........so to speak.  Both the author of this procedure and the gentleman who inserted it into the microfiche files worked at Tek during the 60's and I am just the means by which it made its way to the digital realm.

If readjustment should be necessary, the internal resistor should first be measured to be sure it is in tolerance.  Then the internal variable capacitor may be adjusted to produce a square—cornered, flat-topped response to a square-wave signal on an oscilloscope it is designed to be used with, in the same way as described in the instructions for adjusting scope input capacitance.  The response of the scope to square waves should first be known to be normal, and the input capacitance to the scope should be known to be within about five per cent of its nominal value at the attenuator setting used.  It is normal for input capacitance to vary slightly from one setting to the next to correspond with slight variations of input resistance.  A Tektronix type 130L-C Meter may be used to measure the scope input capacitance.  Measurement should be made while the scope is operating.  When using the Type 130, to avoid a small possibility of a measurement error due to over driving the input, it is a good idea to use an attenuator setting that corresponds to a deflection factor of about 1 volt per division.     

Written by: Leon Orchard
PE/Elec Evaluation
Inserted by: John Mulvey
Tektronix 3-1-67

[RFman]

btw, i did said "manafacturer is installing by defualt 2.6.3 (110225.0) now",
this is not yet bug-free firmware, it is still in principle the same as 2.06.3 (110118.0).

Hanek did changed the PCB revision, so small firmware update was necessary, therefore revision 2.6.3 (110225.0).
It can be downloaded from Hantek website.

No idea if Tekway changed the hardware too, so therefore there is not new firmware on Tekway website.

There are no details on hw modification, we have to wait until someone bought a Hantek with new hw version.
They just started the production, so probably few old hardware revision models still available.

Even if Hantek said "minor modifications only" i can tell you already that
display init is different (same display but connected a bit different),
buzzer driver has been changed (connected to different port - probably pin swap with display),
some FPGA design changes (only for new hw revisions).


The firmware update is of course checking the hw revision, so it will not "destroy" old models
(and will not update buzzer, display and FPGA).

Most interessting thing during hw revision check is the information
which will be read by the "readHWVersion" from EEPROM:
- PCB revision
- LCD revision
- TOUCH revision
- FRONT PANEL revision
- USB revision
- NET revision

PCB and DISPLAY are clear, probably some routing changes to support new functions, NET is know, that's the LAN addon,
USB revision (probably) for the planned SCPI support, TOUCH - yeah, we know there will be touch in the future,
and FRONT PANEL - right, Hantek have on the front panel preparation for DB25 plug - probably LA functionality in the future.

I think i will order a new hw revision and check what inside, however not now - i will wait until the Hantek's LAN addon is ready.

Hi Tinhead,

From your informstion it looks that, if I load the latest Tekway firmware 2.6.3, it won't hurt the oscilloscope because it 1st checks the version of hardware before it updates.

I'm now thinking that it maybe better if I go straight to the 200MHz version and be finished. Is there a prehack upgrade for that as I'm waiting for the UART RS232 TTL interface to come from Hong Kong. I can't do much with my new Tekway as I can't read Chinese...

Yes it looks as if there are a few DST1102b with the old version of software.

Thanks again for all the work that you have put in for us.

[rf-loop]

I have tested with one single scope:

model: DSO5102B
serial T1G/008xxxxxx
soft version: 2.06.3 (110118.0)
hard version: 0x555583e8

This scope accept new FW what is now shared by Hantek.
( dst1kb_2.06.3_15102b_fact(110225.0).up )

But I do not know what is changed. Fast check I can not find any changes - as expected. (later I look more)
(maybe it do not nearly anything change for this scope becouse HW version?)

It seems reset power up counter and FW upgrade counter. Nice.

----------------

alm: thank your opinion comment about Tek "normalizer".

[RFman]

Hi Tinhead,

I have just updated my Tekway DST1102b to a DST1202b. Everything so far is working OK including the English Menu display... . After backing up the memory contents, I used version 2.06.3 - 200MHz update.

So far everything has progressed nicely with my Tekway purchase from Pioneer.

Thanks for your help.

[wjb444]

from china

[tinhead]

from china

look at that, wjb444 is here .. welcome!

For those of you asking what these file are for :

dst1kb_2.06.3_01202b_fc(110225.0)(1).up  - this is the latest original firmware for
Tekway DST1202B (or whatever Tekway model hacked to 200MHz).
Don't use it on Hantek (for Hantek you will find exact the same fw version on Hantek website).


dst1kb_2.06.3_1062ridoflim(110315.1).up - this is somehow broken or template for upcoming firmware 2.6.3_110315.1

wjb444 , any idea for what this file is ? (i see what inside, seems to be like fw template only).

[tinhead]

ups, the Tekway firmware is broken ...

there are all language files (*.lan) missing ... so after update is only chinese available, heh

[spookytooth]

Yes... right. And for me it did not with with the ..(1).up in the end of the filename. Checking this on the console (UART) showed the error "... unexpected character '(' "
  You need to remove the '(1)' from filename so that ist looks like all update files. And - of course you need to copy back the language files...

[Igor]

New design of my Chinese Tekway DST1102B.





   

[tinhead]

Yes... right. And for me it did not with with the ..(1).up in the end of the filename. Checking this on the console (UART) showed the error "... unexpected character '(' "
  You need to remove the '(1)' from filename so that ist looks like all update files. And - of course you need to copy back the language files...

yep ..

i think wjb444 even not recognized that the other lan files are missing as he is using chinese, so what, shit happens

[spookytooth]

One of these ching chong language files contained in the hantek version must be corrupt. The system crashed when I tried to switch language thru one of these funny character settings (don't know which one) ....
So here is an update file that contains Chinese, English, French and German and works fine for me.
And there is on additional (empty) file in the update file set called 'special'. This one is unique to Tekway update set. Anybody any idea what this means?

[tinhead]

One of these ching chong language files contained in the hantek version must be corrupt. The system crashed when I tried to switch language thru one of these funny character settings (don't know which one) ....

yep,indeed (that's why deleted my posting few sec. ago ...)

So here is an update file that contains Chinese, English, French and German and works fine for me.

good work

And there is on additional (empty) file in the update file set called 'special'. This one is unique to Tekway update set. Anybody any idea what this means?

actually it is special file to do special things ... middle of the firmware update procedure this file will be loaded, if exist .. so let's play
to find out what it can do for us.

Actually Hantek fw is also checking for this file, but they just not used it in last fw (probably Tekway shoudl just remove empty file, whatever)

[bilko]

from china

dst1kb_2.06.3_1062ridoflim(110315.1).up - this is somehow broken or template for upcoming firmware 2.6.3_110315.1

wjb444 , any idea for what this file is ? (i see what inside, seems to be like fw template only).

Is the clue in the filename ?

dst1kb_2.06.3_1062ridoflim(110315.1).up

rid of limits ?

[tinhead]

Is the clue in the filename ?
dst1kb_2.06.3_1062ridoflim(110315.1).up
rid of limits ?

you should really not ask so many questions

But yet it is, actually the firmware update procedure is checking for the "special" file, if such file exists
the model version will be not checked and the update procedure simple continue to execute what inside the "update" script.

So if you for example have an firmware update file which contains logotype.dis with "tekway_dst1062b", the script
will update the firmware and on some step replace your logotype.dis with the one from your firmware update.
This works of course as universal hack (well, still one per Hantek/Tekway).

But don't forget, you should also move the /dstxxxxx to the same model as the specified in logotype.dis
If you don't the displayed model (in utility-system information) will be changed, but the DSO itself still not changed.

So yes, it mean "get rid of limits", but of course you have to take care of the /dstxxxxB in root dir too (which was not the case in the fw posted by wjb444).

Here is the prove:
I did changed the logotype.dis content to "tekway_dst1102b", compiled the firmware and updated my DSO.
The bootscreen logo is still displaying DST1202B, but the utility-system info DST1102B.
As you can see i have still 2ns/div, which means it is still DST1202B even if the model version displayed in
utility-system info says DST1102B.

If i add into the "update" script a mv /dst1202B /dst1102B and compile the firmware again (and do update)
the model will be really changed.

So yes, this "special" file is removing check of the "current logotype.dis = logotype.dis from firmware update file",
but you have still to care about the /dstxxxxB file in root dir.

I think probably someone from the Tekway/Hantek was pissed of to care each time they soldered new DSO
to look which update file for which model, so they just added a backdoor.
This kind of protection was a joke, so probably they already working on new protection,
so they probably will not care that this backdoor got exploited.

[wjb444]

Sorry, my English is poor, I use google automatic translation.
Firmware, I did not break people, it comes from the generation of TEKWAY Agents, so, sorry, I can not answer any questions.
But the firmware is easy to use.


??google ??
?TEKWAY??,??
??

[RFman]

New design of my Chinese Tekway DST1102B.





   

Hi Igor,
Your front panels look great. What type of material did you use. Did you use something like the origional panel where the LED shines through for the VO control? What type of printer did you use to print colour? I perfer your origional panel which was close to Tekway's design, where the channel inputs were grouped with their controls. Is it possible for you to produce a PDF of the correct size for your origional design. How did you do the cut outs?

Thanks for for your help.

[RFman]

I have not been able to get Windows 7 to install/update the driver that comes with TTScope in the Driver folder. I've downloaded TTScope from both Tekway and Hentek but the driver will not install in Windows 7. Looks to me as if a different driver is required for windows 7. A search has not found a suitable driver. Has anyone a solution for this problem,

Thanks all.

[Igor]

Your front panels look great.

Thank you!

Is it possible for you to produce a PDF of the correct size for your origional design. How did you do the cut outs?

I have long picked up material for printing. Watched the film and paper. Stopped on paper density 125.
Printed on Pro Minolta laser printer.
Recommend punctured paper hole LED some buttons V0 to laminated. I did not do so, and LED almost not visible.
Then closed the paper on both sides of the matte laminate thickness 75 µm.
Laser cut holes and path.

In the ZIP: PDF files for printing and cutting.

[rf-loop]

Get nearly 200MHz without any modifications!
All tests are made with unmodified Hantek DSO5102B.


I make short test becouse many peoples do not use direct cable with termination to scope from DUT.
Most of times peoples use probes of course. So, what is important: scope BW alone or probe+scope BW?
(of course both are important but in practice normal use is with probes.)

Now I make only one "big trick". I did not use probe GND wire. I use normal universal probe adapter to connect center and GND just on the probe tip. (It is BNC and it have hole for probe and it connect GND and tip)
All (I hope) know that making HF measurements using probe GND wire give very bad effect. So normally we use some kind of tip end what connect GND directly. (With Hantek probes there are not these any included in package, not even small "spring" style what price is 1/100 yuan)

This test do not tell anything about this situation when use "long" GND wire.

For some compare I make also test with Agilent/HP old "not bad" probe with Hantek.
Both probes LF compensation I make for exactly same square shape with scope own 1kHz signal.







Hantek P150 probe (labeled as 100/6MHz).
Now with 1:1 setting. Signal 0,5 - 10MHz.




Same hantek probe 10:1 setting. Signal 1 - around 280MHz
I use step sweep and stepping is stopped after around every 50MHz advanced as "poor mans markers".




same test but now HP/Agilent "150MHz" probe 10:1, sweep stop around 250MHz (sorry, my mistake)




Hantek P150 probe, now mediumfast rising and falling pulses. (pulse generator specs: rise and fall time <=1.3ns not good but enough for this purpose)
(if all is exactly ok, pulse looks symmetric... so there is not perfect frequency response over 0 - 1GHz / But there is scope freq response shape with probe together and lot of transmission line mismatches and also pulese generator itself is not perfect. It is HP8161A with "fast" option (not 50MHz model).



Same test with HP "150MHz" probe. (littlebit better pulse shape but not big advantage. DC/Ac do not make any difference with this pulse freq)


Conclusion: Not bad. In this price class Hantek with own probes is really good. It is not Tektronix, it is not Agilent but... it is reasonable good scope for normal hobby and also professional use for signals under 100MHz.

(also I have make fine adjustment for CH1 and 2 compensation capacitors (inside scope front end) as perfect as I can do but this do not affect these probe tests becouse all they are made using scope lower voltage band.)

Also I test with probe own GND wire. (later pics - but both are really bad (of course) and nearly as same class)


----------
So if you do not really need advantage what mod may give, do not mod a: if you want factory 3 year warranty, b if you are not sure how to do mod. (fail may give lot of troubles - but still it is "repairable" if you follow exactly what all information you can get reading carefully this forum... in this case Rigol is littlebit different)

But remember, samplerate do not go faster.
This is also very good scope in its price class.

(with mod you get 2ns/div but scope is just same speed, it is only "zoom". You can do it with computer... draw nev vlines with 2ns/div and stretch picture with irfan wiev or other. It gives just same effect. Scope real speed is just same before and after mod.)

Understand aliasing specially if use FFT. With this kind of freq response shape there is lot of and littlebit more aliasing.. nearly as garbage collection. And modification make it more bad. Lot of more bad. If want make good, it need do just opposite as modification. It need reject BW and make it more like brickwall shape. (but who cares FFT quality... maybe not so many)

If you use scope in high temperature. Room temperature high or  if air do not flow very free around scope.

Maybe it is good in this case install small fan inside scope. (no need full speed, you can reduce nominal speed example x0.7 - 0.9 (depends fan type and quality) and you get low noise and it still helps airflow inside scope. It need only small extra amount of air. (example double speed if compare natural free air flow and exance rate)  In this case I recommend fan take outside air and flow is to inside. In this case you can put small filter before fan so all dirty do not free travel to inside.  In normal 22Celsius room temp some components work quite high temp. Clearly over 70C case temperature. It is not good to go any more high. They may last but maybe fail rate start grow.

[project]

I have not been able to get Windows 7 to install/update the driver that comes with TTScope in the Driver folder. I've downloaded TTScope from both Tekway and Hentek but the driver will not install in Windows 7. Looks to me as if a different driver is required for windows 7. A search has not found a suitable driver. Has anyone a solution for this problem,

Thanks all.

Didn't see any problem with running on win7 32bit, just use driver under x:\Program Files\TTScope\Driver. If not working for you, install TTScope in xp sp2 compatiable mode.  Don't know 64bit.

[wjb444]

My friends say that this firmware is deceived agents, or were cheated by manufacturers, brush machine, no 2ns file

[rf-loop]

Time difference between channels. May come from timeshared AD conversion and becouse this is not compensated by system (FW).



Mode, average, trig CH1, edge up.

Both channel get same signal but becouse signal split is not as good as it can be (only things what affect timing is ok).
(same cable lenght and same type of cable so trawel time is enough same.)

It may come from timeshared AD conversion. AD samples difference is 1ns but single channel is sampled every 2ns. (2 channel maximum is 500Msa/s for both channels but there are 1ns shift between channel.)

1. sample CH1
1ns time
2. sample CH2
1ns time
3. sample CH1
1ns time... etc

Who can confirm this so that it is exactly this. If not know what is truth, do not confirmate this.

IF it is this reason, I wonder why Tekway/Hantek do this mistake in FW.

It need repair  all ather modes but not Alt trig mode.

Timing is important in oscilloscope and this kind of systematic error is just "bad design".

But also it can solve with short extra lenght cable as poor mans solution...  1ns is not long travel in cable.

[RFman]

Both channel get same signal but becouse signal split is not as good as it can be (only things what affect timing is ok).
(same cable lenght and same type of cable so trawel time is enough same.)

If you swap the signal cables between channels does it change?

[tinhead]

Both channel get same signal but becouse signal split is not as good as it can be (only things what affect timing is ok).
(same cable lenght and same type of cable so trawel time is enough same.)

If you swap the signal cables between channels does it change?

it didn't matter, ch2 is always 1ns behind ch1.

This is actually software error, my DSO (cold) is having 200ps skew time between channels, after warm-up 1ns.

Of course ALT trigger works without skew time, but is not having avg. and long mem mode, so not really a work around.
Even with ext trigger no difference, ext/5 even worse because of bug which has been still not fixed (ext/5 is increasing the 1ns to 3ns or so).

Of course you can create a small loop with RG316 cable for ch1 data and setup trigger on ch2 - hehe, then of course skew time is 0ns.

Skew time is common for many DSOs, but of course every company is reducing it to few ps during manufacturer calibration,
this seems to be not the case anylonger for HanTekway (if i'm not wrong, it wasn't there with very old firmware - but hacked).

[rf-loop]

If you swap the signal cables between channels does it change?

it didn't matter, ch2 is always 1ns behind ch1.

This is actually software error, my DSO (cold) is having 200ps skew time between channels, after warm-up 1ns.

Of course ALT trigger works without skew time, but is not having avg. and long mem mode, so not really a work around.
Even with ext trigger no difference, ext/5 even worse because of bug which has been still not fixed (ext/5 is increasing the 1ns to 3ns or so).

Of course you can create a small loop with RG316 cable for ch1 data and setup trigger on ch2 - hehe, then of course skew time is 0ns.

Skew time is common for many DSOs, but of course every company is reducing it to few ps during manufacturer calibration,
this seems to be not the case anylonger for HanTekway (if i'm not wrong, it wasn't there with very old firmware - but hacked).

RFman: Signal cabeling is same for both channel input BNC point so why cables change make any affect?
(yes of course I also do this but no with any real reason. Becouse cable lenght is same, both cables are good condition and they are exactly same cable type (also manufacturing lot is same...) They are also checked for this accuracy level what even can respect for BNC connector cables. Electrical lenght is much better than +-25ps need. They are checked becouse they are my calibration cables, made for Tektronx and also for some HP scope channel time differencies adjustment (skew calibration). And there need accuracy what is lot of better than 1ns. (travel time is 1ns for 300mm  in empty but in coaxial cable it depends cable and it may be something between 0.9 - 0.7 or something like this so good estimate is around 25cm)


tinhead

This is problem with cheap scopes what have not designed using professionals with long history of experience. (example Tek or HP) Many good scope take high care about it.)
But today it is littlebit different becouse digital scopes. If digitize with 1ns periods resolution is 1ns. Now this 1ns is for 1 channel. In this (and many other scope in same class) two channel maximum samplerate is 500MSa/s. 1 sample every 2ns. can not measure anything with better resolution. Yes, if single shot real time.  Averaging give more. So there must not be human made error in sample and store. Then take care also analog channel itself. This your thermal drift is littlebit strange - why?  (averaging is sometimes nice becouse system have noise and it give some advantage for accuracy. In theoretical ideal signal and sampling this advantage can not get. There are many nice solutions in some systems where adding noise to system give more accuracy )

But in this Hantek case it looks really just like bug. (they have forget..or make some mistake in FW upgrade) I can only quess that memory and capture construction is so made that channels can very easy shift in time axis related to each others.

"Of course you can create a small loop with RG316 cable for ch1 data and setup trigger on ch2 "

why change Trig to CH2. No need.

If delay CH1 signal before scope there is no any reason to think trigger source! It simply means that CH2 signal come earlier as compared to CH1.  Trig do not mean anything. Only meaning is signal timing... becouse we can not put signal travel more fast we can only delay.

I think it do not need picture where  I shift CH2 signal to left so that it is same as CH1 only adding delay to CH1 signal and also so that there is no matter if Trig is CH1 or 2.