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

[Tasman]

I've got a little problem with my Hantek scope.  I bought the DSO5102B back in early 2010, before this thread started. Judging by the condition of packaging when it arrived it had been around for a while before I got it so its now probably getting on for four years old.  During that period it has spent a lot of time unused and recently when I started using it again I noticed that it was losing trigger settings and getting strange offsets and calibration changes when powered off.  It would work ok. after self-calibration then lose it all again next shut down.

  Anyway this gave me the excuse to buy another scope and I shelved the Hantek until I got some spare time.  When I pulled it apart I found that the backup battery was low and decided to change it.  Intended to do this with power on but got called away and forgot to boot up when I came back before changing battery.

Scope now powers up with the standard noises (relay clicking etc) and the usual buttons light up but the screen is dead.  Functions like autoset appear to work, or at least make the right noises and vertical gain relay clicks at about the right point etc. 

Before I make any more stupid mistakes can anyone give some advice on what to do next?  Since I have another scope there's no urgency but I don't like to throw away gear if its useable. 

[tinhead]

I've got a little problem with my Hantek scope.  I bought the DSO5102B back in early 2010, before this thread started. Judging by the condition of packaging when it arrived it had been around for a while before I got it so its now probably getting on for four years old. 

so probably the hw0, which wasn't affected by any known hardware bugs, that's already good.

recently when I started using it again I noticed that it was losing trigger settings and getting strange offsets and calibration changes when powered off.  It would work ok. after self-calibration then lose it all again next shut down.

the selfcalibration is creating files and saving its own results into memory. So that can be file system error, the data can't be saved were the already loaded into memor is giving you working DSO until next reboot.

The first thing i would do is to conect via UART to PC and then see if there are any files in /mnt directory, what is the free space (in principle to check it there is any junk on the file system). It would make sense as well to update firmware, where of course the question is what firmware is currently installed. Direct update should still works, in worst case you will have to copy some files before you run update. The latest firmware is still working on my 2009 mainbaord, so it should as well work on yours (when you have hw0, with hw1005 i don't know, maybe not maybe yes).

Anyway this gave me the excuse to buy another scope

it is always good to have an excuse

When I pulled it apart I found that the backup battery was low and decided to change it.  Intended to do this with power on but got called away and forgot to boot up when I came back before changing battery.

afaik the battery is only used to save the date/time, and afaik the firmware don't care about what set there.

Scope now powers up with the standard noises (relay clicking etc) and the usual buttons light up but the screen is dead.  Functions like autoset appear to work, or at least make the right noises and vertical gain relay clicks at about the right point etc. 

looks like bad contact on the large FPC, so the display is not getting the +5V for backlight. It can be as well the power supply cable on right side. This plug supply +5V for USB and display and +3.3V for the whole digital part of the DSO. So when this is plug issue, you should not measure any voltage on the USB port on the frontpanel. If there is 5V then it is the FPC bad contacting.

[tinhead]

Before I make any more stupid mistakes can anyone give some advice on what to do next? 

btw, check the very first page in this thread, there a links with schematics of these DSOs, so that can be helpful as well.

[Tasman]

Many Thanks Tinhead - you are amazing. Some further information: the hardware is hw0 and I hacked the scope using your original patch on a usb stick.  The faults occurred with all firmwares and 130425.0 is currently installed.

I'll have a look at the schematics and the items you suggest, here's hoping it's not just a box of spare parts.

[tinhead]

Many Thanks Tinhead - you are amazing.

you welcome

Some further information: the hardware is hw0 and I hacked the scope using your original patch on a usb stick.  The faults occurred with all firmwares and 130425.0 is currently installed.

ok, so it will be not the firmware itself, probably only NAND out of space (or bad sectors).

I'll have a look at the schematics and the items you suggest, here's hoping it's not just a box of spare parts.

it is really hard to destroy them, sure, not everybody will manage to re-solder BGAs, but everything else is docummented and can be replaced/repaired.

[Tasman]

Update on the display backlight issue

looks like bad contact on the large FPC, so the display is not getting the +5V for backlight. It can be as well the power supply cable on right side. This plug supply +5V for USB and display and +3.3V for the whole digital part of the DSO. So when this is plug issue, you should not measure any voltage on the USB port on the frontpanel. If there is 5V then it is the FPC bad contacting.

Measured voltage at USB as suggested - 1.8V but at high impedance.  Also checked continuity between +5V pin on plug and USB and found open circuit.

Removed the main board and began tracing the power distribution from schematic then noticed that the p channel mosfet that is shown in the schematic switching power to the backlight was not there. There was no damage to the pcb pads and no remains of tabs - just perfect impressions in the solder of where the tabs had been, with a frosty oxidised appearance like a dry joint.  I remembered that the backup battery had been very tight and I had applied a fair bit of force using a plastic spatula when it suddenly let go and shot out of the socket.  The spatula must have then hit the transistor and knocked it off the board.  Significant point, I think, is that if the soldering had been good either the transistor would have been undamaged or the tabs would have remained in place if the transistor body was broken. 

All well and good this explains the lack of backlight, but why no power to the USB?  With a bit of further tracing I found that the USB power is also taken from the drain of this transistor!  There appears to be an error in the schematic - it shows the USB as directly supplied from +5V (see pdf).

These discoveries might shed some light on the other problems.  I didn't mention earlier that I also had difficulties with the using the USB port.  When a stick was plugged in the DSO would often initially recognise it then immediately announce that it had been disconnected.   When a stick with firmware had finally after several attempts been recognised the firmware update process often required several more attempts before it decided to start.  Perhaps this was due to high resistance at the mosfet terminals dropping voltage to the USB?

Anyway I can't find a suitable mosfet right now so will have to leave it for a couple of days.  Thanks for the help so far.

[tinhead]

right, this is an error in schematic on hw0. An really unusual design to bundle VLCD and VUSB together, controlled by single mosfet (probably a cost savings idea). In hw1005 and hw1007 this seems to be fixed, the +5 from PSU is directly routed to USB, buzzer and other circuis that need +5V. The VLCD is of course then still controled via mosfet.

So yeah, good finding.

Regards your DSO, solder a wire and you fine. Who cares about when the display goes on

[Tasman]

Regards your DSO, solder a wire and you fine. Who cares about when the display goes on

Thanks, I was going to do that but thought perhaps there was some reason why they put the switching circuit there in the first place.  Is there?

[tinhead]

thought perhaps there was some reason why they put the switching circuit there in the first place.  Is there?

sure, normally you would not power up backlight before the display logic is on, that looks sometimes ugly.
For USB it didn't matter here, for sure it could be controlled via i/o pin, but for sure not tgether with display backlight.

So no, if there was a reason to control the 5V, then only in regards of backlight control.

Btw: i'm using wire on my hw0 since loong time, some day i shot that mosfet to nirvana and was to lazy to solder anythign else than a wire.

[tinhead]

Sorry, does anyone have a copy of new DSO5202P firmware? I'm trying to upgrade my 5072p, have no problems with 5102 version, but link to 5202P from the http://www.hantek.com.cn/en/PagesFW_Vzxgj.html doesn't work...

ask Hantek! I told them already 4 weeks ago that this link didn't work, but still not fixed.

uh9lab,

did you got anything from them?

[uh9lab]

Unfortunately, no. Same for Russian dealer, no answer. Maybe that's because I wrote 5072 serial during registration.

[Tasman]

Update on problems:

Have reversed the unscheduled alterations caused by my clumsiness and the scope now has a visible display.  Carried out a more methodical investigation of the original faults with the following results:

There are two faults, which occur only on Channel 2.  Channel 1 has no issues.

Fault 1:  On bootup from cold CH2 will not trigger until 5 to 10 minutes later - time seems to be temperature related.  No further triggering problems until shutdown for more than 20min, when CH2 again fails to trigger for a while.

Fault 2: With CH2, single channel, 4K memory depth the trace is ok at 400ns/div but has a large offset at 200ns/div  (see attachments).  The offset varies over time but so far is not observed to exceed signal amplitude above or below zero.  Problem occurs at all signal amplitudes. No problems with CH1 alone or two channel operation under same conditions.  No problems at all with other memory depths.

Problem appears hardware related but analogue or digital? And why does memory depth affect offset?
Any suggestions?

[tinhead]

Update on problems:

Any suggestions?

it would be nice if you could first update to latest firmware, there are sometimes hugh differences between versions

http://www.hantek.com/Product/DSO5000Series/DSO5202B_Firmware.zip

[tinhead]

Unfortunately, no. Same for Russian dealer, no answer. Maybe that's because I wrote 5072 serial during registration.

ehm, i thought you bought real 5202P and not 5072? I saw some on ebay, but honestly no idea if they really exists or simply stupid dealers. Sure, Hantek is listing that model, but who knows.

[Tasman]

Update on problems:

Any suggestions?

it would be nice if you could first update to latest firmware, there are sometimes hugh differences between versions

Sorry forgot to mention that same problem occured with 130306.0, 130425, and now with 130814.0...

[tinhead]

not getting anything like you got, not on cold nor warm hw0 nor hw1007. To ensure this have nothing to do
with calibration of your DSO, please provide me following files:

/tdc_edge125M
/tdc_overtime125M
/tdc_pulse125M
/chk_base_volt
/mult_adc.log
/tdc.log
/param/sav/chk1kb_091023

[uh9lab]

ehm, i thought you bought real 5202P and not 5072? I saw some on ebay, but honestly no idea if they really exists or simply stupid dealers. Sure, Hantek is listing that model, but who knows.

I think it really exists, it can be pre-ordered from local dealer: http://www.hantek.ru/products/dso5202P.html
But this dealer warns that questions are forwarded to China and it can take a long time to get any solution...

[Tasman]

Problem Solved!

Took a closer look at the PCB and found discolouration of some pins on both top adc's.  Looked like flux residue and cleaned up easily with alcohol and q-tip.  Now have no problems with offsets on CH2 and triggering is much better than ever, rock solid on both channels with no warm-up required.  Previously there had been an occasional trigger jitter on CH1 and a lot more on CH2 at all times.  Will have a look at bottom adc's later.

Perhaps this issue or something similar is cause of trigger problems Dave found in his review?

[tinhead]

Problem Solved!

Took a closer look at the PCB and found discolouration of some pins on both top adc's.  Looked like flux residue and cleaned up easily with alcohol and q-tip

fantastic! Between 400ns and 200ns in single channel the DSO is switching from 4 to 8 ADCs together, so yeah, here it make sense that when some from them have offset the result have offset as well. Wondering only why in 80ns and higher the offset was not there, but as long it works, who cares. Problem and solution noted, knowing how much flux residue was on earlier models i bet others have/will get similar issues.

[tinhead]

I think it really exists, it can be pre-ordered from local dealer: http://www.hantek.ru/products/dso5202P.html
But this dealer warns that questions are forwarded to China and it can take a long time to get any solution...

so it looks more like "produced on request". That item on ebay, 141031651719, seems to be photoshop made. But this one, 271262770701, looks like genuine DSO5202P.

[Tasman]

Problem Solved!

Spoke too soon - when I reassembled after cleaning bottom adc's the problem was back again in all its glory.

Now I find that it comes and goes as I apply pressure to the PCB around the area of the input circuitry so it looks like a dry joint or cracked trace - somewhere.....

[tinhead]

Spoke too soon - when I reassembled after cleaning bottom adc's the problem was back again in all its glory.

Now I find that it comes and goes as I apply pressure to the PCB around the area of the input circuitry so it looks like a dry joint or cracked trace - somewhere.....

i would check relays, offset opamps, voltages on AD8370, these 2 resistor arrays in each input channel (when they broken, and DSO cold, you would have no trigger), the trigger mux LMH6574MA.

[tinhead]

... cracked trace - somewhere.....

in moments like this i always thinking "why the hell i haven't made any PCB picture, with all inner layers".

Anyway, it will be probably dry joint or physicaly broken cap/resistor, i had only once cracked trace on a handheld, where the PCB was "abused" by Hantek engineer before and probably technican/worker who decided to mount the PCB wrongly and "fix it with screews". Any other "non working/DOA" DSO i saw was broken due dry joints or broken parts.

[int21h]

this is crap, someone just bought latest MSO5202D model from Hantek and installed latest firmware from Hantek website. ... and of course the MSO is now bricked !!!

I just did this to an MSO5102D.  It's on it's way back to China.  Not happy 

Got my replacement unit.  Just a quick question: the Single Seq button is translucent, like the Run/Stop and auto buttons, but it doesn't light up.  Is it supposed to?

[tinhead]

Got my replacement unit. 

happy you

Just a quick question: the Single Seq button is translucent, like the Run/Stop and auto buttons, but it doesn't light up.  Is it supposed to?

no, there is no LED for that button.