HobbyKing.com New Products Flash Sale
Reply
Thread Tools
Old Apr 03, 2010, 12:22 PM
Mechatrommer
shafri's Avatar
Semporna, Sabah, Malaysia
Joined Apr 2010
41 Posts
Discussion
Rigol 1052E to 1102D Conversion

Hi. Here i'm seeking help due to problem i encountered during an (brave/stupid) attempt to convert my Rigol Oscilloscope 1052E to 1102D. due to encouragement from few forums:

1) http://www.eevblog.com/2010/03/31/ee...00mhz-ds1102e/
2) http://www.rcgroups.com/forums/showt...663958&page=58
3) http://hackaday.com/2010/03/10/50mhz...pe-conversion/

i've managed to convert my 1052E to 1102D, but with not so smooth process, in the end, my rigol exhibit a few problem mainly on vertical system. and i'm still unable to prove that my 1052e now is indeed a 1102D, coz when i hit the LA button, it still showing "Function not available" and i dont have a Logic Head to test.

the story begin when EEVBlog forum #1 above is published, i followed the step and successfully convert to 1102E. but i think, why not i just dierctly upgrade to LA version, 1102D using same method? is it possible? after hours of reading 58 pages of forum #2 and possibility of upgrading 1052D to 1102D is suggested in forum #3, there is still no concrete evidence that 1052E->1102D has been done. so i think, what the hell, let me give it a try and report it to the world.

as an introductory, here is pics of my converted 1052E to 1102D:
shafri is offline Find More Posts by shafri
Reply With Quote
Sign up now
to remove ads between posts
Old Apr 03, 2010, 12:43 PM
Mechatrommer
shafri's Avatar
Semporna, Sabah, Malaysia
Joined Apr 2010
41 Posts
before i continue with my problem, let me give a summary of what i think that should be done to convert:

1) upgrade your firmware to DS1000D version 02.02.02 sp2 (available at RCGroup forum #2). rename it from DS1000DUpdate.RGL to DS1000EUpdate.RGL,
put in your USB Drive root. put it in your DSO, it should automatically detect and press ok to upgrade (if you already have firmware 02.02.02.sp2 it will ask to upgrade to the same version, just click ok. we are using D (LA) version remember?)

2) follow the step as shown in EEVBlog forum #1 above except:
:INFO:MODEL 1102D (ALT 010)
:INFO:SERIAL DS1EAxxx (ALT 010)

3) restart your DSO and calibrate.

i screwed up when i followed the HackADay forum #3 suggesting 3x linefeed, ie hitting (ALT010) 3 times. my rigol hanged, when i restarted, nothing changed, i'm still in DS1102E model (remember, i made this attempt after i successfully convert to 1102E). when nothing happened, i retype the command using 1x linefeed, exactly as EEVBlog suggest. i restarted, and i succesfully converted it. but when i look carefully, my serial has changed, instead of 114 (at the end), it showed 112 (i dont know whether something screwed it or my careless during the typing). i'm not happy, i retype again, i got screwed up again. after doing *IDN? and :INFO:MODEL 1102D command, dont know what happened, i restarted the rigol, and the graph line (Channel1 and 2) dissapeared, when i checked the info, my model and serial is gone (blank). so there recovery took place and problem start to emerge. i managed to recover my model and serial number by typing :INFO:MODEL 1052E and :INFO SERIAL DS1EDxxx and reconvert by :INFO:MODEL 1102D and :INFO:SERIAL DS1EAxxx method and only 1x linefeed, but the osc graph still dissapear. i managed to get the dancing graph back on line by doing calibration, but now they are dancing differently (more below).

so if anyone dare to try, please dont screw up like i am and do the above step/command in one go, no single mistake is permitted. and if anyone found out the above method is not working, please let me/us know.
and dont blame me if your DSO explodes, do it at your own risk.
please let me know if it is working if done in one go.
shafri is offline Find More Posts by shafri
Last edited by shafri; Apr 03, 2010 at 01:02 PM.
Reply With Quote
Old Apr 03, 2010, 01:33 PM
Mechatrommer
shafri's Avatar
Semporna, Sabah, Malaysia
Joined Apr 2010
41 Posts
The Problem

Now the real problem. I think i broke the vertical system. Even after several times re-calibration, the vertical system/axis not showing the right value. Even if i revert to model 1052E and re-calibrate, this problem still exist and not reversible (i have no way going back anymore, hence i stick and reconvert to 1102D).The horizontal system is doing fine. Testing is done at 10X attenuation. 1X exhibits similar behaviour.

Pic 3.jpg showing the graph i've made when measuring known voltages and the resulting reading from the DSO. 0V means the probe is opened or shorted, 3.3V means the probe is connected to 3.3V source and the same with 5V and 12V. dash "-" means its not being tested due to off the graph result. from the table in 3.jpg, it showed wrong reading and 2V/div and 100V/div are totally broken. when i connected 12V source to the probe, the DSO reads 7V at 5V/div setting. at 2V/div setting no matter what voltages you connect it to, it still showing -2.5V reading.

Pic 4.jpg showing reading when connected to 12V source at 5V/div setting.
Pic 5.jpg showing reading when probes 1&2 not connected to anything at 2V/div setting.
Pic 6.jpg showing reading of 4V at 5V/div when connected to 6V Lead-Acid Battery.
Pic 7.jpg showing reading of -2.5V at 2V/div when connected to 6V Lead-Acid Battery.
Pic 8.jpg showing compensating probe. 2V peak-peak? very funny!
shafri is offline Find More Posts by shafri
Reply With Quote
Old Apr 03, 2010, 01:52 PM
Registered User
Michigan, USA
Joined Aug 2008
390 Posts
Hi, Shafri.

I'm not sure I'm following all the issues you been having since your conversion. Or that there's any need to do so. But I can help you out with some information...

> when i hit the LA button, it still showing "Function not available" and i dont have a Logic Head to test. <

It wouldn't matter if you did have a Logic Head to test with (and the special ribbon cable). Which, BTW, together cost over $400. Even if you opened the front access port and added the internal ribbon cable.

You'd still be missing one vital component, which is not available separately, which is a daughtercard for the LA functions, which contains a 2nd Altera Cyclone chip, along with another 1 MB RAM buffer. It was recently discovered that the 40-pin header on the Rigol mainboard does NOT connect to the front panel, but rather that is where the daughtercard plugs in.

So I'd say the non-LA Rigols are not user-upgradable to LA-models, as some (including myself, at one point) had speculated they might potentially be.

- Mark
Mark_O is offline Find More Posts by Mark_O
Reply With Quote
Old Apr 03, 2010, 01:58 PM
Registered User
Michigan, USA
Joined Aug 2008
390 Posts
shafri,
> Now the real problem. I think i broke the vertical system. Even after several times re-calibration, the vertical system/axis not showing the right value. <

That could be due to having the D-series firmware loaded.

> Even if i revert to model 1052E and re-calibrate, this problem still exist and not reversible (i have no way going back anymore, hence i stick and reconvert to 1102D). <

I'm not sure I understand why you "can't go back"? Have you tried reloading the E-series firmware? That would be where I would start, since the D-firmware isn't going to do you any good.

- Mark
Mark_O is offline Find More Posts by Mark_O
Reply With Quote
Old Apr 03, 2010, 02:02 PM
Mechatrommer
shafri's Avatar
Semporna, Sabah, Malaysia
Joined Apr 2010
41 Posts
Frequency test

For the frequency test. I dont have any wave generator, a step down transformer also nowhere to be found, so the easiest way is i tried to program my existing pic10f202 to output digital pulses (0-5V). As i said (for the test that i've done ranging 12Hz - 200KHz++), horizontal system/axis having no problem and showing the correct timeline. Except the vertical system, showing the wrong value, its like the voltage is being attentuated to a smaller value.

During the frequency test, i found another strange behaviour.ie at lower frequency (around 100Hz and lower), digital pulse from pic10f202 (and any other mcu by right i think) should be "square shaped", right? but no, the converted rigol showing different shapes, curved and trapezoidal shape, depending and the frequency, you can see what i means in pics 12-15.jpg

And again, i think the vertical system is the culprit here.
shafri is offline Find More Posts by shafri
Reply With Quote
Old Apr 03, 2010, 02:10 PM
Mechatrommer
shafri's Avatar
Semporna, Sabah, Malaysia
Joined Apr 2010
41 Posts
Quote:
Originally Posted by Mark_O View Post
That could be due to having the D-series firmware loaded. - Mark
wow! i didnt expect to get the reply this fast! i havent finished reporting yet. thanx alot Mr Mark. for the firmware. i've revert back all to original. firmware E version, model 1052E serial ds1edxxx.... but the recalibration still flawed. i think its non reversible, something (harware or software) must have been broken. and the daughter card? yeah, thats my worst nightmare. i cannot find complete description on that before. now i know i shouldnt doing this at the first place. i thought it just the "ACTIVE" logic head, which i saw the dismantling report by Mr Trevor G.
shafri is offline Find More Posts by shafri
Reply With Quote
Old Apr 03, 2010, 02:26 PM
Mechatrommer
shafri's Avatar
Semporna, Sabah, Malaysia
Joined Apr 2010
41 Posts
Noise Dancing

i should have asked 1st b4 the attempt. . now mr Mark proved that its not worth a dime to convert E to D model.

Anyway, another report is at lower timescale, less than around 20ns, i have the graph more dancing up and down more frequently, i dont know what happened. for larger timescale, this behaviour is not really apparent (similar to my previous non-converted rigol). 20-22.jpg.

Pics 23,24.jpg to show you some of my equipment to do the rest of testing. please dont laugh , i know its crude, and trust me, i can live with it. I'm not EE person, just a simple hobbiest try to make something small. thats why i'm doing this thread to ask from EE gurus, pros and master technician if anything i can do about my rigol.
shafri is offline Find More Posts by shafri
Reply With Quote
Old Apr 03, 2010, 02:42 PM
Mechatrommer
shafri's Avatar
Semporna, Sabah, Malaysia
Joined Apr 2010
41 Posts
The Post Mortem

So, ok, finish with the report. Now the problem arised, i tried to go back, i cant. So how to deal with this? I think my rigol is still usable for most of my digital design (mcu signal monitoring etc), but for analog design (which i seldomly do, but sometime i will), i think this rigol is unreliable anymore, in term of giving the correct voltages. So to repair this, i made a thought for several minutes, maybe the cause is either software or hardware or both. I tried to re-open and disassembled the main board, but i cannot find any sign of explosion, or burnt components. So i divide the possibility to two major parts:

1) Software:
------------
a) is it maybe that some parameter used for calibration burnt/stored in flash or those cpld/fpga accidentally corrupted?
b) is it maybe the rigol company purposedly put a safety program to corrupt itself when a conversion atttempt is detected?
c) or anything that cause the program not working properly?

2) Hardware:
------------
a) Can the components be damaged when the calibration is done and i forgot to unplug my Serial cable connecting PC to rigol? this is happened during the conversion process.
b) Can the conversion process generate a voltage/current spike that cause damage to components?
c) Which part in the mainboard most likely be damaged? (relating to vertical system?
shafri is offline Find More Posts by shafri
Reply With Quote
Old Apr 03, 2010, 02:49 PM
Registered User
Michigan, USA
Joined Aug 2008
390 Posts
> and the daughter card? yeah, thats my worst nightmare. i cannot find complete description on that before. <

Yes, that's something we just discovered about a week (?) ago, on the EEVBlog forum. So it's very recent knowledge.

You mentioned that you did perform the Calibration process after reloading the firmware, but did you make sure you had no probes plugged in at the time? That can mess things up.

> digital pulse from pic10f202 (and any other mcu by right i think) should be "square shaped", right? <

Theoretically, yes. Though since you adjusted your probe compensation, and the tops of the 1 kHz square waves were flat, it's harder to see how the Rigol could be introducing this anomaly. In other words, they might actually have this shape.

One final comment is that if you're using a PC-type power supply for voltage reference signals, they don't always work very well without having a load applied.

I'll let you finish your descriptions before adding more.

- Mark
Mark_O is offline Find More Posts by Mark_O
Reply With Quote
Old Apr 03, 2010, 02:51 PM
Registered User
Michigan, USA
Joined Aug 2008
390 Posts
> Which part in the mainboard most likely be damaged? (relating to vertical system?) <

One thing you'll want to do is check and see if BOTH vertical channels are behaving the same way. If they're not, possibly something is physically damaged. But if they're both the same, it's likely a soft cal setting that's affecting things.

- Mark
Mark_O is offline Find More Posts by Mark_O
Reply With Quote
Old Apr 03, 2010, 03:03 PM
Mechatrommer
shafri's Avatar
Semporna, Sabah, Malaysia
Joined Apr 2010
41 Posts
So here i'm asking help from EE gurus, pros, hardcorer, software and hardware hackers, and master techicians who really familiar with DSO design. How i'm going to repair it? Which part is more likely to be damaged?

i know just minutes ago (Marks reply) that i should not be doing this. I know my rigol is broken. I know i cannot send it to rigol service, i've voided the warranty, its far away in china and postage cost will be like half the price of this 1052e.

But i need to know, is anybody out there that possibly know what to do? what to replace? i'm begging your assistance on this. My main concern are:

1) Which hardware that most likely be damaged? and where it is located so i can concentrate on those area to check and troubleshoot?

2) if its software. How do i hack into the chips software to see if any bytes are broken and out of order? as i understand, there is blackfin, altera and some eeprom flash inside the 1052e. Any datasheet and instruction set or hardware programmer i need to know?

In the mean time, i'll be searching on this besides continue with my other works. there is no need for me to rush on this, things already happened. I know somehow, even though its not perfect anymore, i trust i will manage to get it to work esp on my mcu digital/software design. And i'll find a way to make it work when it come to analog design.

But pleaseeeeeee help me! Esp people who have been contributing to the hacking of Rigol DSO and EE Knowledge on the net (thank you for your contribution), i dont ask to rush, but at least, dont ignore me, thank you.
shafri is offline Find More Posts by shafri
Reply With Quote
Old Apr 03, 2010, 03:22 PM
Mechatrommer
shafri's Avatar
Semporna, Sabah, Malaysia
Joined Apr 2010
41 Posts
"PC-type power supply for voltage reference signals, they don't always work"
i measured before using 1052e, its ok, no false reading. i also tried 12V and 6V lead acid battery, now 1102D (converted) reads only 8V and 4V respectively. and 2V/div and 100V/div is totally broken (no reading at all), graph visibly will jump for a moment and then goes back to -2.5V and -130V reading.

"it's harder to see how the Rigol could be introducing this anomaly"
i have programmed a PWM signal coming out of my arduino mega at 30Hz, DSO showing square wave before, i will retest the signal, but i need time as my arduino already tucked inside my toy car.

"if BOTH vertical channels are behaving the same way"
YES! both behaving the same way except at slightly different value. I showed in the pics that Channel1 has an offset of -2.5V, whereas Channel 2 has -1.8V offset (not connected probe, 10X, at 2V/div), i attached the pic (thought it not that important, sorry). Both are dead at 2V/div and 100V/div (no useful reading)
shafri is offline Find More Posts by shafri
Reply With Quote
Old Apr 03, 2010, 04:03 PM
Mechatrommer
shafri's Avatar
Semporna, Sabah, Malaysia
Joined Apr 2010
41 Posts
"but did you make sure you had no probes plugged in at the time"
Yes. i did as instructed in the service manual. its just i forgot to disconnect the serial cable during the calibration, is it possible to damage if serial is connected?

p/s:i'm converting back my rigol to 1102E model and 1000E firmware, very funny! what a wasteful effort! a daughter card and a logic head might cost me at least another 100++ bucks!
shafri is offline Find More Posts by shafri
Reply With Quote
Old Apr 03, 2010, 05:33 PM
Registered User
Michigan, USA
Joined Aug 2008
390 Posts
> i forgot to disconnect the serial cable during the calibration, is it possible to damage if serial is connected? <

No, you're safe on that count.

- Mark
Mark_O is offline Find More Posts by Mark_O
Reply With Quote
Reply


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Discussion Rigol Digital Storage Oscilloscopes? bigandy DIY Electronics 971 Jul 01, 2014 04:49 AM
Idea Blade mSR conversion to Autogiro dz1sfb Auto Gyros 15 Mar 03, 2010 08:39 AM
Discussion Duel Conversion Crystal is Single conversion RX NX-687 Radios 2 Jan 07, 2010 03:41 AM
Discussion New Rigol CD has wrong DLL and User Guide algolytics DIY Electronics 0 Jul 09, 2009 07:50 PM
Discussion interfacing Rigol mario_re DIY Electronics 0 Nov 16, 2008 05:27 PM