The Rigol DS1052E

[rf-loop]

Mark_O

some comments.

HP pulse generator output is 1,3ns risetime or less. (in output connector)
(after this I have 2m Suhner RG223/U and some cheap quality T where is HP 50ohm terminator (medium quality, not high-end calibrator quality).
I do not know risetime in the oscilloscope input connector! (also Rigol input capasitance is quite high, around 15pF)

We can not calculate this 1,8ns risetime becouse bandwith curve is what it is. Rise is not line, it is curve and curve shape depends many things. I think real comparable risetime in Rigol is around 2,5 - 2,9ns but of course there are also more fast components. It is measured at  around 10 and 90% points as Rigol seems to do. In this point we need remember BW shape.  (and it also depends what is mV/div setting, I use now 50mV/div). Also I have not seen exactly specifications about Rigol "risetime" measurements.

Yes also of course I swap HP outputs. Not big differencies! (this difference in pulse level is not from scope. It is from HP littlebit uncalibrated levels in HP outputs, there are some small difference between channels. (but this we can forget becouse overall accuracy in these test is not so high)

I also look both scopes with Marconi (Aeroflex) 2041.



First I adjust 10kHz level so that 1102 show 300mVp-p (variations becouse noise etc 298-300)
Next I check 1052 with exactly same adjustments. It show 300mVp-p but variable between 300-302)

now... 200MHz (same level as qood as this signal source give, and now I use exactly same cable and same terminator and T for avoid any variations in terminator, T and cable so I use just same and NOT drive scopes parallel with 2 cable... never do this without high-end (and very expensive) splitter. Also good measurement need still do different way... it need calibrated precision terminator, calibrated precision power meter and high end splitter in the scope end of cable.)
But for hobby use accuracy this my measurement is enough accurate for this purpose now. It simpy  show that with these unique scopes modify works extremely well.

Both Equ time mode and average 8, dspl points,  and only one channel use (CH1) and 50mV/div

1052Emod
10kHz ~301mVp-p  ;  200MHz ~190mVp-p


1102E
10kHz ~299mVp-p  ;  200MHz ~174mVp-p

With these unique scopes it looks like BW is littlebit better in 1152Emod
(but remember I have not test frequency response curve shape. Only with these point frequencies. (yes I have fast look with hand sweep and I can not see any big differencies over band. )

note: I use p-p measurement but more important is that both scope get same signalsource with same adjustments. Values are not absolute but comparable between each others.

Later I look risetime also with Tektronix 284 70ps risetime pulser.

[rf-loop]

One warning...

If some day modified scope fails... (I do not mean modify itself is reason for fail)

and if you can not restore original type and serial. (example scope do not start at all)

Then you send it to repair becouse it have 3 year warranty.

You can quess what happend IF they find wrong serial and type...

[jone]

Hmm, I think I'm close - but it's not quite working for me yet..

Changing the serial number from DS1EDxx to DS1EBxx seems to have worked fine, and persists across restarts.

But the model always reverts to DS1052E on restart, any clues what I might be doing wrong?

Thanks in advance!

[JimBeam]

Hmm, I think I'm close - but it's not quite working for me yet..

Changing the serial number from DS1EDxx to DS1EBxx seems to have worked fine, and persists across restarts.

But the model always reverts to DS1052E on restart, any clues what I might be doing wrong?

Thanks in advance!

Yupp, that's because I cursed any other scope on earth to be immune to the change   !

No, seriously - no idea, maybe there's more to it than changing the "first name" of the serial number. I was astonished from the beginning, that nobody had ever tried to change modelnumber and serialnumber before - since the docs about the "secret commands" are already dangling around a while. Perhaps all the fuzz about RIGOL losing money is obsolete, if this "patch" only works by chance...

[Mark_O]

rf-loop,

thanks for all the additional information, and all the testing you've been doing.  Those 200 MHz results are quite impressive.

Also I have not seen exactly specifications about Rigol "risetime" measurements.

The specs in the back of their models indicate 7 nS risetimes for their 50 MHz units, and 3.5 nS for their 100 MHz models.  These are standard industry values.

(but remember I have not test frequency response curve shape. Only with these point frequencies. (yes I have fast look with hand sweep and I can not see any big differencies over band. )

Yes, I understand that, and it's a point worth remembering.  Unless you have a sweep generator, it's hard to examine the entire bandwidth.

- Mark

[rf-loop]

yes, first also I have problem... after restart it is agen 1052... then I try some.....

"maybe" serial numbers...

"maybe" 1102 can not be what ever serial number after ..EB..

......also maybe it is meaning what is scope state (run/stop) as you do it and also... maybe need not soft reset or need... before shut off...  (really I have not test all variations)

After mod need also do selfcal.

[flolic]

I'm not seeing anything that looks remotely like a model number or serial number.  This was from a DS1052E -- anyone else care to do the same so we can compare?

I can do it, but only if you tel me how. Can I dump EEPROM content without desoldering the chip (btw desoldering is not a problem  )?
I have 1102E.

[Mark_O]

Changing the serial number from DS1EDxx to DS1EBxx seems to have worked fine, and persists across restarts.

But the model always reverts to DS1052E on restart, any clues what I might be doing wrong?

Which one did you change first?  I.e., writing the SN may force an update of non-volatile memory, while the Model does not.  If Andreas changed his Model first, then the SN; and you did things in the opposite order, that could explain it.

- Mark

[rf-loop]

rf-loop,

thanks for all the additional information, and all the testing you've been doing.  Those 200 MHz results are quite impressive.

Also I have not seen exactly specifications about Rigol "risetime" measurements.

The specs in the back of their models indicate 7 nS risetimes for their 50 MHz units, and 3.5 nS for their 100 MHz models.  These are standard industry values.

(but remember I have not test frequency response curve shape. Only with these point frequencies. (yes I have fast look with hand sweep and I can not see any big differencies over band. )

Yes, I understand that, and it's a point worth remembering.  Unless you have a sweep generator, it's hard to examine the entire bandwidth.

- Mark

Yes 3,5 and 7ns are "normal". (this have old history where come this 0,35)

I mean that I do not know how exactly Rigol oscilloscope measure risetime. There is no data in specifications. What values are take for 10 and 90% points. (highest value or some value after shoots as I have see some scopes do)
(this question can make specially after drive sine to scope and use scope automatic measurement. In some cases it show totally fast risetimes specially with high frequencies. So this is clear that scope do not use p-p points for 10 and 90% points for calculate risetime.

Most good accuracy can get without sweep gen. But it is hard work...

With sweep gen with digital oscilloscope is very easy go to garbage...

And also it is impossible to sweep whole band at once... (with analog scope situation is littlebit different)
How can read exactly values in separate freq points. My opinion is that only road for accurate testing is do it with measuring single freq points.

[Mark_O]

I mean that I do not know how exactly Rigol oscilloscope measure risetime. There is no data in specifications. What values are take for 10 and 90% points. (highest value or some value after shoots as I have see some scopes do)

Ah, yes. I see what you mean now.  10% and 90% of which data points.  Thanks for the clarification.

- Mark

[jone]

Which one did you change first?  I.e., writing the SN may force an update of non-volatile memory, while the Model does not.  If Andreas changed his Model first, then the SN; and you did things in the opposite order, that could explain it.

- Mark

Good point - I changed serial number first, then model. Will try the other way round after work...

[Simon]

I look at it a little bit differently.

As far as the economics -- this is a bit like like microprocessors, where (e.g.) Intel will make one batch of chips and sell them at three different speed grades, depending on how they test.  This has a couple of benefits for Intel:

what you forget is that intel set out to get fast chips and through the nature of the process get some that won't perform so run them slower and make them still usable, this is clever and good as it lowers the price all round and is less waste, what rigol did was make a 100 MHz scope and then sell a load with a limitation that was artificially imposed. of course there were valid reasons i suppose.

the 2 rigol scopes have the same sample rate, even the 1052 is overclockedelse it would be a 20 MHz scope[/list]

[Simon]

is all this being done with software provided on the original cd ?

from the commands manual I downloaded do I correctly understand that the whole scope can be command line controlled froma pc as it would from it's own control panel ? and of course for doing things like this

[jone]

is all this being done with software provided on the original cd ?

from the commands manual I downloaded do I correctly understand that the whole scope can be command line controlled froma pc as it would from it's own control panel ? and of course for doing things like this

Not sure that there are actual apps on the CD, but there are some programming examples in the PDF which show you how to use the required libraries (I think these were on the CD, else they're on the rigol site).

I used a Linux machine, which has support for USBTMC (USB Test and Measurement Class) in the kernel. As soon as the scope is connected, it appears as a character device which you can read from and write to.

[drieg]

I was astonished from the beginning, that nobody had ever tried to change modelnumber and serialnumber before - since the docs about the "secret commands" are already dangling around a while. Perhaps all the fuzz about RIGOL losing money is obsolete, if this "patch" only works by chance...

Nobody published doesn't mean nobody tried...

As for the LA module - yes, it's also possible to turn DS1000E into DS1000D but you have to build not only external, but also internal LA module with ALTERA FPGA and ISSI SRAM...

[JimBeam]

As for the LA module - yes, it's also possible to turn DS1000E into DS1000D but you have to build not only external, but also internal LA module with ALTERA FPGA and ISSI SRAM...

Ah, this is what I wanted to know! Someone once wrote, that the 40 pin header near the front of the PCB would go directly to the "SCSI" connector on the front bezel - which I doubted, and unfortunately was right... OK, so the 1xxxE -> 1xxxD "modding" is buried for me, as I already own an Intronix Logicport LA (nice part btw.!), it would just have been a nice addition to i.e. have more trigger channels.

[Simon]

is all this being done with software provided on the original cd ?

from the commands manual I downloaded do I correctly understand that the whole scope can be command line controlled froma pc as it would from it's own control panel ? and of course for doing things like this

Not sure that there are actual apps on the CD, but there are some programming examples in the PDF which show you how to use the required libraries (I think these were on the CD, else they're on the rigol site).

I used a Linux machine, which has support for USBTMC (USB Test and Measurement Class) in the kernel. As soon as the scope is connected, it appears as a character device which you can read from and write to.

yes I think there are just examples available on the CD no actual software. what is the easiest software to use ?

[jone]

yes I think there are just examples available on the CD no actual software. what is the easiest software to use ?

I just used a linux machine, and simply echoed the commands to the device port. So that would get my vote for easiest! 

I guess for building the example software under windows, whatever you're most comfortable with (I think there were examples for VB and VC).

[Simon]

hm I could set up ubuntu on an old machine running off a live cd, but then I'll still have to find out how I "talk" to the rigol

[mxmxmx]

yes I think there are just examples available on the CD no actual software. what is the easiest software to use ?

I use Agilent's IO Libraries instead of NI Visa. There is a tool included in the Agilent IO Libraries that allows you to send raw commands to your devices, including the Rigol scope.

[mxmxmx]

Congratulations to the person who suggested that changing the model and serial number actually does convert the scope into a 100 MHz model. It really works! When I found and documented the previously undocumented commands to set those numbers, I would never have thought that it would be so easy to upgrade the scope.

PS: For those of you who say "seeing is believing", I uploaded two screenshots:

100 MHz signal, DS1052E, default setting


Same signal, same scope, modded to 100 MHz bandwidth

[rossmoffett]

Congratulations to the person who suggested that changing the model and serial number actually does convert the scope into a 100 MHz model. It really works! When I found and documented the previously undocumented commands to set those numbers, I would never have thought that it would be so easy to upgrade the scope.

I just assumed you had already tried it, and never made the attempt!

Good on you guys for trying and testing it, guess I'll be soldering my caps back in soon as I get back to Oklahoma.

[Simon]

so does this mean that the hardware mod you made is now obsolete ?

[Mark_O]

is all this being done with software provided on the original cd ?

That is one way, though there are simpler.

from the commands manual I downloaded do I correctly understand that the whole scope can be command line controlled from a pc as it would from it's own control panel ? and of course for doing things like this

Yes, that is correct.  I think you're still making this out to be far more difficult than it needs to be, thinking you need to "program" something.  Assuming you have a serial cable, and HyperTerminal on your PC (comes standard on all WinBoxes), just download these instructions on how to use HyperTerminal to send text commands back and forth to your Rigol.  No special software required... just a simple terminal emulator.  And all the required settings you'll need to use it (setting baudrate, etc.) are described in the Rigol manual referenced.

- Mark

[flolic]

 guess I'll be soldering my caps back in soon as I get back to Oklahoma.

I will not, because my 1102E has even greater bandwidth without them and I don't need BW limit function  

These are shots of 74F08 gate output, i have to find IC with faster rise/fall times:
(EQU sample rate, 4x averaged, 50 ohm terminated resistive input probe)

CH1, with caps removed


CH2, unmodified


Differences are minimal but still visible.
Btw. I still think there is a room for (hardware) improvement...