changing the rigol DS1052E to DS1102E using USB , the dummy guide

[fminne]

Just one more question about the pule generator: from where do you have the PCB?

[RegAV]

Thanks Chet for the link, guess I'll have to read all 49 prev pages first, Ultrascope does not run either with missing dll error, another google required.

You don't need to read the whole story. You need to read page 1, post 1. All the critical information has been continuously updated in that post through the dedicated efforts of Polossatik.

To make things easier, I put all the files together in one package that you can download from
DS1052 Upgrade tools zip file

EXCEPT Ultrascope (which comes from Rigol)

Great stuff torch, have got everything, even Ultrascope now functions ok, cheers mate!

I now have the mode done, many thanks to all.

[torch]

I soldered everything together this afternoon. It's as neat as I could do, but pretty ugly compared to jahonen's piece of art. Anyway, it certainly made a big difference, both by stabilizing the jitter issue I was experiencing and in the performance:



As with earlier today, that's displaying dots with persistence on and the manual cursors set as close to 10% and 90% as I could get. It's coupled directly to the scope, but I don't have an inline terminator so there's just a single 50ohm resistor on the board. I guess I'm going to have to switch it back to an 1102 to verify things now.

[torch]

I guess I have to eat crow. There is a slight -- and I stress the word SLIGHT -- but consistent difference when the Rigol is told it is a DS1152E-EDU vs DS1102E. I have flashed and reflashed back and forth several times, (downgrade to 2.02SP2, change model and serial number, verify new info acccepted, upgrade to 2.04SP1, run self-calibration) and set up the identical configuration -- view dots, persistence on, average on, sinx/x off, normal trigger, 2ns/dif, 2v/div, measure to display max volts and rise time, manual cursors carefully adjusted to 10% and 90%.

As a DS1152E-EDU, the rise time is consistently about 1.84ns, +/-0.04ns. As a DS1102E, the rise time is consistently 1.96ns, +/-0.04ns. The margin of error is based on my manual skills considering the resolution of the display (bear in mind that I am now so, ummm, mature, that my reading glasses now need reading glasses    ) and even the automatic measurement flickers by 0.02ns from time to time. Anyway, here are some representative screen shots of each:

As a DS1102E:


and as a DS1152E-EDU:


I can not explain the difference. It makes no sense to have so slight a difference anyway! But it does seem to be real, as I said, I went back and forth several times to rule out environmental variables like ambient temperature, battery voltage, circuit burn-in, etc.

The only thing I can think of is that the area of memory where the data is stored may be changing as a result of the flash process. I recall reading something about the Rigol having 8 megs of flash ram but the F/W files are all 4 megs or less. The person was theorizing that the F/W is flashed to alternate halves each time (so it does not overwrite the existing F/W) and that the remaining 4 megs may then be used as data storage space. If this is so, then perhaps the different memory areas have slightly different access times which affects the sample rate ever so slightly??? I have no idea if this could be so, perhaps someone with more knowledge about the internals (drieg?) might have a better guess.

[A Hellene]

... or, maybe, your (or, our) lying eyes --it was not in my intentions to be heard becoming personal-- are actually seeing the result of the release of the software filtering of the bandwidth!

Another point is the use of the FLASH memory: It would be bad engineering to keep in the FLASH any variables, which update quite frequently, given that any FLASH memory cell can only be erased (in oder to be re-written) 10,000 to 100,000 times. That's why the variables are kept in EEPROM memories, which can be erased and re-written 100,000 to 1,000,000 times, respectively.

But Rigol engineers are very well aware of that and chose to keep the frequently updated variables NOT in a FLASH nor even in an EEPROM memory; they chose to keep them in an FRAM memory (the "FM24CL04" part, where the user’s settings are probably been stored), which can be erased and re-written indefinitely!

So, it is obvious that when the device is told to become a DS1152 unit, it simply stops any further software filtering of the bandwidth!
By the way, thank you for your insistent efforts to find any differences between the DS1102 and DS1152 modes of operation!


-George

[maze]

  it was so damn simple to modify the header and didn't brick the scope

[torch]

This afternoon I installed the pulser in a nice little metal project box, connected the case to ground and located the circuit board just above the case so it could be a ground plane, coiled the co-ax up nice and stuffed it in the box, buttoned it all up and connected it to the scope to test it. This is what I got when I flipped the switch on:



I'm not sure what to make of things now.

[A Hellene]

198.9 MHz bandwidth!
That's impressive!

[tinhead]

I can not explain the difference. It makes no sense to have so slight a difference anyway! But it does seem to be real, as I said, I went back and forth several times to rule out environmental variables like ambient temperature, battery voltage, circuit burn-in, etc.

The only thing I can think of is that the area of memory where the data is stored may be changing as a result of the flash process. I recall reading something about the Rigol having 8 megs of flash ram but the F/W files are all 4 megs or less. The person was theorizing that the F/W is flashed to alternate halves each time (so it does not overwrite the existing F/W) and that the remaining 4 megs may then be used as data storage space. If this is so, then perhaps the different memory areas have slightly different access times which affects the sample rate ever so slightly??? I have no idea if this could be so, perhaps someone with more knowledge about the internals (drieg?) might have a better guess.

it does not have anything to do with flash, best example - Hantek/Tekway. Input stage like Rigol, flash / Soc totally diff. architecture/size.
It even does not have much to do with the varicap/cap combination between both opamps in input stage, Hantek/Tekway using them only for 60 and 20MHz,
but not 100 (like Rigol does to improve a bit the bw filter), 150 and 200Mhz bw reduction. Everything happens within FPGA design, where FPGA is talking to CPLD and CPLD is changing the gain of AD8370 (combination of digital low pass filter and gain control).
Everythign before AD8370 is flat (to over 300MHz), everything after AD8370 is cut depends on model settingsand not on components values.
Rigol is probably using exact the same way, sure the varicap/cap are used for 100/60 and 20Mhz on Rigols, but that's didin't
really matter (from bw point of view).
So if there is DS1152E model within the firmware, then the firmware is doing "the magic", and not an analog filter.

[flolic]

Ok, one thing is not clear to me.

My scope is original DS1102E with hardware version 05 and firmware 00.02.02 SP2.
I tried to upgrade to 1152E-EDU but somehow it is not possible (or I am doing something wrong). With demoIDN tool I can change model to 1152E, but
i can't achieve that "-EDU" part. Splash screen on startup is same as before, and system info says just DS1152E.

My scope was originally like this:
:INFO:MODEL DS1102E
:INFO:SERIAL DS1EB113*****1

Then I changed it to:
:INFO:MODEL DS1152E
:INFO:SERIAL DS1EF113*****1

But nothing happened, except model name changed. Btw I can change model to 1152E without altering serial number (DS1EB113*****1 works just fine).

Any thoughts on this?  

EDIT:
This is rise time measurement from my scope, LP filter removed.
Equ -time sampling, 8x averaged

[torch]

Did you enter the model number as "DS1152E-EDU" or just "DS1152E"? The former changes the splash screen to:



Adding 40" of RG-174/U in series with a 100 ohm trim pot connected with the core to the transistor collector and shield to ground (other end left unterminated) plus changing the cap that is connected to the collector to 10pf will broaden the pulse so it is more nearly square as shown in my screen shots. This was recommended to me to ensure the scope has time to capture the full voltage of the pulse.

[flolic]

I enter just "DS1152E"

I will connect coaxial cable and try again 

[flolic]

Ok, I managed to change model to DS1152-EDU but performance is exactly as before.

This is with coaxial cable and trim pot connected to pulser:

[torch]

Ok, I managed to change model to DS1152-EDU but performance is exactly as before.

Yeah, I kind of suspected that. Just window dressing

This is with coaxial cable and trim pot connected to pulser:

I see your voltage came up, but is still about half what mine is showing. I wonder why? The pulse shape is pretty similar to mine though. Maybe a difference in a cap value somewhere.

I wish I had access to a high speed scope to test my pulser and confirm it's performance, because either:
a) your scope is faster than mine, or
b) my pulser circuit is slower than yours. Since I saw a difference just closing it in, and since I'm pretty new at all this, my money is on option b. But it could well be a difference in tolerances between the two scopes -- after all, we are pushing them at something like four times what we paid for ;-)

It would be interesting to know if you see a difference between telling your scope it's an 1102 or an 1152 model. In other words, can my results be duplicated or are they an anomaly?

[flolic]

I see your voltage came up, but is still about half what mine is showing. I wonder why? The pulse shape is pretty similar to mine though. Maybe a difference in a cap value somewhere.

My pulser is connected to scope via feed-through 50ohm 10dB attenuator. Also I'm using 1 meter of RG58 coax as transmission line, and finally transistor in pulser is just a relatively slow BC337.

I wish I had access to a high speed scope to test my pulser and confirm it's performance, because either:
a) your scope is faster than mine, or

I would say, yes. I completely removed bandwidth limiting filters in both channels. And I changed resistors around LMH6552 amplifier from 365 to 270 ohms.

b) my pulser circuit is slower than yours. Since I saw a difference just closing it in, and since I'm pretty new at all this, my money is on option b. But it could well be a difference in tolerances between the two scopes -- after all, we are pushing them at something like four times what we paid for ;-)

I really don't know about exact pulser speed. Unfortunately, I don't have access to high speed scope to check what is really going on...

It would be interesting to know if you see a difference between telling your scope it's an 1102 or an 1152 model. In other words, can my results be duplicated or are they an anomaly?

In my case, no difference at all...

[therian]

Ok, I managed to change model to DS1152-EDU but performance is exactly as before.

I have DS1102E (nice to know HW 07 isn't odd, I thought I mess it up) and I cant get model to show us DS1152-EDU in info screen, how did you manage to show up -EDU part ?
by the way I saved some REF of signals on DS1102E and there is tiny but constant difference when I alight them with same signal on DS1152 (no more than a pixel or two in amplitude), line is more jumpy and have sharper corners

[flolic]

I have DS1102E (nice to know HW 07 isn't odd, I thought I mess it up) and I cant get model to show us DS1152-EDU in info screen, how did you manage to show up -EDU part ?

In demoIDN tool, change name to "DS1152E-EDU"

Like this:
:INFO:MODEL DS1152E-EDU

[torch]

I would say, yes. I completely removed bandwidth limiting filters in both channels. And I changed resistors around LMH6552 amplifier from 365 to 270 ohms.

Yeah, that might explain it! 

[fminne]

In the attachment you can see the pulse generator made by Jahonen. He confirmed that this is a factory-made PCB.
My question: where can one buy this PCB-board (or is it a kit?). Can anybody help me, so I can test my modified Rigol DS1052E?

[jahonen]

In the attachment you can see the pulse generator made by Jahonen. He confirmed that this is a factory-made PCB.
My question: where can one buy this PCB-board (or is it a kit?). Can anybody help me, so I can test my modified Rigol DS1052E?

The PCB is designed by me, I have still few PCB's left if you don't mind building one yourself.

Regards,
Janne

[fminne]

This would be great! Tell me how I can order it. Thanks!

[jahonen]

This would be great! Tell me how I can order it. Thanks!

Just PM or drop me an e-mail.

Regrads,
Janne

[fminne]

Hi Ronaldlijs. Been there, done that. I also made this beginners fault . One has to remove first the power button cover before opening the scope.
It took me 3 days to search the internet before I was able to find any information on the switch. Here is what I have got: It is a PS5B-01 push button switch from BEJ Electronics Co.,Ltd, located in China.
But then it took me even much longer to get in contact with these guys. They didn’t react on my honest mails, explaining that such a PS5B-01 got broken and that I wanted to receive a new one. In the long run, I wrote a mail to the CEO, asking for a sample of the switch. As he thought I was going to retrieve large quantities of this switch afterwards, he sent me over such switches with a heavy catalogue. The transportation-cost was for me, so I can assure you that this is the most expensive power switch in a scope (over €100,00).
I still have a spare PS5B-01 available. Since I presume I will not break it an extra time, I can send it over to you. I would appreciate if you just pay the transport-cost of this switch from Belgium to the UK. In any case, this will be cheaper then what I had to pay for that heavy catalogue from China .
Leave me your e-mail address, so we can arrange the details via e-mail, instead of via this forum.

[Mechatrommer]

you can cut something to shape and become a button. but thats just me.

[Chet T16]

For 100 euro i'd be licking my finger and sticking it in there instead