Rohde & Schwarz RTO1024 upgrade ??

[darkstar49]

Hello together,

anyone ever attempted a h/w (mobo) and/or disk and/or OS upgrade of the PC part in an RTO ??

Just got an RTO1024, it works fine, but runs W7 embedded (x64) on a slow 160GB HDD, I'd like to at least replace the disk by an SSD, and eventually the mobo (Core 2 Duo currently, P8400), and also upgrade to W7 Pro, in order to install some additional software.
R&S just said that OS upgrade needs h/w upgrade (in a service center...), which I don't really believe, unless they took specific measures to prevent such an upgrade ??
The question in fact turns to: can the RTO software be user-installed on a blank OS, or is R&S needed for the initial installation ?

And... it has some options I'd hate to loose, anyone knows where these get stored ?? In a file ? Or in some EEPROM outside the mobo (like Lecroy) ?

I'll post some pics and results when I've some time, but would appreciate any info upfront...

[larry42]

I have an older RTE1104 -

I am almost certain you cannot go from W7 Embedded to W7Pro

Win Embedded systems are built specifically by the vendors with certain features added /removed.
I doubt that R&S will upgrade to W7Pro - I think they would offer to upgrade to a Win10 platform (and probably require H/W upgrade due to drivers - probably also drivers for DAQ cards / video cards) - probably not worth the price.

SSD upgrade is possible, but you need a FW CD / USB drive from R&S to boot from and install. I have done this on my RTE - makes the boot time and file saves faster (boot time ca. 1:30 from 2:30)


Anyone know how to enable options on these units?

[jjoonathan]

Hey, could someone with a Windows 7 RTO1000 use HWInfo (or something) to give me the dirty details of the RAM configuration?

Mine is an XP RTO1000, which ships with 2GB RAM as opposed to the 8GB standard on W7 RTO1000s. I've tried putting in new 2x4GB DDR3-1066 SODIMMs but haven't been able to get it to POST with them. For reference, here is what my current configuration looks like.

[Jackob]

I upgraded the mechanical HDD to a SSD in my RTO1004. To do that you do not need anything from R&S - you just pop out the RTO HDD out from the scope, connect it to a PC and clone the HDD to a SSD. For cloning I used Norton Ghost. You can probably use also other HDD cloning softwares.

For the cloning I connected the RTO HDD and the SSD to SATA cables from my PC (when the PC was off), then booted the PC to DOS from an USB stick, run Ghost.exe from the same USB-stick and cloned the HDD to the SSD. I did the cloning via DOS to prevent my PC's OS from writing anything to the HDD to be cloned...

BR, Jackob

[nctnico]

Clonezilla also works very well for cloning. I used Clonezilla for the last couple of cloning jobs and I used to be a die-hard Ghost user.

[jjoonathan]

Ditto. My HDD upgrade went smoothly, I just swapped in a new SATA SSD, cloned, and it worked.

RAM access was trickier. R&S had their resident sheet metal artist design a clever locking system that I was unable to defeat by grinding down hooks I had laying around. I wound up milling keys out of aluminum rod! There was probably an easier way, but that was mine: https://photos.app.goo.gl/M3uLj5EZG8Zk7qXB7

That's not the end of it, though. Apparently the supported RAM modules are a subset of those supported by the chipset -- which is why I'm in here begging for someone with an 8GB RTO to give me HWInfo RAM details :-)

[jjoonathan]

Actually, I'd like to amend that a bit -- I probably should not have asked for HWinfo details, but rather for a SPDTool dump. SPDTool is still a relatively straightforward windows GUI, not a CLI morass, and I'll test + post test screenshots tonight if you want to wait.

Of course, beggars can't be choosers, so I'll take what I can get, be it HWinfo or a full SPD dump of a Win7 RTO :-)

I ask for the SPD because it is conceivable that R&S has been naughty and configured their BIOS to forbid any RAM modules that don't have a magic string in the SPD EEPROM. If that's the case, and this isn't merely a question of RAM geometry, the HWinfo dump alone wouldn't resolve the problem, whereas the SPD dump contains everything.

[jjoonathan]

So I've got Windows 7 and the 3.70.10 RTx software installed on the SSD I upgraded my XP-RTO1024 with. I was able to install both without any apparent hitch -- except that now if I try to run the RTO app it dies before getting to "scope mode" and complains about low RAM on account of only having 2GB instead of 8GB installed.

I downloaded and ran SPDTool for demonstration purposes. It didn't brick my scope, give it herpes, or even require laborious installation, it just spat out the SPD info of my 2 sad 1GB SODIMMs. Those are modules 0 and 2 in the screenshot -- module 7 is a 32MB framebuffer or something. If someone could do this on their 8GB scope and report the SPD of their 2x4GB SODIMMs I'd be much obliged

[MichalZ]

Hi, here you have,  from my RTO1004.
Please change .txt extension to .spd


Maybe some hacks for these baby ?

[jjoonathan]

Thanks! I've ordered those exact RAM modules 

As for hacks, well, one thing at a time  I haven't managed to get into the scope app yet, but when I do, I will be able to see if the options carried over. I did a clean Win7 install, so that would mean they were stored in eeprom. If they don't carry over, I suppose I'll just have to experiment with copying files until they do...

[jjoonathan]

Success! I've upgraded my RTO1024 from 2GB RAM Windows XP to 8GB RAM Windows 7 -- which also lets me upgrade the "firmware" from 1.x to 3.x. It's a big step up. Many of the annoyances I had on the old firmware are gone and there are a ton of new features. Mostly locked, but still :-)

Here is the procedure:
* Ordered, installed a new, empty SSD (easy)
    - I looked for SATA-II compatibility because the chipset doesn't support SATA-III
    - Settled on Kingston 240GB A400 SSD 2.5'' SATA 7MM 2.5-Inch SA400S37/240G
    - It was the first SSD I tried and it worked great
    - The sheet metal screw tray made swapping the HDD super easy
* Got at the RAM (hard)
    - You need keys to slide out the CPU board
    - Here is the lock mechanism and how I made keys for it: https://photos.app.goo.gl/M3uLj5EZG8Zk7qXB7
    - There is undoubtedly another way. Hook tools, something to firmly press in the tabs from the outside... be inventive and share your results!
    - Do you really need 8GB RAM? Yes. I tried without. It was painful and didn't work.
* Ordered, installed 2x4GB RAM SODIMMs (easy)
    - I tried two theoretically chipset-compatible DDR3-1066 SODIMMs that didn't POST (black screen, no boot).
    - The RTO is a picky beast.
    - SODIMMs known to work:
        * 2x HMT351S6CFR8C-PB
        * 2x 16KTF51264HZ-1G6K1
        * 2x M471B5273BH1-CF8
* Installed Win7 Pro onto the empty new SSD (easy)
    - I followed instructions on MS help pages to make a USB stick
    - USB stick is before HDD in the RTO's boot order so it Just Works
    - I created a user "jon" before realizing this was different from XP's "instrument" user, but it didn't matter
* Installed R&S Firmware (easy)
    - Got it here: https://www.rohde-schwarz.com/us/firmware/rto1000/
    - Version 3.70.1.0
    - Install ran without a hitch
* Ran the RTx Application (easy)
    - It flashed a few FPGAs, restarted
    - Then "Error, FEC PicoBlaze not ready, switching to Service Mode > CreateDeviceFootprint() failed"
    - Terminated the Rtx app, ran it again with administrator privileges
    - This time it flashed the PicoBlaze correctly and landed in scope mode
    - The instrument model and old options populated automatically from the smartcard
    - Subsequent invocations don't seem to require admin, but subtly broken features are always a possibility

[myhobo]

hi, my rto1024 upgrade is broken

[jjoonathan]

Good news: I think I know the problem. If so, it's an easy fix 

All of those error messages in your first picture look familiar and normal. They are just part of the EEPROM upgrade process. They should disappear if you open the RTx app again. Once a flash succeeds it does not have to try again. If it tries again and again, that is when you know it is not working. For me, the FEC PicoBlaze messages happened every time I restarted RTx.exe until I figured out that I had to Run As Administrator. Once I did, and after one last flash, the errors went away.

Your "GB Receiver" (gigabit receiver) error is also familiar. I disassembled and reassembled my RTO, taking pictures along the way. I forgot to replace the ADC <-> FPGA cards the first time and I got GB Receiver errors. I put the cards in their proper place and the errors went away. It looks like you get one GB Receiver error for channels 3/4 -- I suspect the rightmost card was poorly seated, dirty, or the two acquisition boards did not have perfect vertical alignment. Easy to fix, so hopefully that is the real problem!

Video of upgrade errors: https://photos.app.goo.gl/yvJuXwGP9jxZReja7

All teardown photos: https://photos.app.goo.gl/c66VTGmrUjgqRzRH9

I hear that google does not work well in China, so I also attached key photos to this post.

[myhobo]

Thank you, I exchange the ADC <> fpga small board, but it also error.

[myhobo]

Do you have service manual?

[nctnico]

Maybe try to clean the contacts of the interconnect boards with alcohol. Don't use contact spray (unless you want to put the oscilloscope in the trash next year)!

[jjoonathan]

To anyone thinking about upgrading but scared by this problem: you do not need to touch the acquisition boards to perform a WinXP -> Win7 upgrade. This is a separate problem. If you go ahead with an upgrade, resist the temptation to poke around at the acquisition boards. They have a mechanically weak connection that is easy to break and will land you in the same boat as myhobo if you do.

Sorry, I don't have a service manual.

However, we shouldn't need one. We know which circuit is responsible for the problem and we can guess how it works. There are 9 differential pairs coming out of each 8-bit ADC. This strongly suggests that the interface is parallel: one differential pair per bit plus one differential pair for data clock (generated by ADC). The problem could theoretically be anywhere in the ADC, but the ADC->FPGA bus is a known weak point. With the front and back screws removed, those three cards are the only mechanical connection keeping the top and bottom boards in vertical alignment. It would be very easy to accidentally break one of them, or, more likely, one of the sockets they plug into.  The fact that the problem is on channel 3/4 suggests it is the rear right card/socket that is problematic. If I understand correctly that you swapped the channel 1/2 card with the channel 3/4 card and the problem persisted, we should strongly suspect the problem is with one of the channel 3/4 sockets. Cleaning the sockets is a good idea but if one of the chn 3/4 sockets got crunched it won't be enough.

Strategy #1: test for connectivity with a multimeter.

Strategy #2: put a full-scale triangle or sawtooth waveform into channel 3/4 and look at the corrupted waveform. If bits (=diff pair channels) are pinned or scrambled, we should be able to tell which bits by looking at the resulting waveform. If the clock is bad, we will tend to see artifacts around bit transitions like 00111111 -> 01000000.

Strategy #3: cut the sheet metal at the back of the machine (it will need to be put back with foil) and use an EMC probe + spectrum analyzer to check each differential pair for activity.

[jjoonathan]

Hey everyone, service passwords for 3.70.1.0   

Enter them in File > Maintenance > Service > Password. If entered incorrectly, the message "Password incorrect" will display and the incorrect password will remain in the entry field. If entered correctly, no message will display but the password will vanish from the field and a new menu will be available according to the password entered.

Password: 894129
    File>Service
        VCO, OCXO DACs
        MSO alignment, thresholds, hysteresis, tests (HW missing)
        Power rail voltages, but checking "Get Values" doesn't actually get values
        AWG/Wavegen (HW missing)

Password: 16052008f
    File>Service 2
        ADC Temperatures, fan settings
        Droop DACs, Gain DACs, RDACs, CDACs, all the DACs you could want and more
        "Pro" calibration -- lots of detail, options (calibration options, not that kind of options, lol)
        Aux Out at frequencies other than  1GHz
        Interleave setup -- I'm not brave enough to poke it
        EEPROM flash


I'm most interested in the ADC temperatures. They aren't exposed to Windows as far as I can tell and I occasionally get overtemp errors on ADC2. Now I can keep an eye on it. It is indeed the hottest, though not by a gigantic margin. I wonder if there's a way to configure the fans to be a bit more aggressive?

[jjoonathan]

Hey, there are team photos in the secret service menu 

[myhobo]

this is my oscillioscope picture.

[jjoonathan]

Here is mine. The temperatures are a bit higher than I'd like. I tried putting new thermal paste on the ADCs. Even though I did find dried out crusty old paste it must have still been doing its job because replacing it with new arctic silver didn't do shit (didn't affect fan speed OR temperature). It was also a big pain in the ass and involved playing chicken with those connectors that gave myhobo so much trouble. Would not recommend.

I suspect the solution to the high temperatures will involve poking some of the fan-related EEPROM constants.

@myhobo is the problem fixed?

[myhobo]

my oscillioscope's 3 & 4 channel is not working. this is rtc1's log

[jjoonathan]

Wow, it looks like none of the traces are connecting properly. Have you beeped them with a multimeter? Do they have DC connectivity?

[myhobo]

Without power-on test, each is turned on, as if each pin is connected together.

[myhobo]

Interface should be conductive