Siglent SDG6000X series 200-500 MHz AWG's

[kladit]

Thomas and other

the brand new sdg6022x is dead.
It does not boot anymore, the Siglent Logo comes up, then a relay clicks, the the Siglent Logo comes up a second time
and then nothing more happens.

Any help is very welcome.
--
Klaus

[TurboTom]

Just checked the bootup sequence on mine: Initially, the splash screen (logo) gets displayed for approx. 7 seconds, then the screen goes blank for a glimpse of an eye, accompanied by a clicking noise (sounds more like it comes from the buzzer than from a relay), and the splash screen appears again. After some more time several relays are switched and finally, after 22 seconds, the instrument becomes operational.

I haven't heard of a broken SDG6000X from anyone else. Since it is new, I'ld just have the seller replace it under warranty. Maybe @tautech can give you some additional advice?

[kladit]

Thomas,

I tried several key combinations in the hope of a general-reset combination, but found none.

When the instument was working I discoverd a C: somewhere at the screen. I hope Siglent
has not choosen Windows as OS, that would be a strong argument to not buying it.

The noise is ok, the analyzers dynamic range is simple not good enough to measure.
Measured with my shortwave receivers noise it is about -100dBc  down at 1 kHz,  these
values are not exact, but in the order of others measurements.

So the SDG6022X is very good as a two-tone-generator for IP3 measurements.

[Performa01]

Thomas and other

the brand new sdg6022x is dead.
It does not boot anymore, the Siglent Logo comes up, then a relay clicks, the the Siglent Logo comes up a second time
and then nothing more happens.

Any help is very welcome.
--
Klaus

This is certainly bad news, especially now at the start of the Christmas/New Year holidays, were I don't think you could get a replacement device from your distributor very quickly.

Your issue sounds like a flash file system corruption - unfortunately not totally uncommon with modern instruments.

There are recovery procedures designed into the bootloader of the Siglent SDG, so there is hope for a quick fix without the need to send the unit back.

Someone from Siglent will try to contact you to guide you through the steps of that recovery procedure.

[tautech]

Thomas and other

the brand new sdg6022x is dead.
It does not boot anymore, the Siglent Logo comes up, then a relay clicks, the the Siglent Logo comes up a second time
and then nothing more happens.

Any help is very welcome.
--
Klaus

Apparently you are being looked after by someone from the factory.
Please report the outcome and/or PM me if you need further assistance.

[kladit]

tautech

I created a fat32 file system with gparted, made it bootable and copied the files sent by Siglent onto it.
After inserting and power up of the SDG6022X I can see that there are some accesses to the stick.
But then the relay clicked and after that same as before, dead and no accesses to the stick (LED off).

I am not sure whow exactly the fat32 has to look. All one partition, a boot and a data permission or whatever.

I do not have windows and I am not famliar with windows.

--
Klaus

[Performa01]

When the instument was working I discoverd a C: somewhere at the screen. I hope Siglent
has not choosen Windows as OS, that would be a strong argument to not buying it.

With that attitude you would miss out a lot of the top-notch gear from Agilent/Keysight, LeCroy and R&S…

Rest assured though that no Chinese brand uses Windows or any other software that requires license fees. All Siglent instruments are based on either Linux or some proprietary framework (for the older/simpler ones).

Measured with my shortwave receivers noise it is about -100dBc  down at 1 kHz,  these
values are not exact, but in the order of others measurements.

This proves that the phase noise and/or dynamic range of the shortwave receiver is the limit here – and that’s a good thing. After all, the signal generator should be good enough to characterize the receiver, not vice versa.

My previous phase noise test in this thread was at 100MHz and phase noise at 1kHz distance from carrier was already better than 110dBc. I did another test at 7MHz and at frequencies that low, the SDG6000X has better than 122dBc phase noise at 1kHz carrier distance.

Even at 10Hz carrier distance we get some 109dBc already.

Phase noise plot attached. In fact the SDG6000X phase noise will be even better than that, at least at carrier distances beyond 1kHz, because what we actually see here are just the limits of my analyzer.

[kladit]

Perfoma01,
thanks for your response and measurements.

"Rest assured though that no Chinese brand uses Windows or any other software that requires license fees. All Siglent instruments are based on either Linux or some proprietary framework (for the older/simpler ones)."

That is good news in general.

Anyhow, with linux used, there would be no need to create a fat32 formatted usb-stick with a file named BOOT.bin  as rescue device.

An the fact that it did not work smells more similiar to windows.
--
Klaus

[Performa01]

Unfortunately I'm not familiar with the recovery process, but still pretty sure you need no fancy partitioning scheme on the USB stick for that.

FAT32 is kind of a de-facto standard by now, just because it's simple to implement. Every OS can deal with it, so not really a hint to anything. Whenever an U-disk needs to be formatted (FAT32), Hackers in this forum regularly suggest using Linux to do that.

For the recovery procedure, the operating system is not even started yet, it's all in the bootloader. On a PC, this would be the BIOS (or UEFI on more modern devices), both have nothing to do with the operating system that is subsequently loaded. So at least for me it's no surprise that the boot files might adhere to some traditional DOS convention (that dates back to the old CP/M systems of the late seventies of the past century).

[tautech]

tautech

I created a fat32 file system with gparted, made it bootable and copied the files sent by Siglent onto it.
After inserting and power up of the SDG6022X I can see that there are some accesses to the stick.
But then the relay clicked and after that same as before, dead and no accesses to the stick (LED off).

I am not sure whow exactly the fat32 has to look. All one partition, a boot and a data permission or whatever.

I do not have windows and I am not famliar with windows.

--
Klaus

In the very few units I have personally recovered I always started with a blank USB stick formatted in FAT32.
No partitions or special permissions are needed.

The latest instructions recommend use of recommend using a 8GB FAT32 Kingston USB flash drive and not 16GB. Larger or smaller drives may work but I do not know for SDG6kX recovery.
I mostly use a cheap 8GB stick.
While a USB stick with a activity light is helpful to watch file installation progress it is not necessary as if the USB stick size, file structure and recovery files are correct the recovery progress is just a minute or so.

The SGD6000X recovery package is larger than others I have seen with 10 Read Only files totalling 43MB plus a small .pdf instructions file.

Depending on when the recovery package was made it may not leave the instrument with the latest firmware and after recovery the latest firmware will need to be installed.
I have helped several owners worldwide and in nearly all cases successful recovery has been simple and straightforward.

Please report progress and/or further problems.

[kladit]

tautech,

no success. I created partition and  fat32 filesystem onto an 8GB USB stick, set lba and bootable flags copied all files sent
to me onto it, except the instructions PDF.

Same behavior as before, accesses visible at power on, no more visible activity past the relay click.

The mounted stick looks like this at linux

dr--r--r--  2 root root      4096 Jan  1  1970 .
drwxr-xr-x 18 bin  users     4096 Dec 26 12:06 ..
-r--r--r--  1 root root   4618780 Aug 20 18:42 BOOT.bin
-r--r--r--  1 root root   2097152 Aug 20 18:42 datafs.img
-r--r--r--  1 root root     15724 Aug 20 18:42 devicetree.dtb
-r--r--r--  1 root root   1966080 Aug 20 18:42 firmdata1.img
-r--r--r--  1 root root   1966080 Aug 20 18:42 log.img
-r--r--r--  1 root root    391736 Aug 20 18:42 logo.bmp
-r--r--r--  1 root root  10465280 Aug 20 18:42 rootfs.cramfs
-r--r--r--  1 root root      3433 Aug 20 18:42 sdg6000x_udiskEnv.txt
-r--r--r--  1 root root  20185088 Aug 20 18:42 siglent.img
-r--r--r--  1 root root   3120904 Aug 20 18:42 uImage

I also tried all files and directories 0755, same behavior as before.

--
Klaus

[tautech]

@ Klaus
The file list and sizes look correct with the list I have here. 
I dunno about these:
dr--r--r--  2 root root      4096 Jan  1  1970 .
drwxr-xr-x 18 bin  users     4096 Dec 26 12:06 ..
Others with Linux experience might help. 

Comments.
NO bootable flags required. The instrument checks for USB files by default when a USB stick is installed at boot.
NO special settings are required as they may be interfering with the instrument OS to access the USB recovery files.
Also, try another USB stick if you have one.

Edit
FYI
Attached screenshot of SDG6000X USB recovery files structure on USB stick in a Windows 10 OS.

[Performa01]

We've seen frequent reports about USB-sticks not beeing recognized by the bootloader, even when they work perfectly fine after the instrument has booted and the main application is running.

I personally have never had any issues, simply because I stick with older USB sticks (no pun intended). The only time I needed U-disk access during boot time (for an OS upgrade, not recovery) I used an old NN 512MB stick and it worked like a charm first try.

You could try and ask member tv84 for advice. He's our reverse engineering star here in the forum and always been very helpful:
https://www.eevblog.com/forum/profile/?u=92796

[tv84]

This theme is somewhat recurring... Almost as "what is the best scope to buy just before New Year's Eve"...

Here I tried to approach it systematically but...

I think there are a number of factors than can contribute to the problems, all related with type of USB disks.

People want the equipments to boot faster so programmers do some simplifications regarding USB disk detection...

My best feeling regarding these latest problems (in Siglent and Rigol) is that: these new USB disks/controllers (due to their size) take more time to bootup (once power is applied do them) and sometimes they end up being ready only after the USB boot detection function has already ran. So, that might explain why the equipment doesnt see he disk. After the equipment is in normal operation, and since the disk is detected in the meantime, there will be no problems with disk access.

With smaller disks (or some faster controllers), the bootup of the disk might be faster and occurs before the detection phase has passed.

Have in mind also that Siglent (and mostly Rigol) sometimes do some raw sector access while booting to detect some special vendor disks... This may also contribute to the timing issues...

The format of the partitions also might contribute to the issue. Try use FAT32 (FAT16 and 12 should also work) and, if you are suspicious, do the formatting in Linux. The number of reserved sectors, etc is different between Linux and Windows (for the same disk format!). So, have in mind that the software that is running in the equipment is Linux.

I'm assuming that there are no errors in the programming...

[tautech]

...............
I'm assuming that there are no errors in the programming...

There isn't !

Just checked my SDG6022X with IQ hacked option and with older FW V6.01.01.29R10.
Totally unrequired Boot failure recovery with USB files as listed above on a cheap Strontium 8GB stick.

Result:
User settings of 'Last' were lost so the unit booted to factory default settings and with latest V6.01.01.35R5B1 FW and the unofficial IQ option remains active. 

Saved me needing to install the latest FW....... 

[tv84]

...............
I'm assuming that there are no errors in the programming...

There isn't !

I wasn't referring to your "recovery disk" process scripts/instructions.

The recovery disk is simply a decompressed FW package with a special setting that signals it to be processed by the bootloader and flashed without the need of anything else. I haven't looked at it but it seems to me that it can be easily done based on any FW package, as long as you've got an .ADS unpacker.

Of course, the bootloader disk detection could be even more stricter regarding boot times...

Losing the settings is a good sign, showing that the factory reset was done. Not loosing the licenses is a better sign, well done by Siglent, since it's assumed that the options wouldn't be creating a boot problem. If they did we could always unlicense them manually with commands.

[Performa01]

I would like to demonstrate how even the internal digital channel combine function can be quite useful for two tone tests as long as we're not dealing with (very) high level components.

This is certainly true for low frequencies like the 7040kHz example. At this frequency, it's no problem to use output levels of +13dBm, which is a quite common maximum level for RF signal generators. In this case, third order intermodulation products are at 68dBc and worst case third order intercept point is +47dBm.

See attached screenshot SDG6000X_IMD_7040kHz_d20kHz_13dBm_int_RBW10Hz and the corresponding peak table:

Frequency	Amplitude	Delta Frequency	Delta Amplitude
7.049999 MHz	13.03 dBm	0.000000 Hz	0.00 dB
7.029998 MHz	13.02 dBm	-20.001200 kHz	-0.01 dB
7.069996 MHz	-55.22 dBm	19.997491 kHz	-68.25 dB
7.009996 MHz	-55.35 dBm	-40.002399 kHz	-68.38 dB

Let's check the situation at a lower output level of -5dBm, which will be used for a practical test later. Third order intermodulation products are now even lower at 78dBc and worst case third order intercept point is +34dBm.

See attached screenshot SDG6000X_IMD_7040kHz_d20kHz_-5dBm_int_RBW10Hz and the corresponding peak table:

Frequency	Amplitude	Delta Frequency	Delta Amplitude
7.049999 MHz	-4.97 dBm	0.000000 Hz	0.00 dB
7.029998 MHz	-4.98 dBm	-20.001200 kHz	-0.01 dB
7.069998 MHz	-82.88 dBm	19.999345 kHz	-77.91 dB
7.009998 MHz	-83.64 dBm	-40.000545 kHz	-78.67 dB

Finally the most critical situation at the upper end of the SDG6052X freqeuncy range, 480MHz have been used for this test. Maximum output level with internal combiner is 6dB lower than without, so we get -6dBm, which happens to suit my praxis test quite fine. Third order intermodulation products are now rather high at 37dBc and worst case third order intercept point is +12.3dBm.

See attached screenshot SDG6000X_IMD_480MHz_d20kHz_-6dBm_int_RBW10Hz and the corresponding peak table:

Frequency	Amplitude	Delta Frequency	Delta Amplitude
480.009893 MHz	-6.03 dBm	0.000000 Hz	0.00 dB
479.989892 MHz	-6.03 dBm	-20.001200 kHz	-0.01 dB
479.969893 MHz	-42.80 dBm	-40.000545 kHz	-36.77 dB
480.029892 MHz	-44.33 dBm	19.999345 kHz	-38.30 dB

Now for the practical test with a Marconi Instruments TF2175 linear amplifier. Here are the relevant specifications:
Bandwidth: 2 ~ 500 MHz +/-1 dB
Maximum linear power: >300 mW (24.8 dBm)
Input for 300 mW output: 0.6 mW (-2.2 dBm)
Gain: 27 dB +/-1 dB
Harmonic Distortion: <30 dBc at 200 mW
3rd Order Intermodulation Intercept Point: +35 dBm typical
Noise Figure: <9.5 dB (8 dB typical)
Dynamic Range: >80 dB

We are interested in the 3rd order intermodulation intercept point. As long as this is lower than the TOI of our two-tone signal source at the required output level, we can get meaningful test results when measuring the TOI of the amplifier.

There's one trap: the specified TOI is for the amplifier output, but it's the input TOI that matter for our test. For the Marconi TF2175, this is Output TOI - Gain = 35dBm - 27dB = +8dBm. This means that the SDG6000X with internal channel combine function should be perfectly usable for measuring the TOI of the TF2175 amplifier over its full frequency range.

First test at 7040kHz.

Generator level is -5dBm, TF2175 output level is +21.6dBm. Third order intermodulation products are at 26.5dBc and worst case third order intercept point is +34.86dBm, which matches the specification quite nicely.

See attached screenshot SDG6000X_IMD_7040kHz_d20kHz_-5dBm_int_RBW10Hz_TF2175 and the corresponding peak table:

Frequency	Amplitude	Delta Frequency	Delta Amplitude
7.049999 MHz	21.59 dBm	0.000000 Hz	0.00 dB
7.029998 MHz	21.58 dBm	-20.001200 kHz	kHz,-0.01 dB
7.069998 MHz	-4.98 dBm	19.999345 kHz	-26.56 dB
7.009998 MHz	-5.07 dBm	-40.000545 kHz	-26.65 dB

Finally the test at 480MHz.

Generator level is -6dBm, TF2175 output level is +20.75dBm. Third order intermodulation products are at 28.5dBc and worst case third order intercept point is +34.99dBm, which once again almost perfectly matches the specification.

See attached screenshot SDG6000X_IMD_480MHz_d20kHz_-6dBm_int_RBW10Hz_TF2175 and the corresponding peak table:

Frequency	Amplitude	Delta Frequency	Delta Amplitude
480.009891 MHz	20.75 dBm	0.000000 Hz	0.00 dB
479.989892 MHz	20.73 dBm	-19.999345 kHz	-0.02 dB
479.969893 MHz	-7.79 dBm	-39.998690 kHz	-28.54 dB
480.029892 MHz	-8.64 dBm	20.001200 kHz	-29.39 dB

[kladit]

Guys,

I tried with two different USB-sticks, none worked, every time same behavior.

So I sent it back to Batronix yesterday.

Very friendly service there, within a minute I've got a return label.

! And they were manned even during Christmas holidays !

Rare in Germany.

Hope they will get it fixed fast. And I am very very curious if they will be able to detect the source of this bug.

Meanwhile I ordered three different 8 GB USB-sticks in the hope next time ( and I am shure
there will be a next time) I can fix the boot failure myself.

Special thanks to Martin Wu from Siglent  for his fast response and help.

Happy New Year to all of you.

--
Klaus

[rf-loop]

no success. I created partition and  fat32 filesystem onto an 8GB USB stick,

Guys,

I tried with two different USB-sticks, none worked, every time same behavior.


Klaus

And they was in native dos FAT32 formatted and in ONE partition and every needed file in root? It have been  long time I have made any recovery for Siglent any machine (so this is perhaps obsolete but...)   but this bit older time it was told that USB stick MUST be in single native partition, real dos FAT32 (not any exfat etc) (even when this machine use Linux (when normal boot))  and if Siglent tell  size of USB recovery media it also need take seriously. And all know also that original FAT32 8Gb default cluster size is 4k and nothing else.
https://support.microsoft.com/sq-al/help/140365/default-cluster-size-for-ntfs-fat-and-exfat

[kladit]

The SDG6022X is back.

Batronix's service is fast and friendly, they simply sent me a new one to shorten my waiting time.
Thank you very much.

Performa01
I can confirm your measurements, see the attached results of my measurement.


Both channels output set to +13dBm, internal combined to the output of CH2.

Analyzer SSA3021X, 20 dB external attenuator used.

IP3 = 13 dBm - (-54dBm/2) = +40dBm

That is not bad I think.

Mistake, mistake ..

IP3 = 13 dBm + (13dBm -  -54 dBm)/2  = 46.5 dBm

That is still not bad, may be even 6.5 dBm better. :-)

Others had to build up more complicated setups, quartz-oscillators, low-pass-filters, amplifiers, combiners,
to get the same results.

--
Again a Happy New Year to all reading this.

Klaus

[TurboTom]

Happy New Year and I'm glad that you received the replacement for your defective AWG.

I agree that in the lower frequency range, TOI measurements are possible with the internal waveform combination, but still better results are possible when using an external combiner. It's more a question of "how good is good enough" for the individual user  .

Have you enabled the measurement functions in your SA? If so, you could use the automatic TOI measurement which makes the whole process a little more comfortable. You've only got to be aware that Siglent confused upper and lower signal / IM band... But this shouldn't make a difference for the measurement itself.

Cheers,
Thomas

[harrywgaul]

I discovered a potential anomaly with the SDG6022X that shows up on the SDG6022X but not on the SDG2122X.  I control these generators with LabView.  If I set up an arbitrary waveform on Channel 1 with a repetition frequency of 1000 Hz and a CW signal at 120 MHz on Channel 2, then I need to set the Phase Mode to Independent.  Otherwise, the CW signal is not at 120MHz but is actually closer to 122 MHz.  This is true for other CW frequencies, but some frequencies are worse than others.

Most of my LabView programs use the SDG2122X and I haven't seen this kind of issue before.  It seems like the SDG2122X simply ignores the Phase-Locked setting if the two channels are at different frequencies, whereas the SDG6022X tries to phase lock the two channels by altering the frequencies.  Note that the default setting for mode upon power up is Phase-Locked.

I am wondering if anyone else has found issues by leaving the SDG in Phase-Locked mode even when they don't need the two channels locked in phase?

[boeserbaer]

HI, I just want to second how good Batronix has been.  I moved to Germany 5 years ago from the US, and Batronix has supplied me with several instrument (the latest was a R&S RTA4004) and they have always been prompt and helpful.  They have also been willing to communicate in English, since my technical German is a bit poor.

--mike

The SDG6022X is back.

Batronix's service is fast and friendly, they simply sent me a new one to shorten my waiting time.
Thank you very much.

Performa01
I can confirm your measurements, see the attached results of my measurement.


Both channels output set to +13dBm, internal combined to the output of CH2.

Analyzer SSA3021X, 20 dB external attenuator used.

IP3 = 13 dBm - (-54dBm/2) = +40dBm

That is not bad I think.

Mistake, mistake ..

IP3 = 13 dBm + (13dBm -  -54 dBm)/2  = 46.5 dBm

That is still not bad, may be even 6.5 dBm better. :-)

Others had to build up more complicated setups, quartz-oscillators, low-pass-filters, amplifiers, combiners,
to get the same results.

--
Again a Happy New Year to all reading this.

Klaus

[DL4RAJ]

Hi all,

I just want to inform all owners that the SDG6000X has a bug,at least in the most recent firmware version.
It is unable store the level settings for noise,neither in its built-in memory nor in an usb stick.
For verification store any noise setting with an arbitrary level within the possible range.
Then recall it and change the level to another value without storing this changed new value.
Recall the stored setting again and you will see that the level has changed to the new value
despite it had not been stored.
If you have created and stored e.g. three different noise settings all three will be changed by the above procedure.
Even if you go to e.g. sine wave mode,change the level there and recall the noise setting(s)
you will notice a certain change of noise level in all stored noise settings.
Batronix was able to reproduce this bug and they will report it to Siglent.
Perhaps it is helpful if as many as possible owners report this bug to Siglent to speed up their debugging
so they will come up with a new firmware as soon as possible.

Greeetings
Clemens

[MikeP]

 A phenomenon has been discovered - voltage changes in one channel affect the second channel. Especially noticeable at DC voltage from +/- 5-10 Volts. Accordingly, the shift of the second channel will occur in the range +/- 1 mV.
 Please confirm my find. I would be grateful if someone gives an explanation.
 Thanks.