Siglent SDG1000X Waveform Generators

[tautech]

Just don't call it "service manual", if the manual does not provide the required information for servicing.

Problem solved

It does.
That IS servicing today.
Board level repair is not part of field service.

As it has been for nearing 3 decades.

Those bitching about it are living in the past. 
They should go look at what is offered in a TDS200 SM. 

[Martin72]

Those bitching about it are living in the past.

I would like to be 20 again.
On the other hand, I'm glad I'm not anymore, age makes you more relaxed, you can look past some things with a smile and don't have to jump over every stick that's held out.
That's it from me.

[slavoy]

My SDG1032X upgraded to 1062X outputs signals that are 15° phase-shifted at 60MHz, even though they should theoretically be synchronized. There's a lock symbol, and I have also tried copy and follow modes. It's always the same. Identical coaxial cables and terminators, I also tried with probes and without changes. Am I missing something here?

[slavoy]

I upgraded the firmware (1062X mod persists, didnt know that). Now I got "only" 3° phase shift between channels. Strange.

[Mortymore]

I don't have any significant phase deviation between channels on my SDG1032X converted to SDG1062X.
I don't have great quality 50 Ohm feed-through terminators so I'm using T's with equal 50Ohm terminators, and my most recent pair of 50Ohm RG-58 cables

SW: 1.01.01.33R8
HW: 02-01-00-24-00

[SilverSurfer]

Hello,
I've just received my SDG 1032x 
The actual firmware is 1.01.01.33R3
I would like to acces to it via Telnet (root) but I would like to know if the process for more older firmware is still working ?

Thanks,

[44kgk1lkf6u]

It looks like the old way does not work anymore with the newest firmware.  But they found an easier way to upgrade. See:

https://www.eevblog.com/forum/testgear/siglent-sdg1000x-waveform-generators/msg4395403/

[44kgk1lkf6u]

I took a look at the encoder problem.  It is due to the firmware reading the encoder too infrequently.  I made a workaround.  I used an ATTINY25 microcontroller to act as a knob that turns just slowly enough.

[coromonadalix]

My modest contribution for a software control,  from model Labview examples  given at SIGLENT

SDG1xxx x upgraded to 60Mhz ....

Far from be complete, basics functions  is enough for me,  im running 2x gens with it


needs 32 bit NI Runtime engine

NB: Always push the smalll white arrow to send command to the gen(s)


pls   dont ask me for updates,  i'm a labview noob

[coromonadalix]

small upgrade to include output impedance 50ohms or High Z

some gui rework in names infos ....   not perfect  loll

[rf-loop]

small upgrade to include output impedance 50ohms or High Z

some gui rework in names infos ....   not perfect  loll

The SDG1000X has a fixed output impedance. Always 50 ohms.
SDG1000 do not have settable output impedance.
As you of course know.

In menu it have two Load values selectable directly. 50ohm and HiZ

But then (some users may not have noticed this) it has also variabl user set value for the load impedance (if the actual load is example 55 ohms, the user can tell it to the instrument, or if it's 666 ohms or 10 kohms, etc... The load impedance value can be set between 50 ohm - 100 kohms and HiZ.

Very useful when load is not 50 ohm or HiZ.
That way user can set more easy right output level values to actual load, also other loads than just 50ohm. (Vpp/Vrms/dBm) or HiZ (Vpp/Vrms).

[44kgk1lkf6u]

The SDG1000X has a fixed output impedance. Always 50 ohms.

But there is a relay on the circuit board to bypass the 49.9 Ω resistor at the output.  Does anyone know why?

[BillyO]

But there is a relay on the circuit board to bypass the 49.9 Ω resistor at the output.  Does anyone know why?

If you reverse engineered this can you provide a schematic?  Is it just a butt simple SPST relay that does this one thing or are there other associated functions?

[tautech]

But there is a relay on the circuit board to bypass the 49.9 Ω resistor at the output.  Does anyone know why?

If you reverse engineered this can you provide a schematic?  Is it just a butt simple SPST relay that does this one thing or are there other associated functions?

RTFM 

2.13.10 Overvoltage Protection
P109
https://int.siglent.com/u_file/download/24_06_07/SDG1000X_UserManual_UM0201X-E01G.pdf

[BillyO]

RTFM 

Nice find.  So it does more than just disconnect the resistor.  It disconnects the entire instrument.

[44kgk1lkf6u]

I had a closer look at the board.  I was wrong about there being a relay to short the resistor at the output.

[44kgk1lkf6u]

Here is the schematic I got by looking at the board.  The silkscreen on the board says "SDY8.007.108D" and "AWGCA_211100".  The user interface says the hardware version is "03-00-00-24-00".

Most voltages are derived from the voltage reference inside the DAC80501 DAC.  The exception is the square wave amplitude derived from the differential LVPECL comparator output.

The power supply for the ADC U44 is separate from that for the other chips.  It may be due to it serving as the reference voltage.  I did not look into how it is generated.

Modulation from the external input is done by having the multiplexed ADC sample the input.

I did not measure the values of the components of the filters switched by RLA7.  I do not have a vector network analyzer.  It is hard for me to measure the components in circuit.

The 2 channels have separate DAC reference voltages.  They are both generated by the DAC U62.

I do not know the part numbers of the opamps for buffering the DAC output.  They are SOIC-8 packaged dual opamps.  They are marked SN1272.  They have the Texas Instruments logo on them.  They may be TLC272.  I do not know why some have their outputs connected to a 4.7 μF capacitor.

The 3 BNC connectors on the back of the device are connected to the main board directly with a 50 mil pitch ribbon cable.

Only the channel 1 signal path is shown.

The resistor at the aux in/out connector R309 is likely there for protection against misuse.  It combined with the output impedance of the 74AHCT buffer is larger than 50 Ω.  The buffer is powered by 5 V.  The sync signal is 1.6 V into 50 Ω.  That means changing it to 0 Ω would make matching better.  My oscilloscope trigger input does not have a builtin termination.  So I shorted the resistor.  It did make the square wave have a flatter top when the oscilloscope end of the cable is not terminated.

The dedicated circuitry to generate square waves is an LVPECL output comparator early in the signal path.  This seems like a hard way.  The have to make sure the rest of the signal path can provide enough bandwidth and also figure out how to tame the output amplitude.

The resistor at the comparator output R210 is not a pullup.  It is there so that the digital input can sit at ground when the comparator output is at VEE.

[blackdog]

Hi,

Recently I modified my SDG 1032X and through this I can show how big the phase differences are between the two outputs of this generator.

This is how things were measured.

Test frequency = 60MHZ
Output level = 500mV TopTop
Cables within 1mm same length and same type, Telon with Huber + Suhner BNC connectors.
Scope is the SDS200x HD from Siglent, full bandwidth. :-)
Channels 1 and 3 are used from this scope.
Channels 1 and 3 are set to 50 Ohm input impedance.
32x averaging was used to see the phase difference well enough.

The test setup


.
This is the output on the scope for a Sinus signal of 60MHz and 500mV TT, only a small phase difference is visible.


.
The same, but now a square wave.


.
Here on the generator I compensated for the phase difference, this was +2 Degrees.


.
The differences are so small that it could be due to cable length, generator differences and scope differences.
And it will certainly be a mix of these three ingredients.

Best regards,
Bram

[blackdog]

Hi,

To get a more complete picture of some function generators, I also made measurements on the SDG 2042-X and the Rigol DG4162.
The measurement conditions are almost the same, only I chose to measure the phase difference between the channels now at 120MHZ for these function generators.

Cable optimization
There were more good cables here and I found two that were the same length within 1mm.
Then of course I still don't know, if the cables were made from the same production run....

Skew optimization
Now how can I make sure that some variables are reduced.
I did that by connecting the two cables coming from C1 and C3 of the scope to a Mini-Circuits ZFRSC-42 power Splitter.
I connected the input of the power splitter to one channel of the function generator.

Then from channel-3 I used the “Deskew” function to make the differences in the cable and scope inputs as small as possible.
Result of the optimization, 60 pSec correction was needed.
But beware, nothing is perfect. ;-)



.
Here is the splitter mounted on the Siglent SDG2042.


.
This is the power spitter with SMA/BNC reducers, both of the same brand to minimize errors.


.
If all goes well, I have now minimized the differences in my measurement system, if I overlook anything, please let me know!

SDG 2042X
The Siglent SDG 2042X, it gives about 10 Degrees of Phase difference uncorrected, according to the scope measurement.


.
And if I correct the phase in the SDG 2042X, it requires 7 Degrees of correction.
Probably the phase measurement in the scope is less accurate,
but it could also be in the generator, which is not easy to check now.


.
Rigol DG4162
First again the phase difference if the generator is not corrected.


.
And this is the phase corrected in the DG4162, this corresponds better to the measurement of the Scoop.
The difference here is only 0.6 degrees.


.
OWON DGE2070
The last generator I have with two outputs, the OWON DGE2070.
This is an inexpensive two channel function generator that I bought specifically because it can be powered from a power bank.
Some measurements I do have a lot of commonmode problems and I then get rid of them when using this generator.

But because it is so cheap, not everything is done very well.
Let's see what the phase difference is and the amplitude of both channels....

But first a picture of this Baby generator.   


.
And this is the phase error and amplitude error of the DGE2070 at the maximum frequency of 70MHz and 500mVpp, not so nice...


.
The correction needed to get the channels in phase at this 70MHz frequency is 104 Degrees!


.

If you know the generator you are using a little bit, phase correction is quite simple to apply.

It gets a little trickier when the phase correction varies over the frequency range.
I don't have time to test all this, these kinds of measurements and the annotation in the pictures take a lot of time.

I hope this measurement session gives you some more insight into the channel differences of function generators with two outputs.

Only the cheap OWON function genrator also gave clear differences in the amplitude of the output signal.
So if this is important for your measurements, the OWON is not suitable for that.

With kind regards,
Bram