EEVblog #806 - Siglent SDG2000X Arb Function Generator First Look

[EEVblog]

Siglent have just fixed the bugs I found, and new firmware is going out to dealers.  Super quick response

[German_EE]

This company continues to impress me.

[TrioTest]

Dave,

to get both edges of the different generators aligned easily, (around 16:35 in the video) there's a function on some analog scopes called "Alternate Channel Triggering."  I think even the old HP1740 you rescued from me has it too.

If you use that function then the scope triggers off the channel it is sweeping so you do not need to try to phase lock the generators to compare the rising (or falling) edges.  You can even have 2 different frequencies and both waveforms will be rock solid.

Those old classics like the HP1740 (and some of us who used and remember them) still have a few useful functions left in them yet!! 

Let me know if you wish to try it and I'll get the SDG2122X back to you with the updated firmware.

[cnkz]

Siglent have just fixed the bugs I found, and new firmware is going out to dealers.  Super quick response

Where can we get this firmware? I didn't find it on the siglent web-page....

[crispy_tofu]

Siglent have just fixed the bugs I found, and new firmware is going out to dealers.  Super quick response

Where can we get this firmware? I didn't find it on the siglent web-page....

Probably will come out soon. 

[alank2]

I'm going to give the SDG2082X a spin.  Hopefully tequipment will get it to me quickly once they get it from Siglent.

The updated firmware has been posted here:

http://www.siglentamerica.com/gjjrj-xq.aspx?id=1472&tid=16

From the revision notes:

1. Supported adjustable noise bandwidth
2. Supported harmonic
3. Added some built-in arbs, so the number of arb types increased to 196
4. Fixed some bugs

[smarteebit]

Dave, thanks for the video.

I think it is a little unfair to compare the SDG2122X with the DG4162 on the 120MHz sine wave, because SDG2122X was outputting a 10Vpp amplitude but DG4162 was outputting a ~4.3Vpp amplitude. I believe if you set the SDG2122X to output 120MHz @ 4Vpp, the waveform will look much better.

Generally in an AWG the last stage circuit is a high-speed current feedback amplifier. It's hard to require the amplifier to output a sine wave with both high frequency and high voltage at the same time. What's addition, amplifiers generally are sensitive to the capacitive load at their output. Large capacitive load could cause unstability, so the cable connected to the output should be as short as possible.

[alank2]

The SDG2082X arrived tonight and it is a keeper for me.  In contrast, I tried the DS1074Z-S a few weeks ago and was very disappointed with just about everything with its signal gen.

My first impression is that it is compact and solid.  Feels really nice.

I checked its internal clock against my GPSDO and it is 0.160 ppm fast which isn't bad at all.  Also, the built in frequency counter is really great because it offers 8 digits and also has a PPM calculation feature which is super.  I had considered getting a dedicated counter like the BK 1856D, but this one has so many features that there is no need.  If they would add a min/max/average to it, it would be awesome.

Unlike the Rigol -S signal gen, the arbitrary waveforms worked really well - at least it would load files and display the right wave, and the controls were responsive and didn't require you to change other controls to become active.

I made up an arb wave with 10 sine waves in it and while my unit is an 80 MHz unit, I could easily take it up to 120 MHz.  It will not however go past 120 Mhz well at all, even 130 MHz:

Output set to 50ohm, 1 Vpp, BNC cable direct to DS4014 CH1, 50ohm enabled

10 Mhz = 1.018 V
50 Mhz = 1.000 V
100 Mhz = 1.006 V
120 Mhz = 0.981 V
130 Mhz = 0.631 V
140 Mhz = 0.200 V

I only did a 256 points per sine wave (2560 points total), but it looks really great at 120 MHz.

It is my first signal gen, but so far I love this thing.

[rs20]

I made up an arb wave with 10 sine waves in it and while my unit is an 80 MHz unit, I could easily take it up to 120 MHz.  It will not however go past 120 Mhz well at all, even 130 MHz:

Output set to 50ohm, 1 Vpp, BNC cable direct to DS4014 CH1, 50ohm enabled

10 Mhz = 1.018 V
50 Mhz = 1.000 V
100 Mhz = 1.006 V
120 Mhz = 0.981 V
130 Mhz = 0.631 V
140 Mhz = 0.200 V

Looks like your oscilloscope and siggen are rolling off together there; after all you'd expect -3dB rolloff from your 100MHz scope at 100 MHz...

[alank2]

A closer look at what I thougth was a bug may not be a bug after all, but the way the DDS chip works.  If you setup a small arb wave with the minimum number of points, and you create one like this:

-10 -10 -10 +10 -10 -10 -10 -10

You will see a triangle type pulse in the middle of it when it transitions from the 3rd sample to the 4th, and then transitions from the 4th to the 5th.  This shows that the DDS does an interpolation between samples which is pretty neat.  But that interpolation does NOT work on the first and last samples.

If you move that +10 to the beginning:

+10 -10 -10 -10 -10 -10 -10 -10

No more triangle with a rising ramp and falling ramp, you get a straight up line and then a falling ramp.  It is not interpolating between the last and first samples.

If you move that +10 to the end:

-10 -10 -10 -10 -10 -10 -10 +10

Now there is a rising ramp and a straight line for the falling.  Again, no interpolating.

Perhaps this is the way the DDS IC is designed?  I haven't looked at the datasheet, but I can't say it is a bug as it works on the front and back of the waveform.

With a decent number of points you wouldn't notice it.  Also, adding the first sample to the end will cause it to transition back to the first sample as well.

[alank2]

In case anyone is interested, the .bin format can be loaded directly with no conversion issues and is simply an array 16 bit signed integers. 0x7FFF (32767) is the highest value and 0x8000 (-32768) is the lowest value.  It can also go as low as 2 samples, not 8 as the spec says.

With the 8M of sample memory, you could load 8*1048576/10=838,860 UART characters and play them back at some baud rate you want.

[tautech]

Link to billfernandez's Siglent SDG2042X thread:
https://www.eevblog.com/forum/testgear/the-siglent-sdg2042x-thread/

[Electro Fan]

The SDG2082X arrived tonight and it is a keeper for me.  In contrast, I tried the DS1074Z-S a few weeks ago and was very disappointed with just about everything with its signal gen.

My first impression is that it is compact and solid.  Feels really nice.

I checked its internal clock against my GPSDO and it is 0.160 ppm fast which isn't bad at all.  Also, the built in frequency counter is really great because it offers 8 digits and also has a PPM calculation feature which is super.  I had considered getting a dedicated counter like the BK 1856D, but this one has so many features that there is no need.  If they would add a min/max/average to it, it would be awesome.

Unlike the Rigol -S signal gen, the arbitrary waveforms worked really well - at least it would load files and display the right wave, and the controls were responsive and didn't require you to change other controls to become active.

I made up an arb wave with 10 sine waves in it and while my unit is an 80 MHz unit, I could easily take it up to 120 MHz.  It will not however go past 120 Mhz well at all, even 130 MHz:

Output set to 50ohm, 1 Vpp, BNC cable direct to DS4014 CH1, 50ohm enabled

10 Mhz = 1.018 V
50 Mhz = 1.000 V
100 Mhz = 1.006 V
120 Mhz = 0.981 V
130 Mhz = 0.631 V
140 Mhz = 0.200 V

I only did a 256 points per sine wave (2560 points total), but it looks really great at 120 MHz.

It is my first signal gen, but so far I love this thing.

Congrats on the new SigGen.  Did you make the arb through the front panel or via a computer?

[alank2]

Congrats on the new SigGen.  Did you make the arb through the front panel or via a computer?

Front panel via USB stick.

[FaultlineAudio]

I order my Siglent SDG2122X this morning.  According to Siglent the newest firmware update adds the following features:
Harmonics Function, 140 or so more ARB waveforms for a total of 193 or so and an adjustable bandwidth to the Noise function.

This firmware update was released on 10/28/15

Siglent should thanks Dave for his review, it was instrumental in my decision to buy the Siglent over the Rigol.

[bson]

Has anyone used this with an external 10MHz reference yet?

[digitaladdictions]

Picked up a SDG2042X in April of 2017. "Upgraded" it to the 120MHz version. I know they released new firmware multiple times by now. Just wanted to point out specifically at 120MHz my sine wave looks better than what Dave was getting in the video on his original release firmware. Maybe not perfect but much better. However, I am only getting 7.24Vpp when set to 10Vpp. 

[tautech]

Picked up a SDG2042X in April of 2017. "Upgraded" it to the 120MHz version. I know they released new firmware multiple times by now. Just wanted to point out specifically at 120MHz my sine wave looks better than what Dave was getting in the video on his original release firmware. Maybe not perfect but much better. However, I am only getting 7.24Vpp when set to 10Vpp.

Can you just check you haven't got the amplitude value set for Vrms ?

[rf-loop]

Picked up a SDG2042X in April of 2017. "Upgraded" it to the 120MHz version. I know they released new firmware multiple times by now. Just wanted to point out specifically at 120MHz my sine wave looks better than what Dave was getting in the video on his original release firmware. Maybe not perfect but much better. However, I am only getting 7.24Vpp when set to 10Vpp.

How you know what is  level in SDG output? Right level is or (mostly) is not on your scope screen. With this kind of frequencies there start grow many errors. Do you have trusted leveled generator for make some compare. Also accurate RF power meter connected directly to SDG input is good way (because we know there is not so much harmonics power).

[nctnico]

I'd be surprised if a relatively low cost generator can reach 10Vpp at 120MHz into 50 OHm. Better check the specs first to see if there aren't any limits on the maximum output level versus frequency. Another problem might be the generator's 50 Ohm series resistor and the scope's input capacitance forming a complex voltage divider if the 1M Ohm input setting is used. 7.2V would indicate an input capacitance of around 14pf.

[digitaladdictions]

I really did not think much of the reduced voltage at the high frequency since Dave was getting similar effects and stated that is just par for the course essentially. Basically, I just wanted to report in that the latest firmware seems to be giving a better sine wave than what he got in the video although the peak to peak voltage has dropped.

I do not have a power meter (it's on the wish list). I was going off what is displayed on the generator and the measurement provided by the scope. I just repeated the measurement right now and am getting 7.36Vpp.  The current screen shot shows the measurement.

The waveform generator has a HiZ and a 50ohm option. I had it set to high impedance but I tried changing it to 50ohm. The amplitude automatically drops to 5Vpp when set to 50ohm but at HiZ it allows 10Vpp. What is odd though is I saw no noticeable difference on the scope either in the reported measurement or the displayed waveform. Despite the generator changing the amplitude from 10Vpp to 5Vpp the waveform stayed the same.

I can change it to 1Vpp and see the scope change. If I put it back to 10Vpp it gives me a warning that it can only do 5Vpp but actually gives me the same 7.36Vpp.  If I go back to HiZ I can change it back to 10Vpp but there is no difference in the output.

[digitaladdictions]

OK I was not changing the scope to match the generator which explains the lack of difference when switching between 50ohm and HiZ on the generator.

Generator set to 5Vpp (max setting) and 50ohm with the scope also set to 50ohm I get 4.2Vpp reading from the scope.
Generator set to 10Vpp (max setting) HiZ with the scope set to 1Mohm I get 7.[2-4]Vpp.

[nctnico]

The high-Z and 50 Ohm setting on the generator only affects the amplitude the generator display. The generato's output resistance stays the same: 50 Ohm. In 50 Ohm mode the generator assumes the load is 50 Ohm so the amplitude across the load is half of what the generator's output amplifier is delivering.