New Rigol 16-bit function generators DG800/900 series

[simone.pignatti]

And id like to see a jitter test.

Set the generator to a weird non round frequency square wave such as 10.123456 MHz. Trigger on one of the edges then move the time on the scope to the next rising and next falling edge.

do you mean something like this?

[TurboTom]

Provide a test unit and it will undergo the same "punishment" on my MDA as the SDG6000X... 

[commongrounder]

I’ve been looking into these generators and noticed a difference between the 800 and 900 series other than the maximum frequency, sampling rate, and memory depth (We’ll leave the color of the case and screen out of this). On the 800 series, 16-bit arb waveform files are truncated to 14-bit, whereas the 900 series reproduces the full 16 bits. That might matter to those looking for the best signal fidelity from their created (and in-built?) arb waveforms.  Of course I wonder if this is a hardware, or software, limitation.

[MT]

Case design is hilarious , i cant buy this even if i needed , want and i had the money!
Which one of the Transformers is it?

[Wolfgang]

They now need to design a special coffee mug that fits on top of this. 

[bd139]

Yes the case is pretty crap. Not quite as rubbish as my ADSL router though. I have yet to find a worse one than that. https://en.avm.de/products/fritzbox/fritzbox-7560/ very MIB.  Rigol is probably misinterpreted Ridley Scott science fiction here.

[bitseeker]

On the 800 series, 16-bit arb waveform files are truncated to 14-bit, whereas the 900 series reproduces the full 16 bits.

Wow, really?

[Wolfgang]

Your router is the most aerodynamic model on the market. I boasts a Cw value of just below 0.2, and has been carfully optimized to minimize wind noises. 

[commongrounder]

On the 800 series, 16-bit arb waveform files are truncated to 14-bit, whereas the 900 series reproduces the full 16 bits.

Wow, really?

Check it out, from DG800 manual:

For the arbitrary waveform file, the voltage values corresponding to each waveform point are stored in binary data format. The voltage value at each point takes up 2 bytes (16 bits). DG800 only uses 14 bits, and the higher 14 bits are used to represent voltage and the lower 2 bits are not used. Therefore, its format of the binary data is from 0x0000 to 0xFFFC.

From DG900 manual:

For the arbitrary waveform file, the voltage values corresponding to each waveform point are stored in binary data format. The voltage value at each point takes up 2 bytes (16 bits). Therefore, its format of the binary data is from 0x0000 to 0xFFFF.

 

[bitseeker]

Nice. The title should've been, "New Rigol up to 16-bit function generators."

[Mr. Scram]

Bonus points for the mechanical power switch. Siglent, take note!

[tautech]

Bonus points for the mechanical power switch. Siglent, take note!

FYI
ALL Siglent AWG's have always had mechanical power switches !

Further:
In the Siglent AWG range, 1kX and old models, 800 and 1k series are 14 bit while SDG2kX and 6kX models are 16 bit.

[Mr. Scram]

FYI
ALL Siglent AWG's have always had mechanical power switches !

Further:
In the Siglent AWG range, 1kX and old models, 800 and 1k series are 14 bit while SDG2kX and 6kX models are 16 bit.

Who cares when the other devices come with soft touch buttons? You don't get brownie points for not messing up across the board.

[Berni]

But the trend of soft power buttons in test equipment is annoying.

There is little to no good reason why such gear would need to have a soft power button. They don't have remote controls to turn them on and they are not working on so much data that the OS could corrupt itself by forced power downs. RAM is cheap now days so simply run your OS from a ramdisk image while keeping only large static files in flash along with a settings file that gets written to in a safe way. Test equipment spends most of its life in the off state. You don't even use your scope for 10 hours a day on average even if you really use your scope a lot. Let alone use some more specialized bit of gear that you perhaps only power on a few times per year.

What makes things worse is that all of this gear uses switching supplies now. Okay they do waste less power than a oldschool transformer at low loads (When properly designed) so that lends them well to soft power off. But these things are also a lot more fragile. Mains power can have all sorts of surges happen and is common to get stuck by lightning. These surges can blow the primary side transistors in switching supplies pretty easily, but a big chunk of copper and iron doesn't really care as long as its less than the few kV of its insulation resistance. If you try pushing too much voltage into a transformer it simply saturates and draws a huge current while the output just rises a bit above normal. Of course it won't save you from a direct hit but i seen a lot of switching supplies blow up from lightning strikes while all transformered things in the same building kept working fine.

The only exception is instruments that have ovenised things in them. Often old gear had a small standby supply that is wired before the power switch to provide just a few watts of power to an ovenised crystal or something so that it drifts as little as possible. So i suppose if you have that then a soft power button is easy to do and saves you having an extra PSU.

[tautech]

But the trend of soft power buttons in test equipment is annoying.

There is little to no good reason why such gear would need to have a soft power button. They don't have remote controls to turn them on and they are not working on so much data that the OS could corrupt itself by forced power downs. RAM is cheap now days so simply run your OS from a ramdisk image while keeping only large static files in flash along with a settings file that gets written to in a safe way. Test equipment spends most of its life in the off state. You don't even use your scope for 10 hours a day on average even if you really use your scope a lot. Let alone use some more specialized bit of gear that you perhaps only power on a few times per year.

What makes things worse is that all of this gear uses switching supplies now. Okay they do waste less power than a oldschool transformer at low loads (When properly designed) so that lends them well to soft power off. But these things are also a lot more fragile. Mains power can have all sorts of surges happen and is common to get stuck by lightning. These surges can blow the primary side transistors in switching supplies pretty easily, but a big chunk of copper and iron doesn't really care as long as its less than the few kV of its insulation resistance. If you try pushing too much voltage into a transformer it simply saturates and draws a huge current while the output just rises a bit above normal. Of course it won't save you from a direct hit but i seen a lot of switching supplies blow up from lightning strikes while all transformered things in the same building kept working fine.

The only exception is instruments that have ovenised things in them. Often old gear had a small standby supply that is wired before the power switch to provide just a few watts of power to an ovenised crystal or something so that it drifts as little as possible. So i suppose if you have that then a soft power button is easy to do and saves you having an extra PSU.

There's a few things you overlook.

Some test gear has a hidden OS (often Windoze or Linux) that must be shutdown correctly so a soft OFF button lends itself well to these needs and fits well with SMPS supplies.
Much equipment is also more portable than those of yesteryear and again SMPS is preferable to save weight and cost rather than brute iron and copper.
Also, we've all seen enough years of SMPS to know their primary failure case is when hard switched rather than left powered for long periods.
And lastly, SMPS lends itself to international markets better than linear PSU's for multi-voltage and multi-frequency needs. At one time marine and aviation test equipment was somewhat specialized whereas today most equipment can handle these rather obscure mains supplies with ease.


Unlike the minuscule walwarts we all have for the likes of our phones etc, equipment with larger form factors do have the room to incorporate proper mains filtering that consists of adequately sized components, spark gaps, air gaps and common mode mains filters that make the modern SMPS quite robust. I've never replaced any in equipment I've supplied and can only vaguely remember one Siglent SMPS failing mentioned somewhere on this forum.
Luck I guess, but it seems not all other brands have shared this luck.

[Wolfgang]

... another exception are larger instruments that need a regular Windows shutdown, like the Keysight S Scopes, VNAs and Spectrum Analyzers.
They have a "hard" power pushbutton, but a "soft" shutdown is performed.

[tautech]

... another exception are larger instruments that need a regular Windows shutdown, like the Keysight S Scopes, VNAs and Spectrum Analyzers.
They have a "hard" power pushbutton, but a "soft" shutdown is performed.

Which of course indicates a PSU that's always ON when plugged in or rear switched. (PITA)
Be it linear or SMPS, smoothing caps still wear out as the years tick by.

[Wolfgang]

I dont know if they do a "hard" shutdown after the "soft" one is completed.
They Keysight stuff (or Agilent, or HP) has a reputation to last for very long, so maybe they use a lot better caps than our Chinese friends from today (see Daves standard rant when tearing down stuff).

[Mr. Scram]

I dont know if they do a "hard" shutdown after the "soft" one is completed.
They Keysight stuff (or Agilent, or HP) has a reputation to last for very long, so maybe they use a lot better caps than our Chinese friends from today (see Daves standard rant when tearing down stuff).

A hard shutdown is simply cutting the power by means of a physical switch. That's what people like, a proper switch with zero power consumption and a non ambiguous function. What you describe sounds like a soft solution.

[Berni]

I did have the PSU in my Tek AWG2041 fail all of a sudden, one day it simply didn't turn on. Its a pretty big thing (Twice as large as a standard ATX PSU) and it had a transistor fail in the PFC that then also blew up the softstart circuit.

And yes a lot of gear now runs a OS back there, and you can indeed screw up the flash if you keep killing power to it as its running. What you simply have to do is move a unchanging golden filesystem image into RAM at boot and then run the OS from ram disk. If you kill the power the RAM is gone anyway so no worries if the filesystem is screwed up. For things you can't loose like the settings you just store that in flash with proper safety (backup copies, minimal writes, syncing changes). Perfectly doable for a multimeter or signal generator. At work we have a bunch of products that run full blown linux and its normal for them to be turned off by cutting power and 1000s of these units keep working for years in the field. All you have to do is keep it from writing to flash all willy nilly.

And yeah PC based scopes like the Keysight Infiniium ones are a different story, obviously you can't just kill Win 7 and running from ramdisk under a hypervisor is not practical for such a big bulky OS. On my particular MSO9000 the power button is simply the motherboards power switch so it works exactly like on a PC (And is pretty much just a PC with a scope board hanging off a PCIe port).

But i do suppose i have not seen many failures of switchmode PSUs on test equipment. I have a lot of test gear and i saw only 1 failure on my Tek. But i do remember switchmode PSU failures being very common in consumer crap (Probably the number 1 cause of failure even).

[Bassman59]

On the 800 series, 16-bit arb waveform files are truncated to 14-bit, whereas the 900 series reproduces the full 16 bits.

Wow, really?

Check it out, from DG800 manual:

For the arbitrary waveform file, the voltage values corresponding to each waveform point are stored in binary data format. The voltage value at each point takes up 2 bytes (16 bits). DG800 only uses 14 bits, and the higher 14 bits are used to represent voltage and the lower 2 bits are not used. Therefore, its format of the binary data is from 0x0000 to 0xFFFC.

From DG900 manual:

For the arbitrary waveform file, the voltage values corresponding to each waveform point are stored in binary data format. The voltage value at each point takes up 2 bytes (16 bits). Therefore, its format of the binary data is from 0x0000 to 0xFFFF.

 

I suppose I could RTFM, but does that apply to standard waveforms such as sine and square waves?

[commongrounder]

On the 800 series, 16-bit arb waveform files are truncated to 14-bit, whereas the 900 series reproduces the full 16 bits.

Wow, really?

Check it out, from DG800 manual:

For the arbitrary waveform file, the voltage values corresponding to each waveform point are stored in binary data format. The voltage value at each point takes up 2 bytes (16 bits). DG800 only uses 14 bits, and the higher 14 bits are used to represent voltage and the lower 2 bits are not used. Therefore, its format of the binary data is from 0x0000 to 0xFFFC.

From DG900 manual:

For the arbitrary waveform file, the voltage values corresponding to each waveform point are stored in binary data format. The voltage value at each point takes up 2 bytes (16 bits). Therefore, its format of the binary data is from 0x0000 to 0xFFFF.

 

I suppose I could RTFM, but does that apply to standard waveforms such as sine and square waves?

I have to assume the "standard" waveforms are 16-bit, since the generator is marketed as 16-bit.  Also, it isn't clear to me whether the 14-bit truncation is just with loaded-in (created) arb files, or the in-built arb waveforms as well.  This info was only stated in the file management section of the manual.  I think it deserves a mention in the published specification.

[Wolfgang]

but the look is like a horror movie ...

What the RIGOL DG800 series needs is a special accessory:

A matching coffee cup that does not risk to fall over when put on top of the unit and spill the coffee thru all the cooling slits on top.
I have a modest proposal here how this could look like  Innovation or Nothing ! 

Another effect is that the coffee will be kept warm by the heat generated within the generator.

[bd139]

That's a great idea. My 33120A appears to have coffee stains on the back of it so this is definitely a desired feature 

[David Hess]

Cases are deliberately designed like that to prevent stacking anything on top when the device has power dissipation problems.  Less than 30 watts and they needed to do that?