Anyone who bought a Rigol DG4062 wish they would have bought a DG4102 or DG4162?

[Electro Fan]

Or vice versa?

- anyone see any notable differences between these models other than the obvious specs and the prices?

On a related question:  Anyone who bought a waveform generator built into an oscilloscope (such as Agilent, etc.) wish they had bought a standalone waveform generator?  (Or vice versa - anyone with a standalone waveform generator wish they had bought a waveform generator built into an oscilloscope?  Seems unlikely, but I thought I'd ask.)

Thanks for sharing any experience/ thoughts.

[jpb]

Given the thread on this forum about hacking the DG4062 up to 200MHz there may be some people who regret buying the higher models. (Personally I probably wouldn't take that route but I don't want to get into a debate which has been already covered at length in another thread).

The higher models main advantage is in higher frequency sine waves but you can get these (at some cost in flatness) by using a ten peak sine wave as an arbitrary wave form. The drawback, as well as the amplitude being uncalibrated (or rather the amplitude drop not being compensated for) is the jitter.

Similarly, the lower rise times of square waves in the DG4062 can be got round by using the arbitrary waveform to produce them but this will limit the frequencies you can use to some extent.

I'm still dithering on a signal generator and the longer I do so the more I realise I perhaps don't really need one - at least not such a good one as I wanted to get.

[AndyC_772]

I sometimes regret getting the DG4062 at all. It has some absolutely stupid limitations which are nothing to do with signal bandwidth, just arbitrary limits which I suspect are imposed by nothing more 'concrete' than the UI itself.

Take pulse generation, for example. If it's got a 500MHz sample rate, why shouldn't I be able to specify that I want a pulse 100ns wide, with square edges, at 1kHz intervals? For some reason the Rigol insists on a linear ramp up and down in Pulse mode which is far, far slower than the 60MHz rated bandwidth of the instrument.

I can get round some of the daft UI limits by defining an Arb waveform, though doing that through the front panel controls is a pain in the arse, and IIRC even that wouldn't accept such a narrow pulse with relatively infrequent repetition. Somewhere inside the box there's either some stupidly broken firmware, or a counter that's not wide enough.

For comparison, here's the pulse I wanted, as generated by the Wavegen feature of my Agilent 3000X, and the closest approximation I can extract from the DG4062 using the 'Pulse' feature, which you'd think would be fit for purpose. Note the difference in horizontal scale.

[jpb]

I sometimes regret getting the DG4062 at all. It has some absolutely stupid limitations which are nothing to do with signal bandwidth, just arbitrary limits which I suspect are imposed by nothing more 'concrete' than the UI itself.

Take pulse generation, for example. If it's got a 500MHz sample rate, why shouldn't I be able to specify that I want a pulse 100ns wide, with square edges, at 1kHz intervals? For some reason the Rigol insists on a linear ramp up and down in Pulse mode which is far, far slower than the 60MHz rated bandwidth of the instrument.

I can get round some of the daft UI limits by defining an Arb waveform, though doing that through the front panel controls is a pain in the arse, and IIRC even that wouldn't accept such a narrow pulse with relatively infrequent repetition. Somewhere inside the box there's either some stupidly broken firmware, or a counter that's not wide enough.

For comparison, here's the pulse I wanted, as generated by the Wavegen feature of my Agilent 3000X, and the closest approximation I can extract from the DG4062 using the 'Pulse' feature, which you'd think would be fit for purpose. Note the difference in horizontal scale.

This article, though rather old, explains the limitations of DDS quite well:

http://m.eet.com/media/1128699/12978-choosing_a_waveform_generator_the_devil_is_in_the_details.pdf


The main problem with the DG4062 is the memory is only 16k, so if you want 100 nsec pulses at 100kHz you are stuck as the mark space ratio is much more than 16k. Because of the fixed clock rate and the DDS architecture you can't do sequencing of different segments. Something like the Tabor WW5061 can easily do it even though it only has a 50MS/s maximum sample rate. Agilent's truform also allows different segments and loops.

Of course arbitrary waveform generators with varying clocks have drawbacks in terms of output filtering and most of all price, and Agilent's truform models are also much more expensive. For pulses the Siglent DG5000 series seems to offer good features but Dave's review indicated some problems with build quality.

[AndyC_772]

Fortunately I have my scope, which can generate these simple waveforms with laughable ease.

I understand the memory limitation, though I'd only expect that to really be an issue for long arbitrary waveforms. A rectangular pulse (or, for that matter, square, triangular and other mathematically simple waveforms) doesn't need RAM to synthesize it, it needs a finite state machine and a counter. If the Rigol is based on an FPGA, this really shouldn't be an issue at all.

[KedasProbe]

I sometimes regret getting the DG4062 at all. It has some absolutely stupid limitations which are nothing to do with signal bandwidth, just arbitrary limits which I suspect are imposed by nothing more 'concrete' than the UI itself.

Yeah there is no technical reason for that, probably marketing reasons, but for short pulses (low duty cycle) with the DG4000 use a square wave with 1 cycle burst.

Edit: I don't regret having the DG4102 instead of the DG4162 (yet).
(I may have regretted it if I had the DG4062 though.)

[rf-loop]

Siglent DG5000 series seems to offer good features but Dave's review indicated some problems with build quality.

What problem in building quality?
If Dave see small amount of rust marks on cutted metal corner this is lot of thumbs down and with big mouth screaming how terrible and horrible  it is. Then same kind rust marks inside Rigol Power supply, not even half word about rust but classified as "High End" laboratory power supply.

Also, there is several other peoples who have looked inside. Also, here in forum is couple of pictures about building details, also real lab tests etc, and really, there is not any markable  quality problems (in price class of course). 

[Orange]

Siglent DG5000 series seems to offer good features but Dave's review indicated some problems with build quality.

What problem in building quality?
If Dave see small amount of rust marks on cutted metal corner this is lot of thumbs down and with big mouth screaming how terrible and horrible  it is. Then same kind rust marks inside Rigol Power supply, not even half word about rust but classified as "High End" laboratory power supply.

Also, there is several other peoples who have looked inside. Also, here in forum is couple of pictures about building details, also real lab tests etc, and really, there is not any markable  quality problems (in price class of course).

Can you please spare use us your propaganda on Siglent and Owon, A rusty chassis is simply BAD

[grego]

Siglent DG5000 series seems to offer good features but Dave's review indicated some problems with build quality.

What problem in building quality?
If Dave see small amount of rust marks on cutted metal corner this is lot of thumbs down and with big mouth screaming how terrible and horrible  it is. Then same kind rust marks inside Rigol Power supply, not even half word about rust but classified as "High End" laboratory power supply.

Also, there is several other peoples who have looked inside. Also, here in forum is couple of pictures about building details, also real lab tests etc, and really, there is not any markable  quality problems (in price class of course).

Can you please spare use us your propaganda on Siglent and Owon, A rusty chassis is simply BAD

No propoganda - a bunch of us have the Siglent units and opened them up after seeing Dave's and none of us have reported a similar issue.  It's likely (as Dave has said on occasion) the "curse of the eevblog".  Not saying it wasn't bad (it was, that should never happen) but it looks like it's an isolated incident.

[Wim13]

I sometimes regret getting the DG4062 at all. It has some absolutely stupid limitations which are nothing to do with signal bandwidth, just arbitrary limits which I suspect are imposed by nothing more 'concrete' than the UI itself.

Take pulse generation, for example. If it's got a 500MHz sample rate, why shouldn't I be able to specify that I want a pulse 100ns wide, with square edges, at 1kHz intervals? For some reason the Rigol insists on a linear ramp up and down in Pulse mode which is far, far slower than the 60MHz rated bandwidth of the instrument.

I can get round some of the daft UI limits by defining an Arb waveform, though doing that through the front panel controls is a pain in the arse, and IIRC even that wouldn't accept such a narrow pulse with relatively infrequent repetition. Somewhere inside the box there's either some stupidly broken firmware, or a counter that's not wide enough.

For comparison, here's the pulse I wanted, as generated by the Wavegen feature of my Agilent 3000X, and the closest approximation I can extract from the DG4062 using the 'Pulse' feature, which you'd think would be fit for purpose. Note the difference in horizontal scale.

Well i have no problems with that on my DG4102...

See picture, of my DG4102, 1 khz puls of 2 pixels of 62 nSec wide.....

Made in arb mode, at 1 khz, 16000 points at 1khz, is pixel wide of 62 nSec...

[Harvs]

I know this has been said plenty of times before on this forum, but I'll bring it up again.  If what you want is a fast pulse generator, then this isn't the tool for you.  In fact it sucks at it compared to 30yr old analog equipment you could pick up off ebay for a fraction of the price.

I've got the DG4062, and this is the first ARB gen I've owned so I can't usefully compare it to other models.

When I first bought it I thought it was going to be awesome.  Then after a month or two once the initial cool factor wore off my feelings were that I'd wasted my money spending $750 on an ARB gen.  Now after having had it for a little over a year and got comfortable quickly putting together a python script to control it, bench DMM and my scope whenever I want to automate some test sequence I simply wouldn't be without it.  Personally I find the dual channels extremely useful as well.

As for the higher bandwidth models, no I simply don't have a need.

[mikeselectricstuff]

I agree that the lack of a proper pulse gen is annoying considering all teh hardware is there to implement it easily, but as usual, Chinese manufacturers lack the imagination to add stuff like this.
 

[Electro Fan]

I agree that the lack of a proper pulse gen is annoying considering all teh hardware is there to implement it easily, but as usual, Chinese manufacturers lack the imagination to add stuff like this.

Sometimes a product manager or product team can be very imaginative, but another good attribute would be the ability to listen to the market and let the market help surface good ideas; it sure seems like Rigol (and other manufacturers) could get many good product ideas by reading the threads here.  Who knows, maybe we'll see the inclusion of a good pulse generator in a future product/release.

[Electro Fan]

I sometimes regret getting the DG4062 at all. It has some absolutely stupid limitations which are nothing to do with signal bandwidth, just arbitrary limits which I suspect are imposed by nothing more 'concrete' than the UI itself.

Take pulse generation, for example. If it's got a 500MHz sample rate, why shouldn't I be able to specify that I want a pulse 100ns wide, with square edges, at 1kHz intervals? For some reason the Rigol insists on a linear ramp up and down in Pulse mode which is far, far slower than the 60MHz rated bandwidth of the instrument.

I can get round some of the daft UI limits by defining an Arb waveform, though doing that through the front panel controls is a pain in the arse, and IIRC even that wouldn't accept such a narrow pulse with relatively infrequent repetition. Somewhere inside the box there's either some stupidly broken firmware, or a counter that's not wide enough.

For comparison, here's the pulse I wanted, as generated by the Wavegen feature of my Agilent 3000X, and the closest approximation I can extract from the DG4062 using the 'Pulse' feature, which you'd think would be fit for purpose. Note the difference in horizontal scale.

Andy, does the DG4062 do anything for you that the 3000X doesn't do?  If the DG4062 went away is there anything you would miss, or could you meet all your function gen requirements with the 3000X?  Is the tradeoff between the two completely a matter of specs, performance, and UI?  Or is there some benefit you find in having an external function generator vs. the built-in generator?  Or maybe the standalone function gen box just takes up bench space?  Thx, EF

[Electro Fan]

I sometimes regret getting the DG4062 at all. It has some absolutely stupid limitations which are nothing to do with signal bandwidth, just arbitrary limits which I suspect are imposed by nothing more 'concrete' than the UI itself.

Take pulse generation, for example. If it's got a 500MHz sample rate, why shouldn't I be able to specify that I want a pulse 100ns wide, with square edges, at 1kHz intervals? For some reason the Rigol insists on a linear ramp up and down in Pulse mode which is far, far slower than the 60MHz rated bandwidth of the instrument.

I can get round some of the daft UI limits by defining an Arb waveform, though doing that through the front panel controls is a pain in the arse, and IIRC even that wouldn't accept such a narrow pulse with relatively infrequent repetition. Somewhere inside the box there's either some stupidly broken firmware, or a counter that's not wide enough.

For comparison, here's the pulse I wanted, as generated by the Wavegen feature of my Agilent 3000X, and the closest approximation I can extract from the DG4062 using the 'Pulse' feature, which you'd think would be fit for purpose. Note the difference in horizontal scale.

Well i have no problems with that on my DG4102...

See picture, of my DG4102, 1 khz puls of 2 pixels of 62 nSec wide.....

Made in arb mode, at 1 khz, 16000 points at 1khz, is pixel wide of 62 nSec...

Wim13, thanks for the post - it is pretty cool.  Just curious, what caused you to purchase the 4102 over the 4062?  Thx, EF

[GEuser]

Siglent DG5000 series seems to offer good features but Dave's review indicated some problems with build quality.

What problem in building quality?
If Dave see small amount of rust marks on cutted metal corner this is lot of thumbs down and with big mouth screaming how terrible and horrible  it is. Then same kind rust marks inside Rigol Power supply, not even half word about rust but classified as "High End" laboratory power supply.

Also, there is several other peoples who have looked inside. Also, here in forum is couple of pictures about building details, also real lab tests etc, and really, there is not any markable  quality problems (in price class of course).

Can you please spare use us your propaganda on Siglent and Owon, A rusty chassis is simply BAD

Thanks for the input , I did not want to post anything at first as it would look like I am defending dave since I have the same flag and rfloop/aghp would certainly capitalize on that , dave can look after himself as he is big enough (but looks short  ) , the propaganda just amazes me and sometimes is not amusing imo .
cheers ....

Moderator: Irrelevant quote from another thread deleted

[KedasProbe]

Just curious, what caused you to purchase the 4102 over the 4062?  Thx, EF

In my case I really wanted to be able to put the carrier wave in the normal radio spectrum.
And sweeping freq as high as possible.
The SDG5000 has much more arb memory (also nice if you want to put some audio signals on it) Also no limits on entering duty cycle is better than DG4000. (other things may be worse though)

[AndyC_772]

Yeah there is no technical reason for that, probably marketing reasons, but for short pulses (low duty cycle) with the DG4000 use a square wave with 1 cycle burst.

Thanks for the tip - I've tried it and it works, and is much easier than setting up an arbitrary wave. What a great shame the mode actually intended for the purpose is so useless, though. I really don't think I should need to use workarounds like this when there's a button on the front panel marked 'Pulse'.

Well i have no problems with that on my DG4102...

See picture, of my DG4102, 1 khz puls of 2 pixels of 62 nSec wide.....

Made in arb mode, at 1 khz, 16000 points at 1khz, is pixel wide of 62 nSec...

Fair comment - but do you feel that a 62ns granularity is fit for purpose on a machine with a stated 500 MHz sample rate? I don't feel there's any valid reason whatsoever why these units shouldn't be able to offer 2ns granularity in the position of a data point. Nor is there any valid reason for needing to set up an arbitrary waveform just to 'get around' limits in the values that the UI will accept.

I agree that the lack of a proper pulse gen is annoying considering all teh hardware is there to implement it easily, but as usual, Chinese manufacturers lack the imagination to add stuff like this.

There's a button on the front panel with "Pulse" written on it. I don't think you can blame lack of imagination - just a lack of testing in real-world cases by engineers who might want to actually use the instrument.

This isn't the only mode that's broken BTW. Want to generate a PWM signal and vary it from 0% to 100%? You can't. A square wave, according to Rigol, has a duty cycle between 20% and 80%. 

Andy, does the DG4062 do anything for you that the 3000X doesn't do?  If the DG4062 went away is there anything you would miss, or could you meet all your function gen requirements with the 3000X?  Is the tradeoff between the two completely a matter of specs, performance, and UI?  Or is there some benefit you find in having an external function generator vs. the built-in generator?  Or maybe the standalone function gen box just takes up bench space?  Thx, EF

I bought the DG4062 first, so it's a bit of a moot point. When I later bought the 3000X, it seemed silly not to get the Wavegen feature enabled.

The 3000X only has one channel, of course, but otherwise both are theoretically capable of meeting my needs. I occasionally need both channels, though not too often.

I am conscious of the difference in cost between the two instruments though... there are circuits to which I'd much more happily connect a DG4062 than an MSO-X3054A. (Normally, of course, the scope has a 10:1 probe in the way to protect it, and its inputs are rated 300V rms anyway - but not its Wavegen output!)

If I managed to blow up the Rigol I'd be annoyed, whereas if I damaged the Agilent I'd be facing a repair bill well into four figures that I could do without.

[Electro Fan]

This article, though rather old, explains the limitations of DDS quite well:

http://m.eet.com/media/1128699/12978-choosing_a_waveform_generator_the_devil_is_in_the_details.pdf


The main problem with the DG4062 is the memory is only 16k, so if you want 100 nsec pulses at 100kHz you are stuck as the mark space ratio is much more than 16k. Because of the fixed clock rate and the DDS architecture you can't do sequencing of different segments.

The article in the link is excellent!  Definitely recommended reading for anyone trying to figure out generator designs.  Impressive how a technology article 15 years old can still be so insightful.  Thanks for sharing it.

[jpb]

The article in the link is excellent!  Definitely recommended reading for anyone trying to figure out generator designs.  Impressive how a technology article 15 years old can still be so insightful.  Thanks for sharing it.

I'm glad you found it useful.

Apart, perhaps, for Agilent's truform approach (which adds a resampling stage) the technology hasn't changed in fundamentals over fifteen years.

From a hobbyist's point of view what was expensive technology being assessed by engineers for professional use fifteen years ago is now much cheaper technology affordable enough for home labs.

[leafi]

Anyone know if the following is possible from the DG4062...

I need to make a dual fixed frequency square wave signals which are out of phase and need to be able to control the Duty cycle of each channel independently yet have their clocks synced such that the signals are not on at the same time? I just got my 4062 and played with it a little bit but I can not trigger the second channel off of the first in square wave mode nor could I link the clocks so I could delay the second channel by half of the period to meet the need. Thanks

[Teneyes]

Anyone know if the following is possible from the DG4062...

I need to make a dual fixed frequency square wave signals which are out of phase

@Leafi

Do you mean like this:
use 'Util' and turn sync on for 'CH2Set'
use start phase on Ch2

[leafi]

YES thank you! I only could spend 15 min playing with it and I did not see a way to sync the channels. Thank you!

[Sylwerdragon]

Well I don't have that scope it cost a lot of money but i think stand alone generator is much better than on scope. In some cases it is better when it is in scope but i think when it is alone usually it is better quality and functions. and you can control it easier. But that is only my opinion.

[Dave Turner]

Is it possible to buy bandwidth upgrades to the 4062 or has Rigol figured out that software upgrades aren't cost effective?