EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: mikeselectricstuff on June 29, 2012, 07:38:17 am
-
Very impressive build quality.
Rigol DG4062 function/arbitary waveform generator teardown (https://www.youtube.com/watch?v=jGXiS4X4Hlw#ws)
-
Great teardown as usual Mike, thanks a lot !
This video definitely teased my nerves here, I might order one based on the build quality it self :P, can't wait for your review !
Btw, related to your other video on dealing with this Rigol warranty seal, that is a really great job you've done there ! <thumbs up> ;D
PS : About those Chinese characters, actually they're color codes for the cables.
Check this color codes list -> http://www.chinese-tools.com/chinese/vocabulary/list/90/colours.html (http://www.chinese-tools.com/chinese/vocabulary/list/90/colours.html)
(https://www.eevblog.com/forum/product-reviews-photos-and-discussion/rigol-dg4062-functionarbitary-waveform-generator-teardown/?action=dlattach;attach=26303;image)
-
Can I ask what version your device gives you (under utility > system > Sys info) because when you did that warranty seal video I looked at my unit and it didn't have the same looking feet.
And I wonder if you can change the device to a higher version by soldering those resistors there.
shameless plug (https://www.eevblog.com/forum/product-reviews-photos-and-discussion/rigol-dg4062-8679/)
-
...Wasn't expecting that high of a quality :-[
-
Very impressive build quality.
Rigol DG4062 function/arbitary waveform generator teardown (https://www.youtube.com/watch?v=jGXiS4X4Hlw#ws)
Where did you get the T-shirt! The text is spot-on.
-
Thank you, Mike!
Nice teardown.
Now, regarding the warranty seal removal without damaging it, I am just heating it up a little with the hot-air rework station set to 100°C from some reasonable distance in order to soften its glue until it peels off graciously.
-George
-
This will also work on ships that have 440Hz AC
-
Very impressive build quality.
Rigol DG4062 function/arbitary waveform generator teardown (https://www.youtube.com/watch?v=jGXiS4X4Hlw#ws)
Where did you get the T-shirt! The text is spot-on.
http://www.spreadshirt.com/ (http://www.spreadshirt.com/) - just type the text on their website and format it - press order and your shirt arrivesa few days later
-
This will also work on ships that have 440Hz AC
400Hz is a standard on aircraft - most SMPSUs and transformer PSUs will work OK on 400Hz, so they might as well include it in the spec to get the odd extra sales.
May also be that they looked at specs from people like Tek & Agilent who sell to military, and said "If they do it we should also include it in our specs"
-
Nice ,very impressive !
-
Very impressive build quality.
Rigol DG4062 function/arbitary waveform generator teardown (https://www.youtube.com/watch?v=jGXiS4X4Hlw#ws)
The resistors will connect to testpads on the bottom so the ATE can read the setting throught the bed of nails contactor.
Early in the video you had the heatsink off one of the output amplifiers. looks like twin SO8 chips with a silver reverse ink printing on them. what are those ? look like AD or TI parts ... later in the video the heatsink was back on.
I had an interesting chat with the rigol people yesterday. we spent about 4 hours going over what i am looking for in a product and they were very carefully taking notes of all the remarks i had to make.
i am looking at a different line of ARB's , the DG5000 series , because they also have a didigtal interface in the back ( you have access to the digital stream going into the DAC as well as advanced pettern generation capabilites like I2C , SPI and others.
Rigol is really serious about their new products. The evolution of the company is that it started about 14 years ago by 3 university students that made little handheld or bare board products . simply because real equipment as too expensive.
Gradually they got better , partnered with some 'big boy' as OEM , and gained a bunch of know-how.
They improved their production and design capability greatly , launched a few other products , partnered up again as OEM but have a full blown desing team. there's about 400 people at Rigol with over 150 design engineers ... They are serious and in it for the long run.
A couple of years ago they scooped up a bunch of engineers from two companies that now sail under the danaher flag ... and those guys brought some 'big guns technology' to the company...
I had a bunch of really oddball questions and their field-apps engineer is going to check out what can be done. The guy was very knowledgable on how the machine works internally . I had questions on their multimeter . Very specific details like the autozeroing sequence , being able to disable it , read the residual etc ... all possible. He knew exactly what i was talking about ( he was one of the 'danaher escappees and very familiar with a range of popular 6-1/2 and 8-1/2 digit benchtop meters.... the rigol machine can do the same things as that one .. )
Same for the scope. A lot of scopes have problems dealing with the following :
you have a 100mVpp signal superimposed on a 10 volts DC signal. Show me the 100mVpp signal , using DC mode , with the highest possible number of divisions on the screen.
There are two scenarios :
- a scope where the offset is injected BEFORE the gain stage. in this case the input preamp is a summing amplifier , you subtract the DC before sending the residual through the attenuator / gainstage and digitize it. Scopes employing this architecture have no problem showing a 5 division high signal ( 20mV div setting ) on the display.
- a scope where offset is injected post attenuation... these cannot show you the signal. they do not have enough range an d you can either see a liitle bit of wiggly , or have the signal clipped.
the first solution requires a very high precision DAC and a good summing amplifier.. costly stuff.
the second stuff is done in software ....
Now, suerely theres' people out there thhat will say : switch the scope in AC mode..
but , that's not doable in all cases. when looking at the switching node of a step-down converter you may want to investigate things like shoot-through , or misfires of a synchronous resctifier... you need to be able to find out what is both the DC level of the signal , and have a realy good view of whatever ringing or other small signal there is superimposed.
So you need a high dynamic range.. only scopes that emply the summing amp pre-attenuator can do this...
-
This is great to know, they could be China's premier MTE makers and I think it shows, in the years we've read about Rigol, Owon, Tekway and Hantek, here on the eevblog, only Rigol IMHO, seems to be consistent and they know where to concentrate; whatever bugs were int he 1052e were not critical, its still a good measuring device.
Your grilling of them really helps get down to the meat of their engineer's capabilities. Hopefully it won't be severely affected by the marketing department.
I wonder which Denaher employees that was? Regardless, I'm sure they've added more to Rigol's Tek-know-how.
Very impressive build quality.
I had an interesting chat with the rigol people yesterday. we spent about 4 hours going over what i am looking for in a product ...
Rigol is really serious about their new products. The evolution of the company is that it started about 14 years ago by 3 university students that made little handheld or bare board products . simply because real equipment as too expensive.
Gradually they got better , partnered with some 'big boy' as OEM , and gained a bunch of know-how.
They improved their production and design capability greatly , launched a few other products , partnered up again as OEM but have a full blown desing team. there's about 400 people at Rigol with over 150 design engineers ... They are serious and in it for the long run.
A couple of years ago they scooped up a bunch of engineers from two companies that now sail under the danaher flag ... and those guys brought some 'big guns technology' to the company...
I had a bunch of really oddball questions and their field-apps engineer is going to check out what can be done. The guy was very knowledgable on how the machine works internally . I had questions on their multimeter . Very specific details like the autozeroing sequence , being able to disable it , read the residual .....
-
Here in Europe you could have shown and let Rigol smell the money, and threaten to actually buy their equipment - you still didn't get any kind of pre-sales support. And of course nothing but lies and silence after-sales fresh out of China.
I was laughing off my ass when Rigol "improved" their presence in Europe by setting up a European headquarter at the end of last year. They really didn't spare any expenses, and hired one manager, one employee for administrative tasks, and, lo and behold, a tech. One. For Europe.
I haven't heard anything from them after they got founded. Maybe someone should check if they are still alive, or if there are three dead bodies rotting away in some warehouse under a bunch of toppled down DG1022 generators.
-
I was laughing off my ass when Rigol "improved" their presence in Europe by setting up a European headquarter at the end of last year. They really didn't spare any expenses, and hired one manager, one employee for administrative tasks, and, lo and behold, a tech. One. For Europe.
MCS told me that there would be a few days' delay in shipping this unit as the person in charge of shipping at Rigol EU was on holiday...
-
I've got a DG4160. I like it quite a lot.
That said when I went shopping for a new scope recently I dropped them from consideration and bought a Agilent MSO7000B. The problem is, earlier this spring I tried to contact them on two separate occasions. On was a simple pre sales question about the DSA815-TG and another about a DG1022 connectivity problem I had. Both times I received a automatic message saying they got my message and will get right back to me. Problem is I never heard from them again. Personally my tolerance for this sort of behaviour is for a product $1500 bucks or less and even then I don't like it. For something more expensive than that I expect some service so Agilent it was.
Other people seem to get detailed responses out of them, so I don't know.
-
Here in the bay area they seem to take things seriously. There is a local rep and he had a guy that flew down from somewhere to do a tour of a bunch of companies here and show their new stuff.
They did a big hamfest a couple of months ago and took like 50 of there new spectrum analysers to that show. In two hours they were all sold...
Ill get some machines on demo and 'torture' them a bit... We'll see.
-
Thanks for the show, interesting to see inside.
You didn't spot/mention the frequency reference which would have been interesting if identifiable.
Another reason for lasering some parts might be to hide the lack of hardware difference between the different versions.
I imagined Rigol were using a lot of common parts between these generators, the 800 series spectrum analyzers and the new scope range. I bet the power supply, LCD, blackfin and a lot of firmware is common to them. I was surprised to see back of the unit, the DSA-815 has 3 BNCs with the ethernet and USB slave above them.
Standby power on my DSA-815-TG measured fluctuation between 0.7 and 2W on 3 different cheap plug in power monitors, operating power about 25W.
Check out the pulse function when you review. I have a TTi TG5011 which basically can position pulse edges with 1ns resolution. That kinda makes a nonsense out of the 1uHz frequency resolution claim shared with the Rigol. If you generate pulses with DDS to get uHz resolution at all frequencies then you get 1 DDS generator clock cycle of jitter - maybe that is where the 500ps jitter spec on the Rigol comes from. The TG5011 doesn't have any jitter I can see which is perhaps more desirable.
-
predator wont dare enter mike's home. he got the FLIR installed. probably mike need to make another review and installation of his tesla coil. will burn the predator to ashes on sight.
-
Thanks for the show, interesting to see inside.
You didn't spot/mention the frequency reference which would have been interesting if identifiable.
It will be the crystal osc below the DAC PLL (7:43 on video) - nothing special but you can connect an external 10M ref, so with one of those rubidium sources, most of the onscreen digits may actually be meaningful, depending how much jitter the PLL has.
-
What do they use as output amp ? You had the heatsink off early in the video. Looked like two so8 chips with silver text.
-
It will be the crystal osc below the DAC PLL (7:43 on video) - nothing special
Just a TCXO and not VCTCXO which could be soft calibrated?
Bit cheap if it isn't and a bit annoying you don't have easy access to calibration if it is. The DSA800 is probably the same and mine runs a pretty stable 2ppm low.
-
What do they use as output amp ? You had the heatsink off early in the video. Looked like two so8 chips with silver text.
numbers were lasered off - doubt it would be hard to ID them though
-
It will be the crystal osc below the DAC PLL (7:43 on video) - nothing special
Just a TCXO and not VCTCXO which could be soft calibrated?
Bit cheap if it isn't and a bit annoying you don't have easy access to calibration if it is. The DSA800 is probably the same and mine runs a pretty stable 2ppm low.
Didn't look too close but didn't look anything other than a standard osc. Hoewever the datasheet claims 2ppm from 18-28 deg, so maybe it is compensated - will take another look and check against a rubidium reference.
-
Can I ask what version your device gives you (under utility > system > Sys info) because when you did that warranty seal video I looked at my unit and it didn't have the same looking feet.
And I wonder if you can change the device to a higher version by soldering those resistors there.
shameless plug (https://www.eevblog.com/forum/product-reviews-photos-and-discussion/rigol-dg4062-8679/)
Software 00.01.02
FPGA 00.01.05
Hard 01.01
Keyboard 04.01
-
Measuring a rubidium reference using the DG4062 counter mode reads 9.999 999 478 MHz, so about 0.05ppm out - not bad, and looks like the osc is at least calibrated and probably temp compensated.
-
Measuring a rubidium reference using the DG4062 counter mode reads 9.999 999 478 MHz, so about 0.05ppm out - not bad, and looks like the osc is at least calibrated and probably temp compensated.
Did you leave that running for a while or was this a one time measurement?
-
Was pretty stable over about 10 minutes
-
yes, i guess it must be a VCTXCO~
-
I can't watch the youtube,who can tell me what's the DAC of DG4062, I doubt it‘s AD9778,Dual 14-Bit, 1.0 GSPS D/A Converter,like AD9739 of Agilent 81160 in Dave's teardown.
-
I can't watch the youtube,who can tell me what's the DAC of DG4062, I doubt it‘s AD9778,Dual 14-Bit, 1.0 GSPS D/A Converter,like AD9739 of Agilent 81160 in Dave's teardown.
AD9781
-
I can't watch the youtube,who can tell me what's the DAC of DG4062, I doubt it‘s AD9778,Dual 14-Bit, 1.0 GSPS D/A Converter,like AD9739 of Agilent 81160 in Dave's teardown.
Hi,stupid maybe you're in china , here is the way to watch the video:
http://f-2-1-o-9-o.ccnn.info/x999/browse.php?u=mmklUp%2FgA2W0FzB%2FoEbJcZqmDUU5eTyB0yYkOJkzYkUzVzXl%2F0E%3D&b=29 (http://f-2-1-o-9-o.ccnn.info/x999/browse.php?u=mmklUp%2FgA2W0FzB%2FoEbJcZqmDUU5eTyB0yYkOJkzYkUzVzXl%2F0E%3D&b=29)
-
Does anyone knows the real dimensions of the DG4062..???
In the brochure it is 313 mm wide and 160 mm high,
in the manual it is 440 mm wide and 218 mm high,
are there different models around...?
-
Does anyone knows the real dimensions of the DG4062..???
In the brochure it is 313 mm wide and 160 mm high,
in the manual it is 440 mm wide and 218 mm high,
are there different models around...?
Just checked -- DG4062 is: 313 mm wide x 160 mm high. (I don't believe there are different world models, but I'm in NA if it helps.)
-
Siglent launched the SDG5000 series, and the parameter is better than DG4000.
Someone used it? How about it?
I used the SDG1000, it's very good.
I think the SDG5000 will not let me down.
-
Learning about the unimplemented buttons/features on the DS2000 got me curious about the DG4000 series - and sure enough, while re-watching Mike's video, it's clear there are 3 unused button positions (2 with illumination) on the keypad PCB (and the corresponding knockouts in the front panel):
(https://www.eevblog.com/forum/testgear/rigol-dg4062-functionarbitary-waveform-generator-teardown/?action=dlattach;attach=45146)
(https://www.eevblog.com/forum/testgear/rigol-dg4062-functionarbitary-waveform-generator-teardown/?action=dlattach;attach=45144)
Edit: Whoops - Chrome crashed while posting that, and lost an image temporarily. :)
-
Learning about the unimplemented buttons/features on the DS2000 got me curious about the DG4000 series - and sure enough, while re-watching Mike's video, it's clear there are 3 unused button positions (2 with illumination) on the keypad PCB (and the corresponding knockouts in the front panel):
Interesting, what is coming there? :) :-//
-
Interesting, what is coming there? :) :-//
Well, judging by the button positions, one would likely be another mode (after MOD, SWEEP, BURST) - but not sure about the other 2 close to the CH selects - perhaps some kind of CH combinations?
-
Learning about the unimplemented buttons/features on the DS2000 got me curious about the DG4000 series - and sure enough, while re-watching Mike's video, it's clear there are 3 unused button positions (2 with illumination) on the keypad PCB (and the corresponding knockouts in the front panel):
Interesting, what is coming there? :) :-//
They might just lay out complete grids of buttons as it is simpler and gives them the potential for future expansion without there being any specific plans.
It's hard to know what they might add as they already have a frequency counter and two channels unless it is something to generate binary patterns or signals.
-
They might just lay out complete grids of buttons as it is simpler and gives them the potential for future expansion without there being any specific plans.
I don't agree with that - considering that just two of the three buttons are illuminated (edit: and the knockouts in the front panel) tends to make me think they had/have plans. Now whether they decided NOT to implement something - or ever will in the future (regardless of plans) - is a different story.
-
The keyboard version (in system info) is 4.01 so they did change it probably more than expected while the hardware version is 1.01
-
I was just looking at the User Manual for the Agilent 33250A:
http://cp.literature.agilent.com/litweb/pdf/33250-90002.pdf (http://cp.literature.agilent.com/litweb/pdf/33250-90002.pdf)
in particular the block diagram on page 304 and it struck me that the Rigol DG4000 series is based on this. The Agilent design dates from 2003/4 and gives quite a lot
of details so it is likely that it would form a template.
The three relays per channel in the teardown that appear to be attenuators are probably two attenuators and perhaps a gain switch on the output amp (going from 20dB to 0dB).
The specs are also similar in format so I suspect that perhaps the pulse and square waves are generated from sine waves and a comparator in the same way which is why
they have different rise times to the arb output see pages 300 and 301. The Agilent design does this above 2MHz whilst I guess the Rigol, if it does it, switches it in at 5MHz because
the specs give a change in jitter at that frequency. I've not looked at the teardown again to see if any of this is likely from the circuit, but the specs seem to be in accord.
Edit : on further thought, it would appear that they use Arb for the pulses at least which can't be frequency swept. For square waves it is puzzling because if they use Arb in the same way why do the different models have different rise times? They can't build them into the arb shape otherwise they would change with frequency and square waves can be frequency swept.