-
Dave teards down the new Rigol DL3021 Electronic load and has a play around with it.
http://emona.com.au/products/electronic-test-measure/electronic-loads/dl-3021.html#.Wb7yY7Kg9hE
-
"teards" that's a new word?
-
If Rigol ever want to be taken seriously as a major T&M company they're going to have to do better than this. OK loads are a relatively niche product, but bad design ( graphic or otherwise) reflects across the whole product range.
The voltage regulator thing is just ridiculous - even they can't design a sufficiently quiet switching reg, they've got a sodding great heatsink in there they could have used.
The graphic design aspect reminded me of the laughably awful Neoden video - seems like the Chinese would rather do a bad job themselves than hire someone that knows what they are doing.
I really can't see any Chinese company reaching the same level of brand reputation as the likes of Keysight, Tek, Keithley, R&S etc. -
Thought this might be an interesting watch, considering that I've been having lots of fun reverse engineering a ds1054z lately.
https://youtu.be/Tk_3rYKWIIg?t=826 looks a lot to me like that is the main CPU of the whole thing -- that's a freescale MC i.MX, the same series of processor that powers the ds1054z. Hopefully a rather newer one ... nope. it's the i.MX283, very very similar to the i.MX286 that powers the ds1054z scope. Freescale considers that a legacy processor, so I'm rather surprised they went with something so old.
There's also both RAM and flash on there. Nicely, they've given us a couple of nicely marked 0.1: unpopulated connectors on the top -- ARM-JTAG and BOOT.
I also wonder a bit if that's a real PCI card. It's not out of the realm of possibility, and it'd be somewhat handy during product development.
Also, note that it's called DigitalBoard in the silkscreen, not IO, or peripherial, or back-panel or the like.
(And there's the same 5-pin diagonal FPGA JTAG connector.)
... yeah, he's coming to the same conclusions as I did, just a bit slower.
Anyway, the design is very similar to the scope, in terms of what chips they are using. The difference is that they've moved to more boards, which means they can mix and match better and hopefully put out products quicker. I am still surprised they've gone with such an old CPU, though. -
Someone beat me to it on YouTube comments, but I'm pretty sure that the problem with the wheel is essentially a PEBKAC. The green underline is selecting which decimal place is currently wheel controlled. Presumably some action (pressing the wheel in?) will cause the selected decimal place to shift (to the left, since the rightmost one is currently selected around 22 minutes in).
If you select the tenths or ones place, perhaps it could actually have velocity or acceleration or whatever and you might have missed it.
-
Damn, and I really wanted one, like a lot. Now not so much....
The reversed italics and the various fonts are horrible. I agree the main binding posts suck, but I do like the sense posts.
I really would like to use something else than CC on the battery mode...and switching the output of when going into store/utility... meh. And again the linear regulators...Then the presumably failed calibration ? -
Watching that - and the video Dave mentioned for a DIY solution, I'm half thinking of building my own - with a microprocessor driving it.
Time to brush up on power mosfets.... -
I'll still consider buying one if they add CP mode to the battery program. They did inform me that the feature will be added to a firmware update at the end of August, but that date came and went and I don't see a firmware update published anywhere yet.
-
I'll still consider buying one if they add CP mode to the battery program. They did inform me that the feature will be added to a firmware update at the end of August, but that date came and went and I don't see a firmware update published anywhere yet.
Same here.
I still trust Rigol more, than one of those copied no name electronic loads.
I really hope, the fix the issues they can in firmware or offer maybe later a free hardware upgrade. -
Shunt: http://www.isabellenhuette.de/fileadmin/content/praezisions-leistungswiderstaende/BVS.PDF
Isabellenhütte BVS, 50PPM/K. It is a cheap shunt.
I'm starting to think, that they laser the opamps for other reasons than IP. Maybe they dont have the rights to import them to china. I mean, designing an amplifier for low side current sensing? Really, that is an IP worth protecting? -
The culture. The rigid educational system that doesn't put any value on imagination or creativity, so there is no significant innovation, just cost-engineering of Western products.I really can't see any Chinese company reaching the same level of brand reputation as the likes of Keysight, Tek, Keithley, R&S etc.
Why? What's unique about China that prevents this from ever happening?
Rigol are generally regarded as the best of the Chinese T&M suppliers yet they're still turning out products with poor software design, and rubbing numbers off chips.
How many Chinese companies can you think of that are leaders in their field? The only one I can think of is DJI.
And in a "professional" rather than "consumer" field? Lenovo maybe, but they inherited a lot from IBM.
-
Well, there is also Xiaomi, with the Yi product line. But something else will happen. Chinese copy something, they sell more. Make more profit with the copy than the original design company. And then buy them. It happened for Segway. We all know the story.
The culture. The rigid educational system that doesn't put any value on imagination or creativity, so there is no significant innovation, just cost-engineering of Western products.I really can't see any Chinese company reaching the same level of brand reputation as the likes of Keysight, Tek, Keithley, R&S etc.
Why? What's unique about China that prevents this from ever happening?
Rigol are generally regarded as the best of the Chinese T&M suppliers yet they're still turning out products with poor software design, and rubbing numbers off chips.
How many Chinese companies can you think of that are leaders in their field? The only one I can think of is DJI.
And in a "professional" rather than "consumer" field? Lenovo maybe, but they inherited a lot from IBM.
It happened to Riot Games, they made a software for 67 million players a month. The copy Chinese software is used by 200 million. Copy software company buys riot games.
Soon they dont need to copy, because they have the IP, and they pay the western engineer to do the work. -
While erasing chip marking sounds funny, how is that worse than using custom chip markings that many non-chinese manufacturers do for their products?
It is there to prevent simple reverse engineering, or rather blatant copying that is reality in China...
Also I would like to point that phrase"poor software design" might mean software that doesn't work up to it's own specs, or to a software that was done to specs, but specs were not well designed.. I would say that most of chinese T&M equipment shows they have good engineers, but have only 10-15 years of experience making it as opposed to 60+ years for EU or US manufacturers... That shows in the fact that are capable of making good hardware, but not being able to "envision" and therefore design feature set, UI and such, to make instrument both powerful, easy to use, and "cool"...
Also in former USSR and China, mentality "function over form" was prevalent. It could have been ugly and hard to use, it didn't matter. If it did the job, operator was supposed to learn how to use it and live with it shortcomings.. So scope is not complicated, it is operator that is not smart enough...
Some of that mentality is still there, but slowly going away...
And I wouldn't say 20 years... More like 10, judging by the speed of learning so far... Also they buy knowledge and technology at a fast pace, to complement their own research...
-
And in a "professional" rather than "consumer" field? Lenovo maybe, but they inherited a lot from IBM.
Huawei comes to my mind. Afaik they pretty rule the carrier network market, ask Ericsson, Nokia or Siemens... ;-) -
The A-Model has more software (remote-control) features enabled out of the box, a better slew-rate and higher frequency range to use.
Banana-Plugs are not specified for 40 oder 60 Amps. Thats why they are not used in these kind of loads.
Regarding the Range-switch at 29:40: the resolution of the internal display is not changed when switching the range of the device. Terrible. So internal measurement is always done in the 40 Amps range (and not out of spec).
Would you like to see why i sent back my rigol DL3021A? Have a look at this:
https://youtu.be/-55cf6NnObo
(for further testing, or if somebody who owns this device would like to do some own testing, it would also make sense to put a serial power resistor of i.e. 10 Ohms between the load and the power supply. In the set up i used with only the wiring resistance the power supply does a lot of damping in the circuit) -
The A-Model has more software (remote-control) features enabled out of the box, a better slew-rate and higher frequency range to use.
But most people won't be using them at full load most of the time. Omission is annoying.
Banana-Plugs are not specified for 40 oder 60 Amps. Thats why they are not used in these kind of loads.
https://youtu.be/-55cf6NnObo
-
Hi,
My big problem is, if Dave is correct, the use of de IRFP250 Mosfet, this Mosfet is NOT DC specified!
I know there are 6 of them, but it gives me the creeps.
I do NOT trust the SOA of this device...
Kind regards,
Blackdog -
Hi,
My big problem is, if Dave is correct, the use of de IRFP250 Mosfet, this Mosfet is NOT DC specified!
I know there are 6 of them, but it gives me the creeps.
I do NOT trust the SOA of this device...
Kind regards,
Blackodg
Waht do you mean with "not DC specified" ? That its made for switching applications only ?
On the Batronix website of the load http://www.batronix.com/versand/last/DL3021.html, it says the min voltage at the inputs must be 1V... Is that true ? I recall Dave discharging a battery down to sub 1V.
-
Hi Rbastlerm,
Yes, this is a Mosfet for switching applications, and yes, you can use it for DC, but when wil it blowup? no specs...
There is always a minimum voltage spec, mostly depended on how many Ampere used .
Kind regards,
Blackdog -
Hi,
I've used tens of thousands of FETs (without the FBSOA in the datasheet) in DC.
My big problem is, if Dave is correct, the use of de IRFP250 Mosfet, this Mosfet is NOT DC specified!
I know there are 6 of them, but it gives me the creeps.
I do NOT trust the SOA of this device...
Kind regards,
Blackdog
You need to carefully select the model, verify the design, and test each and every unit. And use the proper FET technology, some older are better than the newer like Trenchfet and others.
The only ones failing the ones, where isolation was broken or it wasnt properly mounted on a heatsink.
It's 16x TO247 FETs for only 350W. That is about 21W/FET, with typical 0.9K/W junction to sink. I pushted TO220 much harder than this, its not even going to be warm. Unlike the stupid linear regulators they used. -
I think you can use the knob, because it seems that the digits have underscore selector.
-
Hi NANDBlog,
Sorry, but I do think this is a crazy design, and if I see the oscillation phenomena in the schopi68 video
This has to be a joke.
This Electronic Load is never tested the way it should, when is was "designed".
The phase margin is one of the first things you have to test, before production.
8 or 16 Mosfets they like problems there in China *grin*
Use some Linear Mosfets you wil be amazed and you need only two of them, but the price and the Rdson wil be a problem.
But 4 of these wil do the job : IXTH60N20L2
Kind regards,
Blackdog
-
"My big problem is, if Dave is correct, the use of the IRFP250 Mosfet, this Mosfet is NOT DC specified!"
Eh? The specsheet says 180W dissipation. As usual this only applies if it's welded to the side of the Titanic as a heatsink with obligatory iceberg still in contact, but nevertheless I can't see why it's not 'DC rated' what ever that means.
http://www.datasheetcatalog.com/datasheets_pdf/I/R/F/P/IRFP250.shtml
It also has to be remembered that the temperature of a small heatsink on a device dissipating a watt or two can safely be higher than that of a large heatsink on a device dissipating tens or hundreds of watts. This is because the difference between the actual device temp and the external measurable temp will be less than for a device under high power.
LM317 handling a couple of watts - h/s at 80C, chip at 100C, OK.
TO247 Mosfet direct mounted, handling 140W, h/s at 40C, chip at 180C .. about ready to pop.
BTW, Gandalf says whatever you do, don't touch the Eye button. His buddy Saruman did, and... -
Hi IanMacdonald
Download this PDF: http://www.ixys.com/Documents/AppNotes/IXAN0061.pdf
Look @ page 8 figure 10 tipical switching Mosfet, <= Electro-Thermal Instability Boudary
Look at the dashed line and see the horror...
Kind regards,
Blackdog
PS
My Datasheet is from International Rectifier, but Fairchild shows some DC SOA, ST also
-
Hi IanMacdonald
Download this PDF: http://www.ixys.com/Documents/AppNotes/IXAN0061.pdf
Look @ page 8 figure 10 tipical switching Mosfet, <= Electro-Thermal Instability Boudary
Look at the dashed line and see the horror...
Kind regards,
Blackdog
PS
My Datasheet is from International Rectifier, but Fairchild shows some DC SOA, ST also
I understand your point, but woudnt the load limit itself over the maximum power and therefore never get the six ? IRFP250N over that "electro thermal instability boundary" ?
Edit: Shoudnt the mosfets be viewed as one ? Therefore the graph would be extended in the amp axis. Threfore the instability zone would be far away ?
Just some ideas, that came to mind.