Dave reminds me of myself when I was a child. I would get a new RC car and the first thing I did was take it apart, then spent a week putting it back together so I could actually use it :D
+EEVblog -- Noooo! Not another oscilloscope. You and oscilloscopes is worse than women and shoes. When are you getting back to finding the fault of that Samsung LCD TV? Fault-finding is what oscilloscopes were made for.
At 31’ 30” from the video, the oscilloscope was set to 4 channels in operation, the Timebase was 5ns/div, and the sample rate was 250MSPS (sample period was 4ns). Under these settings, waveforms were plotted by sin(x)/x interpolation points. The display seemed to show small periodical noise, which was caused by round-off error from the interpolation point calculation. The minor inaccuracy is far less than the specification of DC gain accuracy (approximately 3%). The related problem from the video has been corrected by beta version firmware. Official version firmware will be released in early March. Dave stopped the above-mentioned waveform and zoomed in to observe it. (Vertically zoomed in 100x, horizontally zoomed in 20x) When Zoom in waveforms and the real sample points are not sufficient to reveal the original waveform, GDS-1000B will adopt sin(x)/x interpolation method to plot the original waveforms. In which, the displayed waveform was expanded from one step to more than one hundred steps (4~ 5 div). Oscilloscope’s AD Converter has only 8 bits resolution. Therefore, waveforms observed under this circumstance are meaningless. With the same operational procedures, oscilloscopes from Tek and Keysight showed similar waveforms. This situation is the same as continuing to zoom in an image to see a mosaic image. GW Instek
Has GW Instek considered offering a upgradable firmware that is open sourced (differing from your main scope software) so that you can still protect your IP, but I would like to develop a new build for dev purposes. Not so much as a scope, but multimeter, or reconfigurable lab instrument etc. taking advantage of the Xilinx SoC. Your scope has an advantage to many in that it's easier to develop when you have to deal with one system IC as opposed to FPGA + ARM..
This is why I always say you should friggen turn it on at least once to make sure everything is working so you can rule out anything you fucked up after taking it apart.
My university still uses GW Instek GOS-6050 oscilloscopes... Cant find any production date of them anywhere, would be interested to know when that model came out.
I can't understand all the whining about the layout and size - I like it - roomy for the fingers, simple and clean and the yellow is rather nostalgic :P Not everything has to be cram packed with the controls so close together that the eyes are always straining to pick one out from the other while the fingers are always hitting an adjacent control. Very nice in my opinion.
+Digger D Yes, this was my feeling too. And people are bitching about the colour of the plastic too. White is VERY ok in a lab environment, but take something like that out into the field a few times, and it's gonna show every single smudge
+TheChipmunk2008 Oddly the product photos on their website seem to have quite a different dark blue/light grey color scheme rather than the dark grey/beige on Dave's GDS-1104B. www.gwinstek.com/upload/website/cc9d1f20eda3d9aa214276327f3eba4d.jpg
A well-designed PCB can reduce the need for external shielding. Ground layers in close proximity to signal traces can work extremely well. There’s a lot of jumping to conclusions and raw speculation in this tear down. The proof is in the pudding, so to speak ;-)
They are using positive air flow to move the air, the fan is only there to increase the air pressure in the cavity. When you take the shield off it's not being used as a heatsink. I bet the only thing that needs cooling is the power supply, and the slight increased air pressure is enough to do it.
Dave, may be it is more reasonable to play around with it before tear down? I'm a little bit worried that the performance can be affected by re-assembly...
Designed very well, IMHO and uniquely. GWInstek has been around for over 40 years and I can't recall in that time, their devices not living up to its spec sheets once released to the market, and built to last. Their main problem had always been functionality vs cost compare to competitors. Before the Rigol and Owons, they battled against Leader, Trio, B&K, Radio Shack, Heathkit etc., and are still around, and many prior competitors, aren't.
A dual-core Cortex-A9 is no slouch, interesting choice. I bet it costs more than a TI Sitara plus separate FPGA, but as you demonstrated the results speak for themselves. Very responsive. I think they've gone with the separate processor board because it's (going to be) used in other products, they can pump out a higher volume of the processor board to potentially lower their overall costs across their entire line of products. Just plug it in to a different device with different programming.
I'm glad I have the Rigol DS1054Z (maxed out, of course), however as you mentioned, it's good to see some competition in this price range as well. I prefer the Rigol because of the matte screen, glossy LCDs are big no for me on test equipment. The only positive thing about the GW are the separate controls for each channel, but even that won't make me sell my Rigol and get a GW instead...
The Rigol has one big advantage over the Instek. All of the software that makes the MUCH more expensive Rigol's in that series much more expensive are already present in that Rigol. They're just software locked. There still exists several websites where the more expensive license key can be generated turning that Rigol into a $1000.oo machine. If you thought ahead and bought a plus version you can also get the 16 Digital channels once they're unlocked if you buy the digital probe cable.
The two chan is only 300usd,. not bad at all! lots of bang for bucks. Seriously considering buying this one, my old one is a Philips PM3267 so yeah overdue for a upgrade (it's actually older then i am..) Heck, the PM3267 goes for around the same amount @ ebay second hand. Oh well, practically free upgrade! Always welcome
For a cheap scope this is fine! I like the larger layout. I hate occidentally nudging nearby pots when adjusting another. I like the separate pots. I like the colour but that scheme is not the only one available. All in all I think this a brilliant scope for the beginner and enthusiast. My first CRO ever was a cheap charlie GW and it was hammered with use for over 30 years and still going strong until stolen.
received my GDS-2102E days ago, an amazing oscilloscope, Amazing large screen, LOT of nice and cool features, super responsive, super fast. I also like the screen saver if the oscilloscope is not in use. Best deal.
I have no idea what you are anyone in the comments are talking about, but I love this channel. I am going to beige watch until I understand! Used to love tearing stuff apart when I was young. As a journeyman machinist who has worked with many an EE this is awesome stuff!
I really like the separate vert controls. And it uses the entire screen for waveform display. Whereas the Rigol wasted the sidebars by not making them hide able. ..All in All I like it !
Very nice teardown. Love it!! When I look at the display side with the "wasted space" then it looks like there is space for a multimeter and logic analyzer extension. The 4 round holes and the one rectangular hole are suspicous. That Zync part costs around $90 at Digikey, wow. Would not have expected such an expensive part in such a low price test gear.
I like the look of the GW Instek to be honest... that 'wasted space' means it's easy to grab the button /knob you need. Depends how pressed for space you are I guess.
+EEVblog - That "processor" runs an ADP5052 for it's local power regulation - a chip specifically designed for running Spartan 6 FPGAs. Spartan 6's also have inbuilt LVDS drivers, so can directly drive an LCD. This should hopefully explain what's going on as between it's inbuilt DDR2/3 memory controller (up to two chips, as shown on this board) and inbuilt LVDS drive, this looks for all the world like it has to be a Spartan 6 on there. Interesting, though, that they got away with using it to power a Zynq..
This was an interesting teardown. A lot was different on this scope then others. Mostly not for the better but interesting with the seperate prosessor board. I was afraid that we would not get to see / know about the prosessor / FPGA but Dave fixed it for us:)
This is the future of complex electronic devices, FPGA and multiple cores on one chip. It will take a strong, diverse development team to bring products to market but the close integration on the same piece of silicon will make this technology the winner in the long run. The next gen Zynq (UltraScale+ MPSoC) should be out soon. Still kind of a solution in search of a problem !! Xilinx needs a version with less cores on more on chip RAM and Flash.
+tesla500 Possibly. But overall (separate assembly runs, both double side load, different BOMs etc) I'm not sure that would really save you anything. Esp if you make the board as small as possible.
The neon is the ARM equivalent to the SSE instructions, and they boost the floating point math speed quite a bit, specially on stuff with vectors and matrixes.
I have and love this scope. I have yet to figure out how to use all of the trigger modes and features and would benefit from a real world application of same. I found this video (for the second time after some years) while searching for how to use in hopes of learning the features I have not yet figured out (or UN learning my mistaken ideas).
I found this video again after searching for a schematic with out success. Not quite brave enough to take mine apart as I am dependent on it. Schematic anyone?
I think one of the latest video game console's uses a similar air flow scheme. The reviewer said they did it that way so they could use a higher air flow volume at lower speed.
Jamicon is a Taiwanese brand cap that is mostly known in the audiophile circle for being the caps loved by Brit firm Musical Fidelity. I've swapped quite a few dried up Jamicons from MF machines before, MF love to run their amp with a lot of bias current, and that really did a number on the Jamicon caps. Not sure if they have managed to keep a better seal now.
I do the whole "leaving the heatsink out thing" on a daily basis, what sucks more is then putting it back together for the second time, only to notice you have a spare CMOS battery on your desk lol.
GW-Instek is much higher in quality compared to the RIGOL, I saw many videos where it shows weakness in measurements for the RIGOL. I am using the GW-Instek since years, Just ordered a new one, arriving in a week. I also have no problem regarding the PSU quality, as long as it works perfectly. so its pointless to talk about the size, or the fan.
Hi Dave, they might be original Nippon-Chemicon Caps, but I know for a fact that Samyoung has used the same logo, even with Samyoung written inside of the logo, so I would not trust them 100% on it beeing actual original Nippon Chemicon caps. (type mentioned here is a Samyoung photo flash cap type CE P 350SV290uF) Jamicon is a fairly standard reasonable quality cap for consumer products here in the Netherlands. I haven't tested them but the products that use them are of reasonable consumer quality. (Amplifiers, function generators (Velleman), Toslink DAC to Cinch etc.) Keep up the great video's, love them!
I'm willing to bet an Esinomed tablet that this is a prototype unit. Production units probably won't have the glitchy display,, they will have the screws in the right place, fan will be pointed the other way and the Nippon caps replaced with Reliable CapXon.
I think GW Instek is hedging their bets. The same core module, especially this single Zynq design, can be used in a very wide variety of things so they can run those in bigger volume. Also since the Zynq have both ARM and FPGA at on place they can effectively attach the ADC directly to the AMBA of the ARM core, allowing better throughput than going through some standard interface.
I have a GW INSTEK GDS1202B, two channel 200 MHz version, and I can't seem to get FFT to work. I'm sure I'm not getting the settings right or something stupid like that but their manual is a bit scant. I also have a TEK 2440 and I like the controls a bit better just a lot less bells like FFT to mess with. Great teardown, loved it.
Good video, Dave It's a fairly clean design, as I have been advocating, you can do super simple layouts these days. And I would argue you could reduce that much further still. All that metal enclosure thinking is quite irrational. They are not building a tank. Of course you can get rid of all the buttons as well and the accompanying boards and just make it the obvious tablet form factor and get rid of all the irrational legacy thinking. I would also like to see isolated channels. ADCs are so cheap now and shared ground is really unprofessional. You could do 1GS/s on the first channel and less on the rest. The power supply is also very old school thinking. That should be a tiny resonant converter on the board and no trafo or lyt caps. Resonant high frequency is so beautiful.
You've got to be joking me. Handicapping a device even further with worse layouts, devices these days with their touch screens are hard enough to use, also the metal is for shielding, resonant power supply? What the hell do you mean with that? Have you got any experience with a how a power supply should even work?
+Barnacules Nerdgasm Thanks, will do. Young kids make it hard to get across the pond unfortunately. I'm always having to turn down paid trip offers :-(
+Tom OConnor htere is nodecision, rigol is so much better, then with the software crack you realy have a nice machine... only for the serial decoding capabilities and rigol wins...
I wouldn't be surprised if that periodic display anomaly turned out to be related to the almost total lack of screening inside the thing. No metal can around the PSU? What were they thinking?
+EEVblog Maybe something weird with the intensity display dithering that they're doing when they're interpolating at such fast speeds. If the samples are really flat and they chose a bad method of drawing the lines that expects a curve instead of a flat line it could be choosing to display weird at the "corners".
+Imre Fabian I don't think the thick finger argument is that compelling. My fingers are thick, I have the Rigol, and I never thought "gee, these buttons are too close". What I do like is the separate buttons for each channel, as the Rigol layout, while functional, is annoying when trying to change the vertical on multiple channels. Not to mention pushing the variable button to select. I hate that.
I really love ya teardowns Dave, and I guess it's your field of expertise, but watching oscilloscopes, signal generators, etc getting torn down just gets kinda samey.
Product review please! :) Also I'm not sure where you get these prices from. This is the price for the 50Mhz model. GDS-1104B goes for about $740 in the US. GDS-1054B cost €880 at element 14. Could you point a seller that holds better prices?
+Evan Cirelli Yeah I saw the price on tequipment.net Sadly my local official GW Instek distributor asks about €550 VAT incl. for this model. Rigol 1054Z goes for €408 VAT inc. here.
+Retep4565 IMO the Rigol will always be the better choice because it can be hacked to higher BW and all other sort of goodies. However if the buyer isn't concerned with hacking the Instek could be a real competitor. One other fact is that this scope's price is much higher in Europe than Rigol's.
Those four plastic bosses below the display at 12:00 look suspiciously like 4mm sockets for a DMM option. The square cutout to the left, digital channels maybe?
Hi dave I was thinking you can do a video series like power supply design, but design an oscilloscope this time i am pretty sure it would be the most complicated and fun thing in the world. keep the good work up!
Uhm, I think Jamicon are actually quite high-quality caps. I've been using them for years because our local electronics shop sells them and I never had a problem with them :)
+Tobias Mädel (Manawyrm) yeah, i never heard of any problems with them. maybe dave just had a brain fart or something they arent tier 1 japanese, but they arent "one hung low" brand either.
+Tobias Mädel (Manawyrm) They're made by a Taiwanese company that's been around since the early '70s, and I can't remember having to replace any of their caps.
Hi Dave, Are you sure that the fan doesn't move the air from the inside out, because that would make more sense. It would draw the air from the side vent over the power supply and out the back. Thanks for sharing...
Hi Ardian, It seemed to me, that if the fan turns clockwise, it would be pushing air toward the outside of the case and the angle of the blades suggest that as well. Thanks for your input.
+Darryl Landry Small fans like this (e.g. PC case fans) traditionally blow away from the "top" of the fan and toward the frame supports. At 8:10 you can see the top of the fan and no supports, so it probably pulls air into the case.
Can you give a basic description of what these testing devices do for us nubs that just like watching your channel? I've never used an oscilloscope, and would just like to know how to use them generally. I get they can be used instead of a multimeter for a lot of applications, but where would you use them instead of a multimeter, and what information do they give that you would act on vs a multimeter?
Solder looks pretty good, solid - even the re-work/ touchup, PCB looks clean, didn't se any micro solder balls or flux on the PCB's. The pulses or bits on the channels are they some type of gnd loop from "under shielding" or a PCB layout? Would a filter on the channel output clean that up?
Would love to see an more in depth practical use review of this scope :) Im curious about that, what i guess could be a interpretive high res feature at ~ 34:35 ? :)
was going to get the rigol 1054z but i will have to take a look at this one hope to see a video on you useing it and your impressions compared to rigol. my b&k 15mhz scope circa 1977 finally died :(
+iron man Second on the rigol, that is a very known quantity (i.e. hackable) whereas who knows how long you'll have to wait to get your free 50MHz on the GW
You spent some time discussing capacitor brands, can you throw some light on how to identify good brands for passive components? E.g. Design calls for 10pf cap. Digikey lists a dozen+brands... How to figure out which are quality?
By reputation of the brand (Rubycon, Chemicon, Elna, etc), and then by reading the "Group chart" that they all publish. The Group Chart is the visual matrix that shows which series of their caps has lower impedance, longer life, lower cost, higher temperature, etc etc and how they compare to each other. You also need to have an understanding of which type of capacitor is best for each application. Ceramic, Mylar, poly, electrolytic, etc each have their advantages and disadvantages. There is no perfect device, each is a compromise, and engineering is the careful analysis and selection of which compromises will be acceptable. The #1 factor for component parts in almost all real-world production environments is always COST, irrespective of the selling price of the final finished device.
Nippon Chemicon caps in the PSU... but secured with that horrible glue that goes conductive after heat/time. I'm surprised they still even make the stuff!
For low price it has bright LED's under the buttons at least so you can see the the state of the buttons under bright light. Could the yellowish front panels be glow in the dark material? That would explain the odd color.
was almost set on the Rigol 1054Z but it seems so slow on the screen update. All the people who use the Rigol...u dont mind that?...twist a knob ...half second later it updates the screen...how can u use that? I would love to see a more in depth review on this one!
most important question: can you software upgrade the 50 to the 100 ? anyways that totally different processor board design looks weird, maybe they have some very generic module that the put in everywhere? or even bought from somewhere else? it even has a different colour
Competition is a GOOD thing ! Processor and FGA one one chip is definitely the way to go ! Rigol is going to have to change their strategy !! I would love to see someone come out with a 4 channel, 12 bit scope.
Those NCC caps in the PSU are bit funny color to be genuine NCC. I've seen lots of fake NCC caps in the past and those have the same crusty look on them. Can't be sure though...
Is this the teardown the size of the bench you were talking about? ;P kidding ... another great teardown! ... im not very happy with the construction and will be sticking with the rigol!
Seems that the apps are downlaodable from their website. They have an DVM "app.", filteringing "app."and a logger "app.". Seems to be an interesting starting-point of some hacking ;-)
hi Dave. I'm going to buy an oscilloscope (400-500$ budget) and stopped my selection on: 1. GDS1102B - because of 100mhz 2. Rigol 1054Z (4 ch?) 3. Analog one (still can't decide which one, Hitachi or Tektronix or which else. Can you please advise and help to decide? P.S: I'm in Baku, Azerbaijan and working with power supplies (for example Alpha series by TDK Lambda), AOM drivers by Panasonic, laptop motherboards and other stuff.
Totally off topic here, but help would be appreciated- Im building a circuit with an LM3914 and the first segment is dimly lit- when I add other segments the first one its still lit in each segment >:( Other than that a circuit works fine. My solution is to put a resistor (4.7k) to positive. It will only add 15mW to the chips dissapation. Will this be okay??? I dont mean to be starting anything on youtube and i know this is a fourm queston I just wanted (perhaps) a fast responce- and what other than youtube!
Hi dave i have , im not a professional in electronics, but i found that most "affordable" oscilloscopes ranges between "50 and 200Mhz", though i found on the internet that most usefull radio frequencies ranges from "300 Mhz to 3 Ghz" (Television broadcasts, LAN, bluetooth, GPS, amateur radios, Mobile phones), or even from "3 to 30 Ghz" like (satellite television broadcasting, LAN, amateur radio), and oscilloscope that has a frequency up to these ranges costs a fortune, does this means that "ordinary people" can't analyze these frequencies or am i missing something.
thank's, but for someone studying radio and tv transmission stuff, wouldn't need a more powerfull oscilloscope, i mean, radios and tv's are not really some advanced stuff, how did they to study their waves in the old days.
Did you check the crosstalk between channels of this scope? That 31:36 is crosstalk due to single adc multiplexing..... all these junks are equal. I can make a better scope at home in that price
hey dave, or anyone who has an opinion, im looking for an entry level scope. Would you recommend a 50mhz with 4 channels or a 100mhz with 2 channels? If it makes a difference I would be using it for audio as well as microcontroller circuits.
Dave reminds me of myself when I was a child. I would get a new RC car and the first thing I did was take it apart, then spent a week putting it back together so I could actually use it :D
+Barnacules Nerdgasm I never grew out of it, now I'm married with two kids and do it for a living :->
+EEVblog -- Noooo! Not another oscilloscope. You and oscilloscopes is worse than women and shoes.
When are you getting back to finding the fault of that Samsung LCD TV? Fault-finding is what oscilloscopes were made for.
+The Kaiser Hehe the Samsung LCD TV was interesting, indeed.
well ain't TH-cam connected
+EEVblog Glad you're not a doctor or biologist, you'd've taken the kids apart...
At 31’ 30” from the video, the oscilloscope was set to 4 channels in operation, the Timebase was 5ns/div, and the sample rate was 250MSPS (sample period was 4ns). Under these settings, waveforms were plotted by sin(x)/x interpolation points. The display seemed to show small periodical noise, which was caused by round-off error from the interpolation point calculation. The minor inaccuracy is far less than the specification of DC gain accuracy (approximately 3%). The related problem from the video has been corrected by beta version firmware. Official version firmware will be released
in early March. Dave stopped the above-mentioned waveform and zoomed in to observe it. (Vertically zoomed in 100x, horizontally zoomed in 20x) When Zoom in waveforms and the real sample points are not sufficient to reveal the original waveform, GDS-1000B will adopt sin(x)/x interpolation method to plot the original waveforms. In which, the displayed waveform was expanded from one step to more than one hundred steps (4~ 5 div). Oscilloscope’s AD Converter has only 8 bits resolution. Therefore, waveforms observed under this circumstance are meaningless. With the same operational procedures, oscilloscopes from Tek and Keysight showed similar waveforms. This situation is the same as continuing
to zoom in an image to see a mosaic image. GW Instek
Has GW Instek considered offering a upgradable firmware that is open sourced (differing from your main scope software) so that you can still protect your IP, but I would like to develop a new build for dev purposes. Not so much as a scope, but multimeter, or reconfigurable lab instrument etc. taking advantage of the Xilinx SoC. Your scope has an advantage to many in that it's easier to develop when you have to deal with one system IC as opposed to FPGA + ARM..
This is why I always say you should friggen turn it on at least once to make sure everything is working so you can rule out anything you fucked up after taking it apart.
+seephor Same thought I had - were those noise issues there BEFORE the teardown? We'll never know.
+seephor Years ago with any new bit of kit it was check it worked then take it apart to check it would continue to work.
+Seephor & Maxiskol S : I totally agree with you. Dave should have done a pretest before the usual teardown - I also think so .
We have come a long way since those green screen oscilloscopes that we used when I was in high school.
+Physics Videos by Eugene Khutoryansky not to mention prices that are amazing those days :)
+robertpk I bet that this thing won't last 5 years.
My university still uses GW Instek GOS-6050 oscilloscopes... Cant find any production date of them anywhere, would be interested to know when that model came out.
+BloodySword I bet it will
+Physics Videos by Eugene Khutoryansky We still use them at the "Materials for Electronics" lab.
11:50 the fuse is the red block next to the thermistor it says T 2A 250V
+Erik Bruijn Agreed.
I can't understand all the whining about the layout and size - I like it - roomy for the fingers, simple and clean and the yellow is rather nostalgic :P Not everything has to be cram packed with the controls so close together that the eyes are always straining to pick one out from the other while the fingers are always hitting an adjacent control. Very nice in my opinion.
+Digger D Yes, this was my feeling too. And people are bitching about the colour of the plastic too. White is VERY ok in a lab environment, but take something like that out into the field a few times, and it's gonna show every single smudge
+TheChipmunk2008 Oddly the product photos on their website seem to have quite a different dark blue/light grey color scheme rather than the dark grey/beige on Dave's GDS-1104B. www.gwinstek.com/upload/website/cc9d1f20eda3d9aa214276327f3eba4d.jpg
I agree other than the lack of shielding. How important ? Ehhh who knows, it realky depends on the environment used in i suppose..
A well-designed PCB can reduce the need for external shielding. Ground layers in close proximity to signal traces can work extremely well. There’s a lot of jumping to conclusions and raw speculation in this tear down. The proof is in the pudding, so to speak ;-)
They are using positive air flow to move the air, the fan is only there to increase the air pressure in the cavity. When you take the shield off it's not being used as a heatsink. I bet the only thing that needs cooling is the power supply, and the slight increased air pressure is enough to do it.
Dave, may be it is more reasonable to play around with it before tear down? I'm a little bit worried that the performance can be affected by re-assembly...
I actually like the styling of this scope. For my taste it "looks" nicer than the Rigol on the outside. On the inside is another story.
Designed very well, IMHO and uniquely. GWInstek has been around for over 40 years and I can't recall in that time, their devices not living up to its spec sheets once released to the market, and built to last. Their main problem had always been functionality vs cost compare to competitors. Before the Rigol and Owons, they battled against Leader, Trio, B&K, Radio Shack, Heathkit etc., and are still around, and many prior competitors, aren't.
New to all of this so I really, really appreciate all the hard work. I will be working my way through lots of videos I am sure. Cheers!
Robert
A dual-core Cortex-A9 is no slouch, interesting choice. I bet it costs more than a TI Sitara plus separate FPGA, but as you demonstrated the results speak for themselves. Very responsive.
I think they've gone with the separate processor board because it's (going to be) used in other products, they can pump out a higher volume of the processor board to potentially lower their overall costs across their entire line of products. Just plug it in to a different device with different programming.
I'm glad I have the Rigol DS1054Z (maxed out, of course), however as you mentioned, it's good to see some competition in this price range as well.
I prefer the Rigol because of the matte screen, glossy LCDs are big no for me on test equipment. The only positive thing about the GW are the separate controls for each channel, but even that won't make me sell my Rigol and get a GW instead...
It is incredible to show us what is inside this device.
The Rigol has one big advantage over the Instek. All of the software that makes the MUCH more expensive Rigol's in that series much more expensive are already present in that Rigol. They're just software locked. There still exists several websites where the more expensive license key can be generated turning that Rigol into a $1000.oo machine. If you thought ahead and bought a plus version you can also get the 16 Digital channels once they're unlocked if you buy the digital probe cable.
Apparently the is way to hack the 50mhz instetek to 100mhz but i didn't find much about it apparently
Gotta love those tear-downs and repairs where you realize you have some spare parts just when you thought you were done.
The two chan is only 300usd,. not bad at all! lots of bang for bucks.
Seriously considering buying this one, my old one is a Philips PM3267 so yeah overdue for a upgrade (it's actually older then i am..)
Heck, the PM3267 goes for around the same amount @ ebay second hand. Oh well, practically free upgrade! Always welcome
For a cheap scope this is fine! I like the larger layout. I hate occidentally nudging nearby pots when adjusting another. I like the separate pots. I like the colour but that scheme is not the only one available.
All in all I think this a brilliant scope for the beginner and enthusiast.
My first CRO ever was a cheap charlie GW and it was hammered with use for over 30 years and still going strong until stolen.
received my GDS-2102E days ago, an amazing oscilloscope, Amazing large screen, LOT of nice and cool features, super responsive, super fast. I also like the screen saver if the oscilloscope is not in use. Best deal.
I have no idea what you are anyone in the comments are talking about, but I love this channel. I am going to beige watch until I understand! Used to love tearing stuff apart when I was young. As a journeyman machinist who has worked with many an EE this is awesome stuff!
I really like the separate vert controls. And it uses the entire screen for waveform display. Whereas the Rigol wasted the sidebars by not making them hide able. ..All in All I like it !
I'm loving the reassembly sound effect.
Very nice teardown. Love it!! When I look at the display side with the "wasted space" then it looks like there is space for a multimeter and logic analyzer extension. The 4 round holes and the one rectangular hole are suspicous. That Zync part costs around $90 at Digikey, wow. Would not have expected such an expensive part in such a low price test gear.
I like the look of the GW Instek to be honest... that 'wasted space' means it's easy to grab the button /knob you need. Depends how pressed for space you are I guess.
+EEVblog - That "processor" runs an ADP5052 for it's local power regulation - a chip specifically designed for running Spartan 6 FPGAs. Spartan 6's also have inbuilt LVDS drivers, so can directly drive an LCD.
This should hopefully explain what's going on as between it's inbuilt DDR2/3 memory controller (up to two chips, as shown on this board) and inbuilt LVDS drive, this looks for all the world like it has to be a Spartan 6 on there. Interesting, though, that they got away with using it to power a Zynq..
This was an interesting teardown. A lot was different on this scope then others. Mostly not for the better but interesting with the seperate prosessor board. I was afraid that we would not get to see / know about the prosessor / FPGA but Dave fixed it for us:)
Yo yo, Jamicon is actually well known in the industry and they're pretty good :)
This is the future of complex electronic devices, FPGA and multiple cores on one chip. It will take a strong, diverse development team to bring products to market but the close integration on the same piece of silicon will make this technology the winner in the long run.
The next gen Zynq (UltraScale+ MPSoC) should be out soon. Still kind of a solution in search of a problem !!
Xilinx needs a version with less cores on more on chip RAM and Flash.
Separate processor board so the large main board can be lower tech (2 layer?)
+tesla500 Possibly. But overall (separate assembly runs, both double side load, different BOMs etc) I'm not sure that would really save you anything. Esp if you make the board as small as possible.
Perhaps they are going for a modular design, and they can use the same processing board across several different products.
+tesla500 Any videos in the foreseeable future? :)
Yep, some more teardowns coming up soon.
+tesla500
I saw this cool gun looking thing on your desk in one of your videos, you should do a teardown on it
The neon is the ARM equivalent to the SSE instructions, and they boost the floating point math speed quite a bit, specially on stuff with vectors and matrixes.
I've seen enough after 15 minutes and I'm getting the Rigol!
I have and love this scope. I have yet to figure out how to use all of the trigger modes and features and would benefit from a real world application of same. I found this video (for the second time after some years) while searching for how to use in hopes of learning the features I have not yet figured out (or UN learning my mistaken ideas).
I found this video again after searching for a schematic with out success. Not quite brave enough to take mine apart as I am dependent on it. Schematic anyone?
I think one of the latest video game console's uses a similar air flow scheme. The reviewer said they did it that way so they could use a higher air flow volume at lower speed.
There's a low hum in the audio. Could be my end but it's definitely there. Thought you might like to know.
Jamicon is a Taiwanese brand cap that is mostly known in the audiophile circle for being the caps loved by Brit firm Musical Fidelity.
I've swapped quite a few dried up Jamicons from MF machines before, MF love to run their amp with a lot of bias current, and that really did a number on the Jamicon caps. Not sure if they have managed to keep a better seal now.
I do the whole "leaving the heatsink out thing" on a daily basis, what sucks more is then putting it back together for the second time, only to notice you have a spare CMOS battery on your desk lol.
GW-Instek is much higher in quality compared to the RIGOL, I saw many videos where it shows weakness in measurements for the RIGOL. I am using the GW-Instek since years, Just ordered a new one, arriving in a week. I also have no problem regarding the PSU quality, as long as it works perfectly. so its pointless to talk about the size, or the fan.
Hi Dave, they might be original Nippon-Chemicon Caps, but I know for a fact that Samyoung has used the same logo, even with Samyoung written inside of the logo, so I would not trust them 100% on it beeing actual original Nippon Chemicon caps. (type mentioned here is a Samyoung photo flash cap type CE P 350SV290uF)
Jamicon is a fairly standard reasonable quality cap for consumer products here in the Netherlands. I haven't tested them but the products that use them are of reasonable consumer quality. (Amplifiers, function generators (Velleman), Toslink DAC to Cinch etc.)
Keep up the great video's, love them!
separate processor board allows them to use it across different models and maybe even different equipment...
The fan is OK as long as the static pressure is enough to supply continues airflow.
I'm willing to bet an Esinomed tablet that this is a prototype unit. Production units probably won't have the glitchy display,, they will have the screws in the right place, fan will be pointed the other way and the Nippon caps replaced with Reliable CapXon.
+z1power I'll take that bet :)
i hate it when they put those useless nippon caps in. Stop being cheap and buy genuine capXon!
I think GW Instek is hedging their bets. The same core module, especially this single Zynq design, can be used in a very wide variety of things so they can run those in bigger volume. Also since the Zynq have both ARM and FPGA at on place they can effectively attach the ADC directly to the AMBA of the ARM core, allowing better throughput than going through some standard interface.
I have a GW INSTEK GDS1202B, two channel 200 MHz version, and I can't seem to get FFT to work. I'm sure I'm not getting the settings right or something stupid like that but their manual is a bit scant. I also have a TEK 2440 and I like the controls a bit better just a lot less bells like FFT to mess with. Great teardown, loved it.
Good video, Dave
It's a fairly clean design, as I have been advocating, you can do super simple layouts these days. And I would argue you could reduce that much further still. All that metal enclosure thinking is quite irrational. They are not building a tank.
Of course you can get rid of all the buttons as well and the accompanying boards and just make it the obvious tablet form factor and get rid of all the irrational legacy thinking.
I would also like to see isolated channels. ADCs are so cheap now and shared ground is really unprofessional. You could do 1GS/s on the first channel and less on the rest.
The power supply is also very old school thinking. That should be a tiny resonant converter on the board and no trafo or lyt caps. Resonant high frequency is so beautiful.
You've got to be joking me. Handicapping a device even further with worse layouts, devices these days with their touch screens are hard enough to use, also the metal is for shielding, resonant power supply? What the hell do you mean with that? Have you got any experience with a how a power supply should even work?
i think the multiple boards is a way to standardize/lower cost between models
2ch 70MHz one sounds VERY good to me (I do mostly low frequency analog troubleshooting stuff) so... yes. TY for the info
If you are ever in the United States specially the west coast, look me up. I'll buy you a pint!
+Barnacules Nerdgasm Dave is not a drinker... it will have to be milk : )
+TheRogerx3 Make it a rice milk protein drink please!
+Barnacules Nerdgasm Thanks, will do. Young kids make it hard to get across the pond unfortunately. I'm always having to turn down paid trip offers :-(
+Barnacules Nerdgasm damn you everywhere I look
+EEVblog glad to see your not an alcholic Dave it seems to be sort of an epidemic amongst youtubers
Which to choose for a low end scope ? The next assessment video might make the decision a bit easier.
+Tom OConnor htere is nodecision, rigol is so much better, then with the software crack you realy have a nice machine... only for the serial decoding capabilities and rigol wins...
Maybe those processor boards might be shared between various scopes.
It looks like a really nice oscilloscope
That fan only moves the still air to keep the air temperature ambient. "To keep it from heat soaking".
I wouldn't be surprised if that periodic display anomaly turned out to be related to the almost total lack of screening inside the thing. No metal can around the PSU? What were they thinking?
+Paul “LeoNerd” Evans No, I think it's a display processing algorithm issue, not a sampling/noise issue.
+EEVblog Maybe something weird with the intensity display dithering that they're doing when they're interpolating at such fast speeds. If the samples are really flat and they chose a bad method of drawing the lines that expects a curve instead of a flat line it could be choosing to display weird at the "corners".
Dave, about the controls layout: I like it as it is, better than the Rigol, because of Ergonomics; I've got thick vingers!
+Imre Fabian I don't think the thick finger argument is that compelling. My fingers are thick, I have the Rigol, and I never thought "gee, these buttons are too close". What I do like is the separate buttons for each channel, as the Rigol layout, while functional, is annoying when trying to change the vertical on multiple channels. Not to mention pushing the variable button to select. I hate that.
Interesting that the diff pairs for each channel to the ADC don't look like they're length matched. May not matter for 100MHz.
I really love ya teardowns Dave, and I guess it's your field of expertise, but watching oscilloscopes, signal generators, etc getting torn down just gets kinda samey.
Product review please! :) Also I'm not sure where you get these prices from. This is the price for the 50Mhz model. GDS-1104B goes for about $740 in the US. GDS-1054B cost €880 at element 14. Could you point a seller that holds better prices?
+Evan Cirelli Yeah I saw the price on tequipment.net
Sadly my local official GW Instek distributor asks about €550 VAT incl. for this model. Rigol 1054Z goes for €408 VAT inc. here.
+1, I don't mind the teardown, but I'd rather like to see a review of this unit and see how it compares to the Rigol DS1054Z first.
+Retep4565 IMO the Rigol will always be the better choice because it can be hacked to higher BW and all other sort of goodies. However if the buyer isn't concerned with hacking the Instek could be a real competitor. One other fact is that this scope's price is much higher in Europe than Rigol's.
Those four plastic bosses below the display at 12:00 look suspiciously like 4mm sockets for a DMM option.
The square cutout to the left, digital channels maybe?
Hi dave I was thinking you can do a video series like power supply design, but design an oscilloscope this time i am pretty sure it would be the most complicated and fun thing in the world. keep the good work up!
Uhm, I think Jamicon are actually quite high-quality caps. I've been using them for years because our local electronics shop sells them and I never had a problem with them :)
+Tobias Mädel (Manawyrm)
yeah, i never heard of any problems with them. maybe dave just had a brain fart or something
they arent tier 1 japanese, but they arent "one hung low" brand either.
+Tobias Mädel (Manawyrm) They're made by a Taiwanese company that's been around since the early '70s, and I can't remember having to replace any of their caps.
Hi Dave,
Are you sure that the fan doesn't move the air from the inside out, because that would make more sense. It would draw the air from the side vent over the power supply and out the back. Thanks for sharing...
+Darryl Landry From the blade contour, I think it's pushing inward.
Hi Ardian,
It seemed to me, that if the fan turns clockwise, it would be pushing air toward the outside of the case and the angle of the blades suggest that as well. Thanks for your input.
+Darryl Landry Small fans like this (e.g. PC case fans) traditionally blow away from the "top" of the fan and toward the frame supports. At 8:10 you can see the top of the fan and no supports, so it probably pulls air into the case.
The design just seems bass ackwards to me! Thanks for your feedback
Can you give a basic description of what these testing devices do for us nubs that just like watching your channel? I've never used an oscilloscope, and would just like to know how to use them generally. I get they can be used instead of a multimeter for a lot of applications, but where would you use them instead of a multimeter, and what information do they give that you would act on vs a multimeter?
Dave: Jump to conclusion then correct himself. Rinse, repeat.
Not always. Sometimes he jumps to conclusion, and doesn't correct himself.
36:10 The artefacts are exactly in position of samples (dots) may be it's intentional.
Solder looks pretty good, solid - even the re-work/ touchup, PCB looks clean, didn't se any micro solder balls or flux on the PCB's. The pulses or bits on the channels are they some type of gnd loop from "under shielding" or a PCB layout? Would a filter on the channel output clean that up?
if you take your pass/fail and put in a noise source wouldn't that give you your waveform updates per second from the pass/fail output?
+Alyx BioHaz Not unless it has zero overhead pass/fail hardware detection (unlikely), otherwise it will be slower and not an indication.
Would love to see an more in depth practical use review of this scope :) Im curious about that, what i guess could be a interpretive high res feature at ~ 34:35 ? :)
was going to get the rigol 1054z but i will have to take a look at this one hope to see a video on you useing it and your impressions compared to rigol. my b&k 15mhz scope circa 1977 finally died :(
+iron man don't boder, just take the rigol...it's so much better...
+iron man Second on the rigol, that is a very known quantity (i.e. hackable) whereas who knows how long you'll have to wait to get your free 50MHz on the GW
You spent some time discussing capacitor brands, can you throw some light on how to identify good brands for passive components?
E.g. Design calls for 10pf cap. Digikey lists a dozen+brands... How to figure out which are quality?
By reputation of the brand (Rubycon, Chemicon, Elna, etc), and then by reading the "Group chart" that they all publish. The Group Chart is the visual matrix that shows which series of their caps has lower impedance, longer life, lower cost, higher temperature, etc etc and how they compare to each other.
You also need to have an understanding of which type of capacitor is best for each application. Ceramic, Mylar, poly, electrolytic, etc each have their advantages and disadvantages. There is no perfect device, each is a compromise, and engineering is the careful analysis and selection of which compromises will be acceptable.
The #1 factor for component parts in almost all real-world production environments is always COST, irrespective of the selling price of the final finished device.
Nippon Chemicon caps in the PSU... but secured with that horrible glue that goes conductive after heat/time.
I'm surprised they still even make the stuff!
For low price it has bright LED's under the buttons at least so you can see the the state of the buttons under bright light. Could the yellowish front panels be glow in the dark material? That would explain the odd color.
Should have bought the GW-Instek. I hate the layout of the Rigol, the multi-function knob, and that you can not hide the menus.
was almost set on the Rigol 1054Z but it seems so slow on the screen update. All the people who use the Rigol...u dont mind that?...twist a knob ...half second later it updates the screen...how can u use that? I would love to see a more in depth review on this one!
cool vid as always dave =)
This is almost double in price compared to the Rigol. How is it cheaper?
so dave, which do you recommend as first oscilloscope this one or the rigol DS1054Z (for beginners) ?
most important question: can you software upgrade the 50 to the 100 ? anyways that totally different processor board design looks weird, maybe they have some very generic module that the put in everywhere? or even bought from somewhere else? it even has a different colour
man i was so keen on this & even looking forward to modding the fan but that noise artifact kinda makes me a bit eeeh :/
38:00 Looks like you can use USB or the internet or something to install those apps. Why else would it ask for a Remote Disk?
11:55 you say there is no fuse protection in it? What is the red plastic thing? "T2A"... seems to be a fuse.
Competition is a GOOD thing !
Processor and FGA one one chip is definitely the way to go ! Rigol is going to have to change their strategy !! I would love to see someone come out with a 4 channel, 12 bit scope.
Any new Rigol comming up or any other crazy value item ? Need to buy a new scope asap.
That warm white plastic on the GW looks awful. Looks like the bromide has seeped out!
+Frank Sandqvist That's an old bromide if I'm not mistaken.
Those NCC caps in the PSU are bit funny color to be genuine NCC. I've seen lots of fake NCC caps in the past and those have the same crusty look on them. Can't be sure though...
Just imagine scopes as thin as TVs are nowadays! New xyz scope, whisbang specs, thin as air!!!!
Is this the teardown the size of the bench you were talking about? ;P
kidding ... another great teardown! ... im not very happy with the construction and will be sticking with the rigol!
Seems that the apps are downlaodable from their website. They have an DVM "app.", filteringing "app."and a logger "app.". Seems to be an interesting starting-point of some hacking ;-)
hi Dave.
I'm going to buy an oscilloscope (400-500$ budget) and stopped my selection on:
1. GDS1102B - because of 100mhz
2. Rigol 1054Z (4 ch?)
3. Analog one (still can't decide which one, Hitachi or Tektronix or which else.
Can you please advise and help to decide?
P.S: I'm in Baku, Azerbaijan and working with power supplies (for example Alpha series by TDK Lambda), AOM drivers by Panasonic, laptop motherboards and other stuff.
Does it do any i2c or spi bus decoding stuff? Maybe thats what the app stuff will be for.
At last a decent scope that I can afford. I will treat myself after Christmas so thanks for the Teardown Dave.
Different boards to avoid digital noise into the analog?!
Totally off topic here, but help would be appreciated- Im building a circuit with an LM3914 and the first segment is dimly lit- when I add other segments the first one its still lit in each segment >:( Other than that a circuit works fine. My solution is to put a resistor (4.7k) to positive. It will only add 15mW to the chips dissapation. Will this be okay???
I dont mean to be starting anything on youtube and i know this is a fourm queston I just wanted (perhaps) a fast responce- and what other than youtube!
Hahahaha, no warranty void sticker.......therefore gw instek approves of these activities!
Yes there is a fuse in the power supply.
Hi dave i have , im not a professional in electronics, but i found that most "affordable" oscilloscopes ranges between "50 and 200Mhz", though i found on the internet that most usefull radio frequencies ranges from "300 Mhz to 3 Ghz" (Television broadcasts, LAN, bluetooth, GPS, amateur radios, Mobile phones), or even from "3 to 30 Ghz" like (satellite television broadcasting, LAN, amateur radio), and oscilloscope that has a frequency up to these ranges costs a fortune, does this means that "ordinary people" can't analyze these frequencies or am i missing something.
+Alfred William "Ordinary project" are almost all about toggling microcontroller pins at
thank's, but for someone studying radio and tv transmission stuff, wouldn't need a more powerfull oscilloscope, i mean, radios and tv's are not really some advanced stuff, how did they to study their waves in the old days.
Did you check the crosstalk between channels of this scope? That 31:36 is crosstalk due to single adc multiplexing..... all these junks are equal. I can make a better scope at home in that price
Dave, you missed the weird looking 7-pin SO-8 chip (19:00).. Whats up with that one?
so Dave, "warranty void if NOT removed"? :D hahahah manufacturers would go berserk with that label
hey dave, or anyone who has an opinion, im looking for an entry level scope. Would you recommend a 50mhz with 4 channels or a 100mhz with 2 channels? If it makes a difference I would be using it for audio as well as microcontroller circuits.
good for a first scope for home hobby use?
+Adrian Higgins Certainly!
+EEVblog Ah, but better than the Rigol 1054Z?
+Mark Dennehy Better but you cannot hack it (yet).
I pine for a scope this size. Mine at work is a big analog, CRT HP scope.