flyingdave my bad about the resistors - I changed 2 and reverted back to the original as they made no difference, so it is just capacitors really.
ATTENTION, DO NOT BOTHER PERFORMING THE MOD BELOW. I just did it and I still have a damn sine wave on my screen when wave out is set to square wave at 6-8MHz (no improvement). How disappointing. This scope is pretty much worthless above 1 MHz. So much for the quad 
Noname, a square wave is not a good merit to justify that the mod does nothing.
After the modification the analog BW rolloff is extended to around 10MHz…
In other words, the effective sampling rate of the quad as well as the analog BW of the front end is not good enough to accurately display a square wave at these frequencies. The “sharp edges” of a square wave (higher harmonics) will be attenuated and you will end up mostly with the fundamental.
If you want to be looking at a fully featured square wave you would probably need a scope with way higher analog BW and sampling rate.
Furthermore, although my B channel still worked after the mod, my A channel was nothing but static. I put C73 back on and this got rid of the static but then there was ZERO signal. Nothing but a flat line no matter what. If Seeedstudio has a spare board or suggestions, please let me know.
The mod was not worth doing, and now my A channel is not working. I will probably need to replace an expensive IC and this device was ridiculously expensive as it is. Is there anyone on the planet willing to bothering opening theirs up and risking likelyhood of the plastic case not sealing completely back up again when they could just sell the damn thing and wait for a MUCH better portable scope somewhere down the road?
As far as any signal “improvement” on the remaining B channel, I see basically none. The limitations at 2MHz and above are obvious before and after the mod. Frankly this leads people to believe it is possible to fix the gaping flaws in this device when it is not. I would almost suspect that bielec may have had his silkscreen labeled with components in different positions, but this seems unlikely.
Noname, I agree that the device specs were overstated - both the analog BW and sampling rate.
However, I disagree that the mod posted by bielec is misleading.
bielec stated:
“I now get about 10MHz sinewave bandwidth on all ranges and good pulse response with minimal overshoot and rounding. Trigger and good display of sub 100nS pulses is now possible. As delivered, the scope had less than 1MHz B.W. and poor risetime which varied depending on range.”
Notice that he mentions sine wave BW, which is what should really be used to justify these results. I already explained the reason why you would not see a nice square wave at 6 or 8MHz.
As a matter of fact I performed very similar mod using 0603 parts and got very similar results - my 3dB was shy of 11MHz. However, I would hardly expact this BW to provide a nice sharp looking 8MHz square wave. My TDS210 would do it with no issue, but the analog BW is 60MHz with 1GS/s sampling rate.
Yes, the QUAD has many flaws - you are correct, but as far as being rediculously expansive I don’t think so. A decent Fluke RMS DVM would cost as much or more and has far less capabilities than even immature QUAD.
In any case, this is just my 2c…
The problem with the quad is that when you are only seeing an decent representation of the waveform you are supposed to see at about 1MHz or less, this is a great limitation to the usefulness of the device. The nano seems to have about the same useful bandwidth, I bought the quad hoping it would be a big improvement over the cheaper nano, it was not. Neither of these devices suit my needs. The wave out on the quad goes up to 8MHz in square wave mode, so out of the box, as is, you would expect it to perform like a good scope should up to at least 8MHz. Sadly, this is not the case. I think the reason why there are no more application or firmware updates to this project is that developers already understand their efforts are better used elsewhere. I wish the merchants selling these (including seeedstudio) would advertise upfront in bold type the 1MHz or less bandwidth limitation so people are tempted into thinking as I had that this might actually be a useful tool for their >1MHz applications.
Some reason stop me to go on with this device:
- bandwidth ,not only the poor limitation but also the non-Smoothness gain ,this make calibration can not be achieved.
- FPGA firmware bug, this case triggering not reliable. No development tools I can get to improve it.
- No way to use a debug tool like J-link to debug MCU sourceware. this make Inefficient to develop an application.
Really hope the bug in Quad can much more less than my in English.
It is now getting to the point where I am going to be asking Seeed for my money back.
This device was clearly misrepresented when sold to me and several months down the line my patience is beginning to wear thin. I believe I have given Seeed plenty of time to resolve some of the very basic problems but they have failed so far to do even that. I have also offered to aid them with suggestions and developed a PC client to make the device more useful.
Perhaps it is not Seeed’s fault that this device is clearly never going to do what it was stated it would but they should have been responsible for making sure of that before selling it to me and many other disappointed clients out there.
For me the most frustrating thing is the fact that they have had considerable amounts of good advice and offers of help to resolve these issues but, for what ever reasons, have chosen not to make use of them.
We seem to be getting firmware updates that replace one bug with another, this is hardly moving forward.
Please Seeed and whom ever else is involved in the development of this product start making some real progress or at the very least inform us clearly of your intent so we (I) can either request a refund or have some time scale for a usable product.
Cheers Pete.
It appears that silence is in fact another form of communication. 
I purposely stayed away for about a month and nothing seems to have changed,
I’ve just read the whole topic and I’m happy that I found this forum. A few hours ago a was very excited about this device and now I see that I was too excited. If there are such a hardware bottlenecks, the product seems not to be mature enough. Now I’m curious if there will be a next hardware release or maybe specification will be updated with the real values.
I am trying to change the opa2354 to LMH6626 in the circuit emulation, (changing opa354 by lmh6624 in tina-ti) for increasing bandwidth. But it somehow does not work. And I can’t find why not.
Is someone willing to look at the datasheet of lmh6626 and circuit and tell me why it is not working, and maybe even tell me how I can get it to work.
this opamp should allow higher than 10mhz bandwidth at gain +10
datasheet: farnell.com/datasheets/1498942.pdf
timbeurskens, LMH6626 is not even remotely close of a substitute for OPA2354 in this case. The specs are all wrong.
For instance, LMH6626 is not rail-to-rail rated op-amp which means that your dynamic range will suffer greatly.
The input bias current of LMH6626 is 7 orders of magnitude greater than OPA2354 which basically makes it unfit for this specific application (front end).
LMH6626 is only stable with gain > 10 and DSO quad uses configurations with gains lower than 10 which will more than likely lead to weird behavior (read oscillations and such)…
Simply putting a wide BWP amplifier without proper PCB layout can be quite interesting (ringing, noise, full out oscillations… you name it). I can only speculate what the PCB layout looks like on the DSO quad.
There isn’t a simple fix for the limited BW of the DSO Quad. It has to be rather redesigned along with couple other things, such as implementing a real trigger…
I hope this was useful 
Mark
!
I really wish I’d seen this thread before buying a DSO203.
Everyone is selling these as 72MHz analog bandwidth.
My analysis of the schematic (hardware 2.6 from the Wiki) is that the last Op Amps just before the ADC are single pole lowpass filters with a maximum -3dB frequency of 4.4MHz. That is about right for 36Msps per channel.
Using the built-in signal generator: On a 100MHz BW Tektronix TDS 220, the signal generator looks like a reasonably good square wave all the way up to 8MHz. With the Quad with only Channel A active, at 4MHz it looks like a sine wave with a minor flattening on the tips. At 6MHz, on the Tektronix it looks like a decent square wave, while on the Quad it is a sine wave with lower voltage. At 8MHz, on the Quad the amplitude is much lower.
So with my unscientific tests and past experience, I’d call this 4MHz analog bandwidth. I consider that useable for audio troubleshooting, and for digital signals no faster than 100kHz.
Any confirmation of bandwidth? I have one that is HW version 2.72. Does that mean it has two 72Msps channels, or is it an older one with two 36Msps channels?
A reviewer at EDN magazine says he measured a 3MHz bandwidth.
http://www.edn.com/design/test-and-measurement/4422846/15/Try-an-oscilloscope-for-under–200
There is my upgrade!
Hi,
My DS203 is harwdare version 2.73.
I just made the modification suggested by bielec:
For ch A.
Remove C73
C11=1.5pF
Across R17 3.3pF
Across R21 22pF
C9=1.5pF
Across R15 3.3pF
Across R19 12pF
For ch B.
Remove C74
C12=1.5pF
Across R18 2.2pF
Across R22 15pF
C10=1.5pF
Across R16 2.2pF
Across R20 12pF
I can confirm a bandwidth of about 11MHz at all input gain stages. This is the maximum possible bandwidth with the used amplifier. At low sensitivity higher bandwidth is possible.
At frequencies higher than 1 MHz there is notable jitter because of unstable trigger I think. Maybe this can be overcome with an improved software.
At the digital inputs I found that rectangle signals with an amplitude of 5V can be measured up to a frequency of about 4 MHz.
Greets
Ingo.
That’s great. For a 72Msps scope, 11MHz is probably pushing it. Although I think later HW versions use two ADCs, can they double up on one channel and do 144Msps? That would make 11MHz bandwidth very usable.