-
Becouse there is old "rewiev" thread what have started with older version and many information and tests there are now obsolete... so I want open new thread for this new generation SDS7102.
These tests here and opinions are made with new factory improved SDS7102V, version V2.4.1
Do not mix old version things to this thread - please! (exept direct explained comparison and well marked what is old version)
This oscilloscope do not need hack or modify for better frequency response.
Frequency response is amazing good in this class. Capturing (ADC) quality is also very good (in this class).
Today I do not know any serious competitor!
Factory have done it allready. So peoples do not need take risk for damage they oscolloscope anymore and also factory have not then "warranty" faild after peoples have broken they scopes. Do not open, you can not do any useful things there!
Also in some places I have seen some information how to make modifications for front end and some other parts inside of scope construction. No need anymore! Factory have allready made these. (no VGA cable inside, adapter board do not go under battery compartment (adapter board is totally new) etc... )
Whole front end construction have changed (and main board). New models of BNC connectors go directly to main board. Solution is mechanically good. Also it is electrically good. Channels have also now better RF isolation. Solution is also more reliable.
But what is also good. It have clearly better frequency response. Lot of better!
Also in this phase they have added LAN (this have not so much advantages exewpt if need use long distange watching or connect many oscilloscopes to teacher board in class room etc. But more or less useful, it is there.
Here first two pictures about risetime. same signal but other have infinite persitence for show that wobbling is not bad and it is amazing stabile trigged in this class of scopes.
Then nice that ALT trig really work and it have also two freq counters.
Then picture about front end construction.
(this is littlebit obsolete - metal pillar what seen top right corner of picture is changed now in production for also improve signal quality (in CH1). Metal pillar affect signal quality - yes!)
And then... cursors are -3dB levels if compare to 1MHz level. (generator have enough good flatness for this test, HP8644B and 1 - 350MHz come with same level as 1MHz. (flatness (level accuracy to well matched 50ohm is inside +-0.1dB. This matching with scope is not ok, this may add some small errors but not markable for this purpose now here.)
Then 300MHz normal realtime sampling.
Then same 300MHz but now infinite prsistence so that it tell littlebit how much there is "wobbling".
And last one 350MHz with infinite persistence.
This is OWON!
Keep fun with modifying other names...
... this do not need!
Later more tests...
Nice risetime in 100MHz chinese oscilloscope! (pulse generator have around 70ps risetime)
-
I would buy one of these Owons but I've seen the insides of these scopes and the varients and I haven't been all to impressed. Not terrible, but not the best.
But these improvements look decent. -
Here first two pictures about risetime. same signal but other have infinite persitence for show that wobbling is not bad and it is amazing stabile trigged in this class of scopes.
The risetime looks slow for the frequency response you indicate, I would expect close to 1 ns for the 300 MHz bandwidth you seem to measure. It appears that something funky is going on with the frequency response, either the roll-off is shallower than -20dB/decade, or they do DSP on the signal to improve apparent frequency response. What's your impression?And then... cursors are -3dB levels if compare to 1MHz level. (generator have enough good flatness for this test, HP8644B and 1 - 350MHz come with same level as 1MHz. (flatness (level accuracy to well matched 50ohm is inside +-0.1dB. This matching with scope is not ok, this may add some small errors but not markable for this purpose now here.)
So the ADC is sampling at just four times the bandwidth in single channel mode, and even less in dual channel mode. This means you have to watch for aliasing again unless the filter roll-off is extremely steep (something at odds with the high rise time), even at the fastest sweep speeds, just like the early DSOs. Making sure that you have no frequencies beyond the Nyquist frequency is all fine and dandy as long as you're measuring signal generators, but not so easy with a real DUT. -
Here first two pictures about risetime. same signal but other have infinite persitence for show that wobbling is not bad and it is amazing stabile trigged in this class of scopes.
The risetime looks slow for the frequency response you indicate, I would expect close to 1 ns for the 300 MHz bandwidth you seem to measure. It appears that something funky is going on with the frequency response, either the roll-off is shallower than -20dB/decade, or they do DSP on the signal to improve apparent frequency response. What's your impression?And then... cursors are -3dB levels if compare to 1MHz level. (generator have enough good flatness for this test, HP8644B and 1 - 350MHz come with same level as 1MHz. (flatness (level accuracy to well matched 50ohm is inside +-0.1dB. This matching with scope is not ok, this may add some small errors but not markable for this purpose now here.)
So the ADC is sampling at just four times the bandwidth in single channel mode, and even less in dual channel mode. This means you have to watch for aliasing again unless the filter roll-off is extremely steep (something at odds with the high rise time), even at the fastest sweep speeds, just like the early DSOs. Making sure that you have no frequencies beyond the Nyquist frequency is all fine and dandy as long as you're measuring signal generators, but not so easy with a real DUT.
With this measuring accuracy and common simple thumb rule (.35 rule) for risetime this measurement is just inline. Also this risetime number is oscilloscope itself measured and with around 40% of 0-255 (8bit) signal level.
I have NOT measured 300MHz -3dB bandwith. With Tektronix originated thumb rules typical 200MHz "gaussian BW" oscilloscope have risetime 1.75ns. (yes I made also risetime measurement with fast "spike" pulse. Oscilloscope show <1ns risetime and this kind of bullshit I do not publice)
Aliasing.
1ch use. Yes if signal includes over 500MHz frequency components it have aliasing and here Owon is not alone. It depends what is over 500MHz signal total attenuation and what is over 500MHz frequency fractions level in measured signal.
2ch use just change 500MHz to 250MHz. This is natural.
But try find oscilloscope what keeps these frequency limits down to 500us/div or 1ms/div. Owon keeps 1GSa/s or 500MSa/s to these horizontal speeds.
If example take Hantek and Owon Single channel use.
Hantek Nyqyuist Nyquist
2ns/div. 500 MHz, Owon 500 MHz
20ns/div. 400 MHz, Owon 500 MHz
200ns/div 400 MHz, Owon 500 MHz
400ns/div 200 MHz, Owon 500 MHz (500ns/div)
2us/div 200 MHz, Owon 500 MHz
20us/div 200 MHz, Owon 500 MHz
40us/div 200 MHz, Owon 500 MHz (50us/div)
80us/div 100 MHz, Owon 500 MHz (100us/div)
200us/div 50 MHz, Owon 500 MHz
400us/div 50 MHz, Owon 500 MHz (500us/div)
800us/div 25 MHz, Owon 250 MHz (1000us/div)
2ms/div 1 2.5 MHz, Owon 125 MHz
Old digital scopes main problem was lack of capture memory. If old 500MSa/s oscilloscope (300MHz analog BW) have 1k capture memory. What is sample speed if it is example 1us/div speed and if it have 20div, maybe 50MSa/s and Nyquist 25MHz and front end analog part widely open up to 300 MHz without markable attenuation - yes there is really easy coming aliasing problem. What is Nyquist frequency, how about aliasing problem if compare example to Owon. Time difference is 20 year and price difference is maybe 30:1.
Imagined "old" 1k memory digital scope if it have single shot 1GSa/s:
And there is really high posibility to meet so bad aliasing that it may destroy all measurements without knowledge how to avoid or check if there is aliasing. Now if this imagined oscoilloscope have gaussian BW with (-3dB) corner frequency 100MHz we can of course draw curve and think how much there are "area" for aliasing with different horizontal speed settings.
(I assume that like typical tektronix displayed 10div and total capture is 20div)
50ns/div Nyquist 500MHz
100ns/div 250MHz
200ns/div 100MHz
500ns/div 50MHz
1us/div 25MHz
2us/div 12.5MHz
5us/div 5MHz
10us/div 2.5MHz
20us/div 1.25MHz
50us/div 500kHz
100us/div 250kHz
200us/div 125kHz
500us/div 50kHz
[/font]
Owon have not this but most of competitors have!
Then, with multi ADC interleaving system there come more problems. Example if there is 10pcs 100MSa/s interleaved ADC's for make "one" 1GSa/s ADC. IF timing is perfect and they are perfectly matched with levels, and only then, there is not problem if also they (and signal roads to them) are well matched over range. But becouse they are not in practice. Every these 100MSa/s ADC Nyquist is 50MHz. If they are not perfectly clocked and matched they produce lot of problems (becouse you do not know what is exatly time point related to other samples + level differents including whole frequency area) . Then it need heavy digital signal handling for cleaning, washing, whitening and dressing and never you exatly know what is truth. Maybe you clean out some truth but maybe not clean out some lies... then you can call it "garbage generator".
Agen for aliasing:
Also this fast rising test edge includes lot of high frequency components! (original risetime on the generator output is around 70ps but with this cable and matching sure there is not these most high frequencies to Oscilloscope BNC. (They all are attenuated in this coaxial cable. If really want connect this fastest risetime to DUT, it need waveguide connection. Not common coaxial cable.
But so or so... This thinking about aliasing is one reason why I show two pictures. Single shot AND multi shot (using persistence to collect many shots together for display).
As we know aliasing do "corner wobbling" in signals what are not sinewave. Most high if signal is just fast rising square wave and specially if it include frequency components over Nyquist.
This Owon do not show high amount of corner wobbling! can you see it, can you tell that there is showed high aliasing problem now?
-
All this discussion on bandwidth is above my knowledge, but I have one question, maybe a little
ingenuous, how are the included probes? How is the frequency response using them?
I think I'll buy this scope (as anyone experience with develissimo?) and I'll use mainly the
included probes often at 1:1 and 10:1 and 1M input.
Also having vga output it would be fantastic if owon and others could use the full
lcd's square centimeters, for example the dso screen for measurements and menu
and vga for the waveforms output, with all the info (full capture and magnified zone, math etc...)
that can be seen all toghether. But I understand that a fancy GUI eats fast the
computing power and the programming and debugging time.
-
All this discussion on bandwidth is above my knowledge, but I have one question, maybe a little
ingenuous, how are the included probes? How is the frequency response using them?
I think I'll buy this scope (as anyone experience with develissimo?) and I'll use mainly the
included probes often at 1:1 and 10:1 and 1M input.
Also having vga output it would be fantastic if owon and others could use the full
lcd's square centimeters, for example the dso screen for measurements and menu
and vga for the waveforms output, with all the info (full capture and magnified zone, math etc...)
that can be seen all toghether. But I understand that a fancy GUI eats fast the
computing power and the programming and debugging time.
Very roughly and badly but quickly out from lab measured with T5100 probe included with SDS7102
Using medium quality HP signal gen, coaxial cable, 50ohm terminator and probe adapter what connect probe tip to center and around tip GND area to cable (adapter) GND (not used probe GND wire and clip)
-3dB point (compared to 1MHz) somewhere around 160MHz with 10:1 setting and around 10MHz with 1:1 setting.
1:1 setting with these kind of probes is mostly with highly rejected BW.
Probe specs can read: (system) BW 10:1 100MHz and 1:1 6MHz.
1:1 setting is "good" low pass filter if need, it is always on the hand
But also it need remember that probe capacitance is high with 1:1.
BW with probe is related to how probe characters matching to signal under test. If probe tip capacitance is example 20pF it can look reactance chart... for 100MHz it is really far away from 10Mohm or 1Mohm (dc resistance). (Capacitive reactance Xc is around 80 ohm!)
Always need also remember that probe change more or less things in device/signal under test.
How to know what is signal without connecting your measuring equipment to it.
So, always more or less measuring is system measuring... can not measure pure DUT. Becouse measuring nearly always affect to DUT. Only with always study and learn more and with high amount of experience, it can make mutual opinion what is maybe meaningful and what is maybe not and still making error is more than easy. -
BTW, just off topic sidenote.
They are fun in develissimo... LAN extra price, VGA extra price... price without VAT...
LAN, VGA 398 eur. VAT20% = 477.
Free shipping make it still good in price.
But how they can do this LAN pricing becouse LAN is factory default. It is today always there.
-
BTW, just off topic sidenote.
They are fun in develissimo... LAN extra price, VGA extra price... price without VAT...
LAN, VGA 398 eur. VAT20% = 477.
Free shipping make it still good in price.
But how they can do this LAN pricing becouse LAN is factory default. It is today always there.
They told me that have still some old units, and will soon update the menu.
It's hard, at least to me, find a better price around, maybe I didn't search enough. -
I have NOT measured 300MHz -3dB bandwith.
250MHz (the -3dB cursors are touching the signal), my bad. 1.4 (theoretical) vs. 1.6 ns is quite close.(yes I made also risetime measurement with fast "spike" pulse. Oscilloscope show <1ns risetime and this kind of bullshit I do not publice)
I wasn't questioning your methods, I'm convinced that you have the correct knowledge and equipment to perform these measurements.Aliasing.
Owon is not alone in being limited by Nyquist-Shannon, but much less of that > 500 MHz signal will make it through the eg. Rigol front-end compared to the large bandwidth Owon front-end. Roll-off is there for a reason, if the -3 dB point is closer to the Nyquist frequency, this is usually compensated by using a steeper roll-off (eg. brick-wall response).
1ch use. Yes if signal includes over 500MHz frequency components it have aliasing and here Owon is not alone. It depends what is over 500MHz signal total attenuation and what is over 500MHz frequency fractions level in measured signal.But try find oscilloscope what keeps these frequency limits down to 500us/div or 1ms/div. Owon keeps 1GSa/s or 500MSa/s to these horizontal speeds.
Agreed, the large memory allows it to achieve its max. sampling rate over a much wider range of horizontal speeds. But with the scopes with a better sampling rate/BW ratio you can at least increase the horizontal sweep speed to check for aliasing. Not so if even at the highest sampling rate, frequencies beyond the Nyquist frequency make it past the filter.Old digital scopes main problem was lack of capture memory. If old 500MSa/s oscilloscope (300MHz analog BW) have 1k capture memory. What is sample speed if it is example 1us/div speed and if it have 20div, maybe 50MSa/s and Nyquist 25MHz and front end analog part widely open up to 300 MHz without markable attenuation - yes there is really easy coming aliasing problem.
I think the problem was worse in the time of the 300 MHz / 100 MS/s or scopes (eg. Tek 24xx), this generation taught people to distrust digital scopes because they are susceptible to aliasing, and to always keep an analog scope on your bench. The high sampling rate, short memory scopes (Tek loved to sell those, while HP sold scopes with long memory and not enough sampling rate) also have issues, but you can check by increasing the sampling rate. Peak detect also helps to prevent aliasing, at least the areas it's unable to sample fast enough are shaded.Also this fast rising test edge includes lot of high frequency components! (original risetime on the generator output is around 70ps but with this cable and matching sure there is not these most high frequencies to Oscilloscope BNC. (They all are attenuated in this coaxial cable. If really want connect this fastest risetime to DUT, it need waveguide connection. Not common coaxial cable.
I would expect a good BNC connector and coaxial cable to be OK up to 100 ps rise time or so.This Owon do not show high amount of corner wobbling! can you see it, can you tell that there is showed high aliasing problem now?
The step response looks good, maybe their interpolation helps? -
How to find out what version of SDS7102 somebody sell???
I presume that there is lot of "old" SDS7102 out in the market. -
How to find out what version of SDS7102 somebody sell???
I presume that there is lot of "old" SDS7102 out in the market.
Some sellers know what they are selling, some do not know or care.
Only seller can answer this, he must know what he is selling. If seller do not know what he is selling maybe it is not best place to buy. Do not make "blind" "buy it now" button purchase if you think it is ipmportant to you what you really want.
But then, also old version is just as good what it have been! And it is amazing good also in this price class.
New and old is classified as 100MHz and new version do not have not any change in its specifications data.
If someone buy SDS7102 "new" version and its frequency response -3dB point is not over 200MHz: In this case, can not claim a refund or repair. No one has formally promised more than the specifications data. It is just as gift in this time - just as long as it is. -
Check this other thread for special noise problem in this new improved unit. I write it there becouse before there was one question and propably this people have this issue. (but only just - maybe!)
Problem have solved in Owon very fast but there may be some units what are "early production" about this improved model and some of these units may suffer small CH1 noise problem and some of these units are now on the end user table.
In this case (if really have this problem AND if it is meaninful to your use of unit) take contact to your seller for special instructions how to do.
Or ask seller they have checked that units have tested for this issue/or updated HW for solve this.
(if unit is new improved model where is LAN and what have this new front end construction also)
SDS71021152xxx -->> do not have this issue (and before this (but improved model) only some units show this problem.
Owon service can do very easy this small HW update. -
Some peoples have asked questions about new SDS7102 oscilloscope CH1 noise issue.
This issue is only in some manufactured new oscilloscopes just after Owon started this new improved version. Late 4Q2011 made improved versions do not have this issue.
Owon find fast this problem and they changed construction just littlebit.
Short time after new improved version start, there was possible that some units have
rised noise level in CH1.
Read both Owon threads to find how to regognize if you have CH1 noise issue.
If you are not with enough experience and skill for do this work please do not open and start this repair!
Best solution is: Send your oscilloscope to Owon accepted service lab!
What ever reason is and you exactly want do it your self, please follow this Guide.
Becouse Owon continuously develop these products, this Guide and pictures may be obsolete what ever time.
If there is some errors, I am not responsible about any problems. You do all with just your own risk!
And REMEMBER: You may loose warranty! You are fully responsible if any problem arises when you open an oscilloscope, and everything you do.
Link: Here is link for Illustrated Guide for disassemble and assemble SDS7102V
This must not use for other versions, only for these improved versions what have default LAN
and what have improved front end construction.
There is also picture for repair CH1 noise problem and possble problem in backside GND connector (screw)
Also there is partially info for disassemple front panel board.
This CH1 special noise problem is ONLY in some new version units what are shipped out before Owon find this problem and make small change in production (this change is here also illustrated - we disassembled one good unit for study new model repair service practices, just for exercise becouse it is our job and then in real situation there do not need keep thumb in mouth, and same time fixed back side GND connector). Also have tested "reverse" so that modify one good scope back to "noisy" construction for fully understand this situation.
(this published partial guide is not full service guide, it have only helpful information related this issue)
If you can not get easy this Owon original nylon pillar you can do your own (DIY).
You can make it. Take around 5.5mm Nylon rod. Exactly 14.9-15.0mm long. Ends cutted exactly 90degree.
Both ends, drill well centered exactly axial direction hole, diameter 2 - 2.5mm, depth around 4-6mm depends screws you use.
(use example small sparse thread screws. Good lenght is around 5-6mm.)
Maybe this DIY solution is as good or even better than Owon original pillar?
If you do not have this material you can also find these from many electronics old stuffs.
Please do not link this guide from any other internet sides/forums etc. Or I need lock it.
-aghp
-
Tested littlebit power supply - just for fun.
Power supply in specs:
100?240 VAC RMS, 50/60Hz, CAT II
-------
What happend with DC.
It works normally starting around 50VDC
Around 45VDC power led light, unit try start but shut off before boot up picture, and it try agen ... continuous loop. (not tested how long it last before damage.. )
around 50VDC start and run normally. Between 50VDC and 125VDC just normal.
Not tested more high DC becouse lack of DC source and higher volatges was not so interesting.
But it only show that it works with 60 and 120VDC system.
(note! this was not full test. The test is not in any way rule out the possibility that the device failure can result from such use. If you try, you are just alone responsible... )
Of course with DC power there is not 50/60Hz trig pulses so "line" trig is not possible.
-
Some tiny test with FFT just for looking what is dynamic and where is noise floor
Here:
Signal source HP8644B
Frequency around 100MHz
Level. For oscilloscope full display but not anything over for avoid clipping distortion.
Signal with coaxial to scope and ended with Tektronix 50ohm.
Scope setting: Input CH1 20mV/div AC.
Capture normal, no average.
Persistence infinity.
FFT, zoomed 5x so that one div is 5MHz, center 100MHz and scale 10dB/div, Hanning.
1. signal itself ( around 72mVrms) (reference level)
2. FFT same signal
3. FFT same settings but signal level dropped 60dB and now around 72uVrms
4. FFT -30dB under first reference level.
persistence used for better visualize noise peak floor.
Result: not bad for cheap oscilloscope.
signal level was set near full screen but still clearly below clipping level. This mean that test littlebit underestimate performance.
Also this lower signal level is still clearly visible over noise floor. But if look real noise peak level it can not go more if need be sure there is signal and not random noise peak. But, what do persistence. Well if there is continuous signal still persitence can help regognize signal from noise. (btw, Owon persistence is not only one brightness level, it also helps when looking display in live. This effect can not well visualize in still picture)
-
What about the SDS8102? Does that support 200 MHz too? I'm on the fence between the SDS8102 and the SDS8202 where the only difference appears to be the 100 MHz vs 200 MHz bandwidth and $100 difference in price. Are they the same inside?
-
As far as i know the most significant difference is the screen reso .
The input circuitry should be similar but the i believe other then that would be different . -
Some kind of religion?
As far as i know the most significant difference is the screen reso .
As far as I do not know anything I do not answer anything. This is good "rule".
Truth:
All these SDS series oscilloscopes have same screen resolution 800x600.
other things:
7102 Have 1GSa/s maximum for 1 channel and 500MSa/s maximum for 2 channel use.
7102 have full speed memory maximum 10Msample per channel.
it means full 1GSa max from 500us/div to 2ns/div
Officially this is named as 100MHz oscilloscope.
With new improved HW -3dB BW have rised over 200MHz. (this is not factory guaranteed but measured reality with many units)
With old HW it was around 140-150MHz.
Also 8102 and 8202 have this improved new HW and I do not know they real BW today.
I have tested only old HW 8102 and 8202.
This time I test (old HW) 7102 was around 140-150MHz -3dB BW
This time I test (OLD HW) 8102 and 8202 they -3dB BW was clearly over 200MHz (220-240MHz but it was old HW!)
As far as I know, I do not know anything about new HW 8102 and 8202 real BW.
(I can believe if I am religious people that it can easy be same or better than 7102 true measured BW.
What I know, there is even possible to control by FW this analog BW by adjusting (digitally adjusted) analog amplifier)
8102 and 8202 have maximum 2GSa/s for one channel and 1GSa/s for two channel.
But this max is usable only with 1 and 10kSample memory. 100k to 10M memory is limited to same speed as 7102.
If need 200MHz oscilloscope what really can use 200MHz BW with two channel simultaneously it need this 8102 or 8202 samplerate. (2 channel use 500MSa/s is enough for 100MHz but not enough for 200MHz and this is not related to Owon, it is related to basic oscilloscope fundamentals)
You can find enough real tests from this forum for 7102.
You may imagine what is different if "7102" new generation front end is littlebit adjusted up and then after front end is 2GSa/1GSa ADC compared to 7102 1GSa/500MSa ADC.
Other features are all same in SmartDS series.
-
*facepalm* Actually i was looking at the MSO8202
-
As far as I know the MSO8202 is the "old old old" hardware and it not based on the SDS8202. I would get a separate Logic analyzer and 7102 or SDS8x02.
-
8102 and 8202 have maximum 2GSa/s for one channel and 1GSa/s for two channel.
But this max is usable only with 1 and 10kSample memory. 100k to 10M memory is limited to same speed as 7102.
Wow, that's disappointing. It's not like I would really need that much memory, but it would be nice to have. -
8102 and 8202 have maximum 2GSa/s for one channel and 1GSa/s for two channel.
But this max is usable only with 1 and 10kSample memory. 100k to 10M memory is limited to same speed as 7102.
Wow, that's disappointing. It's not like I would really need that much memory, but it would be nice to have.
This memory is not important only for "how long capture can make".
Maybe even more important is other reason. In many cases, it does not sufficiently recognize the importance, or its importance is underestimated.
Memory size and speed is highly related to samplerate and becouse this, also to aliasing what was really big disadvantage with older small memory digital oscilloscope. They may have nice samplerate if time/div setting is fast. But just after you drop horizontal speed also samplerate drops, of course.
How about 1k memory? Yes it do same in Owon if set 1k memory. And all oscilloscopes do this with low memory.
If 1k memory and memory is divided to 20 horizontal divs it can do 10MSa/s if time setting is 5us/div. No meaning if in this case there in front panel read Agilent, Tektronix or even Rohde&Scwarz.
time/div 500us samplerate is 100kSa/s (absolute max frequency is now 50kHz ! ALL frequancy components in signal over this 50kHz produce aliases. (it need remember that only pure sinewave have single frequency. If you have square wave it have its base frequency + lot of harmonics and these go more and more high depend of risetime)
with 10k it is 1MSa/s
with 100k it is 10MSa/s
with 1M it is 100MSa/s
with 10M it is 1GSa/s (yes Owon have full speed deep memory)
and it is nearly like this, even if oscilloscope is Agilent/Tektronix or Rohde&Schwarz
So you can think what happend if someone have example littlebit older Tektronix where was 2.5k maximum capture memory.
More memory, more fast it can sample with lower horizontal speeds.
Look what is maximum samplerate with Rigol DS1000E series if setting is example 1us/div or just this 500us/div or same for Hantek DSO5000B series.
It can tell. I do not need long memory but I want high samplerate also for lower horizontal speeds.
And becouse my longlong history I have used tens of years analog oscilloscopes. They are always full BW with all speeds.
This is one reason why I like Owon. It feels littlebit same.
But it is very pity in Owon SDS series 7102 and 8x02 that they have only maximum 1G speed deep memory and using deep memory, samplerate is limited. It is today also clearly stated in newest user manual specifications.
(in this price class it is still very good) So, only 7102 can talk that it have full speed deep memory!
-
Yeah, on page 85 of the User Manual, it spells this out. What's even worse is that the SDS9302, which claims to have 3.2 GS/s sample rate, can only use 10M of memory when operating at <= 800 MS/s single channel and <= 400 MS/s in dual channel! I guess 10k is all the memory they have on their FPGA, and their off board RAM can't go fast enough.
-
Yeah, on page 85 of the User Manual, it spells this out. What's even worse is that the SDS9302, which claims to have 3.2 GS/s sample rate, can only use 10M of memory when operating at <= 800 MS/s single channel and <= 400 MS/s in dual channel! I guess 10k is all the memory they have on their FPGA, and their off board RAM can't go fast enough.
Yes, 2x10k.
It is hard road to transport datastream from ADC to memory. How much can HW total price be. Who can do this whole oscilloscope from scratch to retail boxed ready oscilloscope with USD200 or less (factory "zero" price) including all peoples salaries, taxes, investments, components, and developing costs. How much money you can put to develop some asic type or other special integrated fast solution for this datahandling from ADC. This is bottleneck. With double price it can do, up to some level. But it is not easy case.
Btw. I have try find information about some Agilent oscilloscope model. What are true samplerates with different memory selections. Megazoom adapt automatically optimum settings.. nice and exact?
-
If they can interleave ADCs, they shouldn't have any problem interleaving external RAM. Adding a few extra RAM chips wouldn't add that much cost. Maybe you've looked at the board, but I would guess layout would be the only believable bottleneck with having to get all of those extra traces to and from the FPGA. I think it's not something they really cared that much about, just like with the firmware that I've read complaints about.