DS1052E Up to 150Mhz

[Simon]

well to be honest I'd want to see 10 samples at least for "captures"

[tinhead]

sure, 10x oversampling is better than 5x, and 20x better than 10x ... the necessary oversampling
depends at least on signal form, capture mode.

For continuously signals reduced amplitude due bw limitation is worser than 5x oversampling,
but i'm sure you know this.

[dimlow]

yea, lay your cards on the table or stop bragging 

To be honest who would think they could do 150MHz ? I mean they only went up to a 100MHz model. Maybe the 150MHz version was planned but never came to light. of course we also need to remember that at 100 MHz with 1GS/s (on one channel only) that's the ideal min of 10 samples per waveform, at 150MHz that's 6.7 samples per wave, although still fine on repetitive stuff

The info for doing it was in the first post here, you just change the Model Number as before. There is a 150Mhz version of the scope that is not sold in the rest of the world. That's where i got the info from!

Just search the net for DS1152E-EDU.

Its a no Brainer

[alm]

you can get easily higher than this, 5 samples per wave is common value giving of on single chan up to 200MHz and dual chan 100MHz.
Rigol maybe never did it because of planned and now available 2GSs models, or maybe because the DSP was too slow to handle proper
the data (you remember the two chans bug - running with long mem and 100MHz).

Sample rate closer to the Nyquist frequency does require a steeper filter roll-off of the input filter (eg. brick-wall response) and better interpolation (wasn't there a post about interpolation errors some time ago?). Maybe they compensated for poor interpolation and filtering by increasing the sampling rate?

[saturation]

I could hack my scope to 150 MHz at a later date, once I need to get higher than 50 MHz, under which I currently work, but one best be clear about the sampling artifacts that occur even at 100 MHz. 

The threads discussing it are all over eevblog.  You can overcome the input filter, but the sampling and memory length are still the same, whether its a 50 or 150 MHz scope.

See discussions and threads with posts by rf-loop, jahonen and others.  I think rf-loop is also instrumentvu and posted this over 2 year ago:



What is basically shown is that 100 MHz on a single channel on a 100 MHz rated 1102E, providing 1Gs/s is already tasking the scope, you can see the sine wave wavering about at 100 MHz, and its from a Fluke scope calibrator at that, so the signal source is good.  When he switches on the second channel, it worsens, since the Rigol goes into 500Ms/s.  Then he switches on long memory length it drops to 250 Ms/s, at 2.5x the Nyquist frequency, and finally turns off sinx/x and watch what happens.  Why he got any negative comments for such a clear response baffles me.  My guess, its a language barrier thing.

What is shown won't change if you hack the scope to 150 MHz, those are real time sampling artifacts.   So if you do, even at 100 MHz, the basic 1000E hardware with 2 channels the scope is really best to 50 MHz at the default memory depth, and with long memory, its down to 25 MHz.

Since I have an unmodded 1052E, I have reproduced rf-loops issue by switching on 2 channels with long memory, which turns the scope into 25 MHz, and looking at say a 40 MHz sine wave and it gives the same goofiness as the last images on the youtube link.

The good news is you can overcome some of this by using equivalent time sampling, which isn't shown on the demo, so long as you are measuring repetitive signals.  But if not, one should be aware of these limitations.

[tinhead]

yes it know his review, the first thig i checked on my Tekway was exact same setting with a bit different results. With 100MHz test signal and 4kpts the aplitude didn't change if you enable ch2.
You can't see any influence at all. Of course if you increase the frequency the signal becomes more artifacts, but the overall amplitude stay the same up to 200Mhz.
This of course change as soon you enable 40kpts (or 512kpts), which is basically caused by SRAM speed.
All these designs (Rigol, Tekway, Hantek, UNI-T, Instek) have one big gap, in long memory mode you can't just capture more than 100MHz signals
and if you enable second channel it if even worser (max freq. on Tekway/Hantek is 90Mhz without any signal aplitude lose).

This is probably why Rigol never (really) produced D/E series with more than 100Mhz bw. Becuase of much higher wfrm/s and better filtering Hantek/Tekway are able to do this,
but only with 4pts memory enabled. As i remember Instek and Owon (and probably Rigol in CA series) is interleaving SRAM, that's the way to go with external memory.

In eq. mode or avg. mode signals are of course very clear (talking about repetitive signals), but that's nothing new. For single shot i already accepted that more than
the FPGA integrated memory is not possible with this class of DSOs (just because of end price), which is sad (still comparing to my old TDS754D)
but as always, you get what you paid for (which is still much better than TDS2000C series btw.)

[scrat]

Although this was not sufficiently pointed out about the Rigol, of course if one knows sampling frequency is halving each time you enable a channel or the long memory it is possible to predict those artifacts.
However, another problem with aliasing is the possibility of an (unwanted and unknown) input at frequency over Nyquist being easily read as a low-frequency signal, even considerig the selectivity of input LPF (which someone has shown to be poor). The problem is quite common, however (even where you don't expect it). Let me make an example that may sound h awkward: you frequently have to pay attention to aliasing on a mid-end (5000) Tektronix, too, since its default timebase/sampling frequency ratio was badly choosen (at least for long time windows). The fact is, of course, on the Tek you can rotate the knob to increase sampling rate, while in the Rigol...

[saturation]

Agree fully with you, scrat, folks need to be aware that the hack isn't the end-all be all of getting the scope's bandwidth improved; you need to pay attention to the sampling frequency and memory depth.  Dave's review of the new Agilent Infinivision scopes is what you'd expect in a scope rating, its 2Gs/s single channel, but it remains at 1Gs/s up to 4 channels, for a scope rated at 70MHz, a conservative rating that the sampling and bandwidth will clearly support.  Now, the top scope in that series is 200 MHz, how those scopes maintain the sampling rate on each channel would be good to see.

On the Rigol, it isn't made very clear, but if you press Horizontal MENU button, the sampling rate will appear.  You can "adjust" the scope sampling rate by going to Acquire, and turning sinx/x off or on, and adjust long or normal memory depth, and watch the rate change.  Finally, adjust the time base, you can see the sampling rate change, and try to balance all these controls for the best balance.

Although this was not sufficiently pointed out about the Rigol, of course if one knows sampling frequency is halving each time you enable a channel or the long memory it is possible to predict those artifacts.
However, another problem with aliasing is the possibility of an (unwanted and unknown) input at frequency over Nyquist being easily read as a low-frequency signal, even considerig the selectivity of input LPF (which someone has shown to be poor). The problem is quite common, however (even where you don't expect it). Let me make an example that may sound h awkward: you frequently have to pay attention to aliasing on a mid-end (5000) Tektronix, too, since its default timebase/sampling frequency ratio was badly choosen (at least for long time windows). The fact is, of course, on the Tek you can rotate the knob to increase sampling rate, while in the Rigol...

[saturation]

Yes, tinhead, as always this is great information.  its good the Tekway performs a bit better.  Users need to be aware there's more than the amplitude at issue, its the quality, just be aware of sampling introduced distortion as you approach the bandwith roll off, so there are some other limitations happening that isn't tested by just the rise time.

In equivalent sampling the wave form can appear better, but its still richer in artifacts, you can see this when doing FFT.  You can 'smooth' the artifacts out by using averaging acquisition, but clearly, you are at the scope's capacity limits.

In Dave's current review on Agilent's new Infiniivision 2000X doing 50,000 wfms/s, he was able to calculate the Rigol 1052E wfms/s rate, using an Agilent 'glitch' generator as a guide, the Rigol rates at ~ 800 wfms/s, in single channel mode; it would be good to see what this rate is when 2 channels are engaged.

yes it know his review, the first thig i checked on my Tekway was exact same setting with a bit different results. With 100MHz test signal and 4kpts the aplitude didn't change if you enable ch2.
You can't see any influence at all. Of course if you increase the frequency the signal becomes more artifacts, but the overall amplitude stay the same up to 200Mhz.
This of course change as soon you enable 40kpts (or 512kpts), which is basically caused by SRAM speed.
All these designs (Rigol, Tekway, Hantek, UNI-T, Instek) have one big gap, in long memory mode you can't just capture more than 100MHz signals
and if you enable second channel it if even worser (max freq. on Tekway/Hantek is 90Mhz without any signal aplitude lose).

This is probably why Rigol never (really) produced D/E series with more than 100Mhz bw. Becuase of much higher wfrm/s and better filtering Hantek/Tekway are able to do this,
but only with 4pts memory enabled. As i remember Instek and Owon (and probably Rigol in CA series) is interleaving SRAM, that's the way to go with external memory.

In eq. mode or avg. mode signals are of course very clear (talking about repetitive signals), but that's nothing new. For single shot i already accepted that more than
the FPGA integrated memory is not possible with this class of DSOs (just because of end price), which is sad (still comparing to my old TDS754D)
but as always, you get what you paid for (which is still much better than TDS2000C series btw.)

[scrat]

Since the Rigol doesn't use color grade code for representing "statistically" how much the signal stays on a point (as the Agilent and Tek phosphor-like visualization do), the wfms/s rate is the real graphics refresh rate (graphics, not display). I expect this to be even worse than 800wfms/s, at least in zero persistence mode. This is a great difference, since infinite persistence isn't that comfortable to use, while on good DSOs you can see infrequent signals in normal persistence too.
However, to make an acceptable comparison, many glitch catches should be made...

In Dave's current review on Agilent's new Infiniivision 2000X doing 50,000 wfms/s, he was able to calculate the Rigol 1052E wfms/s rate, using an Agilent 'glitch' generator as a guide, the Rigol rates at ~ 800 wfms/s, in single channel mode; it would be good to see what this rate is when 2 channels are engaged.

.

[saturation]

You're right scrat, good point, the only way to see those glitches with the Rigol was to use persistence mode and its 800 wfms/s shouldn't be based while in persistence mode, my mistake, but its the only way you'll ever see the glitch with a Rigol 1000E series. 

Since the Rigol doesn't use color grade code for representing "statistically" how much the signal stays on a point (as the Agilent and Tek phosphor-like visualization do), the wfms/s rate is the real graphics refresh rate (graphics, not display). . This is a great difference, since infinite persistence isn't that comfortable to use, while on good DSOs you can see infrequent signals in normal persistence too.
However, to make an acceptable comparison, many glitch catches should be made...

In Dave's current review on Agilent's new Infiniivision 2000X doing 50,000 wfms/s, he was able to calculate the Rigol 1052E wfms/s rate, using an Agilent 'glitch' generator as a guide, the Rigol rates at ~ 800 wfms/s, in single channel mode; it would be good to see what this rate is when 2 channels are engaged.

.

[therian]

I want to clarify:
1) model #, change it to DS1152E or DS1152E-EDU ?
2) 5th letter in serial change to F ?

I did the first variation above and my info screen say DS1152E  without -EDU part, please tell me did it work or I have to redo it ?

[rf-loop]

 I think rf-loop is also instrumentvu and posted this over 2 year ago:

This is need correct. I'm not instrumentvu!

I have only tell something about wrong sinx/x issue. Simple, Rigol do not any sinx/x calcualtion. It is only some kind of dsp filtering what try some kind of "sinx/x" like picture. This scope and its way to do Sinx/x is just nearly as "do nice picture". And with some wrong parameters. It can see in my pictures what link is in original thread.

Before and now I tell that I'm tightly with this:

Agilent, Chris Rehorn: Sin(x)/x Interpolation: An Important Aspect of Proper Oscilloscope Measurements
( http://i.cmpnet.com/planetanalog/2009/02/Sin(x)x_Agilent.pdf )

Also I have one OLD tektronix scope (Digi but crt) and It seems do nearly real sinx/x. It can show also highlighted all real samples and then curve what it draw just as sinx/x and there is huge difference between Rigol DSP filtering and Tektronix expensive solution. Rigol sinx/x is only cosmetics. And it breaks all basic principles. Lot of below nyquist it product huge attenuation - really wrong if think measurements and not stupid picture art on scope display.. Alone this tell that it is not anything else but bad filter for cosmetics. Measuring equipments need different thinking and nice free art picture need other.

Also analog channel frequency response is very bad, And also it do not adjust frequency response as samplerate change and also there are not any kind of care about aliasing, not even warning about aliasing situation. ( Even My old -80's HP digi have - but it have no sinx/x becouse it always show only and alone real sample points and not any cosmetic stuffs) Also Rigol is bad gaussian what make aliasing problem really bad.. But if think price and if know limitations... it is acceptable.  Good analog channel for (high-end) digital scope is reaaly difficult to do and expensive - there is not so many who can and even less who want pay it... so.. Rigol is not high-end... not middle-end and it do not need it... but Rigol can littlebit adjust this filter... if they care anything about customers - but this company really do not care anything. They do as they do... and peoples buy if they buy. Simple strategy. If they care customers they do corrections... also other... nearly all what they want do is try make modification more difficult. This is they customer care policy.
But it is not only Rigol... there are some others that make real garbage.

Entertainment display and measuring equipment... measuring is numbers, in this case digitized samples, art is example nice figures... this they have littlebit forget in Rigol company (and many other company)

Most bad is that Rigol destroy only facts... digitized real samples... they are only "true" and these they put to garbage and then draw some kind of nice looking curve where is in worst case not any true points. This is not Sinx/x idea. Drawed curve do not fit these sampled poiints at all. This is real mistake. If they do not understand ... more bad. maybe they get more money if they start making boohoo teddybears. Yes, I can shut off this but I can not understand why they do not repair it. I know, they do not care. And most of customers do not understand... becouse they just look onlu nice pictures, maybe not do any real measurements.

Rigol is ok, with its price and limits. manufacturing quality is not bad...

But in Rigol sinx/x can shut off... HanTekway can not... I hope it is better.

[saturation]

Thanks rf-loop, at least then 2 separate people found added limitations with the Rigol 1000E series that folks should be made aware of, that the hack doesn't fix, and this goes back at least 1-2 years ago.

It would be interesting to see if the Agilent version has the same problems.

...Rigol is ok, with its price and limits. manufacturing quality is not bad...

But in Rigol sinx/x can shut off... HanTekway can not... I hope it is better.

[Mechatrommer]

thanx rf-loop, maybe thats the answer why rigol sin(x)/x doesnt do me any good, i always turn it off, i prefer ugly random polygons, it has the feel of true sampled measurements. and i dont think the sinx/x will improve "enough resolution data" any better, i only feel it will only slow down the capture process.

[therian]

Anyone ? the idea that I done something wrong keep bugging me

[Mechatrommer]

you are allright therian, you dont have to (and should not) include the EDU.

[dimlow]

There is no need to change the serial number, but you do have to send a serial number to the scope after you have changed the model number or else the scope will not store the model number change. And its a capital 'E' you need, DS1152E

[rf-loop]

Maybe all peoples who buy Rigol oscilloscope have not seen this Rigol official statement.

http://www.progshop.com/pdf/oscilloscopes/Rigol/Statement%20of%20Rigol.pdf

"Rigol doesn’t authorize anybody to sell Rigol products directly from China to Europe."

Some sellers still argument that they are selling as "Rigol authorized distributor" (example for selling to EU)

Exactly, they are not. (I mean these sellers what this paper is talking)
Including DX, DeX etc.
If not believe, ask Rigol.

And, what about this:

"Some products from illegal sales channels are not even originally produced by Rigol Technologies, Inc."

[scrat]

"Some products from illegal sales channels are not even originally produced by Rigol Technologies, Inc."

Rigol has so large production numbers and cost reduction knowledge that I think selling a fake Rigol (with about the same features) at a competitive price would be a hard challenge!

[tinhead]

for every (not only chinese) manufacturer there are some necessary steps to be some day "brand name manufacturer"
- you have to design good and cheap products (Rigol did it few yrs ago)
- you have to be visible on local and international market (since some time they known everywhere)
- you have to have some legal issues with ppl coping your products (actually Rigol did copied Instek and not Atten Rigol, or like the court decided - a 1:1 copy of example circuit from components manufacturer app note is not "genuine" product)
- you have to get some "promotion" from other brand name manufacturers (like a product compare [yeah there are some Rigol CA and Agilent DSOX2000 compares on Agilent website but no single word that e.g. Hantek is having 800x480 display or Owon 8" display - in this price category of DSOs - because these companyf "didn't exists yet forAgilent", but Rigol does], or co-operation - like with Agilent in the past, no matter what it costs ...)
- you have to have world-wide support and distributors (that's not the cheap sellers giving you no support)
- you have to release "high-end" product (Rigol's DS6000)

(additionally you have to have at least small office in US to be able to sell there your porducts)

So what, there was no single Rigol clone, maybe only some hacked relabeled devices (but only maybe), all Rigol was to looking for
was to establish distributor connections and global support - like other brand name manufacturers have.

For sure a manufacturer can decide if a specific product will be available internationally and if who can sell/give you support, nothing new - and many manufacturers are doing this (to maximize revenue? probably too). 2yrs ago i bought some products from good knows manufacturer, you can buy them directly in EMEA (but you have to pay 250% of local price) or you can buy in HK, ship them to India (which the manufacturer allowed), from there to Saudi Arabia and from there to Europe ... each of these steps was allowed by manufacturer - and i did saved finally good 150% of the EMEA price. The issue started as i tried to sell them here, i got directly tons of RFPs and court orders. Luckily i never did something wrong, a company policy was very clear (for my local court) " you have to import over official distributors", which i did (more or less) - so finally i did paid anything. The end of the story - this particulary manufacturer changed their polices, so you will be unable now to re-re-re-import their products. Did i care about ? No not really because a distributor have still to earn some money and manufacturer maximize revenues, i just don't like marketing tricks like "ohhh, ebay sellers having cheaper than our distributors, let's destory them and tell everybody that's are clones" - and of course everybody will believe that, because 99% of these sellers was located in china.

Anyway, stay on topic - i believe what rf-loop is trying to tell is nothing else than "who cares, your product have anyway no warranty"

[therian]

There is no need to change the serial number, but you do have to send a serial number to the scope after you have changed the model number

Ok here what I done now
scope was original ds1102e so the serial.
using demoIDN.exe and usb :
:INFO:MODEL DS1152E
:INFO:SERIAL DS1EB1150****7

I do notice a bit more ripple looking at square wave of 10 Mhz
But why this screen-shoot
show model as full DS1152E-EDU
and serial letter F don't belong to 50 or 100 Mhz versions so why it was changed ? 

[rf-loop]

Anyway, stay on topic - i believe what rf-loop is trying to tell is nothing else than "who cares, your product have anyway no warranty"

You find a point.
----------

common comment:
I have seen least one Rigol DS1102B fat was nearly as "sure" modified in China and relabeled. (but I can not anymore proof it was fake) We need also remember that we may think $50 is money what you do not want make lot of work...  it still is more than 300 in china... if you live in china normal life... normal salary is around 1500 per month.. lot of peoples live with less. Not accurate but it is arount so that 1 RMB you there live just as 1 USD. So we may think that 100USD is not so much today. But it is over 600 there.

Who want make copy, what business it is.

Well, ask Atten - Siglent. (but they loose in China law room some time ago: Rigol - Atten, Rigol win 1-0)

How about copy what also outside looks as Rigol... manufacturing this kind of copy is even more cheap than Atten-Siglent have done.
Also money can copy.

Who cares warranty - maybe nobody.

But also we all know, Rigol (in China) have not "customer service". Customer care is commonly known it is really bad.

That is why my next step is HanTekway.

[Stonehedge]

Hi all, my name is Sten, I live in Sweden in the northern part of the European Union and this is my first post here.

I bought a slightly used Rigol DS1052E from an internet auction and received it yesterday. It was running software 00.02.04.01.02 on hardware 58. I have "hacked" the oscilloscope using the instructions found on this site. It is now running as a DS1152E with software 00.02.04.01.02. There were no problems with the "upgrade".

I used these steps:
1) Downgraded to 02.02 using a USB flash memory with the patched file DS1000EUpdate.RGL
2) Used the here kindly provided USB utility to update to DS1152E by editing the data to DS1152E and the serial number to DS1EF12........5
3) Upgraded to software 00.02.04.01. using my USB flash memory with the provided .RGL file with this software

Thank you all that provided files and utilities. You have made a great work. Of course am I not expecting any guaranty services from Rigol.

My main use for the oscilloscope is measurements on audio and microcontroller lab work using PIC and other such controllers to different small projects. These are seldom clocked with more than 40 MHz so I should be fine using this oscilloscope. I also use a Hantek/MI USB oscilloscope with 60 MHz BW.

[saturation]

Thanks, when my warranty runs out I think I'll go your route.  I have the same hardware, but I bought it new from an authorized dealer.

Hi all, my name is Sten, I live in Sweden in the northern part of the European Union and this is my first post here.

I bought a slightly used Rigol DS1052E from an internet auction and received it yesterday. It was running software 00.02.04.01.02 on hardware 58. I have "hacked" the oscilloscope using the instructions found on this site. It is now running as a DS1152E with software 00.02.04.01.02. There were no problems with the "upgrade".

I used these steps:
1) Downgraded to 02.02 using a USB flash memory with the patched file DS1000EUpdate.RGL
2) Used the here kindly provided USB utility to update to DS1152E by editing the data to DS1152E and the serial number to DS1EF12........5
3) Upgraded to software 00.02.04.01. using my USB flash memory with the provided .RGL file with this software

Thank you all that provided files and utilities. You have made a great work. Of course am I not expecting any guaranty services from Rigol.

My main use for the oscilloscope is measurements on audio and microcontroller lab work using PIC and other such controllers to different small projects. These are seldom clocked with more than 40 MHz so I should be fine using this oscilloscope. I also use a Hantek/MI USB oscilloscope with 60 MHz BW.