spend ~700$ for Rigol 2000a or spend 500$ more and get the R&S RTB 2000for ~1200

[tmadness]

Hi guys long time lurker at the EEVBlog forum, first time poster. I want to buy my first new scope. I've owned few scopes in the past, all of them repair/restoration jobs (HP 53501A, Philips pm321 etc.) I'm a college student and my home lab is mostly filled with repaired equipment: a couple of system power supplys,  frequency synthesizer,  few bench multi meters and other smaller kick nacks, so this oscilloscope will be my first big investment (most of the others were free or below 80$). but after working 50 hours every week for 10 weeks as an intern I have earned a decent amount and I want my oscilloscope to be an investment, some thing that lasts me through grad school and hopefully beyond.
I know that the rigol is the best bang for the buck (+hacks to unlock more) but at this point the release date is 4 years behind. Although rigol has been good with its support so far, it will likely decrease with time. The R&S is brand new (~6 months), has a much more appealing look and better usability features, and hence also carries the extra price tag.
So my question is: is the R&S worth the extra 500$?

[ataradov]

Regardless the money, if you want the thing to be useful, you need 4 channels.

R&S is very nice, not sure if it is nice enough to pay that price.

[nctnico]

I'd wait for the RTB2000 firmware to mature (there are still several bugs) and for an offer where you get the decoding options for free. I also second the suggestion for 4 channels.

[2N3055]

RTB 2002 is 70 MHz, 2CH, and not a single one option unlocked for 1200€... To unlock serial protocols and segmented memory you have pay for a 1150€ bundle, so no, it is not 1200€ but more 2400 € for a usable scope. I also wouldn't buy 2CH but 4CH if you do serial buses and such... so base 4CH scope is more, arround 2000$.. Also 70MHz..

Fully unlocked RTB 2004 can reach 7500 USD... While it is very nice scope, it is not worth that kind of money in my opinion....

[markb82]

I really like the RTB2000 series, but having missed out on the introductory pricing spending $2000 USD on a 4 channel 70MHz scope is .... too much.  Although the form factor is excellent (10in screen!!!) and it comes with the 300MHz probes so might be hacked eventually.  That said, if the Siglent SDS1204X-E is actually released at a price near $500 USD it will be the new goto scope.  Sure it will be buggy, but at that price I think you can live with a few bugs.

As an aside:  The RTB2000 firmware available online is encrypted so someone will need to dump the flash for us to start playing with IDA.  From what I've gathered the scope actually runs FreeRTOS not Linux so that could complicate things (more disassembly).  I'm guessing that the RTB2000 has a Cyclone V, and there is a good document on how it boots, but does not mention any security features (like AES FuseKeys on ZYNQs / Xilinx FPGAS) so there is a chance the NAND flash is not encrypted ? [ Ref: https://www.altera.com/en_US/pdfs/literature/an/an709.pdf ]

[Fungus]

if the Siglent SDS1204X-E is actually released at a price near $500 USD it will be the new goto scope.

Yep.

Rigol is also promising new mid-range 'scopes at the end of this year using an entirely new chipset they've been working on (that's why they've been silent for 4 years). It could mean a refresh of the lower-end scopes soon.

Right now I'd say buying a 'scope as an investment is impossible. The world is changing, there's suddenly huge demand from people like yourself so every manufacturer is working like mad to fill the $400-$1200 gap which has been dominated by the Rigols for the last few years.

Also: 2 channels? I wouldn't.

R&S do make nice oscilloscopes though. If you're prepared to pay for them.

[TK]

For around $1000 you can get a good used Keysight/Agilent DSOX2X04A (4 channels) that can be hacked to 200MHz and all the options.  You might get some with warranty on it, as Keysight started offering 5 year and some Agilent scopes I found online still had some warranty left.  They also keep the resell value better than the Rigol.  If you can go up to $2000-$2500, you can get a 3000X series scope with 4 channels with way much better specs.

It is similar to car value, if you buy a new one.  The day you open the box it the price will drop 20-40% (or even without opening the box but reselling as individual on eBay).  If you buy a good well priced used one, you might even make some money when you resell it.

[nctnico]

The problem with Keysight scopes is that they have no memory to speak off.

[TK]

I did not experience any limitation due to having small capture memory.  I just change the Timebase and recapture if I want to see the details.  Actually doing a long capture, zooming and scrolling to find the part you are interested in seeing is a much slower process.  Probably each user and application is different.

[exe]

I almost bought RTB2004, but the price of options pissed me off. I think it's a clever marketing game, the cost of ownership may turn out to be much higher than it appears.

Unless you can buy it on discount. Or buy a bundle with some options pre-enabled. This is is considerably cheaper. Still, it's not a $2k scope, it's more like a $3k+ scope.

[BrianSchmalz]

If you're willing to spend a little bit more ($1500) there's a nice SDS2304X for sale here : https://www.eevblog.com/forum/buysellwanted/(fs-us)-siglent-sds2304x-300mhz-scope-all-options-like-new-condition/msg1251664/#msg1251664

4-channels, all options unlocked, digitals, 300MHz, etc.

(Yes, it's me that's trying to sell it.)

*Brian

[nctnico]

I did not experience any limitation due to having small capture memory.  I just change the Timebase and recapture if I want to see the details.  Actually doing a long capture, zooming and scrolling to find the part you are interested in seeing is a much slower process.  Probably each user and application is different.

That only works if you can recapture exactly the same event and if the same conditions apply otherwise you are comparing apples and oranges. Besides that scrolling can be done fast by changing the time base while scrolling. Long memory has served me well so it is an important feature to me.

[Fungus]

I did not experience any limitation due to having small capture memory.  I just change the Timebase and recapture if I want to see the details.

What if it's a one-time event? Or only happens every 53 minutes? 

[ataradov]

Long memory is the second most important factor after the number of channels. I can even live with lower sampling rate in many cases, but I would not consider a new scope with less than 10Mpoints of memory.

[nctnico]

I did not experience any limitation due to having small capture memory.  I just change the Timebase and recapture if I want to see the details.

What if it's a one-time event? Or only happens every 53 minutes? 

This is not an exaggeration. Sometimes problems only occur with these kind of intervals (or longer). For these you want to get as much detail as you can. And there is also destructive testing where every capture means throwing a couple of hundred dollars down the drain due to the (usually expensive) DUT being destroyed.

[EEVblog]

The dreaded 4CH vs 2CH debate will always be like

Personally, in a bang-per-buck game, I think channel count will always win out.

[tautech]

That said, if the Siglent SDS1204X-E is actually released at a price near $500 USD it will be the new goto scope.  Sure it will be buggy, but at that price I think you can live with a few bugs.

It will be better than you imagine, yes there's a few things that need attention but way less that I'd thought.
Core and most advanced functionality is already very good.

Final price is the unknown.

[TK]

I did not experience any limitation due to having small capture memory.  I just change the Timebase and recapture if I want to see the details.

What if it's a one-time event? Or only happens every 53 minutes? 

This is not an exaggeration. Sometimes problems only occur with these kind of intervals (or longer). For these you want to get as much detail as you can. And there is also destructive testing where every capture means throwing a couple of hundred dollars down the drain due to the (usually expensive) DUT being destroyed.

Why don't you present a specific case and maybe the scope reps in the forum can present how their instruments can help detect the fault.  Maybe Keysight can do it with Segmented Memory where it can wait almost indefinitely for specific events that occur at undetermined time intervals. 

[Fungus]

Why don't you present a specific case and maybe the scope reps in the forum can present how their instruments can help detect the fault.  Maybe Keysight can do it with Segmented Memory where it can wait almost indefinitely for specific events that occur at undetermined time intervals.

The point being discussed is that "just change the Timebase and recapture!" isn't always an option. The event may not be reproducible.

[TK]

Why don't you present a specific case and maybe the scope reps in the forum can present how their instruments can help detect the fault.  Maybe Keysight can do it with Segmented Memory where it can wait almost indefinitely for specific events that occur at undetermined time intervals.

The point being discussed is that "just change the Timebase and recapture!" isn't always an option. The event may not be reproducible.

You are leaving information out: I said "I just change the Timebase and recapture if I want to see the details".  When I post, I speak for myself and my experience.  I use the scopes for hobby and learning electronics and I do not deal with fault finding in a product development environment.  If someone asks in this forum for an advise about what to buy in the $700-$1200 price range, I consider Him/Her to be in a similar situation I was in, so I try to provide my experience to help.  Nobody has the absolute truth.

[nctnico]

I did not experience any limitation due to having small capture memory.  I just change the Timebase and recapture if I want to see the details.

What if it's a one-time event? Or only happens every 53 minutes? 

This is not an exaggeration. Sometimes problems only occur with these kind of intervals (or longer). For these you want to get as much detail as you can. And there is also destructive testing where every capture means throwing a couple of hundred dollars down the drain due to the (usually expensive) DUT being destroyed.

Why don't you present a specific case and maybe the scope reps in the forum can present how their instruments can help detect the fault.  Maybe Keysight can do it with Segmented Memory where it can wait almost indefinitely for specific events that occur at undetermined time intervals.

I already did that. Short memory is one of the main reasons I sold my Agilent DSO7104A and it has 8Mpts (well, Keyisight Mpts so more like 2Mpts in real usage scenarios). For some measurements I need a lot of data which is correlated and for others it is just easier to capture one single shot than dicking around with trigger delay etc. Segmented recording is great but it still needs memory to store the data and that is also where the DSO7104A didn't shine. For tracking down an intermittant I2C problem it could only record 3000 segments on two channels without I2C decoding going wrong.

[2N3055]

With 100k samples you can have 2GS/sec sample rate at 5 uSec/div... Slower timebase than that, you get slower sample rate...

So if you have burst of data 5 mSec long, and you want to look at edges, look for runts and such, or look for higher frequency content in general, you will have 20 MS/sec sampling rate, or one sample every 50 nsec ....

You can work with scope interactively (like with the CRT scope) in which case even 10000 samples is ok.. If you use digital scope to capture a burst for offline analysis, longer the memory, the better...

[Neganur]

I don't get it. If you're looking for a runt then trigger for a runt.
If you're looking for protocol problems an need to record tons of data, use a logic analyzer?

edit: I meant, use both instruments at the same time.

[nctnico]

I don't get it. If you're looking for a runt then trigger for a runt.
If you're looking for protocol problems an need to record tons of data, use a logic analyzer?

Unfortunately the reality of finding problems is more difficult. When starting to diagnose an (intermittent) problem you have no idea where the problem is so you can't trigger on something like a runt or ignore the analog waveform by using a logic analyser.

Making a long recording is also easier. For example: one of my products does a 5 second measurement using various waveforms. If I record the entire 5 seconds with enough detail I only need to start the device once, check various parts of the waveform and be done. If I'd need to take several acquisitions I'd need to do many more manual steps (which introduces more room for error) and makes work less efficient. The volume for this product isn't high enough to do automated testing so that is not a solution either. Sure you can make do with less memory but what counts for me is efficiency and needing as least steps as possible to reduce the chance of errors.

During development deep memory is also easier because you can go from the big picture into the fine details and back without needing the re-measure or remember anything.

[Neganur]

Ok but that sounds to me like you're looking for an awful lot of memory (5 seconds at what Bitrate? I2C is not rally fast) And then you're offloading the data to a computer or are you looking at this much data on the oscilloscope?

I recently had a problem like this, a test box does some 10 seconds of SPI, and even though I could record it on a scope it was highly impractical to browse through it. I ended up using A logic analyzer and had the person who requested the data install the analyzer software to browse through the data. Much easier on the PC.

Anyway, I think that the use case that you come up with won't be solved with either of the two scopes mentioned by the OP.

I don't think too highly of Rigol's 2000-series and would probably recommend the 2000 RTB not only because the whole technology is much newer but also due to the features (not to say you need all of them, but they are possibly nicer to use on the R&S).

If money is the issue, perhaps the Rigol 1000Z series is a good alternative while saving money for a year or two.