Please recommend oscilloscope for robotics research

[Wuerstchenhund]

In the case of GDS, some scopes have an 'A' ending while others have an 'E' ending. For example, 2204A vs. 2204E.

Simple. 'A' was the previous generation, the current one is the 'E'.

The 2104A has a rise time of 3.5ns while the 2204A, 2204E and 2304A has a rise time of 1.17ns. In general, is 3.5ns fast enough?

For GDS, I consider 2102A, 2104A, 2204A and 2204E. Cannot find GDS-2000E on their website. Which is the latest and greatest? I need external trigger function (e.g. press a button and different devices start working) and usb storage of data for plotting in EXCEL or Matlab. Experiments last about 10-30 seconds, no more than a minute or two.

Frankly, I get the very strong impression that you have actually no idea what you want or what you need. What's more frightening is that it looks you don't even seem to know what exactly you want to do (motors and actuators actually make only for a small part of robotics). Which simply makes it impossible to give you a proper recommendation.

Considering all that, maybe the best option for you is really to get a Rigol DS1054z and use that until you know the basics and understand better what you want to do and what you need.

[Fungus]

In general, is 3.5ns fast enough?

Fast enough for what? 

Basically that rise time is just the 'scope's bandwidth written another way.

3.5ns is 100MHz.

2304A has a rise time of 1.17ns.

And that's 200MHz.

[rstofer]

In the case of GDS, some scopes have an 'A' ending while others have an 'E' ending. For example, 2204A vs. 2204E.

Simple. 'A' was the previous generation, the current one is the 'E'.

The 2104A has a rise time of 3.5ns while the 2204A, 2204E and 2304A has a rise time of 1.17ns. In general, is 3.5ns fast enough?

For GDS, I consider 2102A, 2104A, 2204A and 2204E. Cannot find GDS-2000E on their website. Which is the latest and greatest? I need external trigger function (e.g. press a button and different devices start working) and usb storage of data for plotting in EXCEL or Matlab. Experiments last about 10-30 seconds, no more than a minute or two.

Frankly, I get the very strong impression that you have actually no idea what you want or what you need. What's more frightening is that it looks you don't even seem to know what exactly you want to do (motors and actuators actually make only for a small part of robotics). Which simply makes it impossible to give you a proper recommendation.

Considering all that, maybe the best option for you is really to get a Rigol DS1054z and use that until you know the basics and understand better what you want to do and what you need.

Exactly!  The rise times mentioned above are simply indicative of the scope bandwidth.  Faster rise time, more bandwidth hence the difference in model numbers.

One advantage to the DS1054Z is that, even used, it will always be worth $300 or so (based on $400 new).  So use it until it becomes a hindrance (if it ever does) and then either sell it to get funds for a more expensive model or designate it as a sacrificial scope for noobs.

Yes, better scopes are a treat!  More bandwidth is always the right answer and more memory depth, triggering options, decoding options, channels and any other feature I overlooked are always better.  They may not be necessary but they're cool to talk about.

The hacked DS1054Z is a decent 100 MHz scope.  But a 100 MHz square wave is going to look terribly sinusoidal.  If you want to actually display a 100 MHz square wave, you probably need at least a 500 MHz scope (give or take).  So, the 1054 is going to be able to display a 20 MHz square wave pretty well and that's probably good enough.  Most of the processors we use run a lot faster but their IO pins don't toggle at anywhere near the clock rate.

In my corner of the sandbox, four channels is more important than increased bandwidth.  If I run out of bandwidth, I'll upgrade.  In the meantime, I haven't found anything I can't do with the DS1054Z.  But it's early days...

[eeguy]

In the case of GDS, some scopes have an 'A' ending while others have an 'E' ending. For example, 2204A vs. 2204E.

Simple. 'A' was the previous generation, the current one is the 'E'.

The 2104A has a rise time of 3.5ns while the 2204A, 2204E and 2304A has a rise time of 1.17ns. In general, is 3.5ns fast enough?

For GDS, I consider 2102A, 2104A, 2204A and 2204E. Cannot find GDS-2000E on their website. Which is the latest and greatest? I need external trigger function (e.g. press a button and different devices start working) and usb storage of data for plotting in EXCEL or Matlab. Experiments last about 10-30 seconds, no more than a minute or two.

Frankly, I get the very strong impression that you have actually no idea what you want or what you need. What's more frightening is that it looks you don't even seem to know what exactly you want to do (motors and actuators actually make only for a small part of robotics). Which simply makes it impossible to give you a proper recommendation.

Considering all that, maybe the best option for you is really to get a Rigol DS1054z and use that until you know the basics and understand better what you want to do and what you need.

Thanks. So, E is more recent and cheaper?

I always got access to the expensive scopes at well funded research labs. I only used them. I did not have to know the detailed specifications of the scopes. Now I have to buy my own using my own money.

I just checked the specs of a scope I used many years ago. It had 4 channels, 200MS/s sampling rate at 2 channels and 100Ms/s at 4 channels, equivalent sampling 20Gs/s, sensitivity 1mV/div to 5V/div, 5ns/div to 50 s/div, sweep time of 120k word (2 channels), 56k word (4 channels). 2M word max recording length, 150MHz.

[eeguy]

Can you elaborate more on what projects will you use your scope?
I've been using a 100MHz scope throughout my 4 years college career, and I do quite some robotics things, and I feel 100MHz is more than enough.

Common frequencies:
PWM carrier for PWM DAC: 1~2MHz
PWM carrier for power modulation: 0.05~1MHz
SPI: 1~20MHz
I2C: 0.1~3.4MHz
Timer interrupt driven IO: <10MHz (depending on MCU core frequency and ISR entrance/exit latency)
I2S: <12.288MHz (192ksps, stereo, 32 bit, no TDM)
Crystal oscillator (fundamental): <50MHz, usually <35MHz

So, unless you are into communication or you need to do signal integrity analysis, it is highly possible that 100MHz will be more than enough.
When you feel what you are doing calls for a 1GHz scope, it is almost certain that you make more money than what's needed to get a new scope, by then, you can either keep your current one or sell it.
I recommend get a cheap one as your first scope, then keep it even if you decide to get a better one later. God knows when or if you will need a second scope.
Another recommendation is to get used scopes. You lose almost no money buying and selling an used scope. If you buy a new one and sell it as an old one, you lose money.

I plan to make robots and some electronic circuits. A scope is needed during the testing and diagnostic stages. I just try to get a general purpose, future proof scope. I don't think I will need more than 200MHz. I am more concerned about sampling rate and recording length. I guess one good thing about the Agilent scopes is that they can be upgraded as needed.

[rstofer]

Bandwith (in Hz) = 0.35 / RiseTime (in seconds)  so if the Rise Time is 3.5 ns, we have 0.35 / 0.0000000035 or 100,000,000 Hz.  A 100 MHz scope should have a rise time of 3.5 ns.  Give or take...

I just measured the rise time of an Arduino Uno PWM signal as 4.0 ns.  Some part of that is my scope (3.5 ns) and some of it is my crappy probe (just a garden variety 100 MHz probe, not Rigol probes) so, even after getting the magic rise time measurement on my scope, I have no idea where the Arduino rise time really comes out.  I'm actually surprised it is that fast!

I should buy one decent 500 MHz probe to eliminate probe effects from the measurements.

OR, I could be a little more practical and ask the question: "Does 4 billionths of a second make any difference to my application?".  Not likely!

[Fungus]

Thank you for the comments. Most likely I go for either GW Instek or Keysight. Are they quiet? I cannot stand noisy equipment.

Quieter than a DS1054Z, yes.

For GDS, I consider 2102A, 2104A, 2204A and 2204E. Cannot find GDS-2000E on their website. Which is the latest and greatest? I need external trigger function (e.g. press a button and different devices start working) and usb storage of data for plotting in EXCEL or Matlab. Experiments last about 10-30 seconds, no more than a minute or two.

I wouldn't go for 2 channels so that means the cheapest option is the 2104A at $1450 (give or take).

You probably need serial decoding for real robotics work. The Instek "logic analyser" option costs $550 (give or take).

That means you're looking at about $2000 for something to do a job that could be done perfectly well with a DS1054Z.

$1600 for a quieter fan? That's up to you. 

I know some forum members who fitted a much quieter fan to their DS1054Z for <$20.


(Yes, I know a GDS2104A will do more than a Rigol DS1054Z, I'm referring to the feature set you're probably going to use in 'robotics').

[Fungus]

I just measured the rise time of an Arduino Uno PWM signal as 4.0 ns.  Some part of that is my scope (3.5 ns) and some of it is my crappy probe (just a garden variety 100 MHz probe, not Rigol probes) so, even after getting the magic rise time measurement on my scope, I have no idea where the Arduino rise time really comes out.  I'm actually surprised it is that fast!

If it's more than your scope's rise time then most of it will be from the signal. It's when you're seeing exactly the scope's rise time that you realize you need more bandwidth.

I should buy one decent 500 MHz probe to eliminate probe effects from the measurements.

Normal probes are usually good for a lot more then 100Mhz. 300MHz(ish) is where the bad stuff starts to happen. There have been a few threads on this.

[nctnico]

At least the GDS-2000E series has input filtering. I think this is going to be very helpfull for an upcoming project to see what signals are hidden in a PWM and AM modulated signal.

[saturation]

Most any reputable >= 100 MHz scope will meet your needs, its a question of bang for buck and the right bandwidth.

For example the 50 MHz Rigol 1054Z or Instek 1054B will do real time sampling of 1Gs/s 1CH, 500Ms/s 2 CH 250Ms/s at 3 or 4 CH.  Equivalent sampling is no longer specified in newer DSOs as it worked only on truly repetitive signals.  Vertical sensitivity is the same or better and the old measure of 'word' for record length is replaced by megapoints "Mpts".

Both low cost DSO offer more record length memory than mentioned below, and for the Rigol more bandwidth and memory with a successful hack.

I always got access to the expensive scopes at well funded research labs. I only used them. I did not have to know the detailed specifications of the scopes. Now I have to buy my own using my own money.

I just checked the specs of a scope I used many years ago. It had 4 channels, 200MS/s sampling rate at 2 channels and 100Ms/s at 4 channels, equivalent sampling 20Gs/s, sensitivity 1mV/div to 5V/div, 5ns/div to 50 s/div, sweep time of 120k word (2 channels), 56k word (4 channels). 2M word max recording length, 150MHz.

[Fungus]

the Rigol more bandwidth and memory with a successful simple hack.

FTFY. 

[tautech]

I cannot stand noisy equipment.

Well you're going to have to trust somebody's word or really do your homework.

Most modern scopes have a fan and the more functionality that's inbuilt requires more computation and therefore more cooling. Period.
One thing you might want to consider is the size of any fan, larger ones tend to run at lower RPM's and therefore produce less noise.

I just try to get a general purpose, future proof scope.

Good luck with that.
In the years that I've sold scopes there have been so many advances in entry level DSO's but at the end of the day you have to work out what features are of most importance to you.
There's an ongoing battle to add more features to capture market share so who's to say what feature/s might be included in models 6 months out that you won't wish to have too.
Your budget will have an influence on longevity for other than mid and high range equipment most have been built to a price and that includes design, component and build quality and feature set.

[nctnico]

I just try to get a general purpose, future proof scope.

Forget about that. How long do you drive around in the same car? Now let me tell you test equipment depreciates even faster than a car. 

[rstofer]

Another recommendation is to get used scopes. You lose almost no money buying and selling an used scope. If you buy a new one and sell it as an old one, you lose money.

I got involved with recommending a used Tektronix on another thread a couple of weeks ago.  What a mistake!  The buyer got burned with a scope that doesn't function properly.  I felt really bad about promoting the idea.

I got lucky buying a Tek 485 on eBay but I'm going to stop recommending that idea.  Unless they are prepared to write off the cost immediately, I now suggest they avoid used scopes.

So, yesterday I had to troubleshoot the drivers on my CNC mill out in the garage.  Which scope did I take?  The Tek 485, not the Rigol DS1054Z.  The Tek will stand on the floor, I don't need to know any more than the fact that the direction signal changes and the step signal pulses.  Just show me a wiggly line, I don't need details.

Sometimes old school is entirely adequate.  Preferred, even!

[rstofer]

I got involved with recommending a used Tektronix on another thread a couple of weeks ago.  What a mistake!  The buyer got burned with a scope that doesn't function properly.  I felt really bad about promoting the idea.

Buying a used one requires time and luck. My 2 scopes are all bought from this forum, and both works fine.
The first one later died, but KS offered me free service including a new acquisition board and cal cert. Since then, I happily bought 3 years of extended warranty. It was passed on to another forum member at a lower price, but I lost only ~30% of the value for owning it for 1.5 years.
Buying ridiculously old and cheap gears are bound to have trouble. If you buy a <5years old used scope, it is very likely it will be just fine and run strong for another 5 years.

True enough but almost any of the used analog scopes will be a good deal older.  A decent DSO might be a good buy on the used market but I haven't priced them.  At some point I want a warranty and then I turn around and void the warranty by replacing one of the switches.  Duh...

That's another nice thing about the DS1054Z:  If it doesn't work out, I'm not out a lot of money.  So far, I'm liking the scope.  It does so much more than my Tek 485 (except bandwidth).

[Faith]

I just try to get a general purpose, future proof scope.

Forget about that. How long do you drive around in the same car? Now let me tell you test equipment depreciates even faster than a car. 

Yep the speed of the depreciation can be quite funny at times.

Just saw a brand new and unopened Keysight MSOX3104T which was manufactured just three months ago sell on eBay for around five grand when the list price is about three times that amount.

Makes it really difficult to justify buying new.

[Faith]

Just saw a brand new and unopened Keysight MSOX3104T which was manufactured just three months ago sell on eBay for around five grand when the list price is about three times that amount.

I got my all option enabled MSOX6004A for $8.5k, which is 1/6 the full price (base unit, BW upgrade and all options).
But for a low end scope like DS1054z, you do not really get any lower since the scope is pretty sold at BOM plus a thin margin.

Wow! How'd you manage that if I may ask? The options in particular is what interests me the most, since I usually find it harder to find fully licensed units.

[Faith]

The third top prize winner of Scope Month decided to cash his scope, hence I got it. Drove 13 hours to his place, then drove back during the night. Slept in service area for a couple of hours.

Ahh should have guessed. Pretty much every time I see a MSOX3104T on sale it was also acquired from Scope Month.

That must have been quite an exciting drive :p

Congrats!~

[Fungus]

I am concerned about some of the issues mentioned in the video:



Have those issues been fixed?

Yes - latest firmware has sorted that.

[eeguy]

As far as I recall, somebody in the forum mentioned that for Keysight oscilloscopes, the seller might be able to enable more functions for free since they are all software based. Am I correct? What about the GDS-2204E? If they can only enable a few functions, what functions would be good to get?

[nctnico]

There are no software options on the GDS2000E series so no upgrade hacks possible. Also I strongly doubt any Keysight reseller will enable options 'for free' unless there is a promotion going on. Remember the Keysight video where they put one of their scopes on a BBQ? Well they are going to use that same BBQ for any reseller who dares to sell their equipment hacked 

[rstofer]

As far as I recall, somebody in the forum mentioned that for Keysight oscilloscopes, the seller might be able to enable more functions for free since they are all software based. Am I correct? What about the GDS-2204E? If they can only enable a few functions, what functions would be good to get?

Apparently, decoding is a standard feature for the GDS-2204E.  I couldn't find anything that indicates that decoding is an extra-cost item.

Unless there is a promotion, like the current MSO for DSO pricing, I wouldn't expect Keysight to do much of anything for the buyer of one oscilloscope.  Buy a hundred or perhaps even as few as a dozen and they might talk.  One scope?  Not much chance...

http://www.keysight.com/main/facet.jspx?cid=x2015004&pageMode=SO&cc=US&lc=eng&cmpid=20801

[eeguy]

As far as I recall, somebody in the forum mentioned that for Keysight oscilloscopes, the seller might be able to enable more functions for free since they are all software based. Am I correct? What about the GDS-2204E? If they can only enable a few functions, what functions would be good to get?

Apparently, decoding is a standard feature for the GDS-2204E.  I couldn't find anything that indicates that decoding is an extra-cost item.

Unless there is a promotion, like the current MSO for DSO pricing, I wouldn't expect Keysight to do much of anything for the buyer of one oscilloscope.  Buy a hundred or perhaps even as few as a dozen and they might talk.  One scope?  Not much chance...

http://www.keysight.com/main/facet.jspx?cid=x2015004&pageMode=SO&cc=US&lc=eng&cmpid=20801

Yes they mentioned about the MSO for DSO price but I am going to use a Logic Analyzer, I would probably buy an usb one. Do they often have promotion one after the other? If MSO is not important to me, perhaps I wait and see if they have a 200MHz or 300MHz for 100MHz price or 4 channels for 2 channels price?