Tektronix MDO3000 Series Scope

[tekman11]

Hi  all,
Has any one here  used the this series of scope?
Is it competitive with the Agilent 3000 msox  series?

thanks

[don]

I've used mdo4000, and  think other than time correlated SA / time domain, screen size and memory they are very similar.  Mdo3000 does not come with mso by default ($1500 I think) but a promotion now includes 3GHz SA ($2500 option) and a protocol decode ($1100).  SA is cool because it has a 3GHz capture bw vs a swept input. Agilent has a good deal now where you can get all options for ~$800.  The tek is nice because it inherets a lot of scope behavior from the higher end tek scopes if you have grown accustomed to using them. I have not used  msox though.  Is this for general debug?   For my useage the tek has the edge but may vary for you.

[luis garcia]

Tek MDO's have a nice feature of being "bandwidth upgradable" by software!
You may purchase the 100Mhz version and upgrade it to 300Mhz. Of course purchasing the license from Tektronix!. For upgrades beyond 500Mhz you need to send the unit to Tektronix.
The waveform generator, meter, logic analyzer and SA are also software upgrades. For the LA you would need the logic probe though.




 

[Hydrawerk]

You have probably seen these links. https://www.eevblog.com/forum/blog/eevblog-587-tektronix-mdo3000-mixed-domain-oscilloscope-teardown/msg507818/#msg507818
https://www.eevblog.com/forum/testgear/new-6-in-1-scope-from-tektronix/
https://www.flickr.com/photos/eevblog/sets/72157641979331094/

Agilent MSOX3000 is different from Tek MDO3000. It just depends what you prefer. MSOX3000 has unmatched 1000 000 waveforms per second while MDO3000 has longer memory and many useful features.

[Hydrawerk]

I do not like that MDO3000 has uncovered power supply. If a capacitor blows, it will damage the whole digital circuitry. While at MSOX3000, the PSU is hidden behind a metal wall.
Check the competitive fact sheet made by Tektronix guys.
Also see MSOX3000 internals. It is built like a tank. I like it. https://www.flickr.com/photos/eevblog/sets/72157626755861230/

[nctnico]

I do not like that MDO3000 has uncovered power supply. If a capacitor blows, it will damage the whole digital circuitry.

That is a very far fetched non argument. I've never seen a capacitor blow in a power supply like this.

[Hydrawerk]

Here is another competitive comparison in pdf. This time made by Agilent (oh, Keysight) guys.
They say that Tek MDO3000 does not have a segmented memory. You can check it in a datasheet from Tektronix web.
On the other hand, the MDO3000 looks nice.

[Hydrawerk]

I do not like that MDO3000 has uncovered power supply. If a capacitor blows, it will damage the whole digital circuitry.

That is a very far fetched non argument. I've never seen a capacitor blow in a power supply like this.

I believe you but many scopes have some kind of a hidden power supply.

[nctnico]

I do not like that MDO3000 has uncovered power supply. If a capacitor blows, it will damage the whole digital circuitry.

That is a very far fetched non argument. I've never seen a capacitor blow in a power supply like this.

I believe you but many scopes have some kind of a hidden power supply.

Hidden? You probably mean shielded. I guess the input circuitry or the PSU is so good in terms of susceptibility or radiation it doesn't need shielding.

[DanielS]

That is a very far fetched non argument. I've never seen a capacitor blow in a power supply like this.

I have seen some power supplies with blown caps - one of them in a computer my older sister wanted me to fix. That cap was in the gate driver circuit so no real damage was done to the rest of the computer. If one of the output caps blow up though, the amount of junk on the affected power rail(s) may very well fry most of the electronics on them (like Dave's LeCroy scope with weird trigger issues until the 3.3V rail failed and killed the memory controller chips) and moot the mechanical protection argument.

Here is one example of a capacitor failing inside a PSU - blew its top off and all that is left of it on the PCB is bits of paper. Cause or consequence? With the amount of damage all around, I have no idea how to figure that out.

[GreyWoolfe]

I have see just the base left inside failed ELO touchscreen monitors that were supposedly refurbished.  Small capacitors (don't remember values) but big mess.

[EEVblog]

Tek MDO's have a nice feature of being "bandwidth upgradable" by software!
You may purchase the 100Mhz version and upgrade it to 300Mhz. Of course purchasing the license from Tektronix!. For upgrades beyond 500Mhz you need to send the unit to Tektronix.
The waveform generator, meter, logic analyzer and SA are also software upgrades. For the LA you would need the logic probe though.

The Agilent scopes work exactly the same way.

[EEVblog]

The MDO3000 represents great value with the now free spectrum analyser included. Shame it's not time correlated like the MDO4000, but hey, you get it for free.
The biggest difference between the two machine is in terms of UI speed and usability.
The MDO is as slow as a wet week. Turn on the spectrum analyser and it practically locks up the front panel at low RBW's.
In the defense of the Tek it's slow because it searches and computes based on the entire sample memory, not just display memory.
But ultimately it's it's old architecture that makes it slow.
If you want a super fast response scope for stress free everyday use, the Agilent is the winner.
If you want more bang-per-buck the Tek is a good option.

[tekman11]

Hi Dave,
The MDO4000 spectrum analyzer looked impression on you tube.
how does the MD03000 spectrum analyzer compare to the MDO4000,  Is the noise floor similar?
How did you like the protocol analyzer(spi/i2c).

thanks tekman

[Hydrawerk]

(...)I believe you but many scopes have some kind of a hidden power supply.(...)

Hidden? You probably mean shielded. I guess the input circuitry or the PSU is so good in terms of susceptibility or radiation it doesn't need shielding.

Yes, I meant shielded.

[Hydrawerk]

The MDO is as slow as a wet week. Turn on the spectrum analyser and it practically locks up the front panel at low RBW's.
In the defense of the Tek it's slow because it searches and computes based on the entire sample memory, not just display memory.
But ultimately it's it's old architecture that makes it slow.

But the MDO3000 has no competitor in its category.

[jahonen]

The MDO is as slow as a wet week. Turn on the spectrum analyser and it practically locks up the front panel at low RBW's.
In the defense of the Tek it's slow because it searches and computes based on the entire sample memory, not just display memory.

Even if it uses the whole memory for the computation and it takes time, that is ok, but it is absolutely inexcusable to lock up the UI meanwhile the computation is being done. UI should not be ever locked up that way even if some lengthy computation is being done in the background. Multithreaded programming, anyone?

Regards,
Janne

[luis garcia]

MDO's have Linux inside. Unless they have set the SA thread priority to something like real time, it is no usual the GUI gets locked. However it is acceptable to do so if you have a long task to finish and want to present results as soon as possible. When do you perform a SA sweep what do you prefer, wait while you play with the GUI or have the results ASAP ?.

To me the real MDO's are the 4000 series. MDS3000 is a cheap version of the real machine.

[don]

The MDO is as slow as a wet week. Turn on the spectrum analyser and it practically locks up the front panel at low RBW's.
In the defense of the Tek it's slow because it searches and computes based on the entire sample memory, not just display memory.
But ultimately it's it's old architecture that makes it slow.

But the MDO3000 has no competitor in its category.

MDO is not normally slow and really only at low resolution as Dave highlighted.   Running SA with 30Hz resolution BW is pretty extreme and obviously not possible on other oscilloscopes.   A  traditional swept SA can take minutes to sweep at 30Hz resolution bw depending on span (and do not have 3GHz capture BW which is awesome for tracking EMI problems). Doesn't justify laggy UI but realistically all SA's are unusable at low resolution BW until capture is complete. And i think MDO might beat other SA's in total time to capture (at least on one test I saw).  Normal scope usage is fast.  Laggy UI is rare.  Pan/zoom at 10M is fast on MDO.  Measurements can be gated with cursor, screen, or entire waveform.  This impacts computation time. Scope is still  useble (not locked up) while measurements are being computed if you choose to compute based on entire waveform with 10M memory, at least from measurements I've used. .  Protocol decode is fast.  Also it works how you would expect. You can capture a waveform in normal yt or roll mode and enable decode. It will compute decode and display.  Same with cursors / zoom. Turn them on and they work on any mode and consistent across all modes. The cost might be computation time in some cases so choose what's important to you (consistent usage in all modes, waveform computation) or fast response from HW based system but more limited usage .

[kilohercas]

MDO3000 is very nice scope, as i start to say, i only need solder iron and power supply on my bench for normal work.
AFG is very limited , but useful, i was able to lock femtosecond laser to AFG as reference, because laser photo-diode output had some problems with PLL, so that was quite fun debugging.

if some one want to trade MDO3054 (all modules, SA,MSO,AFG) to MSOX3000 series scope, please contact me

[MarkL]

...
Normal scope usage is fast.  Laggy UI is rare.  Pan/zoom at 10M is fast on MDO.  Measurements can be gated with cursor, screen, or entire waveform.  This impacts computation time. Scope is still  useble (not locked up) while measurements are being computed if you choose to compute based on entire waveform with 10M memory, at least from measurements I've used. .  Protocol decode is fast.
...

When I tested the MDO4104 two years ago, it took 50 (correct: fifty) seconds to decode a single 20Mpt capture of I2C data.  Two streams took 100 seconds, and so on.  During decode time the UI was utterly unresponsive and unusable.  Plain SA mode locked up several times and required power cycling.  Feature-wise the spec sheet for the scope reads great, but we sent the eval back along with a list of about 35 other issues and UI bugs.

Have you done any specific protocol testing on the MDO3000 that you can share to support your "it's fast" statements?  Has Tek made any improvements since the MDO4000?  My understanding is that it's essentially the same software as the MDO4000, but I'd like to learn otherwise.

[kilohercas]

Have you done any specific protocol testing on the MDO3000 that you can share to support your "it's fast" statements?  Has Tek made any improvements since the MDO4000?  My understanding is that it's essentially the same software as the MDO4000, but I'd like to learn otherwise.

For 1000 to 1M decoding looks like real time, but longer sampling size will reduce performance even when MDO3000 use hardware decoding ( it does looks so)

[don]

...
Normal scope usage is fast.  Laggy UI is rare.  Pan/zoom at 10M is fast on MDO.  Measurements can be gated with cursor, screen, or entire waveform.  This impacts computation time. Scope is still  useble (not locked up) while measurements are being computed if you choose to compute based on entire waveform with 10M memory, at least from measurements I've used. .  Protocol decode is fast.
...

When I tested the MDO4104 two years ago, it took 50 (correct: fifty) seconds to decode a single 20Mpt capture of I2C data.  Two streams took 100 seconds, and so on.  During decode time the UI was utterly unresponsive and unusable.  Plain SA mode locked up several times and required power cycling.  Feature-wise the spec sheet for the scope reads great, but we sent the eval back along with a list of about 35 other issues and UI bugs.

Have you done any specific protocol testing on the MDO3000 that you can share to support your "it's fast" statements?  Has Tek made any improvements since the MDO4000?  My understanding is that it's essentially the same software as the MDO4000, but I'd like to learn otherwise.

I have not yet used MDO3000, but use the MDO4000 regularly.  My understanding is the MDO3000 is faster than MDO4000 (at least original MDO4000).   I have never encountered anything near 50s to decode a 20Mpt I2C record.   I can't give you a number because I never measured, but seems like an order of magnitude faster than what you were seeing.  Did you have have several full waveform measurements on?  Rough idea of how many I2C transactions?  I would guess I usually have under 100 transactions.  I'll pay attention next time I'm doing some decoding. 

My "fast" statement is subjective.  To put it another way, I'm going to spend A LOT more time parsing through 100 I2C transaction that I will to capture.  Or I'll spend more time waiting for the trigger to occur. Once data is captured and decoded you can zoom and pan and jump from transaction to transaction in real time.   To me, this is critical. Once I capture something, I want to be able to do what I want with captured signal.  I'll wait up front a bit, but I can't deal with capturing a signal only to find I can't zoom or decode after it's captured.  This is where I think the tek scopes excel.   Even a simple thing like being in zoom and having cursors way at the other end of record.  You can probably find a button "click to center" on all scopes, but the tek scope accelerates the scrolling when off screen and in zoom so it's still only a few turns of the knobs to bring it on screen.  Little things like this make it fast in my opinion.

35 issues and bugs?  That's a lot.  Something must be different about how people use scopes.  I tried a siglent SDS2000 and Rigol DS4000 (with the intent of purchase) and found them unusable for my needs.  I find the Tek scopes (including MDO4000 to work without issue.  Maybe even the same firmware you found 35 issues on.  Go figure, right?  Good thing lots of scopes to choose from.

[Omicron]

For those using an MDO4000 or MDO4000B: be sure to grab the latest firmware from the Tek web site. It brings you all the goodies of the MDO3000 (apart from the AWG obviously), including the new UI look. Most notably it gives you the new FastAcq mode and so upgrades the max. wave form capture rate from 50.000 wfps to 250.000 wfps. It also adds the nice color graded display modes. Well worth the small trouble of doing the upgrade! This also shows that there is nothing "faster" about the MDO3000 compared to the MDO4000, it's all just software optimisation that has been done.

A lot of the the early reports about the UI locking up are known issues with the initial versions of the firmware. There have been several updates since then that, as far as I can tell, remedy these types of things. At least I've never had mine lock up (been using it for about a year). Some operations do take time if you have the full capture memory enabled, there is no denying that. That fact that it always uses the full capture memory for measurements however, does mean it can do things an MSOX simply can't. E.g. try measuring the rise time and width of a pulse at the same time on an MSOX. You have to optimise the display for either one or the other, but you can't do both with optimal accuracy. That's the downside of the approach Agilent uses. On the Tek you just select your memory depth so it's running at full sample rate and all measurements you do are at best possible accuracy.

[EEVblog]

But the MDO3000 has no competitor in its category.

True, and have said so in my videos.
That still doesn't make it the best choice for an everyday use scope.
YMMV.