EEVblog
HTML-код
- Опубликовано: 21 окт 2024
- Dave shows how Tektronix digital oscilloscopes work fundamentally different in their Auto trigger mode from other brands.
And also some random ad-hock comparisons between the Tektronix MDO3000, Agilent MSOX3000, Rigol DS2000, and GW Instek GDS-2000A oscilloscopes.
Forum: www.eevblog.com...
EEVblog Main Web Site:
www.eevblog.com
EEVblog Amazon Store:
astore.amazon.c...
Donations:
www.eevblog.com...
Projects:
www.eevblog.com...
Electronics Info Wiki:
www.eevblog.com...
Hey, just found you and I totally enjoy your enthusiasm and sense of discovery.
I am retired now for quite a while with almost 50 years in the test equipment and metrology background. I spent the first half of those working for Tektronix. I was with Tek through The transition years from analog to digital technology in oscilloscopes.
Indeed, the engineer at Tektronix is correct in saying that all of Tek’s digital scopes operate in this fashion. As a matter of fact, all of Tektronix analog scopes operate in this fashion also.
So, as I see things, you should probably be asking a different question. The question should be, ”why do all the other scopes not operate in this fashion?“
Here is my rationale. “AUTO”, “NORMAL”, and for that matter “LINE” and “SINGLE” are SWEEP MODES… NOT TRIGGER MODES!
So let that sink in.
“AUTO mode is short for “Auto Sweep in Absence of a Trigger”.
The “NORMAL” mode is the quiescent (normal) condition of the sweep, in this condition the sweep does not run until the trigger logic has been satisfied.
This is how TEK TDS, DSA, MDO all work, this is how 2465, 465, 453, 7B53a … all Tek analog scopes work too. And it’s always been at about a 20 Hz rep rate.
Why? AUTO sweep provides a convenient way of setting vertical DC levels.
AUTO allows one to set up the sweep triggers for all but the very slowest Repetition signals.
Therefore:
AUTO is not auto trigger at Tektronix, AUTO is Auto Sweep in Absence of a Trigger…pure and simple.
If you want to do some Research on an interesting aspect of auto trigger derivations at Tektronix, look into the purchase of Telequipment by Tektronix back in the 1960’s. It was a Necessary evil to avoid a circuit design infringement lawsuit by Telequipment against Tektronix. It was cheaper to buy Telequipment than pay them for the use of the AUTO circuit design.
I have a agilent DSO3000 and it have a anomaly too. In Auto mode it will not trigger at slow signals like ~2hz.
The Auto trigger mode works the same as it does on the old analog scopes too. The sweep is always driven by the trigger. In the absense of the trigger, it runs at the auto-trigger's self-generated trigger rate. I've seen circumstances where other non-Tek scopes, in Auto mode, can miss a very infrequent event events (like once per second or so). I'm not sure why this is. It must have something to do with the way other scopes switch between auto and triggered mode.
sarowie
Yep, you're probably spot on there. I was bought up on analog scopes. Youngsters coming though who have never seen nor used an analog at school or hobby level would probably have a different appreciation of features.
The Tek "feature" that shows a waveform, but really slowly when not triggered, might well ultimately be useful, but it still gives me the heebie-jeebies. And it doesn't instill confidence in me as it gives the appearance of the scope beign "slow" when in fact it may not be. And the Tek's aren't the fastest things going around as it stands!
I'd bet if Martin were not holding back the foodwaters right now he would be testing his new Tek with the cardio pulse to see if the Tek trigger method can better capture the slow frequency waveform than the other scopes he has.
EEVblog Agreed. In the end, I find being able to see what the waveform looks like in order to setup the proper trigger conditions quite useful.
Rob B I checked out Martin's example on the MDO3K, and had no trouble triggering each time.
was there ever any doubt?
Very interesting demo. Great post Dave.
I have an older 2GHz LeCroy scope that behaves exactly the same as the Tek scope in terms of the auto-trigger rate being considerably slower. It does seem to be intentional, so that you can make out some of the waveforms that are passing through, thereby giving you a chance to set your trigger up correctly in the first place. :)
You wound the wick "out" on the Tektronix. I thought the wick could only be wound *up*! Mind. Blown.
That tek feature actually looks useful, i have a Agilent scope and yeah when its not triggering you only see a big blurry mess, but the tek shows at least something.that you can sort of make out the waveform shape from, but yeah im used to seeing it super fast even when not triggered
On the Tek, if you do a Single vs. hitting stop, you'll get a single trace.
Absolutely great stuff, Dave. I'd love to see more videos like this.
Dave - perhaps the screen is actually only updated 25 times per second. When in Auto mode untriggered, it doesn't bother to acquire any faster than it can actually update the screen. When triggered, it acquires as fast as it can until the next time it is going to update the display and then it performs processing to merge all those acquisitions into one screen update.
The purpose of autotrigger goes way way back, it is to allow you to actually see a waveform as you are adjusting the trigger. One would never actually do measurements with autotrigger on, you only take measurements once you are in normal trigger, as auto trigger actually interferes with triggering. The Tek allows you to see the signal better than the Agilent in auto mode, in the example shown, it seems to me. I learned on the old tube Tek scopes from the 60s, they were almost too heavy to lift. Great work, I enjoy your videos!
I really like these scope videos. There is something about an oscilloscope that is memorizing to watch. I always preferred the Tek scopes but today, I would only buy an Agilent. Much of that decision has to do with watching your scope videos. Thanks Dave for the great info!
I much prefer being able to see the signal so I can setup my trigger.
Also, the Tek should have a "Single" button which will let you do a single shot capture vs hitting the "Run" button.
14:42 - perhaps some aliasing related to spread spectrum vs. missing chunks of triggers in fast acq? Have you tried manipulating the input signal?
Great video, now I like even more my Agilent scope ;)
I think it is intentional. If the free running update rate is very fast then it will not lock on to waveforms with a slow update rate and you would have to turn off auto trigger to see them properly. I haven't had many different oscilloscopes but the ones I have had all seem to have a very slow update rate when free running in auto. I also have an old analog Tektronix that works in a similar way except that the update rate increases as the timebase is made faster, to make sure the trace remains bright enough to be visible. For instance a 5nS/div trace would be invisible at an update rate of 25Hz so the update rate would be increased to a few kHz but for a 100uS/div trace it wouldjust be about 10Hz. With a digital scope there is no longer the need to speed up the update rate as the trace visibility is no longer dependent on persistence of vision.
Dave, the old LeCroy Scopes fore example 9400a have the same issue, it is well known ;). it is the the method of sampling. So is not only Tektronix but don't know for modern LeCroy
Hey Dave, about your mic preamp, how about the Nady DMP-2 for battey powered, and the Behringer Xenyx 302USB for USB-powered, both good + cheap. (sorry, don't have Tweeter...)
i'm just drooling over these scopes while frowning at my old crusty Heathkit.
You Heathkit will still be working in 30 years though. Will these?
Hi Dave, any chance to have an Oscilloscope shootout like you have done to multimeters?
A multimeter shootout take over and hour and a couple of hundred clips, I don't want to imagine what a full oscilloscope shootout will take!
I'd guess the "trigger timeout" is a hold-over from the analog scope days. Back in the day of analog you would definitely want to wait for an event as long as you could. With short sample memory you'd also want to wait as long as you can.
Surely if you ran a scope in free-run without any trigger, at 80,000-250,000 updates/sec, you'd only get a solid block across the screen? Slower updates give you an idea of whats happening to the signal.
No doubt
I agree completely. I see it as a feature, not a problem. Perhaps the other makers should get cracking. Unless it is patented by Tektronix.
yup.
I think what is happening here is:
What you call an "update frequency" is not a screen update rate (that's obvious).
I guess that the seqence for Tektronix is: capture some data -> process it -> wait for next trigger -> start new data block -> capture ... -> display processed data on the screen (that is done 25 times per second). If there is no triggering condition met then all the data is captured in single very long block of data. With that the scope has to process single large block of data and display only few hundred of samples from the beginning of the block (processing is not distributed over time hence the lower update frequency in free running mode).
For other scopes it might be that they divide incoming sample data into blocks regardless of that if trigger condidions were met or not. They probably just take a bit more data that they can display on a screen and then start a new block.
And the jitter might come from the fact that there are probably more processes or threads which are running on the cpu (like front panel handling, interrupts, etc.) simultaneously with data processing.
I prefer Tektronix for all reasons and lower refresh rate is useful in certain circunstances. But the behavior like a analog oscilloscope is much pleasure to see, it's true
It's most likely for energy saving as the signal processing uses energy. Can you test the power consumption of these devices in different settings? I don't have one so I can't test that theory myself.
Yep, you notice it with high-end Tek oscilloscopes. These gets relatively loud when triggering at high sample rate.
I prefer the Tektronix display. You have no idea of the signal you are looking at in the un-triggered (auto) mode on the other Oscilloscopes, unless you stop the acquisition, and then you assume what you see is correct. What would happen if you had no idea of the type of waveform your were trying to measure? Would you trust what you see after you've stopped the acquisition?. On the Tektronix you can get a very good idea of the waveform, even while untriggered, and very easy adjust your trigger level to get a triggered condition. With the Tektronix superior trigger circuit, it is easy to get a triggered waveform. Why would you want to make measurements on an un-triggered waveform? Can you trust what you have captured after you have stopped the acquisition. The normal and only trustworthy measurement is a good triggered acquisition, where you are sure that you don't have aliasing. Furthermore, I prefer the approach where I can select my own record length, with proper search on my acquired data, and gating, to ensure that I measure what I really want to measure. Maybe it is my 30+ years doing calibration on all types of equipment, that made me more aware of capturing the correct information and analyse that information, than trying to assume what I see on the screen is correct.
Do you get to choose any parameters for the ads that run before your videos Dave? I sat through a 4min ad for a product called PooPouri before this video. It was funny, but doesn't seem very relevant to your channel..
By the way, do you still get paid if we skip the ads? I try and let them all play just so you get more revenue, but if you get paid either way..
That's Tek way to show that their signal is "cleaner" when no signal is injected, as noise comply to the Gaussian distribution, the lesser you sample/update, the thinner line you will see on the screen. That's how Tek guy saying that they has a "less noisy" scope compare to others..
I would think that the natural behavior of a storage scope in free running mode would be to stuff signal data into the current memory window as fast as possible. Maybe the Tek automatically enlarges the memory window to max when it does a free run?
memory window = memory depth
Hi there, so which would you recommend out of something like the TBS1022 Tektronix Oscilloscope or a Rigol, my main interest is repairing CB radios 27mhz FM of all things...Fred in England.
I disagree with your comment that the Rigol has a better intensity grading. In my opinion, the Agilent has better intensity grading. Take a look at a composite video signal with movement and you will see better contrast in the grading on the Agilent
I've done a comparison vs an analog scope and IMO the Rigol was better than the Agilent as it was closer to the analog look. It could depend upon the signal type of course.
Just got an MDO! The feature set is amazeballs but the compute speed is pathetic. Sometimes takes 2 seconds to register a button press.
Dave, can you take apart the GW scope and investigate the trigger update rate? I am not sure how difficult will be to find the internal trigger signal, but I think it will be interesting to see if the claimed update rate is only marketing crap. Thank you for the video.
There is an easier way. I can feed in a known glitch and see how long it takes on average to capture it, that will confirm a ballpark figure.
EEVblog Yes, that sounds great. Thanks.
I just measured 80kwfs on the GW using the trigger output, but only on the 500ns range. It's decided to work now for some reason. I did hit the Autoset button :->
EEVblog Thanks Dave, it seems like the spec is relatively accurate.
This is a rather silly comparison. Operating a scope in the untriggered mode does not produce useful info on any scope so update rate is irrelevant.
Love the Tshirt
Hello
Mr.
Please help me.
My Tektrinix TPS2012 broke the screen . Indicate which equipment tektronix have the same screens , because I'm not finding in the Brazilian market the screen with specification of the TPS2012.
Thank you!
maybe it is for saving heat and power, if no signal don't worth to work in full speed. Most of the time an osciloscope is waiting for a signal to be inputed.
reading all the comments rather split views on what is better, I like that TEK scope though.
All my Tek scopes are slow in the auto trigger mode. They seem fastest in normal mode. Hp seem as you say unaffected.
Here one more thing to investigate: Agilent scope (DSO-X 3000) cannot count pulses corectly! It counts different things when zoomed in and out (on the same record). Also, in auto trigger mode very often misses trigger events (it is visible on the screen). Respectively - in normal mode misses single trigger events. Tek's does these thing correctly!
My explanations for the trigger behavior is:
- Tek's trigger circuit waits for event. If no event occurs after some timeout "fake" event is generated and the waveform is displayed. Hence the slow update rate. The positive side - it is very unlikely to miss single trigger event during display update.
- Agilents trigger circuit updates very fast, but it needs some minimal count of trigger events in order to "sinchronize". It resembles the action of PLL circuit (only visualy). IMHO
If it only update by about 20hz when no signal, what if you let it monitor a circuit that make a sudden pulse that's only take 5ns, wouldn't it miss it?
Screen updating every 40ms?
The low (19,000.222Hz) is 19Khz.
The high, (248,000,000Hz) is 248Mhz.
Surely?
Please correct me publicly if I'm wrong, Dave :)
.
You should have a vote, Feature or Fault.
works great perfect !
but what about frame rate of the display?
It may be some sort of power saving mode, like Intel's SpeedStep, perhaps? While doing nothing, it's trying to consume less power? Just a thought.
A few people have said that. That is likely a side effect of this, but I doubt that would have been the design reason. Benchtop test gear makers generally don't give a rats about power consumption.
Considering that nearly all industrial LCD displays update no faster than 60Hz, I can't imagine why you would want an update rate faster than 60Hz in a non-triggered mode. Can anyone explain why that would be desirable?
Sometimes I like a bit of waffle in the morning with a cup of coffee. :)
You have to check out what Ben Heck made! Speak and Dave Jones!!! UNBELIEVABLE!!!
MSOX Scope ... 3000 series ... wow ... That's my personal Ferrari :)
Dave, that 25Hz blanks might be due display update.
Sounds exactly right. Tek scopes seem to have a very slow visual response to inputs, whereas my Agilent (same as Dave's) reponds "instantly" to most any change I make to the settings. The Tek running at 25Hz and the Agilent running at 60Hz pretty much covers exactly that.
From an engineering perspective, I find it interesting that they are using single-ported RAM for waveform storage, especially on the Agilent where they are loudly touting the mega-trigger capability as a means of not missing any glitches. If there's a window every frame when the memory "turns around" and reads out the display data, there's a window where it's missing triggers and this any potential glitches....
That would be disappointing if a scope stopped acquiring because another component (screen driver) had to read data.
Michał Gawron I imagine that the fast acquisition cache for the ADC and triggering subsystem is disconnected from the display memory. I wonder if there's a snapshot process that takes a window of cache at a moment in time then renders it. That blank may be the snapshot request.
Pretty fast updating scope for it's gay!
www.its-not-its.info/
RFC3514 Pretty fast for (it has) gay? lol
9:50
+mbplove haha just read you comment as he said that!
fruedian slip perhaps?
9:54 sorry but lol immature I know...
I almost won one of these tektronix's at a convention.
I just got one
That old tektronix scope said Australian Defense Force on it.
He bought a bunch of decommissioned stuff at auction a little while back :)
yes?
23:34 jeez, how many oscilloscopes can a man have..?
16:10.
On the tek you hit run stop and instead of performing a single capture, it just freezes the screen. I like this functionality. If I want the agilent functionality on the tek i just hit the 'single' button to stop instead of the run-stop button. Problem solved. This way, if the multiple traces look interesting i can freeze the screen, put down the probe and show someone the results. I cant do that on the agilent.
Tek also has the search feature which is really cool. You can stop capture, turn on wave inspector and enter search parameters. It finds all occurrences and you can step through them really fast looking for consistency or oddities. This is great when you have no idea what you are looking for, but you know it looks different than all the rest of the signals. Now, the 'wave inspector' concept is totally retarded. It is just zooming in and scrolling, but now you can scroll through time with a spring loaded knob. whoop tee doo. But whatever it works.
Agilent though does have a nice feature. When you turn the time base knob 4 clicks, it zooms 4 times. On the tek, if you turn the time base knob 4 clicks, it zooms in 3 or 4 times, depending on how the scope feels at the moment :) Yes, it is exactly as annoying as it sounds.
As for the auto trigger update rate. I like the agilent much better. The tek makes me feel like my signal is running at 40hz, but it isnt. Though seeing the signal is nice. It is obviously a purposeful design decision.
What's the difference between grown-ups and children toys?!
Not that I needed it, but here is yet another reason why I'll never buy a Tektronix oscilloscope.
And the tektronix is about 5000$ cheaper. You cant compare supercars to toyota prius.
No it's not. The Tek MDO3000 is a comparable price point to the Agilent MSOX3000 at all the bandwidth levels, it's designed to be that way.
I see what ... 50K USD test equipment ... and lost interest :)
GW instek is crap, got one big ass one at work 6 years ago, channels started to fail,. don't turn on, nor turn on with noise. Chann 1 only show noise, at first we though, we blown the input. Then started to work and the stopped again We take it to the dealer and said it isn't fixable CRaaaaappppppp
CHECK OUT BEN HECK SHOW CHANNEL ASAP!!!!
Ah, it's just all that dang digital circuitry. Analog is better for this purpose.
Why did you delete comments? cloudup.com/cv3x3pHc_ua
Because it was garbage.
Perhaps because this comment is crap? The Agilent is faster, that's a fact. He proved it. So if Agilent really bought Dave to advertise for them: Good job. No marketing wank, no lies.
(sorry, didn't mean to feed the trolls)
I never saw your message and did not delete it. Post it again and you'll see I won't delete it.
Perhaps super-awesome-google-comments system automatically put your comment to spam. Who knows.
Clearly the Tek is wanky. Send it to me and I'll take care of it for you ;)
works great perfect !