BTW #2 performance is just what i would expect from this scope seeing what's going to replace, that is still sold at equal or higher prices, 15 y/o architecture with abysmal wfm/s and 2.5kS memory per channel...
What scope are you assuming this replaces? The TDS2000C?
How is this thing ever going to catch runt pulses ? or glitches ? it almost looks like they are doing software triggering ...
that the screen freezes when you are zooming on or out, ok , but not when you change trigger level ! if i slide that trigger level up or down the machine should keep capturing .. that's how you seek out glitches in the first place. play with the trigger level until it traps something that is not normal. you do not know what you are looking for in the first place. it is an abnormality. so you need to bank on repetitive , fast and reliable acquisition to fish for it. This scope is sleeping 90% of its time , and completely dead if you do anything on screen.
they have a powerful FPGA and multicore processor. why is this not handled in the hardware fabric ?
I used to joke that Tektronix in Dutch was written as Trektopnix (looks like junk). I got bitten multiple times on their TDS540 (i believe it was a 540. it was 4 channel machine with only 2 adc's). equivalent time sampling ... what a joke. And using two adc's for 4 channels. That re-trigger (for the old school scope users : so they effectively run in ALT mode, not chop mode.). So the acquisitions are desynchronized in time. ( the sweep for channel 1 and 3 is taken first , then they throw the multiplexer and sweep channel 2 and 4. you are not looking at the same point in time. if something happens on channel two immediately after you triggered on a channel 1 event you will never see it. (that's what bit me...)
The acquisition system should run independent of the visualization system. That should be easy on that Zync... i wonder how come that is not the case. Are they using this thing just as an application processor ?
Maybe, since this is a prototype, they are still working on the FPGA logic. it could be that this will be fixed in the production machine or a later fpga build. time-to-market could have been a driver. benefit of the doubt. but it doesn't smell right.
How is this thing ever going to catch runt pulses ? or glitches ? it almost looks like they are doing software triggering ...
Lecroy gets away with that just fine but I doubt that Tektronix does triggering in software. They probably do that inside an ASIC.
that the screen freezes when you are zooming on or out, ok , but not when you change trigger level ! if i slide that trigger level up or down the machine should keep capturing .. that's how you seek out glitches in the first place. play with the trigger level until it traps something that is not normal. you do not know what you are looking for in the first place. it is an abnormality. so you need to bank on repetitive , fast and reliable acquisition to fish for it. This scope is sleeping 90% of its time , and completely dead if you do anything on screen.
Just acquire a longer record and repetitive abnormalities are easy to spot. R&S has inverse color grading which is excellent for this purpose. But for anything that occurs with long intervals, triggering is your best option. BTW: If you blink with your eyes on a high waveform update rate scope, you are also likely to miss an abnormality. The same for most of the cases where you are not running the oscilloscope at it's peak update rate. IOW: high waveform update rates are highly overrated.
How is this thing ever going to catch runt pulses ? or glitches ? it almost looks like they are doing software triggering ...
Lecroy gets away with that just fine but I doubt that Tektronix does triggering in software. They probably do that inside an ASIC.
LeCrap ? same class as Trektopnix. I used Z6. Super expensive machines. you needed ear-covering noise cancelling headphones, the fans were that loud. And their logic analysis add-one only lied on screen. it showed stuff that wasn't there and could not maintain a time-lock with the analog channels. Replaced it with an Agilent MSO. And that dinky pluggable frontpanel was nothing but trouble. Nothing but bad contacts. wiggle a knob and the usb disconnect so your scope sat there going pling-ploing all the time (these are windows machines so it chimes everytime a usb device connects/disconnects). and the quality of those plastic knobs on the rotary encoders.. after a year they all split and fell off. move your scope from cubicle to lab and you leave a trail of encoder knobs in the hallway.
Lecroy ? no thank you. i'm done with those guys...
No asic. we are talking this Tek 2 series. there is only that Zync FPGA in there.
Just acquire a longer record and repetitive abnormalities are easy to spot.
it's the non-repetitive ones that i'm interested in. Put the thing in infinite persistence and let it run. That's where update rate is important. The faster the hardware can cycle data the faster you will trap events.
This Tek thing sleeps 90% of its time. and then it does hiccups as well...
No asic. we are talking this Tek 2 series. there is only that Zync FPGA in there.
Look at the teardown. There is another Tek ASIC in there which could have the trigger engine inside. I'm not 100% convinced that this is (only) for the AWG function as Dave & Shariar suggest.
I do fully agree with you about the poor build & hardware quality of Lecroy. It is not what you would expect from such expensive equipment.
Personally I don't see probe detection by means of a ring around the BNC as a big thing.
Me neither - 99% of the time you're using a x10, and if you need anything else, you're fully aware of it and it's no big deal to select the probe manually. Conversely I have had times where a bad connection on a probe ring caused some puzzlement over wierd amplitude readings
Would just like to point out that in my professional career I’ve only ever used scopes on which the probe detection was in some way broken as well. Same applies to the old Tek analogue scopes.
As for probes I do run my scopes with a BNC and T piece termination a lot. But I can press two buttons and not die of exhaustion switching it to 1X mode.
Would just like to point out that in my professional career I’ve only ever used scopes on which the probe detection was in some way broken as well. Same applies to the old Tek analogue scopes.
The only issue I've run into with Tek is that most of them won't correctly read a 20X probe. But as far as other makes go, I'd much rather have no probe sensing rather than a system that gets it wrong and promptly overrides your manual input. I was almost at the point of making some insulating rings until I just lucked into an older non-readout version of my 100X probes.
On Dave's video at 17.38 the runt pulse happens every 838.8ms, so about once per second (per spec), I don't understand how the 3054 is showing it constantly on screen, at what appears to be a much faster rate than that?
Maybe they realised that having advanced triggers that catch a glitch 100% is better than any waveform update rate figure. Remember that high waveform update rates is something that has been hyped by Keysight. IMHO it is not an important number at all.
But its possible to have both. With rigol, the simple triggers will have high wfm/s, but then when you enable the really advanced triggers (mask testing, etc) the wfm/s rate drops dramatically as its done in software.
I don't see how Teks triggers are anything special, maybe I missed some? Shahriar showed off the runt trigger.
why compare backwards to superseded models? What reason to choose 2000A/3000A/3000T instead of 3000G/1x00a/1x00G ? 3000G is same price as 3000T but more features unlocked/included.
It really doesn't matter though, because as you said it, it is the same price.
Same price but different features, so making comparison to older/outdated model is intentionally misleading/cherry picking. But seem to be the flavor of this thread. New scope from Tek competitive with other brands superseded models, completely ignoring where the market has moved to. New scope should be compared with current competitors.
The only comparison I made was for the speed and the number of channels. I think you misunderstand me. I have a MSOX3104T at work, which I selected. I wanted a 3054T, but this was available for practically the same price so I choose this one. I'm not trying to compare it to the 3054G or A, because honestly, I don't even know what is the difference between them. When I say they came out with it, the only thing I saw was that it's ugly color, and yawned. Even if I would pick a scope now, I would probably pick an agilent one, because the Keithley and Tekrtonix was able to let me down in some completely unexpected and innovative ways with every equipment that I bought from them over the years.
Like this one, missing the logic analyser part on launch?
I am just trying to put it's price in context.
IOW: high waveform update rates are highly overrated.
Overrated, perhaps, but I find a reasonably good update rate combined with color grading and medium persistence to be a very good way to find certain types of abnormalities that might be very, very difficult to trigger on even if you already knew they were there. This is one of the capabilities that for me puts the $500 Siglent SDS1104X-E into the outstanding value category. I can't imagine paying good money for a pro scope that doesn't do this as well. Of course, perhaps the Tek
can do it and it just takes a slightly different approach or something. The persistence is clearly very short in the video and no color grading is used. I'd be glad to check it out if Tek wants so send me an a evaluation copy...
IOW: high waveform update rates are highly overrated.
Overrated, perhaps, but I find a reasonably good update rate combined with color grading and medium persistence to be a very good way to find certain types of abnormalities that might be very, very difficult to trigger on even if you already knew they were there. This is one of the capabilities that for me puts the $500 Siglent SDS1104X-E into the outstanding value category. I can't imagine paying good money for a pro scope that doesn't do this as well. Of course, perhaps the Tek can do it and it just takes a slightly different approach or something. The persistence is clearly very short in the video and no color grading is used. I'd be glad to check it out if Tek wants so send me an a evaluation copy...
Yes, no intensity grading is shown in the videos of this new scope so it looks like it doesnt do it at all. So its very slow for some reason, without even doing what the competition manages to do.
BTW: If you blink with your eyes on a high waveform update rate scope, you are also likely to miss an abnormality.
Still persisting with these lies? Waveform update is not equal or related to the display update or persistence, they are orthogonal. Most (all?) intensity graded scope displays have a control over the persistence so that a single outlier event remains on the screen for some significant fraction of a second (or longer if you set it). Blinking does not miss it, they aren't shown for the 1/1000th of a second or just one frame of the display rate.
Update rate always seems to get certain people here hot under the collar. Its a good thing to have and has some specific use cases in the extreme, but there is a middle ground where most products sit (far below the theoretical peak rate) fast enough that it feels fluid and interactive. Thats where the sorts of long gaps in display update that the Tek shows become noticeable and do impact the user, some of the Rigols do it too.
BTW #2 performance is just what i would expect from this scope seeing what's going to replace, that is still sold at equal or higher prices, 15 y/o architecture with abysmal wfm/s and 2.5kS memory per channel...
What scope are you assuming this replaces? The TDS2000C?
It sounds like they are the (unicorn) target market who already own a battery Tek scope, and want a replacement... except the old Tek battery scopes which were very competitive in their day:
TDS3000 500MHz 5GS/s 10k points
TPS2000 200MHz 2GS/s 2.5k points,
isolated input channelsIts that last point, a scope with isolated channels is a compelling thing. This new Tek 2 is not, nothing special or interesting to make it stand out.
Maybe they realised that having advanced triggers that catch a glitch 100% is better than any waveform update rate figure. Remember that high waveform update rates is something that has been hyped by Keysight. IMHO it is not an important number at all.
But if you don't know what the glitch looks like how can you trigger on it ? My old 90's TDS784A with Instavu could see those glitches no problem and then you could set your trigger to catch them
cheers
Me neither - 99% of the time you're using a x10, and if you need anything else, you're fully aware of it and it's no big deal to select the probe manually. Conversely I have had times where a bad connection on a probe ring caused some puzzlement over wierd amplitude readings
I like the Tekprobe interface on my older TDS scopes a lot. In addition to standard 10X probes I also have a couple of different HV differential probes and it's really nice to just be able to plug it in and the scope knows what's there. I wouldn't want to give this up, although there are lots of other ways similar functionality could be achieved these days.
It sounds like they are the (unicorn) target market who already own a battery Tek scope, and want a replacement... except the old Tek battery scopes which were very competitive in their day:
TDS3000 500MHz 5GS/s 10k points
TPS2000 200MHz 2GS/s 2.5k points, isolated input channels
Its that last point, a scope with isolated channels is a compelling thing. This new Tek 2 is not, nothing special or interesting to make it stand out.
I have a TDS3000 and absolutely love it, it's a joy to use, although the memory depth is a bit low by modern standards. It's a shame they don't make something quite like those anymore, although by the later days of the TDS3000C series the price was obscene for an instrument that had been in production that long.
It sounds like they are the (unicorn) target market who already own a battery Tek scope, and want a replacement... except the old Tek battery scopes which were very competitive in their day:
TDS3000 500MHz 5GS/s 10k points
TPS2000 200MHz 2GS/s 2.5k points, isolated input channels
Its that last point, a scope with isolated channels is a compelling thing. This new Tek 2 is not, nothing special or interesting to make it stand out.
I have a TDS3000 and absolutely love it, it's a joy to use, although the memory depth is a bit low by modern standards. It's a shame they don't make something quite like those anymore, although by the later days of the TDS3000C series the price was obscene for an instrument that had been in production that long.
I used TDS3000's plenty in work, a good scope for the era with a not-slow ui. Also, they had intensity grading.....
Something like that with serial trigger/decode and 1M or 10M of memory would have been a reasonable upgrade/refresh. Tek DPO4000s (released 2006) did come to that but the UI slowed down when loaded up.
I'm wondering if the 2 series puts up a better show on wfm/s in segmented ("FastFrame") mode? And what the 3 series "run mode" performance is if ">280000 wfm/s" is in fact a segmented mode specification?
That would suck if it is in segmented mode. Can anyone confirm?
IOW: high waveform update rates are highly overrated.
Overrated, perhaps, but I find a reasonably good update rate combined with color grading and medium persistence to be a very good way to find certain types of abnormalities that might be very, very difficult to trigger on even if you already knew they were there. This is one of the capabilities that for me puts the $500 Siglent SDS1104X-E into the outstanding value category. I can't imagine paying good money for a pro scope that doesn't do this as well. Of course, perhaps the Tek can do it and it just takes a slightly different approach or something. The persistence is clearly very short in the video and no color grading is used. I'd be glad to check it out if Tek wants so send me an a evaluation copy...
Persistence isn't going to help you if the scope doesn't capture the glitch in the waveform to begin with, and that's what's happening here. A slower waveform capture rate means less statistical chance of actually sampling the glitch. If oyu have captured it then you can indeed use persistance to ensure that you are better able to see it visually.
Please note that in my video the scope catches (triggers on) the runt events just fine even when the event never shows up in the screen during regular edge trigger mode. Even though the update rate is slow, the triggering does not seem to be limited by it.
Persistence isn't going to help you if the scope doesn't capture the glitch in the waveform to begin with, and that's what's happening here.
I agree with the first part--and using your test the Siglent SDS1104X-E has 10-15k wfms/s actual throughput, with the burst rate up to 22k and about 15ms dead time every 50ms. Its not a million wfms/s, but it is pretty good at even rare glitches. But in your video you note a little flicker or glitch on the Tek that doesn't show up very well and I'd want to know if the persistence or color grading settings could do a decent job of making it show up with just standard edge triggering.