Trying to avoid shorting out circuits with the test/ground leads, I bet.

??

We have a number of Tek oscopes at work. I don't know what models they are, but I bet they're quite high end. (We design chips that run at very high frequencies.) Regardless, I'll try that out on Monday with one (or more) of the scopes and let you know. Remind me, though. I'm almost certain to forget.

Hi,

If I have time, I will get some waveforms for you too. I have an HP 250 MHz 1GS/s Digital, aa HP 400 MHz 1GS/s Digital, 350 MHz Tek 485, HP 1GHz Digital, and if I have time I might get our my Tek 7104 1GHz Analog.

Ross

Thanks, guys.

I'm mainly interested in results from scopes that cost less than $2000 or so, but a higher quality one or two would be fun as well.

Really, it's been so long since I had a scope, that all I can remember are smooth lines on them. And this scope never has smooth lines, so I don't know whether this is a problem or not.

Cheers

I can tell you now that the screens you've posted largely don't look ridiculously noisy.

I've got a TDS220 on my desk at work, if nobody else has managed to help by monday (and if I remember!) I'll have a look for you.

I think we have a TDS2000 in the Singapore office (where I am now). I'll have a look on Monday (also).

You are possibly starting to see the resolution of the AD converter coming in to play. Typically only 8 or 10 bits (i.e. the VGA resolution actually shows less detail with a 10 bit converter).

You mentioned connecting "nothing" to the probe inputs. That will pick up all sorts of noise and junk. You should always have something connected (even if it's tied to the ground clip which will still act as a loop antenna like Ross mentions).

That 1kHz test signal did look a bit noisy to me though.

Our company had a high end LeCroy scope (about $25000), and we noticed that the probes picked up noise from the scope display. I wouldn't be surprised if the probes were picking up noise from the lamp.

I wonder about that. The scope has 8-bit A-to-D, but how linear is that?

On the 480 pixel tall display, that's perhaps 2-pixels per A-to-D bit of resolution, assuming they map the A-to-D range to the height of the screen. Possibly a bad assumption, but the "noise" is more than 2-pixels worth.

Cheers

Sorry Mark, didn't get a chance to look today. Was really busy and tired because I spent from 1am to 4am catching an escaped hamster that had eat through its "hamster home" - otherwise there would have been tears!

Humn, not just that but what about the accuracy of the A/D (lsb)?

Ahh.. so that would give us +- 4 vertical pixels, but still not enough.

It is within their published specifications, I guess. Those appear to tolerate a 300mV variation on the 5V/div scale, which is pretty much what I'm seeing.

Now where did I think I read that?

Ahh.. their spec says Vertical Accuracy +/- 3%. What I visually observe here is a +/- 150mV (a 300mV noise range) when looking at a +5V signal:

(0.150V / 5.000V) = 3% exactly.

Now we just need to find the spec on a cheap Tek brand scope to compare with this.

Okay, found it here.

The TDS2000 series specifies DC Vertical Accuracy +3% on all models.

That's slightly different: plus 3%, not plus/minus 3%. But close enough for the price, I suppose!

Odd. The 2005 PDF datasheet for my TDS1002 has +/- 3% for the TDS1000 and TDS2000 series.

Ah. Nevermind. The TDS1000B is newer and has a USB port.

Your 1KHz test signal is still pretty noisy but if its within the accuracy range specified then I guess its normal.

Here are two screencaps of one of our oscilloscopes here, a Tek TDS784D. I didn't have a lot of chance to set things up precisely. I stole five minutes when it wasn't being used and I was getting pushed out even then.

How many volts/div was the display set for ?

Thanks.

Heh.. I just noticed some more bugs:

• The grid lines are drawn incorrectly. The horizontal dotted lines have the usual four dots (five sub-divisions) per division, but the vertical grid lines have *five* dots (six sub-divisions) per division..
  
  Makes the mental arithmetic slightly more challenging, that!

• The Quick Measurement functions include an Average voltage display. Except it actually measures something more like average peak rather than plain average. Maybe the firmware guys also got confused by the six sub-divisions when testing their code..

Cheers

Correction: That scope above has only 60Mhz bandwidth, not 100Mhz.

Whoops. Forgot about that part. Here are some more images, both with the 1kHz test signal.

Okay, so they're not that much different at similar settings.

Thanks, Bitt!

Also keep in mind that the one I tested on is like a $30k oscilloscope.

That is SCARILY close to copying the Agilent MSO range's UI, button layout, labeling and screen format...

Hugo

Yeah, but beauty is only skin deep in this case.

Wow, I haven't seen one of those, this is a good hi-res pic and yes it is basically a direct knock-off:

http://www.meilhaus.de/presse/0907_dsomso6000a.jpg

Wow.. there are just so many firmware bugs, that I've now created a web page just to help track them. And since there is no other third-party info available on these products, it may as well be a review page.

My Welec W2022A Review is now up.

Nothing much new that hasn't already been posted here, but now it's collected into an easier to indigest form.

For some odd reason, the Wittig brothers are no longer responding to my emails.

Just for fun, today I downloaded the .flac version of YouScope and viewed it on the Welec W2022A scope.

After inverting channel-A, and adjusting various other controls, it produced a recognizable display, albeit at only 3 frames/sec.

Cool.

That Youscope thing is pretty cool but then an (analogue) CRO is really just a specialised TV... I would never expect a digital scope to do that good of a job of something like that. 3fps though ain't gonna help. That's where analogue scopes still have it to some extent although newer DSOs are trying to bridge that gap.

We actually had to build a complete video generator (reading an image from RAM using a primitive drawing program) and outputting XY back when I was at uni. Was a good practical.

Thanks for your review work Mark.

Here are a couple other bugs that you may want to add to your growing list:-

1. Power on zero drift. If you turn on the unit and observe the zero level, you will find that it drifts quite a bit over time - probably due to the lack of thermal compensation.

2. The video triggering doesn't work at all.

3. Occasionally, the trigger will lose sync and you need to stop and start the capture again for it to re-acquire the trigger.

I have actually e-mail Eric Wittig, whom I was told was in-charge of the code development, about these bugs but have not seen any fixes as yet.

Daniel

Thanks. I have more or less just cut/pasted those into my web page review now.


Ah.. so there really are *three* guys, then: Michael (sales), Thomas (tech), and now Eric (firmware).

But as with many others before us, they generally seem to smile and then not actually fix anything.

Cheers