Rigol scope step responses

formatting link

This is pretty weird. I don't think we are doing anything wrong here.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin
Loading thread data ...

why the Zip file?

Reply to
M Philbrook

How about setting the scope to 1M and adding your own 50 Ohm load at the jack ? Maybe their loading is screwey ?

boB K7IQ

Reply to
boB

There are multiple JPGs and a text file encapsulated in it

Reply to
boB

I added a 50 ohm 10 dB attenuator right at the scope input, with the scope set to hi-z. It only changed the gain, not the shape.

It's so weird that all three Rigols, 200 and 500 and 1G MHz, all have that same shape on the rising edge when set to 1M. It's not just input capacitance; the shape is too goofy.

Other scopes don't do this.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

Old analogue scopes do, and there are many trimcaps/pots to tweak the frequency response. Fettling with them can be strangely relaxing.

Of course I have no idea of similarity/differences between the old/modern input stages nor whether tweaks are possible/required.

Reply to
Tom Gardner

Might be the inverse of the pulse top artifacts of their test setup.

Cheers

Phil Hobbs

Reply to
Phil Hobbs

I wonder if they are doing software FIR filtering to tweak the risetime response. It makes sense to design a cheap front end and then pound it flat digitally. In that case, we may be seeing a systematic software or calibration bug.

It's hard to believe they would ship roughly millions of scopes with this same bug. But I think the problrm is real. The 500 MHz scope is a loaner, to see if we can use three of them in some test stands. Maybe not.

The 200 MHz scope also has weird squirmies on the rise time, apparent Nyquist violations.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

Mind if I share this on EEVblog? I'd be curious to see the responses.

Reply to
JW

Please do. I'd like to see if other people can reproduce this.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

Excellent. Thanks, John.

formatting link

Reply to
JW

This is not entirely surprising. The spec sheet that I saw for the 4054 doesn't seem to list the bandwidth in the 1M mode, only for 50ohms. They don't seem to spec aberrations - even for 50ohms! While I'm not sure about "modern" scopes, the old Tek scopes that I am familiar with had substantially lower bandwidths when using 1M input/attenuators.

There's all kinds of ugly things that happen with switched attenuators, input buffers, and circuit boards at these frequencies and impedance levels. I remember John Addis having lots of fun with the 1GHz 7A29, and that was a strictly 50ohm input system.

John (L), surely you've run into these kinds of problems in your products.

-F

Reply to
Frank Miles

We don't make anything that resembles a scope front-end. We do get down to around 50 ps edges, and sometimes digitize signals at 250 MHz, but it's all 50 ohm stuff.

It's still mysterious that all three Rigol scopes have that same plateau on the rising edge at 1M. The 1G scope also rings impressively at 50 ohms.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

What other scopes have you tested this setup on? Have you tried it on a Tek or HPAgilentKeysight?

Reply to
JW

I tried my Tek DPO2024, 200 MHz, which overshoots a little, as most scopes do nowadays.

The pulse generator is a dual-channel 20 GHz TDR head on a Tek sampling scope, which displays its own output pulse. The pulse generating channel can be looped back into the other channel to see the step at the end of the cable, and it's a beautiful flat 30 ps step on both ends... all 50 ohms of course.

We have a new 500 MHz Tek here that I need to try. We were hoping to use the 4-ch Rigol in the new test sets, as it's about 0.4 times the price of the 2-ch Tek.

Really, a little digital filtering code could fix this.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

Only a couple of replies

formatting link

Any thoughts? Have you tried it with the Tek?

Reply to
JW

Here's a direct comparison: same setup, only the scopes changed.

formatting link

formatting link

Both scopes have that little plateau on the rising edge in 1M mode, but the Rigol is much worse, bigger and longer.

The cable is a semi-hardline that TDRs nicely, 51 ohms and 480 ps one-way. I trust the step pulse.

All three Rigol scopes have that step. My Tek DPO2024 doesn't.

formatting link

We may be able to use the Rigol in 50 ohm mode to make the critical measurements in our test sets, but we'd have to be careful about probing. Fet probes should be OK.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

I can't seem to read it, what kind of frequency is this ? Or better yet the time base setting ?

Reply to
jurb6006

The TEK_D jpeg looks OK to me. Are you looking at it on a phone?

The TDR pulses at 100 KHz. The timebase setting on the DPO2024 is 2 ns/cm.

The indicated rise time is 2.34 ns, which corresponds to about 150 MHz. I paid for 200 MHz.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

No, my eyes are that bad. Something about the colors just made it too hard to read. Anything other than pure black and white gives me alot of trouble.

Do they spec that so many dB down, or percent or what ? If not specified I generally assume that the stated limit is the -3 dB point. Isn't that how i t used to be ? But then, if you poked around their website enough you might find they use a different spec, with a "because it's digital" copout maybe . I started a thread on a forum about how they lie about audio amps. They s tarted lying in the early 1980s about speakers. Now we got amps that say th ey are 3,500 WPC when they aren't even 100 WPC.

Can't trust any company. Used to, but now you can't have auto bill pay, at least around here. The friggin water company makes error to the point where you lose your house. Someone can push a few buttons on your remote control and order a forty buck movie. And 900 and 976 numbers screwed alot of peop le, and there were always sex lines and chat lines, you just needed a credi t card. They made it so you can use someone else's phone and stick it to th em. I know a way out of that and I know a way out of a few other things, bu t alot of people don't.

So there is a possibility that Rigol never expected anyone to really fully measure the performance of their product. Sorry to think that because I hav e heard good things about them. It has been said that Tektronix has forgott en how to build scopes and HP finally learned. But Rigol has been said to g ive you much more bang for the buck.

Maybe that bang is not as big as they said it is. Maybe it is time to conta ct Rigol about this. You seem to have covered all the bases and you got two of them, while the Tek works much better, or at least predicably at those frequencies. You know, if you know the limitations of your equipment you ca n adjust, compensate or whatever but I am not sure what to make of this.

Have you tried sine waves ? I know it is hard to get flat response out of a generator up there but with a detector probe you should be able to tell. I f you have flat response to the rating (or -3 dB) this is some sort of digi tal artifact. And the fact that it is the same on two different models supp orts that supposition. What's more it indicates a flaw in the design which is probably in every model they made with that particular topology.

Maybe just get RAs on both of them and go to a surplus joint or eBay and ge t an old Tek CRO. I would at least talk to their customer service departmen t if possible. For the kind of money those things cost you should be able t o talk to someone in English.

But thanks for making this pubic. Looks like I won't get a Rigol. Just find some elcheapo or stick with my old junk.

As far as I can tell you covered all the bases including cable impedance. I remember having trouble with that a while back. My buddy bought some surpl us BNC to BNC cables, and it rang like a bell. But he had different lengths and they rang differently. Upon further examination I see the cables were made with RG-59U, which is 75 ohms, not 50. But you said the cable length d id not affect it. And we know it is not the generator because the Tek reads it differently.

I think you got a damn good case for a full refund, or maybe keep them for what functionality they do have with a partial refund, or something. I doub t you want to just keep them and eat it, so the only ones who can do anythi ng are the manufacturer.

Reply to
jurb6006

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.