strange sine wave

formatting link

That pic was sent by a customer who is testing our capacitive level sensor simulator. He has a 10 KHz sinewave generator, tee connected to the input of our box, and then to the scope ch1, yellow trace.

The waveform is jitterey and has distinct stairsteps. He's sort of blaming our box, but we have a 100K input impedance and the waveform doesn't change if our gadget is powered off. It's unlikely that his old Wavetek is stairstepping; weren't they all analog?

Note that the trigger is 723 us off screen.

Could the explanation be:

This scope doesn't have enough memory to support a pretrigger delay this big, so it downsamples. That quantizes the waveform in time hence in amplitude, making the stairsteps and somehow also causing the period jitter.

Has anyone seen a scope do this? I think my Rigol just refuses to pretrigger farther than memory allows; gotta verify that.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin
Loading thread data ...

Nope, but I can't see the stair steps on your pic. Is it just the limited horizontal resolution.. A sine wave will look steppy on my DSO 'cause of that. Ask 'em to change the time base.

700 uS with 50 uS/ div doesn't seem like that much... my TEK TDS1001 B will do 50 mS of delay at 50 uS/div

No idea about the jitter... maybe it's the wavetek?

George H. (customers often have goofy ideas about what might be wrong.) I think my Rigol just refuses to

Reply to
George Herold

My customer just called. I got him to set the timebase delay to zero and the jitter went away. I think the ancient Wavetek has a lot of FM (dried-out filter caps?) and that got magnified by the trigger time delay. I think the stairstepping is just the quantization of that scope, although it looks kinda big still.

Some weird interaction of the Wavetek FM and the scope delay might cause stairstepping, but that's at handwaving level just now.

This guy is a Fellow of one of the bigger aerospace outfits, so I have to let him down easy.

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

Take a look at it using line sync. If it is power supply related, it will be apparent.

Reply to
tom

Strange, on the right side of the scope the sine wave has more time variation than on the left. One would then conclude on an analog scope that the frequency changes, is FM modulated. Hard to tell on them digital ones.. Why is so much pre-trigger needed?

I see so many things wrong with that Tek screen's color layout I would not want that scope around for free. Bit like Apple .. Best times over. You may well be right, and switching your box off proves it is not at fault. Disconnecting it would be the ultimate proof.

Reply to
<698839253X6D445TD

Maybe the FM modulation mode switch is turned on. That would happen on an old sig gen we use to resell.

GH

Reply to
George Herold

Right. That ancient Wavetek has a lot of residual FM. (He asured me that sweep was off.)

That was my observation, which I noted in a few emails. I finally talked him through the scope front panel and we found the horizontal delay "set to zero" function. That fixed his problem. On my Rigol, I just press the hor position knob to plop the trigger at mid-screen.

Some things are never fully explained because some of the observations are incorrect.

I had a Tek DPO2024 as my bench scope, but it was slow and buggy. I have a Rigol now.

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

The trigger is a separate menu, how do you figure 720u of pretrigger??? Tha t 720u is the marker measurement. You're dealing with a digitally synthesiz ed Wavetek. That's why the bottom waveform sweep is jittering, the scope is not synced so the trigger threshold crossing is jumping around with those amplitude steps.

Reply to
bloggs.fredbloggs.fred

I was going to suggest making fun of his scope, but I guess discretion is the better part of profitability.

-- john, KE5FX

Reply to
John Miles, KE5FX

I got him to set the timebase delay to zero

Yes.

Reply to
makolber

He's a really nice guy, software genius, but sort of helpless with electronics. He's also a weight lifter and hugs me hard whenever he meets me. If I stop posting here, that's probably why.

(IMHO girls are for hugging.)

--
John Larkin         Highland Technology, Inc 

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

I'm pretty sure the Wavetek is an ancient analog box. Here's a video that the customer sent:

formatting link

I'm guessing that the Wavetek has some 120 Hz FM that's aliasing against the scope acquisitions or something obscure like that. Maybe the stairsteps are just crummy scope digitizing, but I don't see them on the blue trace.

--
John Larkin         Highland Technology, Inc 

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

That 720u is the marker measurement. You're dealing with a digitally synthe sized Wavetek. That's why the bottom waveform sweep is jittering, the scope is not synced so the trigger threshold crossing is jumping around with tho se amplitude steps.

I'm pretty sure Wavetek introduced a digital line something like 35 years a go, that's pretty ancient too. The yellow trace certainly looks like DAC st eps.

Reply to
bloggs.fredbloggs.fred

What's the scope display resolution? The blue trace looks fuzzy enough to mask that.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

te:

to

ce

?? That 720u is the marker measurement. You're dealing with a digitally syn thesized Wavetek. That's why the bottom waveform sweep is jittering, the sc ope is not synced so the trigger threshold crossing is jumping around with those amplitude steps.

rs ago, that's pretty ancient too. The yellow trace certainly looks like DA C steps.

That crummy TEK TPS2014B is good for 2500 samples per sweep (at 8-bit sampl ing resolution). Both channels are 5V/DIV. So 8DIV x 5V/256=156mV=0.03D IV sampling resolution, steps look like 0.25DIV. My rough count shows less than 250 of those steps, so it's not the scope, unless he overrode somethin g, or TEK uses sequential trigger overlay or something like that. Bottom tr ace is not FM it is time translation, and the translation looks to be about the same time displacement as those amplitude steps. That points to those unsynchronized steps dithering the scope sweep trigger.

Reply to
bloggs.fredbloggs.fred

e:

rote:

l

d to

rm

s
y

ence

r??? That 720u is the marker measurement. You're dealing with a digitally s ynthesized Wavetek. That's why the bottom waveform sweep is jittering, the scope is not synced so the trigger threshold crossing is jumping around wit h those amplitude steps.

m

ears ago, that's pretty ancient too. The yellow trace certainly looks like DAC steps.

pling resolution). Both channels are 5V/DIV. So 8DIV x 5V/256=156mV=0.0

3DIV sampling resolution, steps look like 0.25DIV. My rough count shows les s than 250 of those steps, so it's not the scope, unless he overrode someth ing, or TEK uses sequential trigger overlay or something like that. Bottom trace is not FM it is time translation, and the translation looks to be abo ut the same time displacement as those amplitude steps. That points to thos e unsynchronized steps dithering the scope sweep trigger.

It's the horizontal resolution that gives steps. There's ~500 bits across the screen. (Well that's what it looks like on my older DSO.) For a good FFT on it, you can only use so many time steps per cycle. Nyquist.

George H.

Reply to
George Herold

I seriously doubt that a scope display has 2500 horizontal pixels. VGA is more like it--general purpose scopes have 8 bits vertical resolution at best, so those would be some seriously oblong pixels.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

The issue of 'jitter' with DDS ARW sig generator technology in general came up in the massive 69 page EEVblog thread on the "cheap 'n' (not quite so) nasty" Feeltech FY6600.

In this case it was a matter of the +/- 4ns jitter you got with the DAC

that's a deficiency I can live with. :-)

However, the one main deficiency I'd prefer to sort out is to replace the THS3002 dual opamp with a pair of single THS3491 opamps to get a vastly improved sine output beyond 20MHz - something Feeltech should have done when they raised the sine wave upper frequency limit from 15 to

25MHz, let alone when they upped the limit to 40 and finally 60MHz.

It's not as if Feeltech hadn't considered the need for such a hardware upgrade path in later production runs since the main board has the necessary traces to allow the dual opamp to be replaced by two singles. I guess their purchasing manager must have placed an order for the 3001s way in excess of requirements and they're still trying to disperse this overstock into the later 25MHz and upwards model production runs.

Incidentally, my first thought when I saw that dropbox image was, "Bloody Hell! The build quality looks no better than my recently purchased Siglent SDS 1202X-E !!!" In fact, but for the Tektronix Logo, it could have been any one of a dozen cheap Chinese DSOs for all its appearance was going to indicate. I can't help feeling that my cheap Siglent has a slightly more polished look than that rather generic looking (yet so outrageously overpriced) example of the DSO genre.

Indeed, being curious about the perceived "goodness factor" of Tektronix scopes, I googled for pricing and was staggered to see it was still being sold by Farnell/Element14 and others at a whopping pre-VAT price of

better price comparison would be to pit it against Siglent's SDS1204X-E

decently specced modern Chinese equivalent 'scope.

I took a look at its data sheet and the quality of the screenshots reinforced my surmising that when they describe the display system as:

"Display type 1/4 VGA Active TFT Color(sic) LCD display"

they must have actually meant it was just a mere 320 by 240 resolution screen (no figures were given) - no wonder the displayed waveforms and the screen fonts looked so s**te!

Admittedly, the Siglent is only two channel but apart from that, the basic specs are comparable but the Siglent offers enhanced display options that go beyond a display resolution of 800 by 480 and more triggering features including common serial protocols with protocol decoding as standard (rather than as a time limited free trial of a priced optional extra).

Would I be right in thinking this is a 'vintage' model of 'scope? The presence of a CF slot, its condition and the piss poor display resolution and lack of 'standard features' rather suggests this is the case but I'm still getting over the insane pricing (and its availability as a brand new item) compared to a "cheap Chinese DSO" at less than a 7th of the price.

Quite frankly, I can't see Farnell/Element14 getting rid of the last one they have in stock any time soon. However, there must be a significant number of Tektronix fanbois driving up the price of second hand refurbished offerings in the mistaken belief that in the case of the later DSOs such as this model, where you could have sandpapered the brand name off and replaced it with Hantek or whatever making its looks more in tune with expectations, they're still worth paying way over the odds for.

There was a time when Tektronix could justify their astronomical pricing back in the days before high speed DSP chips and cheap LCD colour display panels squeezed all the joy out of designing, building and lining up a vacuum tubed (thermionic valved) CRO but that time is long since passed and it seems that the brand name has become yet another victim of tank badge engineering that could be applied to any anonymous DSO package that you'd expect to see sporting brand names such as Rigol, Hantek and Siglent.

Tektronix no longer has the gravitas it once commanded back in the day (Literally! That old klunker only weighs a mere 6 Lbs, just half a pound more than my 2 channel Siglent).

Harking back to the topic, I've managed to get my Siglent 'confused' to the point of displaying patently fictional waveforms on several occasions over the past week or so. The last time it happened I'd been chasing the "25MHz" output from an XCO module, which I'd rescued off a scrapped PC motherboard, with the FY6600 sig genny.

The waveform from the XCO had displayed ok when the scope was being triggered from it. Intrigued, I thought I'd try and match the frequency of the ARW generator to it so I could sit back and admire the waveforms lazily drifting past each other.

Fat chance! I guess the problem was my croc clipping a 4 AA cell battery pack with part used cells totalling 4.77v[1] with no decoupling of the supply pins, upsetting the stability of the XCO. That and, I'm assuming, a slowly dropping voltage was aggravating the frequency drift issue making my quest a hopeless one.

Eventually, after half an hour or perhaps longer (the time just flies when you're having such 'fun', don't ya know!), I found I couldn't get a stable trace established with either signal source. I power cycled the scope the first time this happened which restored a stable display but when the scope got its knickers in a twist yet again, I dared to press the ominously named "Auto Setup" button which after a second or two, restored the display back to normal.

The point I'm making is that that old clunker of a DSO might just have been suffering digital indigestion, requiring simply a press on the button labelled "Auto Set" (presumably the same function as "Auto Setup") to let it sort itself out.

I think I was rather loath to resort to the 'cheat button' when I (or rather the scope) got 'lost in hyperspace' simply because whenever I used to do the same with my "Boat Anchor", I knew I couldn't simply reboot my way out of the situation, I'd still have to fix my cocked up settings on my own regardless. "Auto Setup" is like a fast reboot version of power cycling the scope out of hyperspace.

It seems a benign enough button to press in those circumstances but I think I really should have another RTFM session just to make sure of this and identify *the* button (or buttons) one should avoid pressing in the blind faith that 'nothing bad will happen'[2] to a set of carefully crafted settings.

[1] The XCO had run just fine off a single 3.7v LiIon cell on my initial test - the thought (and sight) of a short to the can on the positive croc clip had put me off using the LiIon cell to favour the softer 4 cell AA battery option. I'd only wanted to verify that my rescue mission had not been all in vain.

The messing around with the signal generator to sync it to the XCO manually was just a bit of 'amusement' that got out of hand. Mind you, even mere amusements can prove educational. Still, that's an hour or so that I won't be getting back which could have been put to more pressing use. C'est la vie.

[2] Well nothing 'bad' in the sense of damaging the 'scope so much as my self esteem in finding myself being a little inconvenienced for 'pressing the wrong button'. :-)
--
Johnny B Good
Reply to
Johnny B Good

Call your customers back and tell him to retake the picture without shaking the camera. :)

BTW: The paperclips in the terminal blocks is a nice touch.

Reply to
mpm

e:

rote:

l

d to

rm

s
y

ence

r??? That 720u is the marker measurement. You're dealing with a digitally s ynthesized Wavetek. That's why the bottom waveform sweep is jittering, the scope is not synced so the trigger threshold crossing is jumping around wit h those amplitude steps.

m

ears ago, that's pretty ancient too. The yellow trace certainly looks like DAC steps.

h

ampling resolution). Both channels are 5V/DIV. So 8DIV x 5V/256=156mV=0 .03DIV sampling resolution, steps look like 0.25DIV. My rough count shows l ess than 250 of those steps, so it's not the scope, unless he overrode some thing, or TEK uses sequential trigger overlay or something like that. Botto m trace is not FM it is time translation, and the translation looks to be a bout the same time displacement as those amplitude steps. That points to th ose unsynchronized steps dithering the scope sweep trigger.

Those step things are the Wavetek DAC steps. The digitization clock period is not an integer submultiple of the synthesized waveform period, which mea ns the samples "roll" through the waveform from period to period, they're n ot sampling the same phase in consecutive periods, however you want to put it. This is causing the scope sweep trigger jitter. The scope should have the capability to zoom into the acquisition file for examination in microscopic detail, with a net horizontal resolution somewha t in excess of 2500 points per sweep. The scope display is not an aliasing issue as long as the display waveform frequency is in agreement with the Wavetek setting.

Reply to
bloggs.fredbloggs.fred

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.