Author Topic: Siglent SDS2000X Plus  (Read 715013 times)

maxspb69 and 7 Guests are viewing this topic.

Online Performa01

  • Super Contributor
  • ***
  • Posts: 1594
  • Country: at
Re: Siglent SDS2000X Plus
« Reply #3475 on: June 16, 2022, 05:18:55 am »
What leads me to my statement are the possible causes of incorrect calibration:
  • Timing errors. These are solely determined by the system clock, which usually cannot be adjusted and is not affected by self-cal.
  • Gain errors. We don't expect any substantial gain changes over temperature from the OpAmp in the LF-path of the input buffer, nor the bootstrapped source follower in the HF-path. And a typical integrated PGA has a temperature coefficient of about 0.001 dB/K, so the temperature could change about 100 °C until we see an additional 1% gain error (and still be within the specification).
  • Offset errors. This is about the only thing that actually benefits from self-cal in practice - if only because of the high sensitivity of modern low noise instruments, where a few hundred microvolts offset error can get very obvious. Offset specifications are  1.5%*full scale+1 mV. That is more than one division at 1 mV/div - a vertical gain setting that is very unlikely to be used in any industrial process where certified calibration is required. For the more common ranges 5 mV/div and up, offset simply is a non-issue as well.
My lab in has seen temperature changes in excess of 10 K. Yet I've never experienced offset errors exceeding a couple hundred microvolts - this is why I'm inclined to neglect even the offset error. I do not run self-cal very often - but only because it takes time...

It is certainly good practice to run a self-cal after FW-update. Yet I think Siglent clearly states it in the update instruction whether a self-cal is actually required. Usually we can make do without.
 
The following users thanked this post: Martin72

Offline oz2cpu

  • Frequent Contributor
  • **
  • Posts: 850
  • Country: dk
    • webx.dk private hobby and diy stuff
Re: Siglent SDS2000X Plus
« Reply #3476 on: June 16, 2022, 11:07:16 am »
>That is really a non-issue and is a result of the limitations of the sinx/x interpolation with not very many samples.

thanks a lot, i was hoping the scope could reset / sync, its sample time, to trig event,
this way none of this will happen.
Radioamateur OZ2CPU, Senior EE at Prevas
EMC RF SMPS SI PCB LAYOUT and all that stuff.
 

Offline bdunham7

  • Super Contributor
  • ***
  • Posts: 7729
  • Country: us
Re: Siglent SDS2000X Plus
« Reply #3477 on: June 16, 2022, 02:58:51 pm »
thanks a lot, i was hoping the scope could reset / sync, its sample time, to trig event,
this way none of this will happen.

You wouldn't actually want that.  A lot of strange looking artifacts can appear when a signal's frequency is very close to a multiple of the sample rate and less synchronization, not more, is the fix.  Older scopes often deliberately shifted the sample timing (Equivalent Time Sampling or ETS) and could build up a waveform display from multiple acquisitions.  The fix for what is shown in the video is to simply display the actual samples (dot mode) and not use the sin/x vector display.
A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 
The following users thanked this post: oz2cpu

Offline bdunham7

  • Super Contributor
  • ***
  • Posts: 7729
  • Country: us
Re: Siglent SDS2000X Plus
« Reply #3478 on: June 17, 2022, 03:55:28 am »
My lab in has seen temperature changes in excess of 10 K. Yet I've never experienced offset errors exceeding a couple hundred microvolts - this is why I'm inclined to neglect even the offset error. I do not run self-cal very often - but only because it takes time...

I've also observed that offset is the main issue, but I have also observed small gain/linearity errors at higher v/div settings, typically less than 1 LSB so not much to worry about.  They always improve or at least don't worsen with self-cal.  Thus my objection (and Tek's) to the notion that self-cal is incompatible with a calibrated instrument.  However, given the huge tolerances in the specifications, this is not likely to be an issue either way.

The large tolerances aren't just at sensitive settings.  For an example I set up an SDS1104X-E with a 10V DC input.  Before self-cal at 5V/div it read 9.97V and at 10V/div it read 9.70V.  All seemingly within the commonly assumed (but wrong) notion that scopes are good specified for about +/-3%.  After self-cal it was 10.00V and 9.75V respectively.  Note that 1 LSB @ 10V/div appears to be 400mV--although checking this was confusing and it seemed to be 200mV in some ways.  Anyway, if you look up the combined specifications for reading 10V @ 10V/div, I get a tolerance of +/- 1.502V or about 15% of the reading.  Yeah, it will probably meet that with or without self-cal.
A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 

Online Performa01

  • Super Contributor
  • ***
  • Posts: 1594
  • Country: at
Re: Siglent SDS2000X Plus
« Reply #3479 on: June 17, 2022, 05:00:41 am »
The large tolerances aren't just at sensitive settings.  For an example I set up an SDS1104X-E with a 10V DC input.  Before self-cal at 5V/div it read 9.97V and at 10V/div it read 9.70V.  All seemingly within the commonly assumed (but wrong) notion that scopes are good specified for about +/-3%.  After self-cal it was 10.00V and 9.75V respectively.  Note that 1 LSB @ 10V/div appears to be 400mV--although checking this was confusing and it seemed to be 200mV in some ways.  Anyway, if you look up the combined specifications for reading 10V @ 10V/div, I get a tolerance of +/- 1.502V or about 15% of the reading.  Yeah, it will probably meet that with or without self-cal.

The visible screen height on any SDS1000X-E corresponds to exactly 200 LSB, i.e. 25 LSB/div.
At 10 V/div, 1 LSB is inevitably 400 mV.

Probably because of the automatic measurments with their bunch of decimal places, people tend to forget that this is still only a low resolution 8-bit system. The mean value in the measurement statistics can provide a lot more resolution though, hence is the preferred measurement whenever some accuracy is required. The measurement of the mean voltage in itself is already an averaging across the screen and has correspondingly higher resolution.

Self-cal will minimize the gain error, and even though it's usually just a matter of +/-1 LSB, the effect can get quite obvious, especially when only a small portion of the input range is used, like a measurement of 10V at 10 V/div, where the visible range is 80 V and the ADC full scale is even higher at 102.4 V. But then again, it has been a basic rule since elctrical measurement devices exist, that in order to get decent accuracy we need to get as close as possible to full scale.

Apart from that, I once demonstrated how the actual offset accuracy is much, much better than specified. See the attached screenshot of 206 V measured to 0.053 % with just a humble SDS1104X-E.

EDIT: According to the specifications, the tolerances in this scenario should be 200 V x 1.0 % + 8 x 2 V x 1.5 % + 2 mV + 6 V x 3.0 % = 2 V + 240 mV + 2 mV + 180 mV = 2.422 V, this would be a total error of 1.175 %!

In fact, the offset is the main reference for self-cal and the very loose specifications especially for the multiplicative error are not really plausible.
« Last Edit: June 17, 2022, 05:26:15 am by Performa01 »
 
The following users thanked this post: 2N3055, Martin72

Offline gf

  • Super Contributor
  • ***
  • Posts: 1132
  • Country: de
Re: Siglent SDS2000X Plus
« Reply #3480 on: June 17, 2022, 08:11:44 am »
SDS scope question ??

what kind of funny clock skew or free running issue are we seeing here ??
why is this not stable ?
is it the scope ? or the generator ? that is causing this ?

https://youtu.be/AQZ0Y1W4-ZI

That is really a non-issue and is a result of the limitations of the sinx/x interpolation with not very many samples.  As the actual point of the samples varies in the capture, the sinx/x interpolation comes out a bit different on different traces.  Usually this shows up as a widened, jumpy trace, but if the sample rate is almost an exact multiple of the signal frequency, the dots can 'crawl' instead of moving around quickly so you see the squirming effect.  If you have the same equipment, you can do a series of single captures and switch back and forth between dots and vectors (sinx/x) and you'll see what I mean.

I'd like to add: It's eventually an aliasing artifact. An exact unambiguous reconstruction of the original signal from the samples via sinc interpolation is not possible if the captured signal was not band-limited to < 1/2 sample rate. These artifacts are not supposed to happen if the signal were properly band-limited (and the video also shows that it disappears when the rise time is increased, which reduces the occupied bandwidth).

[ Btw, an exact sinc interpolation is of course not feasible either, since the sinc function has inifinite support. So it needs to be truncated to a finite length in practice, introducing some error, too, ]
 

Offline Johnny B Good

  • Frequent Contributor
  • **
  • Posts: 811
  • Country: gb
Re: Siglent SDS2000X Plus
« Reply #3481 on: June 19, 2022, 01:10:10 am »
SDS scope question ??

what kind of funny clock skew or free running issue are we seeing here ??
why is this not stable ?
is it the scope ? or the generator ? that is causing this ?

https://youtu.be/AQZ0Y1W4-ZI

 You can rest assured that what you're seeing is entirely down to Siglent's use of their "Innovative Easy Pulse" technology to hide the DDS jitter on pulse waveforms in the SDG rather an issue with the DSO.  :)

 BTW, thanks for that YT video. I've only seen Feeltech's version of this feature in action on their latest FY6900. They call it "Magic Pulse" (or something like that) and that video was reminiscent of the weird looking sidestepping effect when triggering from a square wave with a 30mHz frequency offset from the other channel at around 5MHz as far as I recall.

Feeltech posted a YT video demonstrating it (square waves only)

Evava posted some screenshots and a short video file demonstrating a similar effect

 https://www.eevblog.com/forum/testgear/re-feelelec-new-arrival-fy-6900-signal-generator/msg2587782/#msg2587782
John
 

Offline bdunham7

  • Super Contributor
  • ***
  • Posts: 7729
  • Country: us
Re: Siglent SDS2000X Plus
« Reply #3482 on: June 19, 2022, 01:23:46 am »
You can rest assured that what you're seeing is entirely down to Siglent's use of their "Innovative Easy Pulse" technology to hide the DDS jitter on pulse waveforms in the SDG rather an issue with the DSO.  :)

I don't think there should be any DDS jitter at exactly 1kHz...

...and I can reproduce the same effect on my DSO with a non-DDS edge input.

https://youtu.be/g6the-b412o
« Last Edit: June 19, 2022, 03:40:07 am by bdunham7 »
A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 
The following users thanked this post: Johnny B Good

Offline Johnny B Good

  • Frequent Contributor
  • **
  • Posts: 811
  • Country: gb
Re: Siglent SDS2000X Plus
« Reply #3483 on: June 19, 2022, 04:55:06 pm »
@bdunham7

 That's an interesting, if unexpected, obsveration. I've fired up my 2104X+ to compare the FY6600 and the SDG2042X square wave outputs at 1KHz. I had to wait for the DSO to complete its calibration (I couldn't get it to display the FY6600's 1KHz square wave outputs on either channel). Hopefully just a hiccup in the DSO for lack of any use these past 6 months or so. I've been using the SDS1202X-E since it suffices for monitoring 24/7 the frequency drift between the GPSDO and my Rubidium reference project using infinite persist (it saves burning up the extra 32 watts to do the same task with the 2104).

 Problem solved, I'd lost sight of the decimal point in the FY6600's display and landed up outputting a 1mHz square wave :palm: The problem again being 'lack of use'. :-[ Ah well, I got the setups sorted in the end (even figured out why the SDG's square wave was double what I'd expected (too many amplitude/impedance settings options - I'd preempted the 50 ohm setting but missed the + and - amplitude option landing up with a 10vpk-pk at 50 ohm setting instead of the 5v pk-pk I'd been aiming for).

 Anyway, that peculiar 'wiggle' is an artefact of the SDG after all! :o The FY6600 shows no sign of any wiggle, just the slight over (and under) shoot when looking at an edge with rise and fall times of 7ns. I reduced the frequency to see if I could detect any wobble but there was no sign, just a slight (one pixel) amplitude jitteryness that became apparent by the time I'd dropped the frequency to 10Hz.

 I didn't bother exploring lower frequencies with the SDG since I saw the wiggle issue straight away at 1KHz which is enough to verify my hypothesis regarding it being due to Siglent's  "Innovative Easy Pulse" technology. Incidentally, I noted the longer (and varying!) rise and fall times of the SDG's square wave, just over 8ns for each case.

oz2cpu can rest assured that his SDS2000X+ isn't the culprit after all in this case. :)

« Last Edit: June 20, 2022, 10:39:55 pm by Johnny B Good »
John
 
The following users thanked this post: Bad_Driver

Offline bdunham7

  • Super Contributor
  • ***
  • Posts: 7729
  • Country: us
Re: Siglent SDS2000X Plus
« Reply #3484 on: June 19, 2022, 06:48:19 pm »
That's an interesting, if unexpected, obsveration.

It's really not unexpected and it is a form of, or at least closely related to, aliasing as mentioned earlier by gf.

Quote
Anyway, that peculiar 'wiggle' is an artefact of the SDG after all! :o

You may be able to reproduce a similar artifact from an AWG using interpolation to generate square waves or pulses, but that isn't the only cause of an effect like this.  There's a couple of differences and ways to tell them apart.  I don't get any such artifact from my SDG at exactly 1kHz and I would not expect to.  If you do, then perhaps yours is special!  However, I do get a similar but discernably different artifact at 1.0001kHz, which I do expect since you can't get that exact frequency by dividing the clock.  You can see that the rise time varies, but there are no significant varying 'waves' beyond the transition as there are in my first video.

https://youtu.be/Nr_ySvSPa0U

The difference between this and my first video and the one from oz2cpu is that here there are plenty--30 or 40 at least--samples that cover the event in question.  That is enough so that even if you shift them around, all of the sinx/x interpolations will be virtually identical.  The other two only have 3-5 samples for the entire event which is not enough to accurately capture the overbandwidth components of the edge and shifting them around results in significantly different sinx/x interpolations for each capture.

So some of the ways that you can tell the difference between the effect I'm describing and the one you describe are:

1) switch from vector to dot display and use a little persistence if you need to--your effect will remain, mine will not
2) consider the number of samples and the BW/rise time of the various instruments.  In the YT video oz2cpu posted there are 4 samples per division and IMO the varying sine wiggles reflect that, not any actual signal.
3) use a non-DDS source like I did in my first video to see how the scope can perform with a known consistent fast edge.
4) look and see if the rise time alone is shifting or if the whole thing is wiggling like a snake or at higher wfm/s, showing a wide trace at and around the tops and bottoms of the rising edge.  The latter is due to the effect I describe.

Quote
oz2cpu can rest assured that his SDS2000X+ isn't the culprit after all in this case. :)

Well obviously I completely disagree and I've shown the video previously of the same effect where it clearly isn't the source.  My fast edge is much, much faster than the scope, so there's no doubt that the effect is all scope.  In oz2cpu's video, I suppose it is less clear since the AWG and scope rise times are in the same ballpark if the scope has been upgraded and is not the original 200MHz.  But as I observed earlier, I wouldn't expect any Easy Pulse interpolation at 1kHz.  If the scope has been upgraded to 500MHz+, then perhaps some of the effect could be AWG interpolation at the frequencies that don't divide evenly.
« Last Edit: June 19, 2022, 06:53:28 pm by bdunham7 »
A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 
The following users thanked this post: Performa01

Offline mawyatt

  • Super Contributor
  • ***
  • Posts: 3194
  • Country: us
Re: Siglent SDS2000X Plus
« Reply #3485 on: July 01, 2022, 09:33:58 pm »
Not sure where to place this, if there's a better place we'll move it.

The new SDS2000X HD has shown some very capable Bode Plots by Performa01, the 455KHz IF filter was especially interesting and an excellent device for a Bode measurement. So we decided to task the little brother SDS2104X+ with a task with an old popular passive Twin T Notch Filter, these were (maybe still are) very useful in reducing narrow tone interference in many applications.

Grabbed a handful of old cheap 10nF film capacitors and made some quick measurements with a LCR meter to select 4, then grabbed 4 1% 10K resistors and plugged everything into a white proto board. The network is shown in the LTspice schematic, and we included the ESR for the caps gathered with the capacitance readings from the LCR meter. An ideal 10nF cap and 10K resistor should yield a notch at 1/(2*pi*R*C) or ~1.5915KHz.

The SDS was setup with 10 bit mode and using an external AWG via LAN and a pair of 10X scope probes. No attempt was made to "tweak" the data other than make the scale factors on the simulations and measurement the same. Two sweeps are shown, one with log scale from 10Hz to 1MHz, and the other with linear scale from 500Hz to 2.5KHz to better resolve the "notch".

The results speak for themselves!!!

Best,



 
« Last Edit: July 02, 2022, 12:04:46 am by mawyatt »
Curiosity killed the cat, also depleted my wallet!
~Wyatt Labs by Mike~
 
The following users thanked this post: Performa01, tv84, 2N3055, Coder69

Offline mawyatt

  • Super Contributor
  • ***
  • Posts: 3194
  • Country: us
Re: Siglent SDS2000X Plus
« Reply #3486 on: July 02, 2022, 12:06:57 am »
Edit, couldn't add these to above, so added another post.

Added simulations and plots with a 500Hz BW around 1.6KHz with a 1Hz resolution. Note the simulation predicts a -55.86dBV with -6.068 Degrees @ 1.5965KHz for the notch, and the measured values are -55.20dBV with -6.00 Degrees @ 1.6KHz. The left edge simulation at 1.35KHz is -21.525dBV with -84.3198 Degrees and the SDS measured value of -21.40dBV with -84.00 Degrees!!

Best,
Curiosity killed the cat, also depleted my wallet!
~Wyatt Labs by Mike~
 
The following users thanked this post: rf-loop, Performa01, tv84, 2N3055, Coder69

Offline thinkfat

  • Supporter
  • ****
  • Posts: 2150
  • Country: de
  • This is just a hobby I spend too much time on.
    • Matthias' Hackerstübchen
Re: Siglent SDS2000X Plus
« Reply #3487 on: July 02, 2022, 08:06:01 am »
[ Btw, an exact sinc interpolation is of course not feasible either, since the sinc function has inifinite support. So it needs to be truncated to a finite length in practice, introducing some error, too, ]

Which is fine, as long as the error is hidden in the overall noise of the instrument.
Everybody likes gadgets. Until they try to make them.
 

Online tv84

  • Super Contributor
  • ***
  • Posts: 3212
  • Country: pt
Re: Siglent SDS2000X Plus
« Reply #3488 on: July 02, 2022, 10:07:42 am »
Edit, couldn't add these to above, so added another post.

Added simulations and plots with a 500Hz BW around 1.6KHz with a 1Hz resolution. Note the simulation predicts a -55.86dBV with -6.068 Degrees @ 1.5965KHz for the notch, and the measured values are -55.20dBV with -6.00 Degrees @ 1.6KHz. The left edge simulation at 1.35KHz is -21.525dBV with -84.3198 Degrees and the SDS measured value of -21.40dBV with -84.00 Degrees!!

Best,

Uhmmm...  ::)  I think the HDPlus is a "simulator".

Edit: I meant Plus.  |O
« Last Edit: July 02, 2022, 06:10:12 pm by tv84 »
 

Offline mawyatt

  • Super Contributor
  • ***
  • Posts: 3194
  • Country: us
Re: Siglent SDS2000X Plus
« Reply #3489 on: July 02, 2022, 12:54:08 pm »
Uhmmm...  ::)  I think the HD is a "simulator".

The green plots are from LTspice. In my case yes the HD is, wish I had the real hardware tho  ::)

So will have to make due with the X+ for now, which has been "augmented" with your kind (and others) help  :clap:

Best,
« Last Edit: July 02, 2022, 01:33:10 pm by mawyatt »
Curiosity killed the cat, also depleted my wallet!
~Wyatt Labs by Mike~
 

Offline bayjelly

  • Regular Contributor
  • *
  • Posts: 64
  • Country: us
Re: Siglent SDS2000X Plus
« Reply #3490 on: July 02, 2022, 05:44:38 pm »
Simulator?
 

Offline mawyatt

  • Super Contributor
  • ***
  • Posts: 3194
  • Country: us
Re: Siglent SDS2000X Plus
« Reply #3491 on: July 02, 2022, 05:54:07 pm »
Simulator?

Should clarify, all the Full Green Plots are simulations with LTspice, the Magenta/Green Plots are the actual measurement readings from the SDS2104X Plus in Bode Plot mode using the 10 Bit Mode. The capacitors were measured with a LCR meter with ESR as shown and included in the schematic above.

Edit:

Just for fun cascaded two Twin "T" Networks together and here's the simulation and actual measured results. The SDS2104X+ produces a respectable performance wrt the notch depth at -88dBV!!

Just a note, cascading networks together like this doesn't create the product of the transfer functions since the overall network response includes the source impedance and load impedance, and the first Twin T section has a load of the input to the second Twin T section which "sees" the first Twin T section as it's source. This is why a buffer amp between filter network sections is often employed.

Best,
« Last Edit: July 03, 2022, 01:53:47 am by mawyatt »
Curiosity killed the cat, also depleted my wallet!
~Wyatt Labs by Mike~
 
The following users thanked this post: 2N3055

Online Martin72

  • Super Contributor
  • ***
  • Posts: 5680
  • Country: de
  • Testfield Technician
Re: Siglent SDS2000X Plus
« Reply #3492 on: July 02, 2022, 08:05:39 pm »
The green plots are from LTspice. In my case yes the HD is, wish I had the real hardware tho  ::)

Send me your filter... ;)

Offline mawyatt

  • Super Contributor
  • ***
  • Posts: 3194
  • Country: us
Re: Siglent SDS2000X Plus
« Reply #3493 on: July 02, 2022, 11:05:24 pm »
The green plots are from LTspice. In my case yes the HD is, wish I had the real hardware tho  ::)

Send me your filter... ;)

I'll swap the filter for your HD  :)

You have a good LCR meter I recall, grab 4 10K 1% Resistors and select some 10nF caps and build up the filter shown in our earlier post today (#3486). Try and match the cap values, and the two parallel shunt caps to 2 times the cap values. If you have an external AWG then set it to 20Vpp if possible for the best Bode DR. The Dual version of the Twin "T" was just two sections of this network connected together.

Bet your HD will produce a superb Bode Plot :-+

Here's a quick dual Twin "T" Network with nominal 10K Resistors, ~10nF and ~470nF Capacitors which should produce notches at 33.86Hz and 1.5915KHz.

Best
« Last Edit: July 02, 2022, 11:55:50 pm by mawyatt »
Curiosity killed the cat, also depleted my wallet!
~Wyatt Labs by Mike~
 
The following users thanked this post: Martin72

Online Martin72

  • Super Contributor
  • ***
  • Posts: 5680
  • Country: de
  • Testfield Technician
Re: Siglent SDS2000X Plus
« Reply #3494 on: July 03, 2022, 07:03:07 pm »
Quote
and select some 10nF caps

Thank you for this - It reminds me that i didn´t got any caps under 100n at home.
But not for long, PP-Cap set will arrive on tuesday.. 8)
Generator is no problem, I got a SDG1062X, LCR is ET4410.

Martin

Offline mawyatt

  • Super Contributor
  • ***
  • Posts: 3194
  • Country: us
Re: Siglent SDS2000X Plus
« Reply #3495 on: July 03, 2022, 10:51:56 pm »
You can use another size cap (or R), for instance if you have 10K R and 100nF then expect the notch to be at 159Hz, or if 220nF then 72Hz, or 470nF then 33Hz, since the notch should be ~ 1/(2*pi*R*C). Measure the C and ESR close to the desired notch frequency for better fidelity.

Best,
Curiosity killed the cat, also depleted my wallet!
~Wyatt Labs by Mike~
 

Online Martin72

  • Super Contributor
  • ***
  • Posts: 5680
  • Country: de
  • Testfield Technician
Re: Siglent SDS2000X Plus
« Reply #3496 on: July 03, 2022, 11:15:56 pm »
Hi,
Yep but it was the best time to buy other caps... ;)



Online Martin72

  • Super Contributor
  • ***
  • Posts: 5680
  • Country: de
  • Testfield Technician
Re: Siglent SDS2000X Plus
« Reply #3497 on: July 05, 2022, 07:23:29 pm »
Grabbed a handful of old cheap 10nF film capacitors and made some quick measurements with a LCR meter to select 4, then grabbed 4 1% 10K resistors and plugged everything into a white proto board. The network is shown in the LTspice schematic, and we included the ESR for the caps gathered with the capacitance readings from the LCR meter. An ideal 10nF cap and 10K resistor should yield a notch at 1/(2*pi*R*C) or ~1.5915KHz.

Did you really measure ESR between appx 46ohms like mentioned in the LT pic ?
Seems high for a film cap.
Today my cap-set arrived, I measure somewhat between 7...8Ohms on 1khz(no stable value).
Maybe your caps are very old.



Offline mawyatt

  • Super Contributor
  • ***
  • Posts: 3194
  • Country: us
Re: Siglent SDS2000X Plus
« Reply #3498 on: July 05, 2022, 07:47:13 pm »

Did you really measure ESR between appx 46ohms like mentioned in the LT pic ?
Seems high for a film cap.
Today my cap-set arrived, I measure somewhat between 7...8Ohms on 1khz(no stable value).
Maybe your caps are very old.

Yes that's what was measured, these were some old cheap film types in a bag (green types) that had been laying around for many years (likely came from eBay). Just grabbed one and checked again, it reads 9.9248nF with 47ohms at 1.6KHz. Checked with the TH2830 and the IM3536 and both LCR meters agree, if the frequency is 10KHz then the ESR drops to ~12 ohms, even on the DE-5000 at 10KHz. At 1KHz all 3 LCR meters read ~57 ohms ESR.

Found some new stock 10nF metalized film type (yellow WIMA type), all three LCR read @1KHz ~67 ohms, @10KHz ~15 Ohms and the TH2830 & IM3536 @1.6KHz ~52 Ohms.

Hope this helps.

Best,
« Last Edit: July 05, 2022, 08:26:17 pm by mawyatt »
Curiosity killed the cat, also depleted my wallet!
~Wyatt Labs by Mike~
 
The following users thanked this post: Martin72

Online Martin72

  • Super Contributor
  • ***
  • Posts: 5680
  • Country: de
  • Testfield Technician
Re: Siglent SDS2000X Plus
« Reply #3499 on: July 05, 2022, 09:26:52 pm »
Thank you so far.
I don´t want to "stretch" this thread with these things too much, my results will be posted in the 2000HD thread.
And for this thing I´ve "explore" while measuring the caps concerning the ESR, I´ll open up a new thread.

Martin
 
The following users thanked this post: mawyatt


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf