Author Topic: The Siglent SDG2042X Thread  (Read 346582 times)

0 Members and 1 Guest are viewing this topic.

Online tautech

  • Super Contributor
  • ***
  • Posts: 21164
  • Country: nz
  • Taupaki Technologies Ltd. NZ Siglent Distributor
    • Taupaki Technologies Ltd.
Re: The Siglent SDG2042X Thread
« Reply #1375 on: April 01, 2021, 01:39:12 am »
@nez
AFAIK the same fan is used in SDS X-E models and I suspect this as noise levels are very similar to SDG and Siglent would obviously simplify their inventory by using a single fan for as many models as possible.

Back in 2018 in another thread about DSO fan noise the then spec for the fan was:
CD-Fan: Sleeve bearing, 60mm DFM6025S, 0.11A, 3300RPM, 17.41CFM, 2.56mH2O, 24.64dBm

http://www.szdosense.com/eng/products-detail.asp?cpid=177
Avid Rabid Hobbyist
 
The following users thanked this post: nez

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 20927
  • Country: nl
    • NCT Developments
Re: The Siglent SDG2042X Thread
« Reply #1376 on: April 01, 2021, 08:16:13 am »
Using the medium or low Noctua fan speeds using one of the adapters, the airflow and static pressure might be too low compared to the stock fan.

My understanding is that in tight spaces with lots of air resistance, static pressure can be particularly important.
The latter is true. You have to compare static pressures because airflow alone says nothing. Personally I like to have a pressure versus airflow graph of the original and replacement fan to compare the pressure at half the airflow rating. IMHO that gives a better ballpark figure on how the fans compare when they are pushing air through a device (and whether a potential replacement fan is good enough or not).
« Last Edit: April 01, 2021, 08:21:25 am by nctnico »
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline Ghislain

  • Regular Contributor
  • *
  • Posts: 73
  • Country: be
Re: The Siglent SDG2042X Thread
« Reply #1377 on: April 01, 2021, 03:11:17 pm »
Using the medium or low Noctua fan speeds using one of the adapters, the airflow and static pressure might be too low compared to the stock fan.

My understanding is that in tight spaces with lots of air resistance, static pressure can be particularly important.
The latter is true. You have to compare static pressures because airflow alone says nothing. Personally I like to have a pressure versus airflow graph of the original and replacement fan to compare the pressure at half the airflow rating. IMHO that gives a better ballpark figure on how the fans compare when they are pushing air through a device (and whether a potential replacement fan is good enough or not).

Absolutely! No idea what the original Siglent fan's ratings look like but these are the ones for the potential Noctua replacement....
 

Online tautech

  • Super Contributor
  • ***
  • Posts: 21164
  • Country: nz
  • Taupaki Technologies Ltd. NZ Siglent Distributor
    • Taupaki Technologies Ltd.
Re: The Siglent SDG2042X Thread
« Reply #1378 on: April 01, 2021, 06:57:31 pm »
Using the medium or low Noctua fan speeds using one of the adapters, the airflow and static pressure might be too low compared to the stock fan.

My understanding is that in tight spaces with lots of air resistance, static pressure can be particularly important.
The latter is true. You have to compare static pressures because airflow alone says nothing. Personally I like to have a pressure versus airflow graph of the original and replacement fan to compare the pressure at half the airflow rating. IMHO that gives a better ballpark figure on how the fans compare when they are pushing air through a device (and whether a potential replacement fan is good enough or not).

Absolutely! No idea what the original Siglent fan's ratings look like but these are the ones for the potential Noctua replacement....
Link to fan graphs provided 2 posts ago and here again:
http://www.szdosense.com/eng/products-detail.asp?cpid=177
Avid Rabid Hobbyist
 

Offline Blue

  • Regular Contributor
  • *
  • Posts: 109
  • Country: nl
  • Fighting for peace is like fucking for virginity
Re: The Siglent SDG2042X Thread
« Reply #1379 on: April 01, 2021, 08:37:07 pm »
originial installed: PAAD16025SL
and googling around: https://www.eevblog.com/forum/testgear/siglent-1202x-e-noise-issues/25/
[attachimg=1]
« Last Edit: April 01, 2021, 08:43:21 pm by Blue »
 

Online tautech

  • Super Contributor
  • ***
  • Posts: 21164
  • Country: nz
  • Taupaki Technologies Ltd. NZ Siglent Distributor
    • Taupaki Technologies Ltd.
Re: The Siglent SDG2042X Thread
« Reply #1380 on: April 01, 2021, 08:57:51 pm »
and googling around: https://www.eevblog.com/forum/testgear/siglent-1202x-e-noise-issues/25/
FYI SDS1202X-E fans have been changed or run at much lower RPM than early units so these now are somewhat quieter than SDG's or Siglent's other small form factor SDS DSO's.
Siglent's large DSO's run an 80mm fan at lower speeds and some models also use temp sensing fan speed control.
Avid Rabid Hobbyist
 

Offline nez

  • Contributor
  • Posts: 38
  • Country: us
Re: The Siglent SDG2042X Thread
« Reply #1381 on: April 02, 2021, 01:31:15 am »
originial installed: PAAD16025SL

For the stock fan vs Noctua one (Edit: with lowest speed adapter), looks like the airflow is similar but the difference in static pressure is pretty large.



If the layout in the AWG permits, maybe adding a second quiet fan to have one as intake and one as exhaust would be suitable. That's up to your personal amount of interest and budget.

Actually I was in the same boat with the same 60mm Noctua fan for something else -- I bought the Noctua fan but was kind of dissapointed at its noise level at full speed.  In the end, I picked the middle speed-adapter on the single Noctua as a compromise at my own risk.  I can still hear the fan with the middle speed, but I got rid of some kind of resonance/tonality that was driving me crazy with the stock fan on the other equipment!

So now it's a soft white noise I can put up with ok.  However, I still logically know it may be running hotter on some components due to lower static pressure compared to the original fan -- I don't have a thermal camera (not that using one with case off tells me the real in-use temps) or the inclination to otherwise try to check the temps.
« Last Edit: April 02, 2021, 04:01:35 am by nez »
 

Offline hpw

  • Contributor
  • Posts: 36
  • Country: ch
Re: The Siglent SDG2042X Thread
« Reply #1382 on: April 02, 2021, 12:01:28 pm »

I am currently voting for some Accessories given from Rigol & Siglent to get ride of...  :-DD

While protect your hearings as a tinnitus will be after years your next friend  :horse:
 

Offline JohnG

  • Frequent Contributor
  • **
  • Posts: 353
  • Country: us
Re: The Siglent SDG2042X Thread
« Reply #1383 on: April 02, 2021, 01:33:47 pm »
I just turn on my HP6030A or Lecroy SDA 6020 and then I don't hear anything else...

More seriously, there are people whose hearing is particular sensitive to certain types of sounds even at relatively low to moderate volumes. I personally am not especially bothered by fans, unless they are loud (I wear hearing protection when using the above for any length of time, for example). But, there are particular sounds, like intermittent chirps, or very high frequency tones that change as you move about, that can get on my nerves.

But I've spent many years in labs, and most new equipment is so much quieter that older equipment that it hardly bothers me. On the other hand, bright blue LEDs and anything that blinks at more than a few Hz, I find really bad and headache-inducing.

John
"Those who learn the lessons of history are doomed to know when they are repeating the mistakes of the past." Putt's Law of History
 

Offline ebastler

  • Super Contributor
  • ***
  • Posts: 3877
  • Country: de
Re: The Siglent SDG2042X Thread
« Reply #1384 on: April 02, 2021, 02:43:44 pm »
While protect your hearings as a tinnitus will be after years your next friend  :horse:

You already chose the right icon there, beating that dead horse. You keep ranting about tinnitus, but you know that fan noise on the level we are discussing here has nothing to do with that, right? You may, of course, still be annoyed by the fan noise; that's very subjective. But please keep it real here.
 

Online 2N3055

  • Super Contributor
  • ***
  • Posts: 3514
  • Country: hr
Re: The Siglent SDG2042X Thread
« Reply #1385 on: April 02, 2021, 05:42:22 pm »
Hearing damage occurs when SPL is more than 85 dB continuously. That is a freight train at 30 meters..
Everything else is just how much it bothers you on perception level.
And white noise is actually sometimes used to mask tinnitus, also using perception tricks.
 

Offline Roger Need

  • Contributor
  • Posts: 41
  • Country: ca
Re: The Siglent SDG2042X Thread
« Reply #1386 on: April 02, 2021, 10:00:34 pm »
The SDG2042X permits frequency entry to fractions of a Hz. like 1.000 000 05 MHz. and the generator will output correctly at this frequency.  However the display will only read 1.000 000MHz.  Is there any way to see the other digits? It would also be nice to be able to change the fractional Hz. with the spinner knob like all the other digits.

If this is not possible maybe Siglent will consider this in a future firmware upgrade. 

Roger
 
The following users thanked this post: Johnny B Good

Offline Bad_Driver

  • Regular Contributor
  • *
  • Posts: 162
  • Country: de
Re: The Siglent SDG2042X Thread
« Reply #1387 on: April 03, 2021, 11:13:37 am »
It’s the same with the signal generator in the SDS2000X + scope. It accepts more valid digits than shown in the display.
Good to know that it works with the SDG as well! Some fine tuning for the UI will be appreceated!
 
The following users thanked this post: Johnny B Good

Offline Dwaine

  • Frequent Contributor
  • **
  • Posts: 292
  • Country: ca
Re: The Siglent SDG2042X Thread
« Reply #1388 on: April 03, 2021, 02:23:39 pm »
It's been a while since the last firmware update.  I'm still hoping they add a graph screen for the frequency counter.

 

Offline Johnny B Good

  • Frequent Contributor
  • **
  • Posts: 493
  • Country: gb
Re: The Siglent SDG2042X Thread
« Reply #1389 on: April 11, 2021, 02:25:16 am »
The SDG2042X permits frequency entry to fractions of a Hz. like 1.000 000 05 MHz. and the generator will output correctly at this frequency.  However the display will only read 1.000 000MHz.  Is there any way to see the other digits? It would also be nice to be able to change the fractional Hz. with the spinner knob like all the other digits.

If this is not possible maybe Siglent will consider this in a future firmware upgrade. 

Roger


 I too was similarly optimistic of seeing such 4 or 5 year old UI issues being addressed, especially after exchanging a few emails with Siglent's European Sales manager last October. Quite frankly, I've been using mine these past 5 months or so only as a plinth to park my FY6600-60M AWG on when I was working on a temperature regulated LPRO-101 Rubidium frequency reference project where I used it as a proxy for the RFS to dial in uHz offsets (usually no tighter than 10uHz, more typically 30 to 50uHz step adjustments - the 1uHz resolution is, quite frankly overkill... for the time being).

 It grieves me to see an AWG, capable of being locked to an external 10MHz 'Atomic Standard' so crippled by its woefully inadequate UI as to render it less useful than a cheap toy AWG such as an FY6600-60M cured of of its original frequency instability with a relatively cheap OCXO upgrade with an external 10MHz reference input socket by which to lock the OCXO's phase (typically with a PLL but, in my case, using a frequency injection circuit - don't ask!).

 Rather than work my way up to yet another rant, I'll simply paste the bulk of what proved to be my penultimate response in that email exchange over a faulty SDG1032X I'd returned as "unfit for purpose" by way of an extremely jittery internal clock.

=============================================================================================
[Quote from an email reply to Siglent's European Sales manager last October]

  The answer is yes, your understanding of the test setup is absolutely correct. Although the trigger source is the sine wave output from a cheap FY6600-60M, it has been modified with an OCXO which is injection locked to my GPSDO 10MHz lab reference. The RFS on CH3 provides a 'sanity check' on both its stability and its low jitter by virtue of its complete independence of any of the other signal sources (no correlated behaviour to mask jitter of any sort in any of those other traces).

  I've since run similar tests on the SDG2042X to verify that its internal clock reference is only showing the low rate of frequency drift to be expected of any reasonable quality XO whether it's a VCXO or just a common or garden type XO. The one in the 2042 doesn't appear to be a VCTCXO as other users have assumed. In my opinion, it seems to suffer too much warm up drift and variation in frequency with temperature compared to a cheap 50MHz TCXO I'd been able to trim to within 30ppb of accuracy and stability for it to be a temperature compensated VCXO.

  However, the fact that my 50MHz TCXO board was used inside my FY6600 as an airflow deflector immediately above the small fan I'd installed into the base of the case to overcome the non-existent ventilation may have helped by reducing temperature variations to essentially just that of the room air, may have given it an unfair advantage over the one used in the 2042X. Do you know whether the 10MHz VCXO used in the SDG2000X series is actually a temperature compensated one or not?

 I've been rather preoccupied of late digging around for answers as to why my much modified FY6600-60M should seemingly outperform the SDG2042X in regard of the Sinc pulse waveform. I've had that question quite nicely answered by an EEVBlog member, posting under the pseudonym of "rf-loop" and now have a better understanding of why this should be so.

 Even so, Feeltech seem to reached a better compromise between the actual mathematical description of the Sinc pulse and a practical version that can be generated up to a useful maximum frequency than that chosen by Siglent (10 ripples versus the 20 used by Siglent). In effect, you don't hit the Nyquist limit with the Sinc pulse until Feeltech's version reaches a 12.5MHz fundamental whilst Siglent hits that limit at just 7.5MHz when thereafter they start falling apart when aliasing products start to intrude. Incidentally, the Feeltech one isn't really usable at 20MHz as I'd previously stated since, at this extreme, the 4ns clock jitter finally puts in an appearance. In the Siglent's case, you can't see any of the 3.333ns jitter since it has degenerated into a horrible mess long before reaching its own arbitrarily imposed 20MHz limit.

 The main reason why I concentrated on the Sinc pulse performance was simply because it was the only pulse waveform that the FY6600 could generate which was free of this 4ns jitter around the 10MHz frequency I was using to compare against other 10MHz sources. It stood out nicely against other sine waveform traces and it was easy to spot which way it or the test waveform was drifting - when both are the same sine wave shape, it's all too easy to lose track at the very slow drift rates I was monitoring (1mHz or less frequency difference - circa a 17 minute period).

 The SDG2042X pulse waveform being low jitter means I'm no longer reliant on the Sinc pulse as I had been with the Feeltech AWG so this has become just a matter of academic interest to me now. However, that still leaves the issue of what I (and other users) have noted about just how user hostile the front panel UI is in all of the Siglent range of AWG function generators. I'm sure that if Siglent put their mind to it and look to how Feeltech implemented their keypadless front panel UI, they could gain a considerable marketing advantage over their A and B brand rivals, never mind Rigol.

 For a start, picking on the easy target, they could enhance the frequency entry system into a less user hostile interface by shortening "Frequency" and "Amplitude" into their industry standard abbreviations of "Freq" and "Ampl" to make space for another 6 digits of resolution to allow 1uHz offsets to be applied throughout the whole frequency range DC to 120MHz in the case of the SDG2122X models and above, by using the "Universal Knob" to fine tune previously entered values by virtue of making the digit cursor sticky in between setting other parameters or performing other side tasks.

  Also, the under utilised push button on the Universal Knob could provide a "remember current user settings" function with a long (1.5s) press similar to the long press function of the channel enable buttons to toggle the output impedance settings between Hi-Z and 50 ohms which I only came across when I read through the whole firmware revision history. As for the rest of the front panel UI features and functions, quite frankly I just don't know where to begin since my day job was not, and still isn't, a UI developer but I feel certain, tricky a task as this undoubtedly is, that there's considerable scope for improvement here.

 If they do nothing else, just sorting out the frequency settings interface along the lines I've suggested will go a considerable way to improving their AWG product line's usability in my humble opinion.
=============================================================================================
[End Quote]

[EDIT] I just spotted the figure of "6 extra digits" in the third from last paragraph (which is fine for DC to 100MHz) but I must have meant to say 7 extra digits for the 120MHz and higher frequency models. I know abbreviating to Freq and Ampl only eliminates 5 characters but being a graphical rather than a character based display panel, there's sufficient room to modify the format to accommodate the extra 1 or 2 digits required to match or exceed the digit count in the humble FY6xxx's frequency display. ::)

 A major concern which I hadn't wanted to go into detail regarding the "Other Issues" with the UI was with regard to the nonsensical way control of the bi-directional external reference port had been implemented. Basically, when running off its internal reference, by default it outputs this low grade reference out of the external reference socket to pollute a daisy chained 10MHz distribution setup or back feed into an amplifier in a star based distribution setup - both are rather nasty situations.

 Since you can't get it to accept an external reference source before plugging said source into this socket beforehand, you're forced to connect first and then tell it to use the external reference. There's simply no option to disable outputting its own reference in order to comply with the need for a suitable 10MHz to be pre-connected without this pollution hazard - None! :wtf:

 If you think, after having successfully negotiated this annoyance, you're troubles are over, think again. The obvious act of just unplugging the external reference when you no longer need such precision for any of many good reasons to free up the 10MHz feed, results in a disabling of output to the DUT and an error message about the loss of the external reference, requiring you to manually tell it to go back to its internal reference (and resume outputting it onto the external reference socket -Gahh!). The error message warning is fair enough but the nonsense of disabling its output until you actually take some form of remedial action is totally unacceptable.

 As you can see from the above, I had good reason not to go into detail about the "Other Issues" in the UI. ::)

 I've attached an image of one of my screen shots that he'd edited to annotate it with his understanding of my description to which my response is in the very first paragraph of the quoted text of the email above.

[EDIT] I've just noticed that he'd erroneously linked the gpsdo to the LPRO-101. That wasn't the case, only the FY6600 was locked to the GPSDO - the LPRO-101 was running completely independently of any other interference.

 He did say he was passing on the information to the relevant managers in Siglent's organisation for them to consider but, as far as I'm aware, there's been no sign of any firmware updates addressing these non-trivial issues so I've lost my initial optimism in this regard.

 It's a great pity that high profile tech kit reviewers like Dave Jones fail to spot such glaring deficiencies. They have an excellent opportunity to shame Siglent into getting their act together by making comparisons with, of all things, those much despised (even reviled) FY6xxx series of AWGs that Feeltech / FeelElec have sprung upon the hobbyist market these past few years. They could have contributed so much more towards the development of improved UIs on these Siglent AWGs if they had dared to make such comparisons but, as things stand, we have to put up with a quite frankly user hostile UI. :(

 For now, I'm experiencing the rather unexpected situation where the expensive AWG is reserved for the "Less Demanding" work whilst that cheap 'toy' FY6600 is my go to AWG for the more demanding tasks.  ::):wtf:
« Last Edit: April 11, 2021, 03:36:37 am by Johnny B Good »
John
 

Offline ebastler

  • Super Contributor
  • ***
  • Posts: 3877
  • Country: de
Re: The Siglent SDG2042X Thread
« Reply #1390 on: April 11, 2021, 06:34:37 am »
You are expecting a lot from that poor European Sales Manager. ::)

Frankly I could not follow. Are you complaining about a "woefully inadequate UI", about lack of stability of the built-in XO, or something else? While there is value in a detailed description, I would recommend that you try and describe the core issue in two paragraphs at most if you want to be heard by Siglent.
 

Online 2N3055

  • Super Contributor
  • ***
  • Posts: 3514
  • Country: hr
Re: The Siglent SDG2042X Thread
« Reply #1391 on: April 11, 2021, 07:35:50 am »
You are expecting a lot from that poor European Sales Manager. ::)

Frankly I could not follow. Are you complaining about a "woefully inadequate UI", about lack of stability of the built-in XO, or something else? While there is value in a detailed description, I would recommend that you try and describe the core issue in two paragraphs at most if you want to be heard by Siglent.

I must agree I'm a bit lost too.. It resembles a bit "The musings on generators of signals whimsical Siglentian" by William Shakespeare.  :-DD
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 20927
  • Country: nl
    • NCT Developments
Re: The Siglent SDG2042X Thread
« Reply #1392 on: April 11, 2021, 11:07:03 am »
You are expecting a lot from that poor European Sales Manager. ::)

Frankly I could not follow. Are you complaining about a "woefully inadequate UI", about lack of stability of the built-in XO, or something else? While there is value in a detailed description, I would recommend that you try and describe the core issue in two paragraphs at most if you want to be heard by Siglent.
UI is a matter of taste but the general message is that the SDG2000X series is not very useful for use in time & frequency transfer (which requires stable operation with uHz frequency offsets) testing for various reasons. I already reached the same conclusion a long time ago. In order to fix this, Siglent will basically have to go over all the math again, check the implementation and get rid of accumulating rounding errors.
« Last Edit: April 11, 2021, 11:10:29 am by nctnico »
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline ebastler

  • Super Contributor
  • ***
  • Posts: 3877
  • Country: de
Re: The Siglent SDG2042X Thread
« Reply #1393 on: April 11, 2021, 01:21:59 pm »
... the SDG2000X series is not very useful for use in time & frequency transfer (which requires stable operation with uHz frequency offsets) testing for various reasons. I already reached the same conclusion a long time ago. In order to fix this, Siglent will basically have to go over all the math again, check the implementation and get rid of accumulating rounding errors.

That may or may not be what Johnny B Good is talking about...
I read him discussing, in a single email to Siglent,

- unhealty jitter on the Siglent,
- something to do with temperature drift, where he has modified the Feeltech,
- bad choice of Sinc parameters by Siglent,
- not caring about the bad choice of Since parameters since he has low-enough jitter on the Siglent,
- various UI complaints which seem quite unrelated to any of the above.

If I were the Siglent sales manager at the receiving end of this, my conclusion would be: "Probably a knowledgeable guy, but with way too much time on his hands. I don't have the time to try and understand all of this, and I don't trust his judgement whether all or any of this is really important to people who do not have too much time on their hands." Hence my recommendation to shorten the message and focus on the key complaints.
 

Offline Johnny B Good

  • Frequent Contributor
  • **
  • Posts: 493
  • Country: gb
Re: The Siglent SDG2042X Thread
« Reply #1394 on: April 11, 2021, 04:04:01 pm »
@ebastler
@2N3055

 I'm afraid I have to agree with you both on that point (excessive Shakespearian like ramblings)  :palm: I only posted that extract from my penultimate reply to demonstrate that I had made some attempt at providing feedback to Siglent when I was given an opportunity to communicate directly with their head of European sales as a result of their difficulties in understanding a jitter issue with the internal reference oscillator in the SDG1032X I'd purchased and ultimately returned to Telonic for a refund so I could take my business elsewhere to purchase that underused SDG2042X now doing a sterling job as a plinth for my much modded but ageing FY6600.

 I've no real complaint against Telonic since they had expedited my return and request for a full refund in a timely fashion once I'd decided to take that route to resolve the deadlock created by Siglent themselves. Telonic seems to have been as much stuck in the middle as I had been. We were both victims of this weeks' long process created by Siglent's "technical staff"'s lack of experience of what now appears to be (thankfully!) an extremely rare jitter/phase noise fault in the SMD (VC?)XO used in the SDG1000X series.

 The fault in this case being jitter/phase noise with a frequency in the tens of Hz (flicker noise modulation) which can only be demonstrated either by use of an extended timebase delay in hundreds of ms which can be zoomed to the 10ns/div resolution at a point some 50 to 500ms after the trigger event on a 'scope with an excellent timebase stability (which the SDS2104X Plus possesses in abundance) or else, utilise "The bleedin' obvious" strategy of syntonising another oscillator source to the DUT for comparison using any old 'scope you happen to have on hand, capable of displaying such frequency sources (even if they happen to be a little beyond the official upper bandwidth limit of said 'scope) where the issue becomes blindingly obvious.

 I had spotted this problem straight away when I'd tried to compare the 10,000,000Hz sine wave output against my GPSDO and LPRO-101 sources to quantify the accuracy and stability of the SDG1042X's frequency output as one does (doesn't everyone play this game of "Chase Will o' the Whisp" when checking out a newly acquired signal generator?). Anyway, this was when I discovered the shameful shortcomings in the frequency settings UI which I soon realised also afflicted every single model of Siglent's AWG product line.

 You might wonder why I chose to spend yet more of my hard earned on the equally crippled SDG2042X in the full knowledge of this shortcoming. The answer is quite simply that I knew this could readily be fixed with a timely (if radical) firmware update. Unless Siglent had managed to paint themselves into a corner, the underlying hardware could hardly be inferior to that used in the FY6xxx range of AWGs from Feeltech, could it?. All the indications suggest this is just a matter of incompetence in the UI development team rather than one of hardware limitations.

 In short, I was optimistic that such a long overdue firmware upgrade must be be 'just round the corner' and this "Sow's Ear" would finally be given its long overdue "Silk Purse" makeover. Unless Siglent are about to prove me wrong, this doesn't appear to be the case. >:(

 I've been meaning to gather my thoughts on the UI defficiencies and put together a calm and concise list of the issues that need addressing to forward to Siglent's technical director but, like everyone else it would seem who also have reason to complain, I've been putting this off since it's hard to avoid a level of "Rantiness" creeping into such a list (there's just so much that is insanely wrong with the UI and its features, such "Rantiness" seems almost inevitable).

 I may eventually get 'a round tuit' and actually present such a complaints list but don't rely on my efforts alone when you've all got reason to complain and make some contribution to the "Squeeky Wheel Effect" that seems the only way we're gonna get this particular wheel oiled.

« Last Edit: April 14, 2021, 03:54:22 pm by Johnny B Good »
John
 
The following users thanked this post: ebastler, 2N3055

Offline Johnny B Good

  • Frequent Contributor
  • **
  • Posts: 493
  • Country: gb
Re: The Siglent SDG2042X Thread
« Reply #1395 on: April 11, 2021, 07:53:27 pm »
You are expecting a lot from that poor European Sales Manager. ::)

Frankly I could not follow. Are you complaining about a "woefully inadequate UI", about lack of stability of the built-in XO, or something else? While there is value in a detailed description, I would recommend that you try and describe the core issue in two paragraphs at most if you want to be heard by Siglent.
UI is a matter of taste but the general message is that the SDG2000X series is not very useful for use in time & frequency transfer (which requires stable operation with uHz frequency offsets) testing for various reasons. I already reached the same conclusion a long time ago. In order to fix this, Siglent will basically have to go over all the math again, check the implementation and get rid of accumulating rounding errors.

 I was unloading all of the original SDG1032X's woes upon 'that poor European Sales Manager" simply because I'd been given an opportunity to communicate with him and he was, at least, in a position to forward all the evidence (screenshots and, I think, a few short movie clips) and my conclusions to the appropriate department.

 The original issue had been the outrageous flicker like phase noise in the internal reference of the SDG1032X which I'd returned for a refund by that point in time. The SDG2042X was only mentioned to indicate that its own internal reference was entirely free of the issue that had plagued the SDG1032X I'd returned, observing that it merely exhibited the usual behaviour expected of a good quality SMD VCXO (not in the same league as an OCXO, or even a TCXO of course) and that it was well within its given spec.

 I'm not sure about the existence of "rounding errors" but  do recall seeing a posting about an issue with FM which looked like the result of a rounding error. As far as setting a specific carrier frequency was concerned, what rounding errors to be had were, like the Feeltech, vanishingly small.

 When I finally added an external reference socket to injection lock the OCXO I'd installed into the Feeltech unit, one of the tests I did was to compare its 10,000,000,000,000uHz output to the GPSDO's 10MHz reference to see how accurately it could synthesise this "non-magical" frequency. It took a day's run time to detect with reasonable certainty that this amounted to a 1 second error's worth of phase shift in 1.8 million years ::)

 One of the other tests I did, believing it could only be done using the same technique, was to answer the question as to whether making a 1uHz adjustment had any validity. After struggling with the 200ps or so pk-pk jitter to discern a 1ns drift, I hit upon the zero jitter technique of setting both channels to 10MHz and dialling the 1uHz into either of the channels to look for a 100ps of drift now so clear to observe in just a matter of 16 or 17 minutes. There was a slight discrepancy in the timings between a positive and a negative 1uHz offset, hinting at the first signs of rounding error creeping in.

 I was so impressed by these findings it never occurred to me to try and figure out a way repeat these tests at the 'magical frequencies', namely 12.5, 25 or 50MHz by way of a 'sanity check'. I've since come to realise that I could have simply used the 'scope's hold off feature to trigger every fifth pulse of a 50MHz sine wave test signal to compare against the 10MHz reference. I'll give this a try later on to confirm whether this would be a reliable work around to making such a 'sanity check'.

 Anyway, that was all about how good the Feeltech unit had proved to be. It says nothing about the Siglent kit but one would reasonably hope this would be at least as good as, if not better than a cheap toy AWG :-//

 Incidentally, it has now occurred to me that this limitation on the frequency setting UI might just simply be an inheritance from earlier AWG models that lacked any external reference socket where even the 8 digit resolution would have been considered 'overkill' for the needs of even a 5ppm rated XO.

 I may be wrong (ICBA to check right now) but it would neatly explain this issue as simply an oversight in revamping the UI to match the improved capabilities of the hardware. The cack handed implementation of the function of a bi-directional reference input/output socket would suggest a lack of experience on the part of the UI developer(s?) in such matters.

 These models were released to market over 5 years ago. In view of their firmware updating record on their other kit, you'd think Siglent would have gotten off their backsides long since to address these issues. It would appear that Siglent's AWG division are seriously lacking software development talent or even minimal support. The hardware is obviously capable of a lot more than the UI firmware allows (most probably by a very wide margin) and one just has to wonder why Siglent has allowed such an egregious situation to carry on for so long without any redress.
« Last Edit: April 12, 2021, 01:49:25 am by Johnny B Good »
John
 
The following users thanked this post: TurboTom, Jacon


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf