Siglent General Wish List

[Electro Fan]

We know that a 10 Mhz Ref input is not that common - that’s why some users would like to see it become more common

  Could you (and/or other posters) provide some scenarios where you would find a ref in to be be helpful / necessary when using an "entry level" scope?  How much extra would you be willing to pay - in absolute dollars or as a percent of base scope price - for a ref in?  And would this need to be a dedicated connector or would it be acceptable to share the ref in connector with other functions (like trigger in)?

I can ask our scopes product manager about this, but I'm curious to get feedback from the forum as well.   Thanks!

Thanks for the questions.  And Thank You Very Much for the beautiful educational material you produce.  I appreciate both the outstanding information you provide and the way in which you ask for and encourage user feedback.

As a preface, my scenarios regarding the 10 MHz Ref input feature are much more likely to be in vicinity of the “helpful” category than in the “necessary” category.  Reason being is that my scenarios are based on want rather than need.  Reason being is that I’m an electrical enthusiast (“Electro Fan”) vs an Electrical Engineering pro.  Fwiw, though, I’m VERY enthusiastic about electro things (both analog and digital).  I think my electro enthusiasm virus was fed and fueled by reading tens of thousands of EEVblog posts and writing over a thousand.  Having said all that, I get that any one user is just a datapoint, which might or might not be indicative of an addressable marketplace. 

Regarding the “market”:  if you look at the EEVblog member stats and sort by posts, I might be an “early adopter” at least in terms of enthusiasm if posts are any indication.  My post numbers get me to the first page of 1,000 members, but there are 64 more pages of forum members (ie, ~64,000 members).  If you go to page 32 you find that the median post per member is about 3.  And my guess is that there are many more people out there beyond the current 64k who are also interested in figuring out what’s what, and what causes what, and how to learn and how to get involved at the intersections of electricity and the future. 

So I realize it’s sometimes hard to know what the needs and wants are for a marketplace, even if a manufacturer tries to survey the market.  One of the reasons that uncovering new market needs and wants can be difficult is that some potential users don’t know what’s available or possible, or even if they do, why something might or might not have value, or how much value it might offer.  If we would have asked people in 1900 what their scenarios would be for air travel they might have said “limited” – both the enthusiasts and the pros too.

What all that means is 1) I’m just an enthusiast and not a pro EE, 2) we (I) don’t know for sure what the distribution is here between pros and enthusiasts, and 3) we (I) don’t know if my interests are representative of any other enthusiasts, much less the pros.  I get it.

Nonetheless, I would say that I am HIGHLY enthusiastic about the 10 MHz Ref input feature.  While I don’t “need” almost any of the test equipment that seems to have taken over my “lab” one of the “system” features/functions that that I find most interesting and compelling is the 10 MHz Ref capability. 

As a frame of reference, among the approximately 4 oscilloscopes, 3 signal generators, 3 power supplies, 3 counters, 1 spectrum analyzer, 2 GPSDOs, and dozen plus DMMs that have found their way onto my bench, the 10 MHz Ref capability that connects a GPSDO to a distribution amplifier to 2 counters, the Spec An, and 1 oscilloscope is among the functions that gives me the most “enjoyment.”  (I have no clue about how to measure enjoyment - but I’m certain I would get even more enjoyment out of the overall investment of time, money, and space if all or most of the other pieces of test equipment on the bench also had 10 MHz Ref inputs.)

One way I look at all this is that there is an iterative process in learning and science (hypothesize, test/measure, repeat).  Of course, what all the gear does is enable me to measure stuff including frequency, amplitude, and phase.  The oscilloscope being the “king” of TE is particularly cool because (along with the spectrum analyzer) it gives a wonderful visual insight to signal properties in addition to quantifying them with measured numeric values.   What does the oscilloscope measure and enable me to visualize?  In my mind, near the top of the list is frequency – which is pretty much related to phase and amplitude.  Frequency is arguably (at least metaphorically) the heartbeat of “waves”.  So why wouldn’t I be interested in measuring frequency with as much accuracy as possible?  And while accuracy is different than resolution, more resolution might be nice too.  Of course, there are skeptics who will say “do you really need resolution?”, and “you do know, don’t you, that more resolution does not mean more accuracy, right?”  and “btw, how much accuracy do you really need?”  And “if you need to measure voltage (amplitude) or maybe even frequency with high accuracy and high resolution the oscilloscope is the wrong tool.”  To which I would say “really?” and to which I would further say “sure is interesting how much user engagement the TE forum is getting these days with the introduction and adoption of “HD” oscilloscopes.”  Why do we care about “HD” and moving from 8 bits to 12 bits if we don’t care about accuracy and resolution?

So maybe the question (“why would anyone want a Ref input?”) shouldn't be directed at or limited by this or that specific use case.  Rather, the question might be better asked “why would anyone not want a 10 MHz Ref input?”  Of course this leads to questions about NRE, COGS, and price.

Apparently, we could add a high performance reference in each piece of equipment.  According to one of the links above we could do it for about $1199.  (Such a deal.)  Maybe if it wasn’t designed as an add-in and it was incorporated from the get-go we could get it down to some lower price per unit.  But at the end of the day a user would need to multiply the incremental price increase times the number of pieces of test equipment adding a reference quality oscillator. 

Or, we could put one high quality reference oscillator on the bench in the form of a GPSDO and then drive maybe 6-8 pieces of equipment.  With a ~$200 double oven GPSDO and a ~$100 distribution amplifier and some pretty inexpensive coax cables we could provide relatively higher accuracy to 6-8 pieces of equipment.  And btw, all those pieces of equipment would know what time it is in the rest of the world, AND they would all measure pretty much in unison.  True, they might all be off by some amount, but to a very great extent they would be off by the same amount.  Would you rather have one yardstick that is off by a little, or 8 yardsticks that are all off by differing amounts?  Which would be more useful?  Of course you would prefer to have one clock that is accurate vs several that are less accurate and that vary from one to another.

So the only real question left is how much NRE and COGS is it going to take for Siglent or R&S or X, Y, and Z to add a 10 MHz Ref input?  Once we know that we can get some idea of how much it will increase the price of a unit of Test Equipment.  As pointed out earlier in this thread we have under $400 signal generators that manage to include a 10 MHz Ref input (and a 10 MHz Ref output capability).  So I think by the time it’s all said and done, the price impact on a $400-$500 oscilloscope is going to be in the realm of “not much.”  At which point, for about $300 an enthusiastic user could distribute a GPSDO signal to the oscilloscope and a half a dozen other pieces of equipment. 

Maybe some other users here will come up with some specific professional or enthusiast use cases for a 10 MHz ref input, but I think it might be like asking people in about 1920 what would they do with a plane or a car.  Back then (ie, still somewhat early in the development of the future marketplace) some people would have had some ideas in mind and some early adopters would have already gained some experience with a plane or a car, but to what extent would the overall feedback been indicative of where the market was headed?

PS, this discussion reminds me a little of the digital camera market, and maybe the CNC market.  There is a time in almost every market when people ask “do you really need this feature or that function”?  In the photography market some of the pro’s pros would say “nah, if you are a manly man you can get a good photo with what you already have” and “Ansel Adams would tell you that the best camera is the one you have with you” and yada yada.  The reality, as we all know, is that the world is not just moving rapidly, but accelerating, fast.  And anyone here who doesn’t think there is a market for more accuracy and resolution, and for easily synchronized system-wide interoperability might be missing some opportunities.  But hey, I get it, there are manly men who can chop with an axe and only measure with a micrometer as needed.  They might have missed the market for CNC machines but with their knowledge, skills, and experience they can still build some beautiful stuff without a CNC machine, no doubt.  But I’m betting if they ever did add a CNC machine to their workshop, they might want to hook the CNC machine to a computer, and maybe connect the computer to the Internet.  The possibilities are endless – but they often start with frequency, amplitude, and phase – and electricity - which leads to needs and desires for increasing levels of accuracy and resolution to help examine, understand, and manage the ever increasing uses of frequencies and bandwidth. 

PSS, it’s just metaphors, I don’t have a CNC machine, or even make much out of wood or metal 😊
           

[pdenisowski]

Thanks for the questions.  And Thank You Very Much for the beautiful educational material you produce.  I appreciate both the outstanding information you provide and the way in which you ask for and encourage user feedback.

Thanks for the kind feedback and the detailed reply!

So maybe the question (“why would anyone want a Ref input?”) shouldn't be directed at or limited by this or that specific use case.  Rather, the question might be better asked “why would anyone not want a 10 MHz Ref input?”

Good point.  Except for the increase in cost, there's really no downside to having a ref in - negligible impact on the rest of the instrument.

we have under $400 signal generators that manage to include a 10 MHz Ref input (and a 10 MHz Ref output capability). 

My first thought when I read that is "how good is the ref out on a $400 function generator?"  I could measure this myself (since I have both cheap function generators and a high-end phase noise analyzer), but I also know from experience that there can be significant differences in the "quality" of 10 MHz Ref out signals, even those produced by so-called A-list T&M instrument manufacturers. 

So after reading your post (again, many thanks) my thoughts are:

1) A ref out is really not as important as a ref in.  Plenty of people have access to an acceptably "good" reference already, and the cost of including a "good" reference might be quite high compared to a ref in

2) A ref in is at least a "nice to have" that might get used more widely if it were available on lower-tier scopes. The question is mostly "how much would it cost to add one?"

I'll ask around internally and see what people say.  Quite honestly, I don't expect us to start adding ref ins (or ref outs) to lower-tier scopes anytime soon without some kind of business case, but it would be nice to be able to know the reasons in any event

And if anyone else has suggestions or feedback on the Ref in/out topic, please do let me know!

[Electro Fan]

Thanks for the questions.  And Thank You Very Much for the beautiful educational material you produce.  I appreciate both the outstanding information you provide and the way in which you ask for and encourage user feedback.

Thanks for the kind feedback and the detailed reply!

So maybe the question (“why would anyone want a Ref input?”) shouldn't be directed at or limited by this or that specific use case.  Rather, the question might be better asked “why would anyone not want a 10 MHz Ref input?”

Good point.  Except for the increase in cost, there's really no downside to having a ref in - negligible impact on the rest of the instrument.

we have under $400 signal generators that manage to include a 10 MHz Ref input (and a 10 MHz Ref output capability). 

My first thought when I read that is "how good is the ref out on a $400 function generator?"  I could measure this myself (since I have both cheap function generators and a high-end phase noise analyzer), but I also know from experience that there can be significant differences in the "quality" of 10 MHz Ref out signals, even those produced by so-called A-list T&M instrument manufacturers. 

So after reading your post (again, many thanks) my thoughts are:

1) A ref out is really not as important as a ref in.  Plenty of people have access to an acceptably "good" reference already, and the cost of including a "good" reference might be quite high compared to a ref in

2) A ref in is at least a "nice to have" that might get used more widely if it were available on lower-tier scopes. The question is mostly "how much would it cost to add one?"

I'll ask around internally and see what people say.  Quite honestly, I don't expect us to start adding ref ins (or ref outs) to lower-tier scopes anytime soon without some kind of business case, but it would be nice to be able to know the reasons in any event

And if anyone else has suggestions or feedback on the Ref in/out topic, please do let me know!

Thanks for read and distilling.  Yes, it would seem that Ref in is more useful and valuable than Ref out but somehow Siglent figured out how to get both into a ~$359 SDG1032X - so maybe someone could save a few dollars by leaving out the out, and adding the in, and then Bob or Dave's our Uncle.  With the availability of affordable but high quality GPSDOs that rival or in some respects exceed what used to be expensive rubidium standard reference clocks users can BYO whatever GPSDO meets their needs and extend it to multiple devices.  Customers that don't need a Ref in won't use it; wouldn't be the first time some users didn't use some (or a lot of the features) in a piece of test equipment but this feature and capability is kind of central to enhanced performance.  If it really just adds a modest incremental cost and price to an oscilloscope, let's do it.  If it adds too much to a $300-$500 scope, then maybe it's not feasible on an entry scope but given the 1032X example it would seem like on a ~$500-$1k oscilloscope it should feasible and worthwhile.  I predict it would be a meaningful competitive advantage at any price point over any similar scope without it.  Thanks for looking into it.  

[BillyO]

Quite honestly, I don't expect us to start adding ref ins (or ref outs) to lower-tier scopes anytime soon without some kind of business case, but it would be nice to be able to know the reasons in any event

Not sure how your company works, but the best business case for any company I had ever worked for was "It's what our customers want."  Knowing what your customer wants is usually cash in the bank if you are willing to act on it.  In this case, for the cost of a connector and a switch, it should be an easy action.

Where did I get my parts count from?  I have personally modified 18 pieces of equipment for myself, and folks I know, to take external 10MHz references and those parts, along with a small length of wire (you have a PCB so can do away with the wire bit) was all it took.

[2N3055]

If device bases it's own clock from the beginning from 10MHz than adding external Ref in might not be so hard, BOM wise. But what people always forget is that developing, characterizing any added feature also cost money.
Therefore features are not added just because. It is not software, and even in software, developing and testing a function cost money and computer resources.
If device bases it's own clock on, for instance, 25Mhz TCXO than adding external 10MHz is more involved.

I still wait for actual use case that requires external clock in in general use. I know there are many possible specialistic scenarios. But on one side we have people swearing you don't need anything better than CRT scopes with 1% timebase accuracy, and on the other side, those that would like to have all the features of the 30000 USD LeCroy on a 500 USD scope, "because, how hard can it be..."

Inexpensive scopes are pinnacle of engineering optimization, not cutting edge features. Or consumer like thinking where they just add all kinds of buzzwords to one up a competition and users never even know what it means.

[KE5FX]

My Agilent MSO6000-series scopes have built-in counters with 1-Hz resolution, so an external 10 MHz reference is occasionally useful to verify that a signal is close to the correct frequency.

Can't see why I'd care about a 10 MHz reference input on a traditional DSO otherwise.  It doesn't belong on the lower-end models, certainly.  I'm surprised that Agilent added external-ref support to the Infiniivision lineup TBH.

[2N3055]

Quite honestly, I don't expect us to start adding ref ins (or ref outs) to lower-tier scopes anytime soon without some kind of business case, but it would be nice to be able to know the reasons in any event

Not sure how your company works, but the best business case for any company I had ever worked for was "It's what our customers want."  Knowing what your customer wants is usually cash in the bank if you are willing to act on it.  In this case, for the cost of a connector and a switch, it should be an easy action.

Where did I get my parts count from?  I have personally modified 18 pieces of equipment for myself, and folks I know, to take external 10MHz references and those parts, along with a small length of wire (you have a PCB so can do away with the wire bit) was all it took.

I'll be blunt.  Customers want everything, and want to pay nothing for it. That is just the way it is.
And I'm not judging here. It is their right to negotiate value for money.

But trying to design a device that would cater to all customer wishes is impossible...
Or shall I say differently: I offer them good product that serves the purpose for good price. They came back to me with a "wish list" of "tiny additions" they would like.
I send them revised price for what is realistically 8x more work.
They order original device as initially spoken.

I speak from many years of experience.

And that works nicely when you are making bespoke stuff where you negotiate 1x1 with customer.

But making a device that is going to be sold to 100000 people, that is an art. You have to strike a balance and that is very hard.

[2N3055]

My Agilent MSO6000-series scopes have built-in counters with 1-Hz resolution, so an external 10 MHz reference is occasionally useful to verify that a signal is close to the correct frequency.

Can't see why I'd care about a 10 MHz reference input on a traditional DSO otherwise.  It doesn't belong on the lower-end models, certainly.  I'm surprised that Agilent added external-ref support to the Infiniivision lineup TBH.

In current Infiniivision lineup only 4000B and 6000 has it. 3000T/G does not despite being 15000 USD scope.

[Antonio90]

Repeating a request from user @mawyatt:

Pretty please, implement math on Bode plot traces. I suppose this is far from a trivial feature to implement, but it would be great to have.

Edit: typo.

[BillyO]

I still wait for actual use case that requires external clock in in general use.

Do you consider ham radio general?  Or networking or any other pursuit that deals with a lot of cabling?  Time domain reflectometry requires an accurate time base. 

[2N3055]

I still wait for actual use case that requires external clock in in general use.

Do you consider ham radio general?  Or networking or any other pursuit that deals with a lot of cabling?  Time domain reflectometry requires an accurate time base.

TDR?  If clock on scope has 20ppm absolute error , how much is that going to influence measurement of the length of the cable? That is 20m on 1000km length type of error. I assure you cable dielectric will have much greater span of propagation velocities than 20 ppm.

As for HAM and accurate frequencies, if 7 digits frequency readout is enough that could benefit from better accuracy, that would be good example.

[Martin72]

I can well imagine that a highly accurate clock and the ability to synchronize other devices to it has its special justification.
In general, however, I have certain doubts.

[radiolistener]

This is current supported for 4ch X-E DSO range as an option......these are 200 MHz max DSO's operating well below any 2.4 GHz WiFi.

They produce very wide bandwidth noise from 10-20 kHz to 5 GHz (for 5GHz protocol) and above. On SDR receiver I clearly see noise floor raising for about 5 dB when I enable WiFi on my router, I also see noise from neighbors WiFi routers.

I also can see 20-30 dB spikes from DC up to 300 kHz when my WiFi router transmits the packets (in idle state it produce pulse with about 0.5-2 Hz rate), the same I can see about 2-5 dB spikes from my neighbors WiFi routers in 0..300 kHz band.

WiFi router has about 2 Watts TX power at 2.4 GHz, or even more for some models, which is pretty high to affect your measurements.


You may not notice these interferences in usual oscilloscope mode on display, but it don't means that it is missing. You can found it if you record sample and analyze it on computer.

[temperance]

If the ref. clock is important, you will probably also need an appropriate trigger system and some other requirements related to to delays in the ADC when also considering that the ADC are in interleaved mode when only 1 ch for 1 ADC is active.

Who has the ability to measure for example aperture delay jitter when the ADC's are working in interleaved mode? If you happen to have that kind of equipment available I can only wonder why you are toying around with a budget scope. For those interested I can dig up a book on high speed ADC's from Analog device and scan some pages if I'm not too lazy with all formulas you need to perform such a "simple" measurement for which you will probably need equipment of which many in here (including me) don't even know the price.

As someone else said:

I'll be blunt.  Customers want everything, and want to pay nothing for it. That is just the way it is.

[2N3055]

This is current supported for 4ch X-E DSO range as an option......these are 200 MHz max DSO's operating well below any 2.4 GHz WiFi.

They produce very wide bandwidth noise from 10-20 kHz to 5 GHz (for 5GHz protocol) and above. On SDR receiver I clearly see noise floor raising for about 5 dB when I enable WiFi on my router, I also see noise from neighbors WiFi routers.

I also can see 20-30 dB spikes from DC up to 300 kHz when my WiFi router transmits the packets (in idle state it produce pulse with about 0.5-2 Hz rate), the same I can see about 2-5 dB spikes from my neighbors WiFi routers in 0..300 kHz band.

WiFi router has about 2 Watts TX power at 2.4 GHz, or even more for some models, which is pretty high to affect your measurements.

I agree with you.

Problem is not 2.4GHz but demodulated artefacts.
It is same problem like when your can hear interference in cheap computer speakers when phone is close...

[BillyO]

TDR?  If clock on scope has 20ppm absolute error , how much is that going to influence measurement of the length of the cable? That is 20m on 1000km length type of error. I assure you cable dielectric will have much greater span of propagation velocities than 20 ppm.

As for HAM and accurate frequencies, if 7 digits frequency readout is enough that could benefit from better accuracy, that would be good example.

I'll be quite honest.  I personally don't have a specific need for this feature.  I was just throwing some stuff out there.  From my perspective I want this feature.  I like all my stuff to sing the same song.  If I set my FG to 10,000,000 I want my counter to read 10,000,000 and if my counter is reading 104,000,000 I want my scope to read 104,000,000.  I'm just a hobbyist at this point and certainly do not need .01 ppm accuracy, put I have it and I like it.  I have a nice GPSDO and a distribution amplifier so all my FGs and my FCs agree on what they are doing.  It would also be nice to have my scope(s) that way to.

As someone else said:

I'll be blunt.  Customers want everything, and want to pay nothing for it. That is just the way it is.

If I could have gotten a version of my SDS804X-HD with a external reference input I would have paid a reasonable premium for that.  Let's say <= $50.

Do I need it?  No.  Heck, I don't even need to do electronics or ride dirt bikes or do nature photography.  I want to do them and I am willing to, and do, pay a small fortune to do them.

A small fortune is all I have 

I'm not a total nut.  At least I don't spend $15,000 for speaker cables, $10,000 for an "audiophile grade" network switch or $5,000 for a headphone amplifier.

[KungFuJosh]

I'm not a total nut.  At least I don't spend $15,000 for speaker cables, $10,000 for an "audiophile grade" network switch or $5,000 for a headphone amplifier.

Why would you when you could build them for a fraction of the cost? 😉

[BillyO]

Why would you when you could build them for a fraction of the cost? 😉

Exactly.  We electronics nerds are way, way smarter than your average nerds!

[tautech]

This is current supported for 4ch X-E DSO range as an option......these are 200 MHz max DSO's operating well below any 2.4 GHz WiFi.

They produce very wide bandwidth noise from 10-20 kHz to 5 GHz (for 5GHz protocol) and above. On SDR receiver I clearly see noise floor raising for about 5 dB when I enable WiFi on my router, I also see noise from neighbors WiFi routers.

I also can see 20-30 dB spikes from DC up to 300 kHz when my WiFi router transmits the packets (in idle state it produce pulse with about 0.5-2 Hz rate), the same I can see about 2-5 dB spikes from my neighbors WiFi routers in 0..300 kHz band.

WiFi router has about 2 Watts TX power at 2.4 GHz, or even more for some models, which is pretty high to affect your measurements.

I agree with you.

Problem is not 2.4GHz but demodulated artefacts.
It is same problem like when your can hear interference in cheap computer speakers when phone is close...

This, never seen any WiFi or phone interference on a scope at any time however when the analyzer comes out the phone is left somewhere else......just once doing sensitive analyser measurements the level warning was triggered when the phone did a handshake with a cell tower and none are close. 

Further TP Link TL-WN725N USB wireless dongle is the subject of this discussion and its everso powerful 100mW of 2.4 GHz output.
https://www.tp-link.com/us/home-networking/usb-adapter/tl-wn725n/

Maybe one day when I have a SDS7404A or higher BW scope I might discover otherwise but for now, WiFi impacting on scope measurements, it's an old wives tale ! 

[tautech]

10 MHz Reference input, and probably a 10 MHz output too.

Most HF dependant instruments already offer this.
For most AWG's and analyzers plus SDS7000A 10M In/Out is already supported.

With reason to get our SHA852A out today thought I had better see what 10 MHz reference options it offers...

Internal of course, External and autoSense plus a GPS reference when its GPS feature is activated. 

[Electro Fan]

The suggestion is to add a high performance paddle with an entry level paddle board (800HD) or at least with intermediate paddle boards (1kHD, 2kHD), we know Siglent can make really nice sailboats, powerboats, and probably yachts. 

[nanopico]

It would be nice for Siglent to provide users with a decent set of 50 ohm attenuators (3-6-10-20 dB).
For entry-level oscilloscopes, such as SDS804X, the performance of the attenuators at 1 GHz is not important.
Even for the 50 ohm feed through terminator, I have to use one from Rigol - ADP0150BNC 
Almost no distributor has the common BNC to BNC coaxial cable in stock. It seems that all users use Junkosha or, at least, Amphenol cables.
If I have to wait a week or two for the distributor to send me a Siglent cable, I will definitely order elsewhere.

[Electro Fan]

Back on the subject of the cost feasibility for adding 10 MHz Ref in to an oscilloscope....

Looks like not only does a Rigol DHO4000 provide 10 MHz Ref in and out, but even a 2 channel Rigol DHO1000 for about $699 (on sale for $599) provides 10 MHz Ref in and out (along with a 10.1'' 1280*800 HD display and HDMI). 

Before someone gets defensive, I see some good advantages with Siglent and I'd like to try a Siglent for my next scope but maybe a 10 MHz ref and maybe a tad more resolution and HDMI might be on the list of possible features in a 2000X HD "Plus"?  (I'm not holding my breath given the X HD's are pretty new, but maybe someday sooner than a lot later.)

[BillyO]

Back on the subject of the cost feasibility for adding 10 MHz Ref in to an oscilloscope....

Looks like not only does a Rigol ..

One then has to be happy with what Rigol leave out to give you the 10 MHz reference.

I don't want to start yet another scope war, but the first time I tried to use a Rigol it was like I hit a brick wall.  The UI made no sense to me.  The first time I used a Siglent it was like I had been using it for years.

I really do want that 10 MHz reference input, but would not switch to a Rigol to get it.  I still need to be able to use my scope.