NEW 100MHz Keysight Scope Release on 1st March 2017

[serggio]

The input capacitance of the oscilloscope is probably changing as it warms up which is a bad sign.  Is the compensation consistent as the high impedance input attenuators are switched in and out?

Do you mean, is some changing in the signal for 10x after switching from 10x to 1x and back 10x? No, signal remains same. Both probes have same behaviour on both channels.
I assume that LF compensation is nothing more than matching output probe capacitance to the oscilloscope input probe capacitance. It should be equal and adjusting only by variable capacitor placed near BNC connector at the probe.  Only passive components involved in this chain.
I give my scope warm up 20 min and more and after that checking compensation matching. I use also use hi-res and averaging acquisition modes for make fine adjustment.
If I switch off scope, in give him rest and start using again, probably next day, after checking compensation signal form not seems so fine adjusted. It always seems just slightly undercompensated.
May be I want fine precision, may be I just cavilling at the scope, probably it's ok for scope, I do not know.
Would bee good to check it with external signal generator, but I currently do not have it.

[kcbrown]

Prices should have gone down over the past 20 years. Salaries are irrelevant in this discussion.

Is that so?

You're making an economic argument here.  Salaries are directly relevant to that.   And clearly you haven't thought the economics all the way through.

Trace every single transaction all the way back to its endpoint sources.  Guess what all of those sources involve?  The answer is simple: the source of every component of every economic transaction is a payment for someone's time.   Even when you're paying for raw materials, what you're actually paying for is someone's time to retrieve those raw materials from their source, transport them from that location, and package them up in some usable form.

And guess what a person's time is exchanged for?   A salary or wage.

Salaries and wages are directly involved in the real cost of goods and services.  They have to be, because they are how people are compensated for their time, and time is ultimately what you're paying for when you pay for a good or service.

If the Chinese manufacturers can do it, why not the American manufacturers? Everything is produced in China anyhow.

But not everything is designed and developed in China.  The firmware for Keysight scopes, for instance, is almost certainly written by people in the United States.

In fact Chinese scopes are not subject to the basic flaws that were discovered recently on this American scope. And no, this is not a software problem. This is a plain hardware problem. Way to go Keysight!

Is that so?   Then pray tell, where's the logic analyzer on the Siglent SDS-1202X-E?  Where's the logic analyzer on the Instek GDS-1000G series?

[mikeselectricstuff]

If the Chinese manufacturers can do it, why not the American manufacturers? Everything is produced in China anyhow.

Because unlike Chinese manufacturers they value quality and support over lowest possible proce

Chinese scopes do have some software issues, but at least these can all be fixed in software.

Except chances are they won't be
 

[serggio]

Pascal_sweden, I'm really can't understand what do you try to say us about Keysight and price.
If you want to be part of army free beta testers for Chinese manufacturers, you can easy join it for small amount of fees. Other people choice companies with good support, well done documentation, key studies, application notes, trainings materials and videos available from this companies. 
At least Keysight pay salary to Daniel, who haunt here and give us response for our questions. Where somebody from Rigol, Hantek and etc..

Where I need to go with Rigol for performing calibration for my device? I know where I can do it for Fluke, Keysight, Tek...
That all cost money, that is because not able to do price comparison for Chinese brands and Americans brands.

[rsjsouza]

Ok, I'll bite.

I am not saying that the MSOX3012 should be 20 times cheaper than the HP54645D,
but it should not have exactly the same price tag either.

Wrong. The 1997 US$3.5k is valued today at US$5.3k due to inflation, which buys you today a MSOX3014 - twice the channels.

If the Chinese manufacturers can do it, why not the American manufacturers? Everything is produced in China anyhow.

The cost of tangible items comes down with technology, but the intangible part of the oscilloscope (firmware) grew in importance over the decades but not its cost. That is one element that keeps the prices from going down significantly.

In fact Chinese scopes are not subject to the basic flaws that were discovered recently on this American scope. And no, this is not a software problem. This is a plain hardware problem. Way to go Keysight!

Chinese scopes do have some software issues, but at least these can all be fixed in software.

And although the probes can be a bit flimsy, at least they don't scratch over time. Way to go Keysight!

I would take scracthy probes all day long when compared to the DP832A overheat, Project Yaigol or the other user that had the ADC issues on his DS4000 due to the absence of a heatsink. My DS4000 came with NO heatsinks, and was purchased straight from them!

No brand is free from issues, as the time to release a product reduces the amount of testability by a significant margin - like it or not, that is today's world for you where we are all beta testers up to a certain degree. The important aspect here is how you recover when put in emergency mode - something I can clearly guarantee you that established brands perform better and more consistently.

Other people choice companies with good support, well done documentation, key studies, application notes, trainings materials and videos available from this companies.  At least Keysight pay salary to Daniel, who haunt here and give us response for our questions. Where somebody from Rigol, Hantek and etc..

When I worked closely with customers back in Brazil that was my neverending battle: some of them wanted the world in terms of information and support but always wanted to lowball us with competition that had no presence in the country. I had to explain this very same story over and over. 

Where I need to go with Rigol for performing calibration for my device? I know where I can do it for Fluke, Keysight, Tek...
That all cost money, that is because not able to do price comparison for Chinese brands and Americans brands.

Not only calibration but also repair and maintenance agreements - the oscilloscope becomes a very risky investment if one can't find options to keep it in the long run. This tends to cost money.

[pascal_sweden]

I would take scracthy probes all day long when compared to the DP832A overheat, Project Yaigol or the other user that had the ADC issues on his DS4000 due to the absence of a heatsink. My DS4000 came with NO heatsinks, and was purchased straight from them!

You deliberately take out the probe compensation issue? Or is that issue just a minor glitch in your opinion? Really?

- Even been perfectly compensated, after some time probes not give me same fine picture after some time of using scope (probably 1-2 hour) and after checking probe compensation again it seems just slightly  under compensated. I tried to do fine compensation for both probes after 20 min warm up my scope and repeat check it again after one hour of using scope, I seen just slight under compensation for my probes. Probably input chain is slightly unstable, I don't know exactly 

And you really have no issue that Keysight is still building further on a 15 year old MegaZoom technology, which cripples the memory size to 1 MPts? The Education series of the 1000 X-Series even has less memory. Cripple the students ... cripple the future engineers that are going to drive further the technological evolution in this society. Yes ... why not Keysight?

Dear Keysight Product Management ... MegaZoom was great back then 15 years ago ... but come on... don't build further on to it, until infinity ... 1 MPts doesn't do the job anymore anno 2017, we know it, you know it, everybody knows it!

Hard to believe in fact that they haven't amortized the investment yet for that 15 year old ASIC.
Spinning a new ASIC is no rocket science these days, when using High-Level Synthesis tools.

[ebastler]

pascal_sweden -- I really don't understand what mission you see yourself on?

Feel free to buy something else. Feel free to apply for a marketing job at Keysight, so you can let them see the light. But apart from that, what are you trying to achieve here? Obtain a court ruling that Keysight be banned from the market for continuing to build on dated technology platform??

As others have said here, market acceptance is the only force that will drive price vs. feature set decisions. And I am pretty sure that the folks at Keysight are monitoring their sales, as well as the development of the market in general, quite closely, and are drawing their conclusions.

[TK]

You deliberately take out the probe compensation issue? Or is that issue just a minor glitch in your opinion? Really?

It might be an issue with one particular unit, as my unit does not show this problem

[David Hess]

The input capacitance of the oscilloscope is probably changing as it warms up which is a bad sign.  Is the compensation consistent as the high impedance input attenuators are switched in and out?

Do you mean, is some changing in the signal for 10x after switching from 10x to 1x and back 10x? No, signal remains same. Both probes have same behaviour on both channels.

I mean the oscilloscope has one or more switchable high impedance attenuators between the input BNC and the high impedance buffer.  If the input capacitance is drifting, then it is unlikely to be drifting by the same amount for each high impedance attenuator (1) so the compensation will not track as the volt/div is changed.

I assume that LF compensation is nothing more than matching output probe capacitance to the oscilloscope input probe capacitance. It should be equal and adjusting only by variable capacitor placed near BNC connector at the probe.  Only passive components involved in this chain.

That is basically correct or at least close enough for this discussion.  The compensation adjustment could be either in the x10 probe's body or at the BNC connector depending on the x10 probe's design.

I give my scope warm up 20 min and more and after that checking compensation matching. I use also use hi-res and averaging acquisition modes for make fine adjustment.

I do the same thing but without any warm up period because the compensation does not noticeably change on any of my oscilloscopes.  On my analog oscilloscopes, I enable their 20 MHz bandwidth limit to reduce noise.

If I switch off scope, in give him rest and start using again, probably next day, after checking compensation signal form not seems so fine adjusted. It always seems just slightly undercompensated.
May be I want fine precision, may be I just cavilling at the scope, probably it's ok for scope, I do not know.

This is the part which seems odd; I would normally consider warm up drift of the compensation a design or manufacturing flaw.

Would bee good to check it with external signal generator, but I currently do not have it.

This is probably not a bad idea; maybe the probe compensation signal is drifting as it warms up.

An external signal source using a BNC to probe tip adapter is also a good way to get a very clean compensation signal.

(1) Each high impedance attenuator has its own input capacitance and compensation adjustments so any combination of attenuators always produces the same input capacitance.  Modern designs might be so precise that the input attenuators do not need to be trimmed.  Adjusting the input capacitance and compensation of these attenuators is very much like adjusting the compensation of a x10 probe and a x10 probe can be used for the procedure if an input capacitance normalizer is not available.

[rsjsouza]

I would take scracthy probes all day long when compared to the DP832A overheat, Project Yaigol or the other user that had the ADC issues on his DS4000 due to the absence of a heatsink. My DS4000 came with NO heatsinks, and was purchased straight from them!

You deliberately take out the probe compensation issue? Or is that issue just a minor glitch in your opinion? Really?

So far this is an isolated case which may be a faulty unit.

Conversely, you deliberately failed to acknowledge that your previous assertions about Chinese brands are simply baseless.

(...)
In fact Chinese scopes are not subject to the basic flaws that were discovered recently on this American scope. And no, this is not a software problem. This is a plain hardware problem. Way to go Keysight!
(...)
Chinese scopes do have some software issues, but at least these can all be fixed in software.

Look, I understand how it seems frustrating that a higher price should give you a flawless and ultimate experience, but not only issues can creep into the production (that is what warranties are designed for) but there are also many details that come into consideration when the products evolve and fill specific market needs.

The HP marketers of 20 years ago would not have guessed they would create a joint venture with a local Chinese manufacturer only to see them as competition and having to create a product that addresses that market need, despite lacking the feature they mostly praised in the article. Also, their vision of 20 years ago certainly guided the path to modern days (they still have MSOs on their product line, for example), but they couldn't predict that a critical sales pitch of the future would be focused on the availability of serial decoding protocols, memory depth, software enabled options and others.

[ebastler]

Would bee good to check it with external signal generator, but I currently do not have it.

This is probably not a bad idea; maybe the probe compensation signal is drifting as it warms up.

But would a drift of the compensation signal really be critical, and be able to make previously adjusted probes appear uncompensated? I have always been working under the assumption that the details of the compensation reference signal don't matter much. In fact, my (low-end) scopes have compensation signals with pretty lame slopes...

[ralphrmartin]

Prices should have gone down over the past 20 years. Salaries are irrelevant in this discussion.

If the Chinese manufacturers can do it, why not the American manufacturers? Everything is produced in China anyhow.

Your logic is faulty. It is the high salaries of US engineers (relative to Chinese engineers) which keep the US equipment prices higher. Much of the cost of an oscilloscope is in the design and development and testing, not in the price of the parts. Many options offered by many scope manufacturers have basically zero parts cost.

[David Hess]

Would bee good to check it with external signal generator, but I currently do not have it.

This is probably not a bad idea; maybe the probe compensation signal is drifting as it warms up.

But would a drift of the compensation signal really be critical, and be able to make previously adjusted probes appear uncompensated? I have always been working under the assumption that the details of the compensation reference signal don't matter much. In fact, my (low-end) scopes have compensation signals with pretty lame slopes...

For probe compensation the slope and overshoot of the compensation signal is less important than the low frequency settling time.  For example a compensation source which relies on bipolar Vce saturation for a flat pulse is likely to have problems.

But it is more likely that there is a problem with the vertical amplifier high input impedance circuits.  Printed circuit board "hook" problems can look like poor compensation and they are something which could vary from production run to production run.  I can just imagine some purchasing manager thinking, "One FR4 board supplier is as good as another, right?  RIGHT?"

[AndyC_772]

I have always been working under the assumption that the details of the compensation reference signal don't matter much. In fact, my (low-end) scopes have compensation signals with pretty lame slopes...

You're quite right. Probe compensation is to do with frequencies much lower than you might expect, and there's no need for the compensation signal to have 'fast' edges anywhere near the bandwidth limit of the scope or probe.

[serggio]

Well,
I did probe compensation more than one hour ago and last hour signal remain the same.
I'll do same measurement tomorrow after cold start my device.
I'll also marked issue with time format, hope Keysight will correct it in next FW release and will be possible to setup Time/Date format by user.

I also do not have ground wire in my supply line, inside my home phase and neutral only in sockets, but I suppose this is not problem for probe compensation.

[nfmax]

It might be worth looking at the effect of the temperature of the probe, as well.

[serggio]

I did exactly same measurements for probe compensation (same setup for acqusition, same room temperature, and etc..) after cold start.

As I told before, probes seems undercompensated.
Interesting, that this issue exactly same for both channels. Yesterday second channel probe was also exactly compensated after warm up, matched ch.1 probe.
Well, would be interesting to know what DanielBogdanoff will say about his testing.

[Keysight DanielBogdanoff]

The probes team is still looking into it, and getting some fresh probes from manufacturing to check in on things. It'll be at least a few days before we get our hands on a fresh set of probes.

[serggio]

Hi Daniel,

Do you think that this is could be related with probes itself?
Well, seems I did pretty stupid testing, but I tried to logging temperature inside BNC connector.
I used my Fluke 287 with thermocouple probe putted inside BNC itself, because probes end, close to BNC seems to be "slightly" warm. This is chassis warm up, and to be honestly, I do not believe that passive components inside probe can be such unstable with 5-10 C degrees warm up.
After 5 min rest thermocouple probe inside the BNC (just to check that temp is stable), I turn on power of my scope. As you can see, temp was rice more than 7 C degrees at 25 min, after that thermocouple has been extracted.
At beginning, probes also seems been undercompensated, but with warm up it been more and more close to past fine compensation. After one hour probe connected to  Ch.1 seems been fine compensated.

Well,.. I connected  second probe, that rested on the table before, at Ch.2 BNC and after 15 min in seems to be approaching to Ch.1

I do not know, is it really make to sence, but hope it will help (if your team really see problem. Just may be all this does not matter for scope work)

[TK]

The probes team is still looking into it, and getting some fresh probes from manufacturing to check in on things. It'll be at least a few days before we get our hands on a fresh set of probes.

I am experiencing a similar problem with the 70MHz probes included with the EDUX scope.  I compensate them, and I find them uncompensated many days later, but not clear the circumstances... sometimes is right after cold startup and it does not change with the scope warmed up.

[serggio]

Thank TK,

Because seems, only one who watching how I suffer from chicken-feed with this probes right here. Unfortunately, she telling nothing   

I discovered this issue when I tried to setup my scope right from the box after arriving. I did initial probe compensation but after some time I found signal changing and repeat this procedure twice or more

[Keysight DanielBogdanoff]

Hi @serggio,

It's definitely possible it's a probe thing, but I'm going to hold off on making any guesses until the R&D folks do their investigating.

[TheAmmoniacal]

Scrapped my toy..

Finally found a way to buy this thing, turns out a nearby electronics shop resells a lot of Farnell stuff - so I got them to stock the DSOX1102A. For those in Norway, it can be purchased here: https://www.digitalimpuls.no/WebPages/Produkt/ProduktInfo.aspx?plid=155058

First impressions are very much positive, looks small and compact, feels heavy and solid, nice and large screen, responsive and intuitive (G)UI. All I would expect from Keysight  .

You can message me all the license options when you have time Daniel, won't tell anyone 

[kcbrown]

Thank TK,

Because seems, only one who watching how I suffer from chicken-feed with this probes right here. Unfortunately, she telling nothing   

I discovered this issue when I tried to setup my scope right from the box after arriving. I did initial probe compensation but after some time I found signal changing and repeat this procedure twice or more

Are you sure the issue isn't this:



 

[Keysight DanielBogdanoff]

Preliminary findings on the probe comp issue:

R&D says it appears that the compensation change is correlated to probe temperature. We still have a very small sample size for testing the issue and are waiting for a shipment of more to further characterize and evaluate.

For now, R&D's guidance is that if highly accurate measurements are required, warm up both scope & probe together then compensate.