How is the new Keysight 1000 x series scope compared with other low cost scopes?

[fishandchips]

Any opinion on this?

[pascal_sweden]

The probes are very low quality and sensitive to temperature. So you will have to calibrate the probe over and over.

[fishandchips]

Thanks for your experience. I am consider the new Keysight 1000x series, GW Instek GDS-2204E and the Rigol DS1054Z. How are they compared?

I think I read a post mentioning that the new Keysight 1000x series could be hacked like the 1054Z. What extra features will that bring? Is it easy to hack? If I intend to hack it to get better and more features for free, is it better to buy it as soon as possible while Keysight still allows that? Some people think that this is done intentionally by Keysight but who knows for sure?

[nctnico]

I'd add the MicSig TO1104 to the list with choices as well. Probes can be replaced so meh but what could become a problem is the extremely short memory in the Keysight 1000X series and the fact it only has 2 channels.

What do you need in an oscilloscope? What kind of circuits do you work on?

[fishandchips]

I'd add the MicSig TO1104 to the list with choices as well. Probes can be replaced so meh but what could become a problem is the extremely short memory in the Keysight 1000X series and the fact it only has 2 channels.

What do you need in an oscilloscope? What kind of circuits do you work on?

Extremely short memory even after the Keysight scope is hacked?

I do motor control, simple filters, Op Amp, mechatronics, arduino/maker kind of stuffs.

[JPortici]

which hack?

IMHO short memory is relative, when you're used to work with 2.5 kpts and 48 kpts.
you have listed scopes in very different price ranges and with different capabilities. for example, the rigol doesn't decode nor trigger can bus, if at some point you'll need it you're out of luck. the keysight while beying pricey and with a shorter memory does have a buil-in waveform gen and the FRA app. If you don't care for it you may want to instead invest in another scope

[nctnico]

As a rule of thumb you can use only up to 25% of the memory Keysight specifies and it gets less with reference traces and digital channels enabled. For 1Mpts that comes down to 250kpts per channel. I don't think you can hack more memory into the 1000X series and even if it would be possible in the future then there will be a hard limit of 4Mpts in total (1Mpts per channel) because the hardware simple doesn't have more memory.

[pascal_sweden]

Thanks for your experience. I am consider the new Keysight 1000x series, GW Instek GDS-2204E and the Rigol DS1054Z. How are they compared?

I don't have the scope myself, but read about a lot of issues in another thread on this forum.

[kcbrown]

As a rule of thumb you can use only up to 25% of the memory Keysight specifies and it gets less with reference traces and digital channels enabled. For 1Mpts that comes down to 250kpts per channel. I don't think you can hack more memory into the 1000X series and even if it would be possible in the future then there will be a hard limit of 4Mpts in total (1Mpts per channel) because the hardware simple doesn't have more memory.

The 1000X doesn't have digital channels (unless the trigger input channel counts -- it can be used for SPI protocol decoding so maybe it counts here?  Dunno), so I wouldn't think that would be much of a concern.

So if you stop the scope, or the scope stops automatically in one-shot trigger mode, you only get 250K points worth of buffered memory with the 1000X?   Does someone here have a 1000X that they can verify this with?  This should be straightforward to verify.

Also, I'm amused that this scope has "a lot of issues", seeing how it's an A-brand scope from a well-respected manufacturer, not some Chinese B-brand scope that is usually the target of ridicule...

[nctnico]

Unfortunately the actual  memory depth is well hidden on Keysight scopes and probably for a good reason. Deep memory is a very nice feature to have. Sure you can make do with less but IMHO deep memory is a feature which serves so many purposes that it should have a high priority when choosing an oscilloscope.

[TK]

The 1000X doesn't have digital channels (unless the trigger input channel counts -- it can be used for SPI protocol decoding so maybe it counts here?  Dunno), so I wouldn't think that would be much of a concern.

DSOX1000-X series has 1 digital channel that is shared with the external trigger BNC port.  It is not a simple external trigger, you can visualize the signal and add it to the analog BUS with CH1 and CH2.  EDUX1000-X does not have the same capability, but can be modded.

[TK]

Thanks for your experience. I am consider the new Keysight 1000x series, GW Instek GDS-2204E and the Rigol DS1054Z. How are they compared?

I don't have the scope myself, but read about a lot of issues in another thread on this forum.

Can you list the issues, besides the ones already commented on this thread?

[TheAmmoniacal]

And let's not forget that the probe compensation issue was one person, my probes are fine. They do feel cheap, like $15/pair cheap, but they do the job - and the lead is silicone.

[TK]

And let's not forget that the probe compensation issue was one person, my probes are fine. They do feel cheap, like $15/pair cheap, but they do the job - and the lead is silicone.

I also posted that I experienced a similar problem with the probe compensation, but I could not reproduce the problem.  I think I experienced the problem when I used a metallic pointed probe compensation tool, but I compensated again using the tool provided by Keysight and has been working fine.

[TK]

I'd add the MicSig TO1104 to the list with choices as well. Probes can be replaced so meh but what could become a problem is the extremely short memory in the Keysight 1000X series and the fact it only has 2 channels.

What do you need in an oscilloscope? What kind of circuits do you work on?

Extremely short memory even after the Keysight scope is hacked?

I do motor control, simple filters, Op Amp, mechatronics, arduino/maker kind of stuffs.

The EDUX can gain the most by hacking (it requieres adding some HW components as well), but on the DSOX the only hack possible is 70MHz to 100MHz bandwidth improvement (I think it was tested up to 200MHz by Dave Jones).

[kcbrown]

Unfortunately the actual  memory depth is well hidden on Keysight scopes and probably for a good reason. Deep memory is a very nice feature to have. Sure you can make do with less but IMHO deep memory is a feature which serves so many purposes that it should have a high priority when choosing an oscilloscope.

The actual usable memory depth should be trivial to determine.  You know your timebase and your sample rate, and you can see when the scope is stopped how much of the capture is visible on the screen -- and if you zoom out enough once the scope is stopped, you can see how much of the screen the entire capture occupies.  Then you just do the math.

You won't get any argument from me about the usefulness of deep memory.  It's why I've argued that capture memory really should be much deeper than it is, given the cost and availability of DDR3 memory.

[fishandchips]

I read that the DS1054Z's fan is very noisy and the GDS-2204E is a bit quiet. How about the Keysight 1000x series? In terms of the GUI, looks like the Keysight 1000x series is the best. Do we see slow motions and lags in the 1000x series? I read that there were lags with the DS1054Z but the problem seems to have been fixed.

[TK]

I read that the DS1054Z's fan is very noisy and the GDS-2204E is a bit quiet. How about the Keysight 1000x series? In terms of the GUI, looks like the Keysight 1000x series is the best. Do we see slow motions and lags in the 1000x series? I read that there were lags with the DS1054Z but the problem seems to have been fixed.

Keysight (Agilent, HP) scopes are the best in terms of UI responsiveness

[mikeselectricstuff]

I'd add the MicSig TO1104 to the list with choices as well. Probes can be replaced so meh but what could become a problem is the extremely short memory in the Keysight 1000X series and the fact it only has 2 channels.

What do you need in an oscilloscope? What kind of circuits do you work on?

Extremely short memory even after the Keysight scope is hacked?

I do motor control, simple filters, Op Amp, mechatronics, arduino/maker kind of stuffs.

The EDUX can gain the most by hacking (it requieres adding some HW components as well), but on the DSOX the only hack possible is 70MHz to 100MHz bandwidth improvement (I think it was tested up to 200MHz by Dave Jones).

The hardware is definitely capable of 200MHz, but I suspect some software hacks will be needed to make it work properly.

[nctnico]

Unfortunately the actual  memory depth is well hidden on Keysight scopes and probably for a good reason. Deep memory is a very nice feature to have. Sure you can make do with less but IMHO deep memory is a feature which serves so many purposes that it should have a high priority when choosing an oscilloscope.

The actual usable memory depth should be trivial to determine.  You know your timebase and your sample rate,

And where does Keysight show the samplerate? Last time I used a Keysight scope that was well hidden as well so it isn't easy to 'just do the math'. And why should I need to do the math in the first place? Setting up segmented recording without knowing length per segment and samplerate is a real PITA because you need to know the samplerate to sample a signal at a high enough rate and the length to know how long each segment is.

BTW: I find the GDS2204E a lot easier to drive compared to the Keysight scope I used to have. The buttons at the bottom and side on the GDS2204E make it much easier to setup compared to going several levels deep into sub menus on the Keysight.

[TK]

Unfortunately the actual  memory depth is well hidden on Keysight scopes and probably for a good reason. Deep memory is a very nice feature to have. Sure you can make do with less but IMHO deep memory is a feature which serves so many purposes that it should have a high priority when choosing an oscilloscope.

The actual usable memory depth should be trivial to determine.  You know your timebase and your sample rate,

And where does Keysight show the samplerate? Last time I used a Keysight scope that was well hidden as well so it isn't easy to 'just do the math'. And why should I need to do the math in the first place? Setting up segmented recording without knowing length per segment and samplerate is a real PITA because you need to know the samplerate to sample a signal at a high enough rate and the length to know how long each segment is.

BTW: I find the GDS2204E a lot easier to drive compared to the Keysight scope I used to have. The buttons at the bottom and side on the GDS2204E make it much easier to setup compared to going several levels deep into sub menus on the Keysight.

On the 1000X series, sample rate is shown on the right side of the screen, under Acquisition

[nctnico]

But once you exit that menu it is gone and thus not available when you need to know it to setup something else 

[kcbrown]

And where does Keysight show the samplerate? Last time I used a Keysight scope that was well hidden as well so it isn't easy to 'just do the math'.

That was apparently answered as I type this.

And why should I need to do the math in the first place?

You mean as opposed to having the scope tell you how much total memory is being used?  Yeah, I prefer the scope tell me these things as well.

But when you're using the scope, what are you really interested in as regards the memory depth?   The actual amount of memory used?  Or the amount of time captured?

Setting up segmented recording without knowing length per segment and samplerate is a real PITA because you need to know the samplerate to sample a signal at a high enough rate and the length to know how long each segment is.

Yep.

In Keysight's defense, at least they're now transparent about how and when memory is reduced.  From page 80 of the user manual (http://literature.cdn.keysight.com/litweb/pdf/54612-97001.pdf):

The maximum number of data points depends on these things:
• Whether acquisitions are running. When stopped, data comes from the raw
acquisition record. When running, data comes from the smaller measurement
record.
• Whether the oscilloscope was stopped using [Stop] or [Single]. Running
acquisitions split memory to provide fast waveform update rates. Single
acquisitions use full memory.
• Whether only one channel of a pair is turned on. (Channels 1 and 2 are one
pair.) Acquisition memory is divided among the channels in a pair.
• Whether reference waveforms are on. Displayed reference waveforms consume
acquisition memory.
• Whether segmented memory (available on DSOX1000-Series models) is on.
Acquisition memory is divided by the number of segments.

Of course, that doesn't make your main point any less valid in the slightest.

BTW: I find the GDS2204E a lot easier to drive compared to the Keysight scope I used to have. The buttons at the bottom and side on the GDS2204E make it much easier to setup compared to going several levels deep into sub menus on the Keysight.

I do wish I could comment on this (the GDS2204E looks like a really nice scope, and I wish someone would do a video review of it to show what it's really capable of).   

[Relaxe]

I do motor control, simple filters, Op Amp, mechatronics, arduino/maker kind of stuffs.

I would definitively go for a 4-channel if you intend to work on BLDC, PMSM, ACIM or any kind of brushless motors.
Having one channel per phase is a very good thing.

[Someone]

But once you exit that menu it is gone and thus not available when you need to know it to setup something else 

You have that back to front, the menu is always there showing the default "summary" information about the channels and the horizontal mode and sampling rate:
https://i.ytimg.com/vi/lZULfMz9nlI/maxresdefault.jpg
Its covered by menus when you are changing something more detailed, unlike the larger 2000/3000/4000/6000 x series which have their soft buttons along the bottom and keep the summary pane on the right visible under more conditions.