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

[JPortici]

I read that there were lags with the DS1054Z but the problem seems to have been fixed.

fixed in the sense that by enabling something like decode it's not awfully slower than it usually is? kind of. almost. it's still _very_ slow

[Fungus]

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.

Also for "Arduino" - two channels isn't enough for that.

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

And probably some hardware hacks as well.

[mikeselectricstuff]

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.

Also for "Arduino" - two channels isn't enough for that.

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

And probably some hardware hacks as well.

I don't think so. Simply by changing the option links you can get ch1 fully working at 200MHz but ch2 has issues - this is clearly to do with how the firmware is configuring the hardware,  so it should be possible to tweak it to set the hardware up correctly for 200mHz

[serggio]

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.

Really?

Just do measurement with next settings on warmed up scope.

ANALOG
Ch 1 Scale 100mV/, Pos 300.00mV, Coup DC, BW Limit On, Inv Off, Imp 1M Ohm
     Probe 10.000000 : 1, Skew 0.0s

TRIGGER
Sweep Mode Normal, Coup DC, Noise Rej Off, HF Rej Off, Holdoff 60.0ns
Mode Edge

HORIZONTAL
Mode Normal, Ref Left, Main Scale 200.0us/, Main Delay 0.0s

ACQUISITION
Mode Averaging, # Avgs 8, Realtime On, Vectors On, Persistence Off.

Take are screen shot your compensation signal.

Give your scope rest couple hours.
Then start up again and make new screen shot after cold start.

For this money scope itself really cool, menu very fast, no lags, no delay. All functions work well.
But included probes - bull shit. I can make photo how probe's tips was scratched by their included hook tips after two week of use. Right place for this probes in trash.

As for fan, yes, it's probably noisy, I cant compare with Rigol right now, because I don't have it for side by side comparison, so I think this is compatible by noise based in my past experience, may be slightly less.

While I work with serial buses, I also do not feel any inconvenience with 1 Mpts memory in my scope, could be great if follower of "huge deep memory" and long single mode acquisition time, will explain me for what in 70/100 mHz scopes them need more. Use it as logic analyzer?

[nctnico]

Deep memory is handy to look at details in long captures. No need to try and figure out how to trigger on part of a signal and retry several times. Also data in one long capture can be related and two seperate captures may not be so useful. I'm currently working on an RFID system. Being able to capture the entire write transaction in great detail is very helpful.

[fishandchips]

For this money scope itself really cool, menu very fast, no lags, no delay. All functions work well.
But included probes - bull shit. I can make photo how probe's tips was scratched by their included hook tips after two week of use. Right place for this probes in trash.

Can we get a better/much better pair of probes at a cost of under US$100?

[GeorgeOfTheJungle]

While I work with serial buses, I also do not feel any inconvenience with 1 Mpts memory in my scope, could be great if follower of "huge deep memory" and long single mode acquisition time, will explain me for what in 70/100 mHz scopes them need more. Use it as logic analyzer?

You have a cpu, a power supply current sense signal a program and a serial port and want to know what and where is using how much energy. So you put a bunch of printf()s to the serial port here and there and voila, can correlate on the scope screen power in and code execution.

Typical case the cpu wakes up does something and goes back to sleep, max sample rate would then be mem depth/capture time in seconds, so you'd only have 333kSa/s (or worse) max for a say 3 seconds capture which may very well be too few data points per baud to decode properly the serial data => decoding doesn't work.

Most things these days use 115200 baud by default, for that you'd better have 115200*16= 1.8 MPoints per second of capture to be safe or at least half that (8 samples per bit slot might be enough perhaps, idk).

Those infiniivisions have a hardware decoder so (I guess) none of this applies 

But even so, 100 Kpts mem depth nowadays such as in the EDU1000X is a 

[serggio]

How long you'll be rewind all your captured data for looking for something interesting you? 
I have CCTV cameras that can do 24/7 capture, but I use movement detector on it  Do you know why? I prefer do not spend my time when I need focusing on the problem.
Yes, I use trigger, I like trigger. It let me focusing on the problem with signal, it let me compare fast and slow changing signal at same point, and I don't want to wait 5 second to fill my scope  memory, then rewind all tape, then change setup of my scope and repeat this procedure again and again till I'll not looking for something really important in the signal..
 

[TK]

For long time capture, I prefer to use an USB logic analyzer with protocol analyzer.

[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.

Really?

You are right, on those settings my probes are not properly compensated. I will recompensate the two sets of probes I have and check them again in a few hours.

[nctnico]

How long you'll be rewind all your captured data for looking for something interesting you? 

The thing is you don't look at everything but if you have enough data you can dig straight into a problem area without needing to recapture data. It simply is a much more efficient way of working.

[serggio]

How long you'll be rewind all your captured data for looking for something interesting you? 

The thing is you don't look at everything but if you have enough data you can dig straight into a problem area without needing to recapture data. It simply is a much more efficient way of working.

If you don't look at everything, why you need everything then? 
Setup right trigger, use segmented memory and etc..
Ok,... What memory use Rigol 1054z  from all available memory at 1 GSa/s (I suppose It not support 2 Gsa/s)?

[nctnico]

How long you'll be rewind all your captured data for looking for something interesting you? 

The thing is you don't look at everything but if you have enough data you can dig straight into a problem area without needing to recapture data. It simply is a much more efficient way of working.

If you don't look at everything, why you need everything then? 
Setup right trigger, use segmented memory and etc..

Why do things in a complicated way? Also it is not always easy/possible to redo the same test. BTW some oscilloscopes have a search function where you can let the oscilloscope look through the captured waveform to loop for anomalies.

[serggio]

How long you'll be rewind all your captured data for looking for something interesting you? 

The thing is you don't look at everything but if you have enough data you can dig straight into a problem area without needing to recapture data. It simply is a much more efficient way of working.

If you don't look at everything, why you need everything then? 
Setup right trigger, use segmented memory and etc..

Why do things in a complicated way? Also it is not always easy/possible to redo the same test. BTW some oscilloscopes have a search function where you can let the oscilloscope look through the captured waveform to loop for anomalies.

What anomalies you talking about? If you know nature of this anomalies, you'll able to setup trigger. If you don't know what you looking for, you'll seat front of th scope and discover all recorded data oneself. What is difference? 1-20 Mpts...  We talking about nothing. All this 14-54 Mpts seems like toys when we talking about 1-2GSa/s scopes.

[fishandchips]

In the past, many forum users recommended the Rigol DS1054z when somebody asked for suggestions to purchase a scope. Given that the new Keysight 1000x is out, which one is a better choice? It seems to me that the GDS-2204E is better than the DS1054Z. If 2-channel is OK for me, which do you recommend?

If we could hack the Keysight to get 200MHz and more features, shall I get the cheapest version like the 50MHz and hack to 200MHz?

[TK]

There is no hack yet to get to the full option 1000X buying the cheapest EDUX 50MHz scope.  The most you can get from the EDUX 50MHz is to DSOX with 70MHz and you need to switch a couple of resistors and purchase and solder some components to the EXT INPUT section.

Also if you need WAVEGEN, you need to purchase the EDUX-G or DSOX-G model, you cannot get the WAVEGEN later if you purchase the EDUX-A or DSOX-A models, as the WAVEGEN components are not populated in the PCB.  Bode Plot is only available on the G models, as you need the Wavegen functionality.

Here is the 1000X hack thread: https://www.eevblog.com/forum/blog/eevblog-978-keysight-1000x-hacking/msg1154923/#msg1154923

[fishandchips]

There is no hack yet to get to the full option 1000X buying the cheapest EDUX 50MHz scope.  The most you can get from the EDUX 50MHz is to DSOX with 70MHz and you need to switch a couple of resistors and purchase and solder some components to the EXT INPUT section.

Also if you need WAVEGEN, you need to purchase the EDUX-G or DSOX-G model, you cannot get the WAVEGEN later if you purchase the EDUX-A or DSOX-A models, as the WAVEGEN components are not populated in the PCB.  Bode Plot is only available on the G models, as you need the Wavegen functionality.

Here is the 1000X hack thread: https://www.eevblog.com/forum/blog/eevblog-978-keysight-1000x-hacking/msg1154923/#msg1154923

Thanks.  I have a function generator already. Is there a point to have the WAVEGEN component?

[TK]

There is no hack yet to get to the full option 1000X buying the cheapest EDUX 50MHz scope.  The most you can get from the EDUX 50MHz is to DSOX with 70MHz and you need to switch a couple of resistors and purchase and solder some components to the EXT INPUT section.

Also if you need WAVEGEN, you need to purchase the EDUX-G or DSOX-G model, you cannot get the WAVEGEN later if you purchase the EDUX-A or DSOX-A models, as the WAVEGEN components are not populated in the PCB.  Bode Plot is only available on the G models, as you need the Wavegen functionality.

Here is the 1000X hack thread: https://www.eevblog.com/forum/blog/eevblog-978-keysight-1000x-hacking/msg1154923/#msg1154923

Thanks.  I have a function generator already. Is there a point to have the WAVEGEN component?

The only point I can think of is to use the built in Frequency response analyzer (Bode Plot) on the G models

[Fungus]

In the past, many forum users recommended the Rigol DS1054z when somebody asked for suggestions to purchase a scope. Given that the new Keysight 1000x is out, which one is a better choice? It seems to me that the GDS-2204E is better than the DS1054Z. If 2-channel is OK for me, which do you recommend?

The Siglent.

[GeorgeOfTheJungle]

The Siglent.

No way. Not unless/until they fix the (too many) bugs in the FW. I've had one sds1kx and it's a pleasure to use (not more than the keysights though) but you can't rely on something that refuses to work properly all too often. So I had to return it.

[mmagin]

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

Hmm, that's weird.  I wonder if good probes use an air dielectric trimmer for compensation but cheap probes use something very sensitive to temperature changes?

[Fungus]

The Siglent.

No way. Not unless/until they fix the (too many) bugs in the FW. I've had one sds1kx and it's a pleasure to use (not more than the keysights though) but you can't rely on something that refuses to work properly all too often. So I had to return it.

The Keysight firmware isn't much better and that's on your list, so...

PS: The Rigol firmware is quite good now. 

[Fungus]

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

Hmm, that's weird.  I wonder if good probes use an air dielectric trimmer for compensation but cheap probes use something very sensitive to temperature changes?

There's plenty of cheap probes that aren't affected by temperature. It looks like this is purely Keysight trying to save $0.02 on their new models.

[GeorgeOfTheJungle]

The Keysight firmware isn't much better and that's on your list, so...

Miles better, I know for sure I use them daily.

[mikeselectricstuff]

The Keysight firmware isn't much better and that's on your list, so...

You are talking complete bollocks.