Oscilloscope Buying questions

[nctnico]

I did useful things with my Black&decker cordless drill too but life got so much better with the Makita cordless drill! 

Maybe so, but you were comparing owning the Black&decker with owning no drill at all (or at best, only owning a hand drill).

It says Makita and Black&decker. Where does it say no drill??

[Fungus]

I did useful things with my Black&decker cordless drill too but life got so much better with the Makita cordless drill! 

Maybe so, but you were comparing owning the Black&decker with owning no drill at all (or at best, only owning a hand drill).

It says Makita and Black&decker. Where does it say no drill??

Does your $400 GW-Instek have serial decoders?

[rf-loop]

Serial decoders which only decode what is on screen are utterly useless so they don't even count.

What equipment can decode data what is not captured.

Example in Siglent. ALL captured data lenght is visible on the screen and it  decode whole lenght of captured data.

What data this your dream machine decode if not captured data. Example, Siglent can not decode at all anything what is out from screen. Because there is nothing. What scope can, I have not seen any.

I think you have used too long time oscilloscopes what have only very narrow window in the screen to captured data and lot of captured data is unvisble out from screen. Why need keep captured data unvisible out from screen and then tell that it can decode also this part of data and then tell machine is good because it can decode also out from screen. Perhaps this is some special kind of humor but I do not know where need laugh.

May I ask what clever is there if keep big part of captured data outside from screen and then tell that scope can decode also out from screen. Oh well it is nice machine - is it.

Can you understand: Independent of captured lenght,  screen width = capture lenght.  Zoom is for details. So simply.

[Howardlong]

Serial decoders which only decode what is on screen are utterly useless so they don't even count.

While it is indeed a significant restriction, it's definitely better than no decode at all. Software based decoders are generally slow, so I can see why they did it that way. On the MDO3000 you can be waiting a _long_ time for a decode. The problem is if you've been used to hardware decode on Keysights, everything else seems a let down. The trick on the Rigols is to set up a serial trigger which unlike some decodes (on I2C) appears 100% reliable. To be fair, on any scope, including the Keysights, setting up a serial trigger is probably the way to go anyway for a reasonably long serial stream.

[nctnico]

Serial decoders which only decode what is on screen are utterly useless so they don't even count.

While it is indeed a significant restriction, it's definitely better than no decode at all. Software based decoders are generally slow, so I can see why they did it that way. On the MDO3000 you can be waiting a _long_ time for a decode. The problem is if you've been used to hardware decode on Keysights, everything else seems a let down.

I'm starting to sound like a broken record but the decoding on the GD2204E is quick due to the hefty amount of CPU horsepower under the hood. Only at 10Mpts it needs a moment to catch up.

@Rf-loop: ofcourse you use zoom to look at more details (capture a whole bunch of messages in single shot mode and then zoom in) but if the decode gets confused because it only decodes part of the memory which is on the screen it is of no use because you can't zoom in to look at problems at the bit-level and for long messages you'll lose track of where you are in a message IF the oscilloscope can even determine the start of a frame (a given number of bits; usually a byte) within a message which is unlikely when looking at SPI or I2C busses. So yes it is important an oscilloscope with decoding can decode the entire memory and show partial messages/bits if necessary but also keep track of where you are in a message.

[rf-loop]

@Rf-loop: ofcourse you use zoom to look at more details (capture a whole bunch of messages in single shot mode and then zoom in) but if the decode gets confused because it only decodes part of the memory which is on the screen

Also single shot, Siglent decode whole capture lenght.
On the screen is whole capture lenght and whole capture lencht is decoded. Zoom window can use where ever in this whole lenght for details.

[Fungus]

I'm starting to sound like a broken record but the decoding on the GD2204E is quick due to the hefty amount of CPU horsepower under the hood.

Well of course it is on your super-duper, three-times-the-price GD2204E

But... you were talking about the GDS-1054B before you started weaseling.

The GDS-1054B loses to the DS1054Z in a shootout because it has no serial decode at all.

[nctnico]

I'm starting to sound like a broken record but the decoding on the GD2204E is quick due to the hefty amount of CPU horsepower under the hood.

Well of course it is on your super-duper, three-times-the-price GD2204E

But... you were talking about the GDS-1054B before you started weaseling.

The GDS-1054B loses to the DS1054Z in a shootout because it has no serial decode at all.

Without hacking and the serial decode options the DS1054Z has no decoding either. You where the one who wrote you'd buy the DS1054Z even if it where not hackable! GW Instek just follows a different licensing scheme. They basically have 2 model lines where one has decoding and the other one doesn't. It's simple as that.

[Fungus]

Without hacking and the serial decode options the DS1054Z has no decoding either. You where the one who wrote you'd buy the DS1054Z even if it where not hackable!

Yes I did, and yes I would. I think it's worth $400 of anybody's money.

On the other hand the Rigol is hackable and everybody knows it.

What would I choose in a shootout? A $400 scope with no frills or a $400 scope with twice the bandwidth plus serial decoders, extra memory, etc.? It seems a no-brainer. The GW-Instek loses even if the Rigol does have a couple of hard-to-find bugs.

[Howardlong]

Serial decoders which only decode what is on screen are utterly useless so they don't even count.

While it is indeed a significant restriction, it's definitely better than no decode at all. Software based decoders are generally slow, so I can see why they did it that way. On the MDO3000 you can be waiting a _long_ time for a decode. The problem is if you've been used to hardware decode on Keysights, everything else seems a let down.

I'm starting to sound like a broken record but the decoding on the GD2204E is quick due to the hefty amount of CPU horsepower under the hood. Only at 10Mpts it needs a moment to catch up.

A video or it didn't happen ;-)

[alsetalokin4017]

Without hacking and the serial decode options the DS1054Z has no decoding either. You where the one who wrote you'd buy the DS1054Z even if it where not hackable!

Yes I did, and yes I would. I think it's worth $400 of anybody's money.

On the other hand the Rigol is hackable and everybody knows it.

What would I choose in a shootout? A $400 scope with no frills or a $400 scope with twice the bandwidth plus serial decoders, extra memory, etc.? It seems a no-brainer. The GW-Instek loses even if the Rigol does have a couple of hard-to-find bugs.

Buy a Rigol DS1054z yourself and use it for something other than displaying pretty colored lines from a function generator. Then you can see for yourself how "hard to find" the "couple" of bugs are.
FYI: in English the word "couple" generally means "around 2". For you constantly to represent that the DS1054z has a "couple" of bugs that interfere with practical use of the scope is misleading, false and untrue. It's easy to find 5 that can even be displayed at the same time with a perfectly ordinary setup. A set of Measurements that fail without the user noticing it, requiring reboot to restart any measurements at all is a pretty darn severe bug in my opinion -- of course you may disagree but I think the first time that such a fault bites you, even you might change your mind. Trying to find the maximum slew rate of a signal and discovering that the "diff" math function is useless and returns garbage on-screen may leave you feeling like you've been... er.... shortchanged a little bit, to say the least.

Sure, the scope is a good value for the money, but it has some (many more than a "couple")  _severe_ problems that Rigol really should fix, and to brush these problems under the carpet and ignore them is doing future purchasers and current users a disservice.

[Fungus]

Sure, the scope is a good value for the money, but it has some (many more than a "couple")  _severe_ problems that Rigol really should fix, and to brush these problems under the carpet and ignore them is doing future purchasers and current users a disservice.

Not as much of a disservice as telling them it's a useless piece of junk that no sane person should consider buying.

...how "hard to find" the "couple" of bugs are. ... Trying to find the maximum slew rate of a signal and discovering that the "diff" math function is useless and returns garbage on-screen

Because ... we all spend our lives trying to find the maximum slew rate of signals using the diff function, right? 

it has some problems that Rigol really should fix

I agree that they should be releasing more fixes. The next update is well into the 'overdue' category now.

They have a good product, it must surely be earning enough to pay for its own support team. All the animosity in these forums could be silenced by being a bit more active with the fixes.

[JPortici]

features i don't use in my 1054z
- math: useless. seems like you're using 4 bit arithmetics, not even 8 bit. differentiation only gives back garbage
- x/y: scope becomes extremely slow. much much much much slower than enabling the fft and i can't conceive a reason for it. much slower than an older 1052 or a tek tps (and that one is REALLY slow)
- serial triggering. just doesn't work. actually none of the other trigger modes than edge that i've tested seem to work
at least i didn't have to waste money for those
and in general it's hard for me to get a stable trigger for something that isn't the built in calibrator where other scopes i use/have works just fine

they'd have a good product if things worked. it can be a good product, it's OK for now

[wraper]

Example of Rigol design and build "quality". RP-3300A probe at the BNC connector side.  Central conductor living by it's own in the hole and shorting to the metal. Already repaired both of those I received with DS-2072 by bending the wire. But they failed again, guess will need to fill the hole with silicone to prevent them shorting again. Also you can see quality construction with 2 resistors soldered on top of each other, one of them upside down. Hook attachments never worked properly too. When you press it, it's very scratchy, often gets stuck and the gook on the end is no good too, very hard to attach to the measuring point. Piece of crap in one word.

[Fungus]

none of the other trigger modes than edge that i've tested seem to work at least i didn't have to waste money for those and in general it's hard for me to get a stable trigger for something that isn't the built in calibrator where other scopes i use/have works just fine

I never had a problem with triggering, and if it was as bad as you say then the anti-DS1054Z crowd would be all over it

(obscure bugs in the slew rate calculation are nothing compared to triggers that don't work!)

The triggering is actually done in hardware by special circuitry. Maybe your 'scope has a hardware fault.

[nctnico]

He says none of the other trigger modes than edge works. So make a video of how your DS1054Z triggers on a specified character send by a UART.

[rstofer]

Boy, you guys had me going!

So, I set up an Arduino to send the character string "ABCD" followed by a 2 mS delay (just because).  I set up RS232 decode and, voila', I can trigger on any of the 4 characters.  It took me a moment to realize that the character was to the left of the trigger mark.  But of course!  You can't trigger on something until you have seen it all.

I also tried 115200 baud and no delay - it works even better.  And I put a very long string after the "ABCD" thing and it still works!

Code: [Select]

void setup() {
  // put your setup code here, to run once:
  Serial.begin(115200);
}

void loop() {
  // put your main code here, to run repeatedly:
  Serial.print("ABCD now is the time for all good men to come to the aid of their country");
  ;delay(2);
}


No, I'm not going to try to photograph the screen much less try to download an image.  Try the experiment yourself.  It works!

I have not tried I2C or SPI but, if I get bored later today, I might.

[nctnico]

Now make it trigger on A or C 

[Howardlong]

- serial triggering. just doesn't work. actually none of the other trigger modes than edge that i've tested seem to work
....
and in general it's hard for me to get a stable trigger for something that isn't the built in calibrator where other scopes i use/have works just fine

I can't say I've ever encountered either of these problems. Can you demonstrate how to reproduce them?

[rstofer]

Now make it trigger on A or C 

As I said, I can make it trigger on ANY of the 4 chars.  All I have to do is turn the little knob to set the trigger value.

[Fungus]

Now make it trigger on A or C 

As I said, I can make it trigger on ANY of the 4 chars.  All I have to do is turn the little knob to set the trigger value.

I think the weird little face means he wants a boolean "OR" function, ie. 'A' OR 'C'

[JPortici]

Not really, the scope just doesn't trigger in any mode that isn't edge.
(Also i was playing around with a wall wart transformer that outputted a 18Vpk not so clean ac waveform on an old analog scope i just got. i measured it with the 1054 and at 50V/division it wasn't able to trigger on zero.. found out that by chance and it's ofncourse a very singular case that should not be taken into account in this scope's bashing. Anyway, depending on the timebase i adjusted the finetune scale until it triggered and it also changed between 40 and 35 v/div. It's a borderlone case because there is a different input circuit at that scale... and being sincronous with AC might also explain why the trigger was so difficul but that made me wonder if i actually have a faulty scope.. because a 30 years old analog scope will trigger with no problems at all)
I wonder if i indeed have a faulty scope. I'll make further tests and eventually contact rigol

[rstofer]

Now make it trigger on A or C 

As I said, I can make it trigger on ANY of the 4 chars.  All I have to do is turn the little knob to set the trigger value.

I think the weird little face means he wants a boolean "OR" function, ie. 'A' OR 'C'

Ah...  Skipped right over my head.  The 'Evil' emoticon means BOOLEAN - good to know.  Upping my skills, one day at a time!

I'm not sure how to trigger on something with "don't care" bits.  I'm not even sure I could do it with my logic analyzer.  Of course, I couldn't trigger on the square root of 2 to 17 digits either.  Yup!  The scope must be defective.

The challenge was to trigger on an RS232 char and that absolutely works.  Nobody said anything about a square root of 2 challenge.  The original statement was that NONE of the triggers, other than edge, worked.  I don't think that is true, I have proven to my satisfaction that it isn't true yet it will forever be written as true.

Unfortunate...

This stuff just confuses the new guys!  Buy the scope?  Don't buy the scope?  Does it work?  Does it not work?  Who says so?

[nctnico]

There is always a day some functionality comes in handy. Did I mention the GW Instek serial protocol decoding can trigger on don't care bits using up to 10 bytes so you can filter on very specific messages?

[rstofer]

Not really, the scope just doesn't trigger in any mode that isn't edge.
(Also i was playing around with a wall wart transformer that outputted a 18Vpk not so clean ac waveform on an old analog scope i just got. i measured it with the 1054 and at 50V/division it wasn't able to trigger on zero.. found out that by chance and it's ofncourse a very singular case that should not be taken into account in this scope's bashing. Anyway, depending on the timebase i adjusted the finetune scale until it triggered and it also changed between 40 and 35 v/div. It's a borderlone case because there is a different input circuit at that scale... and being sincronous with AC might also explain why the trigger was so difficul but that made me wonder if i actually have a faulty scope)
I wonder if i indeed have a faulty scope. I'll make further tests and eventually contact rigol

You should limit the scope of the statement highlighted above.  Your scope may not trigger, you may not be able to get it to trigger, but that doesn't mean the entire set of 1054 scopes don't trigger.  From my admittedly limited testing, my scope seems to work just fine.