Advice on buying used oscilloscope...

[plore90]

Hello everyone,
I am new to this very interesting forum and I have read some posts regarding oscilloscopes. I state that I do not have experience in oscilloscopes, having worked over the years only in the IT field (software and systems). Now that I am retired and I devote myself for passion to projects with arduino and small digital systems I would like to have an oscilloscope that could help me in my projects and that would also allow me to analyze the various serial protocols (I2C, RS232, USB etc. ). I state that I already have some used HP measuring and power-source instruments which is my favorite brand.
My budget is not high and I was focusing on the used model Agilent 54622D 100 MHz, 2 + 16 Channel. I would like to have some comments and suggestions from you on which model to use.
Thank you and greetings

[james_s]

If you want to analyze serial protocols that eliminates most of the older scopes, you'd probably be better off with one of the Rigol, Siglent, etc low cost 4 channel DSOs.

Then again, for decoding serial protocols a logic analyzer is really a better tool so you could use just about any scope. I've never used a HP/Agilent DSO but I've heard good things about them and it does look like the 54622D can be had for ~$200 or less which is pretty cheap for a usable scope.

[plore90]

Thanks for your answer James.
I will follow your advice and also will evaluate Rigol, Siglent, etc low cost 4 channel DSOs

[TK]

The 54622D can trigger on serial data, but not decode

[Fungus]

Then again, for decoding serial protocols a logic analyzer is really a better tool so you could use just about any scope.

That depends on what you're doing. An oscilloscope will show you the shape of the pulses and let you verify that all is well, electrically speaking. If it can do a bit of decoding you can also verify that the right data is being transmitted.

For analyzing long sequences of data though, an oscilloscope will fall short. It doesn't have the right user interface (eg. you can't type CTRL-F and search for a sequence of bytes) and memory will usually be too small compared to the gigabytes available in a PC.

Now that I am retired and I devote myself for passion to projects with arduino and small digital systems I would like to have an oscilloscope that could help me in my projects and that would also allow me to analyze the various serial protocols (I2C, RS232, USB etc. ).

My budget is not high and I was focusing on the used model Agilent 54622D 100 MHz, 2 + 16 Channel. I would like to have some comments and suggestions from you on which model to use.

I'm going to suggest looking at the Analog Discovery II instead. It will sit on the table alongside your Arduino and do all sorts of useful things.

https://analogdiscovery.com/

[plore90]

Thanks Fungus
Thanks Th
I will do

[Elasia]

+1 for the AD tool, its a neat gadget to have around that's cheap to get started with

When you want an upgrade look at either siglent or rigol if new and want more logic/serial decoding, the other brands cost big bucks

If no decode there are tons of used scopes that work just fine, usually something up on the buy/sell forum here and better odds you are getting good kit vs random whatever on ebay

[Gandalf_Sr]

The Rigol DS1054z is a good, cheap, 4 channel scope that (I think) now comes with all the decode options for free.  It can easily be hacked up to 100 MHz and to add all the options (Google "Riglol" - not a typo). In the US, you can get one for $325.

This really is a great starter scope, in the US it currently has a 5 year warranty as a bonus offer from Rigol.

[stafil]

DS1074Z Plus is a hackable 4 analog channel scope that gives you an option for digital as well.

Pair it with a DG1022Z hackable weve generator, add a cheap Linear power supply and a cheap electronic load and you have a complete lab for less than $1000

[james_s]

The Rigol scopes are great, but they are $350 which is significantly more than $200. If you have the budget then it's a great deal, if you don't then the older DSOs are still very capable instruments. Any DSO can trigger on and capture data from the various serial protocols, so you can use a scope to look at the data integrity type issues and then use a logic analyzer or something like a Bus Pirate to decode the actual data. It's a feasible option if that's what the budget allows.

I have several DSOs and none of mine can do serial decoding, in practice it has not been a big limitation as I also have a logic analyzer and a Bus Pirate.

[georges80]

At least for the rigol 1054, the decode is software and fairly slow. The scope needs to'see' enough of the waveform on the screen to decode it, so zooming/panning can break the decode. I have access to both the 1054 and an Agilent msox2024A (hardware decode) and the Agilent decode capability is miles ahead. Of course the cost is also miles ahead

Nothing wrong with the 1054, just expect some decode frustration and learning what to do to get it to decode what you want on the screen.

cheers,
george.

[james_s]

It's a low cost scope and has limitations for sure, but it does offer a lot of bang for the buck. In the absence of infinite budget you're going to have compromises.

[Fungus]

At least for the rigol 1054, the decode is software and fairly slow. The scope needs to'see' enough of the waveform on the screen to decode it, so zooming/panning can break the decode.

In practice I never use mine for massive decodes. I mostly use it to look at the shape of the signals, if the baud rate is correct, if the shape of the signal is good (eg. am I using the right pullup resistors on my I2C bus?), etc. The other thing I've done a few times is use my 'scope to figure out unknown RS232 baud rates.

The decoding of a few bytes of known data is a useful crosscheck when you're doing all that so that's where the serial decode option comes into play. Once the data is flowing correctly it's the wrong tool for the job IMHO and I inevitably switch to something that knows a bit about the data structure and can break it it up and format it for easy human reading (ie. I can sniff a bus with my Arduino, split it into packets and echo the data to the serial console on my PC screen).

Is the rudimentary decoding on my Rigol DS1054Z holding me back? Would I decode more data if my 'scope was better at it? I suspect not.

[2N3055]

At least for the rigol 1054, the decode is software and fairly slow. The scope needs to'see' enough of the waveform on the screen to decode it, so zooming/panning can break the decode.

In practice I never use mine for massive decodes. I mostly use it to look at the shape of the signals, if the baud rate is correct, if the shape of the signal is good (eg. am I using the right pullup resistors on my I2C bus?), etc. The other thing I've done a few times is use my 'scope to figure out unknown RS232 baud rates.

The decoding of a few bytes of known data is a useful crosscheck when you're doing all that so that's where the serial decode option comes into play. Once the data is flowing correctly it's the wrong tool for the job IMHO and I inevitably switch to something that knows a bit about the data structure and can break it it up and format it for easy human reading (ie. I can sniff a bus with my Arduino, split it into packets and echo the data to the serial console on my PC screen).

Is the rudimentary decoding on my Rigol DS1054Z holding me back? Would I decode more data if my 'scope was better at it? I suspect not.

This. 100%.

[Gandalf_Sr]

At least for the rigol 1054, the decode is software and fairly slow. The scope needs to'see' enough of the waveform on the screen to decode it, so zooming/panning can break the decode.

In practice I never use mine for massive decodes. I mostly use it to look at the shape of the signals, if the baud rate is correct, if the shape of the signal is good (eg. am I using the right pullup resistors on my I2C bus?), etc. The other thing I've done a few times is use my 'scope to figure out unknown RS232 baud rates.

The decoding of a few bytes of known data is a useful crosscheck when you're doing all that so that's where the serial decode option comes into play. Once the data is flowing correctly it's the wrong tool for the job IMHO and I inevitably switch to something that knows a bit about the data structure and can break it it up and format it for easy human reading (ie. I can sniff a bus with my Arduino, split it into packets and echo the data to the serial console on my PC screen).

Is the rudimentary decoding on my Rigol DS1054Z holding me back? Would I decode more data if my 'scope was better at it? I suspect not.

This. 100%.

This plus another 100%

[nctnico]

The decoding of a few bytes of known data is a useful crosscheck when you're doing all that so that's where the serial decode option comes into play. Once the data is flowing correctly it's the wrong tool for the job IMHO and I inevitably switch to something that knows a bit about the data structure and can break it it up and format it for easy human reading (ie. I can sniff a bus with my Arduino, split it into packets and echo the data to the serial console on my PC screen).

That approach doesn't work for cases where decoding lots of data using an oscilloscope is very useful: intermittant bugs which can have their origin in the analog domain. A logic analyser is not able to show the analog domain so you can only see when something went wrong but not how and why.

[james_s]

That approach doesn't work for cases where decoding lots of data using an oscilloscope is very useful: intermittant bugs which can have their origin in the analog domain. A logic analyser is not able to show the analog domain so you can only see when something went wrong but not how and why.

I can't say that I've ever found myself in that situation. Surely you could set up a trigger to capture a glitch? You don't really need to decode a long stream of data in order to see a glitch that occurs in the analog domain, all you need is the area around the glitch.

In my own experience though this sort of stuff either works or it doesn't. The times I've had issues with digital interfaces it has been obvious signal integrity issues, incorrect levels, excessive reflections, crosstalk, incorrect speed, that sort of thing. If it's right on the edge it may result in intermittent failures but in that case even when it works it's not going to look right.

[Fungus]

The decoding of a few bytes of known data is a useful crosscheck when you're doing all that so that's where the serial decode option comes into play. Once the data is flowing correctly it's the wrong tool for the job IMHO and I inevitably switch to something that knows a bit about the data structure and can break it it up and format it for easy human reading (ie. I can sniff a bus with my Arduino, split it into packets and echo the data to the serial console on my PC screen).

That approach doesn't work for cases where decoding lots of data using an oscilloscope is very useful: intermittant bugs which can have their origin in the analog domain. A logic analyser is not able to show the analog domain so you can only see when something went wrong but not how and why.

That's true.... but it's a rare case and I'm not sure that better serial decoding would be much use in finding the problem either, not unless you can set it up to trigger on "data corruption".

Better serial decoding makes it more likely you'll sit there pressing "single" over and over then wading through thousands of bytes of data looking for the needle in the haystack. Which could take days.

Me? I'd try to automate the process, eg.:
1) Attach an Arduino to sniff the bus to watch for data corruption (eg. Checksum error, parity/framing error, ...) and generate a trigger signal for the oscilloscope if it sees any.
2) Send the same packet of data over and over with the oscilloscope in pass/fail mode then go for lunch while it works.

nb. Neither of those relies on the serial decoding abilities of the 'scope.

[Fungus]

+1 for the AD tool, its a neat gadget to have around that's cheap to get started with

Yep, I'm sorely tempted to get one myself after revising the web site.

For the work I've been doing lately it looks like it would be much more useful than an oscilloscope.

[2N3055]

Well, Nico has a valid point. He has a lot of experience with this.
Once you setup and verify basic signal integrity (analog) and verify and debug messages (decoding), if you have a bit more complicated system you need to run it "for real" to verify there are no problems when it runs real world code and load.
CAN bus is a good example, if you have problems with nodes trying to access bus with even a minimum overlap, things are going to get funky, and in decode you will see funny things, but won't be able to figure out which one is culprit without analog domain look into it...Problem is that it might be weird interaction, node responding to message you get once an hour, or only if some weird set of circumstances are valid..

But that is when working with pro stuff.. If you're doing arduino and two sensors, it much easier...

[tggzzz]

Hello everyone,
I am new to this very interesting forum and I have read some posts regarding oscilloscopes. I state that I do not have experience in oscilloscopes, having worked over the years only in the IT field (software and systems). Now that I am retired and I devote myself for passion to projects with arduino and small digital systems I would like to have an oscilloscope that could help me in my projects and that would also allow me to analyze the various serial protocols (I2C, RS232, USB etc. ). I state that I already have some used HP measuring and power-source instruments which is my favorite brand.
My budget is not high and I was focusing on the used model Agilent 54622D 100 MHz, 2 + 16 Channel. I would like to have some comments and suggestions from you on which model to use.
Thank you and greetings

Welcome.

I strongly advocate the approach noted above: use a scope to look at analogue waveforms, but use a logic analyser or protocol analyser or printf statement for digital signals.

The AD is a good starter tool with significant benefits, but check the input voltages are suitable for your circuit.

Make sure you understand how to safely connect a scope or LA to a circuit. Make sure you understand the different types of probe and when to use/avoid them.

[jxjbsd]

I recommend seleae USB logic analyzer. It works well with computer.His analysis software is very easy to use.

[David Hess]

The Agilent 54622D has peak detection so I have no objections to it but as others pointed out, it lacks serial decoding capability.

[plore90]

Thanks a all!!

[mindcrime]

For decoding serial protocols you could consider picking up one of those $10.00 chinese logic analyzers[1] from Ebay / Aliexpress, and use Sigrok[2] (as long as the bus frequency isn't too high). Note that Sigrok can also work with some oscilloscopes as the data collection component, but you're obviously limited to a fairly small number of channels. But for serial, if you have a 4 channel scope, that's probably going to be enough 99.9% of the time.

[1]: https://www.ebay.com/sch/i.html?_nkw=24mhz+logic+analyzer

[2]: https://sigrok.org/