Scope with Data Logging for monitoring Automation Control signals

[Lucas86]

This is my first post here, so greetings everybody.

I have a degree in Electrical Engineering / Radioelectronics but currently am working as an Automation and Control / Controls Systems Engineer.

I have been looking for a device (preferably a scope) that I could hook up to automation control circuits and log data for at least several hours but preferably couple of days. It can also be a device that I would be reading with some sort of a script running on a connected to it PC.

Now, the signals I work with shouldn't have high frequency components. Of course there will be some transients, but I think sampling with few hundred Hertz should be sufficient. Although 1 kHz or higher sampling rate could be useful in some cases.

before starting this conversation I used the search engine and found the following topic:
https://www.eevblog.com/forum/testgear/scope-with-data-logging-questions/msg1456621/#msg1456621

Even though this topic above is fairly similar, when I wanted to replay to it, I was suggested to start a new thread, so here it is.

Now, my budget for this scope is on the low end (around $1000) and I would like to have digital inputs also.

At first I was looking closely at the Rigol MSO 5000 series (however I am not certain I could read it with a script) but after reading the above thread I am not sure should I go look into Sigilent or GW Instek scopes instead of Rigol?

Thanks for all the help.

[Rerouter]

Yeah, still need more info here, What voltage ranges, both analog and digital, and how many of each type of channel

my first thought is something like a logic8 or logic16 by saleae, as the modern ones have analog inputs, though they are voltage constrained.

[dzseki]

Just to be clear, logging one channel at 8 bit at 1ks/s for one full day takes up 86Mbyte data, what are you going to do with that? Perhaps you are looking for some odd transient response or jitter? then there are more clever ways to do that other than brute force.

[nctnico]

Oscilloscopes from GW Instek (the modern ones are the GDS2000-E, MDO2000 and MSO2000 series) and Yokogawa have data logging features. GW Instek can dump data directly onto a shared network folder on a PC. For Yokogawa I'm not sure what the most modern oscilloscopes can do. The old ones could send data onto an FTP server but I assume the newer ones have more choices.

I'd try to avoid doing too much scripting and get an oscilloscope which does logging out of the box. If you do want to go for scripting then avoid needing to install the NI package. It is bloated and complicated to install. There are many NI versions and the latest don't support older operating systems while the instrument may need a more recent version then your OS supports. I never got the NI package to work with any instrument. You better use a lightweight LXI library instead which talks to an oscilloscope directly over the network or USBTMC.

[Lucas86]

Well, in Europe Controls and Automation systems use mostly 4(0) - 20 mA analog signals or sometimes 0 - 10 V. Digital signals are 24 V or 230 V. I understand that with a scope I would have to measure the voltage drop on some kind of resistance for the current analogs.

When it comes to the inputs, like a typical electric actuator for a valve could have like 6-10 digital signals associated with it's control circuitry. Other devices can have even little more than that but I don't need to monitor all of them probably. 4 digital inputs would be essential, 8 should be enough and 16 is probably a small overkill. When it comes to analogs, most automation and control devices use 1 or 2 (set point and measured process value). Maybe in like two stage controllers or something you might need 3rd analog channel. You also might want to measure the "error" signal. However, I think 4 analog channels should be sufficient for most purposes.

Well for sure capturing transients will be a thing of interest and I understand you can set a scope to sense those changes and capture it. I don't think I need a scope that could capture like an entire day in it's own memory but it would be nice to be able to read it from a PC with a script if necessary to log an extended period of time. I understand that I won't be able to get high sampling rate when reading the scope from a PC.

[Rerouter]

your requirements are sounding more and more like the job of a PLC that has some datalogging mode, A scope for current measurements would likely need differential measurements, and there digital inputs generally only good to +-20V

[Lucas86]

Well, I need to monitor a PLC. PLCs talk to SCADA but not all signals are transparent in the SCADA system. Also SCADA reads the PLC periodically, so you not necessarily can relay on the time relations between different signals... Of course I could make my own data-logger on a PLC, a micro-controller or even search for some kind of evaluation board but I really think there should ba a Test Equipment on the market that would have such a feature for a reasonable price.

[nctnico]

I think every decent PLC manufacturer has a tool to log data from their PLCs into a PC realtime. I think there isn't much use for an external device. Where an oscilloscope does come in handy is to look for glitches on a signal which may fool a PLC but you'll need to use analog channels for that otherwise you'll be just as blind as the PLC.

[Lucas86]

You mean like using the TIA Portal to log the data? Well that could be definitely useful but then you also have converter, relays etc. Not everything that a PLC gets is generated directly by the object device in question... I think ultimately measuring directly is best you can do.

[nctnico]

To be honest it is not quite clear to me what you try to achieve. Do you have a specific problem you are working on or are you trying to get as much useful tools in your toolbox so you have these available?

[Lucas86]

I work in a heat and power plant. A lot of the control systems have almost 20 years. Problems occour quite frequently and it is to expected to be even more frequently in the future, before we modernize our heat sources. At this time there isn't any particular problem that I am trying to solve with this tool but I got like a $1000-$1500 put into 2019 budget for a device that would help me solve problems. For example, I have a pump that is driven from a PLC (S7-400) analog output via ABB inverter. This circuit used to work ok, but recently the inverter would go to a maximum frequency right after it was turned on. This was happening regardless of the analog signal going out of the PLC being like 6-8 mA. I am still not sure what is the cause of that but putting a 4-20mA / 4-20mA separator solved it for the time being. However, this is merely a bandage not eliminating the cause of the problem. Of course to my knowledge nothing was changed intentionally. So now I am not sure is the PLC current source in the AQ loosing efficiency and is not able to drive via this length of the line or maybe the inverter input is draining too much current? On the other hand, some new interfering signal could occur from other circuitry that was changed or added.

In essence, I got some money in budget and am trying to make best use of it to expand my toolbox.

[nctnico]

I know that there are devices specifically for testing 4-20mA interfaces (both sink & source). These aren't very expensive and could be useful to at least see the PLC still drives the right signal levels and a device accepts 4-20mA. Also keep in mind that 90% of the problems are wiring issues. You could buy an isolation and continuity tester as well. The isolation tester will find leaks to ground for example.

[Rerouter]

To clarify, you didnt need to necesarily use the diagnostic mode of the existing PLC, just one your compforatble you can get the data from, as its already got inputs that can handle the ranges you would be hooking it up like any data logger,

But if that is off the table, dirt cheap digital logging would be a saleae clone with another board to divide down the voltage and possibly rectify a little to keep logic states for AC signals. Cheap as dirt, but a little bodgy,

For playing with inverters, that would likely be a 2 channel scope minimum, with properly rated scope probes for the voltage, not that expensive, and if you leverage the triggering to what kinds of faults your looking for, then most options can store hundreds or thousands of fault waveforms, 4 channels is always nicer, but it comes down to how the money stretches,

If you want it wrapped up all into one, I know yokogowa used to make essentially multichannel printer modules, but they started at about $3500 last I looked for a 10Hz update rate. but you may be able to find the devices if you hunt on the big manufacturers websites a few minutes,

[Lucas86]

I have a Beamex calibrator that can source and sink 4-20 mA. We also have isolation tester as that is required by law to measure from time to time. The pump case, I used the Beamex calibrator and everything was ok. It could drive the ABB inverter over the line instead the PLC. The PLC also could drive an amp meter over the line. I also have a DC milliamp clamp meter (KYORITSU KEW2500) and it would read fine but the ABB inverter would still go to maximum frequency. Then I put the 4..20 mA / 4..20 mA separator and everything works as intended. Maybe I should check the isolation. There could be a leak somewhere, even in the inverter itself. The overarching problem in all of this is also that I can't keep testing everything over and over, because a lot of the time the plant needs stuff working asap. I could really use something that I can connect to a circuitry and would log signals.

Yes, I have done a monitoring via not used inputs on the PLCs but than it requires modification in the PLC program and in a power plant you really want to keep that to a minimum. Of course I could program some inputs and have them ready in case I need them but there is still problem of the SCADA system not really reading the PLCs memory in real time. I am not 100% sure but it can take fraction of seconds or even single seconds between reading different PLCs. I am not sure where would I log data into elsewhere than the SCADA servers. In theory I could get a cheap PLC (maybe I even have one) like the Siemens S7-1200 and dedicate a PC to read only that PLC. I am sure this would be somewhat optimal but this is kind of an ad-hock solution. I would prefer something more compact and sort of out of the box, that would look little more professionally (not saying using a PLC is lame).

And yes, Yokogawa has a device that does what I need but it is well over my current budget. Maybe, if I could prove my point (for example with a scope) that a device like that is needed I could get a fully dedicated device such as the https://tmi.yokogawa.com/us/solutions/products/data-acquisition-equipment/high-speed-data-acquisition/dl350-scopecorder/ in the future.

[L_Euler]

Have you considered a LabVIEW setup.  LabVIEW (at least in its original form) is specifically a data collection platform that can store date directly to PC files.  LabVIEW also has numerous interface options to look at signals directly or interrogate PLC data files.

[Lucas86]

I think a LabView setup would be expensive. I am talking $1000-$1500 scope here...

[_Wim_]

One of the best tools you can have for service is a software package like plc analyzer (https://www.autem.de/products/plc-analyzer-pro-6/) or simular (remark: plc analyzer from Autem is not the most stable, but still very usefull). This allows you to trend internal tags and IO's quite fast. Best is to record many tags during all operation modes when the equipment is working correctly, so you have a reference how it is supposed to be when it breaks down. Log as much as possible, but minimum all IO's (both analog and digital), and all phase or mode information (be sure to perform this for ALL modes, and pay special attention to regulating loops, as these are the most difficult to fix when they become unstable)
Typically the most difficult task when an older machine process breaks is figuring out how it is supposed to be.
If you want to log an industrial network like profibus or profinet, buy a dedicated analyser for this network. Scopes can be useful to check bus health, but to analyse messages, they are pretty useless.

[sotos]

I think you are looking for something like this.

http://www.graphtec.co.jp/en/instruments/gl240/index.html

http://www.graphtec.co.jp/en/instruments/gl980/index.html

https://www.datatec.de/labor-messgeraete/datenlogger-multiplexer/datenlogger/gl240-eu-kit-10-kanal-datenlogger-kit-mit-koffer-akku

https://www.althensensors.com/instrumentation/graphtec-data-acquisition-systems/compact-data-loggers/5756/gl240-compact-data-logger/

[threephase]

Yes the Graphtec are good little instruments, Hioki also manufacture similar items and also individual modules for 4 to 20mA, 0 to 10V and RTD signals.

https://www.gmciuk.com/multi-channel-data-loggers

There are also items such as CMC-141 from Omni Instruments that are digital versions of the old paper chart recorders.

https://www.omniinstruments.co.uk/display-control-products-signal-amplifiers-psus/large-digit-and-multi-channel-displays/multicon-cmc-141.html?gclid=Cj0KCQiA-8PjBRCWARIsADc18TJsXlRaN_l7jrNPNojbp6x6baQkFlxjmuh7rjATZjOfyNfyRssfeUwaAhsjEALw_wcB

[Lucas86]

I have decided on getting a scope. It is pretty much down to GwInstek GDS-2074A or Rigol MSO5074, any suggestions? I have looked into the Users Manuals, but am not 100% sure can I write continuously live waveforms to a USB drive or a LAN folder or FTP. It seems that the MSO5000 can write to a USB, but I am not sure is that a waveform that's already stored in the scope's internal memory or can it dump the waveform contentiously until there is no free space on the USB drive.