Entry Level Ocilloscope Rigol DS1052E will it do my job

[linc1]

Hi all

Whilst new to the site and not a electronics technician i am an electrician with post trade qualifications, i also play with Arduino boards when i have time

My questions are basically 2  1, is this instrument the right unit for me, the main reason i need such an instrument is not for my hobby electronics but more towards my working needs

What i really want the instrument for is to check governor responce of diesel and gas generator engines
Measuring option 1, i have been toying with was to directly measure the main output at 240 Volts RMS and the frequency of 50 HZ ( representing engine speed) the issue with this is i need to get a single line representation of the frequency change over time showing ringing and recovery to steady state after load changes etc  (Nominal rpm is 1500 and 30 RPM = 1 Hz)

Measuring option 2, Most of these engines ( but some dont) will also have a magnetic pick up measuring flywheel speed the output of the mag pick up can vary due to tooth shape proximity of the pick up to the flywheel tooth etc but generally you will see an output voltage of 20-30 Vac at approx 90 KHz square wave

Result i wish to see is a  capture the variances over 5 to 10 second period at load changes, now i have tried with a Uni-T UT81 instrument to do this by taking the 90 KHz square wave rectifying it to DC but the waveform is not a true representation and is not smooth

If i take the 50hz signal and bunch it up by changing the time base, this sort of works if i look at the top or bottom edge of the waveform

Question 2 , Am i attempting to do this correctly or should i look at freqency transducer to obtain this signal ( PS i dont want the cost of obtain such a transducer which is most likly cost prohibitive and may not be acurate)

Anyones comments would be appreciated & Dave thanks for your Tube reviews

regards Linc1

[amspire]

Is the 90KHz squarewave just a pulse stream with a frequency proportional to the motor speed, or something else?

Anyway, the 240V can be used.  A Rigol could be used if you were capturing the 50hz period to a file, but if you have a frequency counter that can capture data to a computer, you will get more accurate results. I am not sure how regularly you can log period from a Rigol - someone else may know. If it is not several time a second, it may not be ad0equate.

A good counter lets you get the period over several cycles for more accuracy (if that is what you need), but the Rigol will be better then a cheap frequency counter.

Don't try and connect the Rigol directly to the mains. Just get a cheap transformer - 240V to 12v, 15v or  whatever.

This is something that might be done better with a small Arduino project. The code would be very simple, as long as you can learn how to use interrupts and the internal counters. All you need to do is measure every period in terms of some clock cycles, subtract it from the number of clock cycles in an exact 50Hz waveform period (20mSecs), and output the difference from a PWM output or Digital to Analog converter. Then you could use a digital scope, like the Rigol to view the period change waveform.
The 90kHz will probably show the frequency changes in more resolution, but if it is not on all machines, using the 50Hz sounds better. If using the PWM is too much trouble, you want a Digital to Analog Converter output.  You could use a shield with a serial DAC, or make your own from a handfull of resistors:

http://www.ikalogic.com/dac08.php

To output to a Rigol, you can live without the output amplifier as long as you use a 10x 10Mohm probe.

This is the kind of project you could set up on a Arduino prototyping shield, and then if you like the results, you could make your own dedicated board.

An analog frequency to voltage convert is very simple - you just need a circuit that generates fixed with 5uSec pulse on the leading edge of each 90KHz pulse or or a 10mSec pulse on each 50hz cycle , and you use a RC filter to smooth out the result. The output voltage will be proportional to the frequency.  The only problem is with generators, you are often looking at very small variations around the 50Hz, and at the end of the day, a microcontroller project + a Rigol would give the more accurate results then a quick and dirty analog circuit - especially if you used a 555 timer for the monostable.


The thing about a custom project is that with mains, you often want to see very small variations around 50Hz (or 90KHz). With the digital solution, you can view the frequency changes as precisely as you need.

Richard

[linc1]

Is the 90KHz squarewave just a pulse stream with a frequency proportional to the motor speed, or something else?

When you say a pulse stream engine speed is directly proportional to frequency when the engine is cranking the voltage amplitude is low as engine speed increases the voltage amplitude probably starts stabilise and the frequency changes with speed

1.8 KHz change would equite to 1 hz so if the no load frequency was 52 Hz 1560  RPM, 93.6Khz,  full load would be  50 Hz 1500 rpm 90 khz 

i will look at what you suggest, generally after a speed change induced by load the engine should settle back down within a few seconds if there is a continual hunt or ringing somthing is wrong.

again thanks for your responce

[saturation]

The work environs described is industrial and at worst, near the rated safety limits of the Rigol.

You need a true portable oscilloscope that is safety rated for mains use and robust to take outdoor use.  Take a look at mikeselectricstuff's review of his Owon or find a good used portable Fluke scopemeter in NZ, even an older model is within the capability you need.

https://www.eevblog.com/forum/index.php?topic=3844.0

[alm]

Re title: if there was an affordable machine available to do your job, your boss would already have replaced you with said machine.

Agree with saturation, something like a Fluke Scopemeter would be more suitable. Safety, isolation and portability are going to be much more important than raw bandwidth or sample rate. Don't remember how the Owon performed regarding safety or isolation, did it offer isolated channels like all but the oldest Scopemeters? Note that you need special probes made by HCK, also shipped by Fluke, for use with floating ground. The insulation on the 'ground' lead needs to be much more substantial since it can be at dangerous potentials. Don't remember what probes Owon shipped.

The reason why isolation is useful is because isolating the DUT is not always feasible when you're talking about industrial equipment. To measure mains with a normal scope like a Rigol, you either need to measure through a transformer (how will this influence your measurement?) or need a differential probe that will cost as much as the scope.

[ejeffrey]

A good counter lets you get the period over several cycles for more accuracy (if that is what you need), but the Rigol will be better then a cheap frequency counter.

I disagree.  Even a cheap frequency counter will do better at measuring frequency than any oscilloscope.  A hand-held DMM with the Hz feature will do fine for this and if you get one from a reputable brand it will be well enough protected to safely use with a 240 VAC.  Most of them can also count the 90 kHz pulse train directly, although IIRC 100 kHz is about where some switch to logic level counting only.

If you use a benchtop frequency counter instead of a DMM make sure its inputs are rated for the voltage you apply, or use a suitable high voltage probe for the 240VAC/50Hz or a 10x probe for the 30 V / 90 kHz pulse train.

[amspire]

Measuring the frequency in this case is not the point.  The aim is to get a graphical display showing the real time frequency changes and oscillations under various conditions.

Cheap counters usually just don't have the flexability and communication sophistication to do this well, particularly if what is required is, say, to output the period of 5 cycles every 100 mSecs. The reason you may need to do this is that if there is a lot of noise present, the accuracy of measuring each single cycle may not be enough to see small frequency purturbations. Using the counter in frequency mode for capturing fast frequency variations at 50Hz would be quite useless, but in reality, any half-decent modern counter measures 50 Hz by measuring the period of a fixed number of cycles, and then it converts the averaged period into the frequency display.

Cheap counters are vary capable of showing accurate numbers in the display, but in this kind of application, numbers shown on the display of a counter are useless. It is all about configurability, data capture and PC communication abilities.

I haven't looked but are any of the DMMS with counters equipped, but I would be impressed if you could set the actual number of cycles per period measurement, and then capture a stream of continuous readings without dropping cycles.

[linc1]

Thank you all for the responces amspire is correct it is more the graphical solution i wish to capture, that owan instrument looks good,  my UT 81 scope multimeter does some similar data capturing but not as good as that and is very slow.

in regard to the rigol maximimum voltage input was 300 Volts RMS , i would have thought that 240 volts rms would be ok and would not be a saftey issue ! am i missing somthing?
Max Input Voltage
All inputs 1M??15pF 300 V RMS CAT?
 
 

 

[saturation]

IIRC the Rigol 1052e is only CAT 1, check the pdf manual please to confirm,  and given the inductive loads you'll use it in plus there are very likely to be transients on the line in your workplace, the minimum your test gear should be CAT II,  but it would be better at CAT III.

Thank you all for the responces amspire is correct it is more the graphical solution i wish to capture, that owan instrument looks good,  my UT 81 scope multimeter does some similar data capturing but not as good as that and is very slow.

in regard to the rigol maximimum voltage input was 300 Volts RMS , i would have thought that 240 volts rms would be ok and would not be a saftey issue ! am i missing somthing?
Max Input Voltage
All inputs 1M??15pF 300 V RMS CAT?
 
 

 

[alm]

Don't forget the overvoltage rating of the probes. Probes can also shock you or fry the instrument if used beyond the rated voltage.

In case you've never used a scope, keep in mind that every measurement on a normal bench scope will be referenced to ground. A small PCB will often have a floating power supply, like a battery or lab supply. Otherwise it can easily be powered from an isolation transformer. The only way around this for these scopes is a differential probe, good ones are expensive.

This is quite different from a DMM. Ignoring this will result in fireworks or worse. This is why an isolated scope like a Scopemeter is superiod for mains-related work. The flparing channels can be used like a DMM.

I don't remember how well the Owon did on safety. didn't he have reservations about using it for dangerous voltages? Or am I confused with another cheap portable scope?

[amspire]

IIRC the Rigol 1052e is only CAT 1, check the pdf manual please to confirm,  and given the inductive loads you'll use it in plus there are very likely to be transients on the line in your workplace, the minimum your test gear should be CAT II,  but it would be better at CAT III.

Thank you all for the responces amspire is correct it is more the graphical solution i wish to capture, that owan instrument looks good,  my UT 81 scope multimeter does some similar data capturing but not as good as that and is very slow.

in regard to the rigol maximimum voltage input was 300 Volts RMS , i would have thought that 240 volts rms would be ok and would not be a saftey issue ! am i missing somthing?
Max Input Voltage
All inputs 1M??15pF 300 V RMS CAT?
 

The Rigol DS1052E is Cat I to 300VRMS 1000V peak and Cat II to 100VRMS 1000V peak, and it should be reasonably safe if you are using Cat III 100:1 probes or a Cat III differential probe - like the Fluke dp120 which cost about double the price of the Rigol scope. Cat III would give an extra degree of safety, but cautious use of the Rigol with two Cat III 100:1 probes used in differential mode probably makes you safer then anyone using lower rated consumer devices plugged into the 240V.

But in this case, there is no need if you only have to monitor the frequency output. If you isolate the 240V with a safely wired low voltage output transformer, then you can use almost any instrumentation safely.

It is hard to judge the safety issues with the 90KHz output. It is probably pretty safe, but I wouldn't know.

[vk6zgo]

Since the OP is only looking at fairly slow(compared to 50/60 Hz) variations,using a transformer with a low  voltage secondary would be the obvious way of doing this,as there wouldn't be any worry about inadvertently filtering out any HF stuff.
The only question is if the Rigol can do the job.
I don't know enough about them to say if it will work .Back in the day,the instrument of choice would be a chart recorder,with some kind of detector/interface circuit.

Two possibilities are a PLL at 50Hz,recording the correction signal,or a ratio detector,which if I remember correctly,just needs a tapped transformer.

VK6ZGO

[IanB]

My naive thinking about this is that a logging multimeter like the Fluke 289 would be the way to go, assuming that it could:

• Sample 10 or more times per second
• Record frequency measurements
• (Ideally) have a relative delta feature to zero out the base frequency

I have no idea of the 289 is actually capable of this, but it seems like an ideal industrial application. If it could do it, you could record the period of interest and then display a graph of frequency vs time and see just what you were looking for.

Trying to log a 50 Hz frequency too many times per second is going to be hard, but the 90 kHz signal should be less of an issue. How long does it actually take a high end meter to sample and lock onto a frequency for capture or display?

[vk6zgo]

Another thought is that this might well be in the range of a PC based voltmeter setup.
There are a number of sites on the net which had designs which interface to the PC via the parallel or series inputs of older PCs ,& there may be more modern designs as well.The software is usually provided.

You would still need a frequency/voltage converter or FM detector to get a low frequency signal which varies with the engine RPM,but a PLL should do the job (there is probably not enough inductance in a power transformer secondary for a tuned circuit at 50Hz,so perhaps not a ratio detector).
Another way is to put a video camera in front of an analog RPM meter!

VK6ZGO

[alm]

My naive thinking about this is that a logging multimeter like the Fluke 289 would be the way to go, assuming that it could:

• Sample 10 or more times per second
• Record frequency measurements
• (Ideally) have a relative delta feature to zero out the base frequency

I have no idea of the 289 is actually capable of this, but it seems like an ideal industrial application. If it could do it, you could record the period of interest and then display a graph of frequency vs time and see just what you were looking for.

Trying to log a 50 Hz frequency too many times per second is going to be hard, but the 90 kHz signal should be less of an issue. How long does it actually take a high end meter to sample and lock onto a frequency for capture or display?

A good frequency counter should be able to count 90kHz with a fairly short gate time, but don't count on a handheld meter sampling more than a few times per second (don't know the 289 specs by heart, see datasheet). Bench meters go up to a few thousand samples per second (at least for V/I/R, not sure about f), but cost more than a cheap scope, and aren't really designed for an industrial environment either (CAT II 300V or so, somewhat fragile like most bench instruments).

[saturation]

You're right alm, I've added some photos to illustrated your points.  Here's the derating curve of the RP2200 probes shipped with Rigol and a sample of the scopemeter inputs; note the posted maximums are different between floating terminals and from each terminal to ground.

I don't think mike touched much of the safety of the Owon more than it was capable of doing the measurements he required.  In the end the safety of these devices are cannot be tested non-destructively and we rely on 3rd party safety testing like UL, ETL or other such labs.

For the kind of the work the OP suggest, a used older Fluke scopemeter goes anywhere between 300-600, which are good to at least 3 MHz, based on the sampling rate, such as the this:

http://cgi.ebay.com/FLUKE-97-50-MHz-ScopeMeter-MINT-/160633060771?pt=BI_Oscilloscopes&hash=item256679fda3

The oldest ones may not be officially CAT rated, but they were made for those settings; if that's a concern then search for one that has specific CAT ratings.

Don't forget the overvoltage rating of the probes. Probes can also shock you or fry the instrument if used beyond the rated voltage.

In case you've never used a scope, keep in mind that every measurement on a normal bench scope will be referenced to ground. A small PCB will often have a floating power supply, like a battery or lab supply. Otherwise it can easily be powered from an isolation transformer. The only way around this for these scopes is a differential probe, good ones are expensive.

This is quite different from a DMM. Ignoring this will result in fireworks or worse. This is why an isolated scope like a Scopemeter is superiod for mains-related work. The flparing channels can be used like a DMM.

I don't remember how well the Owon did on safety. didn't he have reservations about using it for dangerous voltages? Or am I confused with another cheap portable scope?

[linc1]

well a few interesting coments , i can see its not that easy as first thought and has been doing my head in

i have a recording multimeter that reads frequency, its data swample rate is to slow

yes a chart recorder would be the way to go but i still need to get frequency into a voltage output that varies with frequency

[vk6zgo]

A phase locked loop should do it,but its correction loop may not be fast enough.
(You would read the correction voltage which would change with mains frequency).

An alternative would be to sample the 50Hz signal with faster pulses,then put the result through an integrating circuit.
One problem is that you would probably get a long term variation of the resultant dc signal,as well as the transients you are looking for.

Traditionally,mains frequency was read with reed meters,where the 50Hz reed reached its maximum vibration amplitude when you are on frequency. One of these & a video camera,instead of camera & RPM meter as I suggested earlier?

You could divide a crystal down to 50Hz,then heterodyne it against your 50Hz from the alternator.That will give you a very low frequency ac signal to play with.
Theoretically,you could "zero-beat" the two in a mixer,so when the alternator was running ,there would be no output from the mixer.The transient frequency variations would then show up as an output signal from the mixer.
In practice,you would not be able to maintain "zero-beat",as the alternator frequency isn't stable enough.
It still may be easier to do something with the few Hz frequency derived,though!

You could trigger your oscilloscope off a stable 50Hz source,& display the signal from your alternator on the screen.
If they are very close,there will be very little sideways drift of the display.Any variation will show up as an increase in sideways drift.
Again,you are stuck with converting this into a readable graphical form.

VK6ZGO

[saturation]

There are many ways to accomplish your need, but all the substitutes, such as data logging multimeters or frequency counters, are simply trying to show you using an alternative device for what an oscilloscope is better at doing.

The basic industrial DMM like a Fluke 87V or the new Agilent 1272a can do the tasks of logging and frequency counter, and it has the input protection you need.  Also, as an intermediate between a scope and a DMM, there is the Fluke 289 or the prior older versions.  They are also in the $200-400 range.

http://www.fluke.com/fluke/usen/digital-multimeters/fluke-289.htm?PID=56061

well a few interesting coments , i can see its not that easy as first thought and has been doing my head in

i have a recording multimeter that reads frequency, its data swample rate is to slow

yes a chart recorder would be the way to go but i still need to get frequency into a voltage output that varies with frequency

[vk6zgo]

I'm not sure this would work,but how about if you could save several seconds worth of 50Hz on your Rigol,(If it is capable of this),then zoom in & measure the time between zero crossings on each cycle of the saved signal.
The change in this,versus total time elapsed will give a measure of signal phase,rather than frequency change,but they are closely related.

You could then list the results,& using Excel,produce a graph of signal phase change versus time,& by a
bit of mathematical manipulation,frequency-v-time.

It wouldn't be real time,but if it works,it would give you a good idea what was happening.

VK6ZGO