Function Generator Amplitude and Zero Offset change

[juppees]

Hi there,

I was wondering what a typical function generator does when changing the aplitude or zero offset while the output is active?

To the background:
I am a mechanical engineer (with a quite strong affinity for electronic tinkering) and I'm working in a university's materials testing lab and we have a few servo hydraulic fatigue testing machines. Some allow to use an external +-10 V signal as a control signal - so for example -10 V is mapped to cylinder position -10 mm and +10 V to +10 mm. Because I have to conduct tests that are not possible to control with the provided test software and the vendor did not provide direct access to the machine control, I have to use the mentioned analog auxiliary input. My idea was to use a computer-controllable function generator like the RIGOL DG812 to generate a sine wave (1...100 Hz). Since I have to adjust the amplitude and zero offset multiple times during the test, manual operation would be very tediuous. But even after consulting various manuals and looking a few videos I was not able to figure out, how a change in aplitude and zero offset is handled by the device (unfortunately, we do not have one at our institute...). Since the input signal is soft wired to the hydraulic cylinder, a step change or phase shift might cause severe damage to (at least) the specimen. Hence, we can only use a device that provides a soft transition and not only interrupts the current signal and starts over from beginning, when a parameter change is requested.


Thank you very mouch in advance!
Best wishes,
Josef

[Vovk_Z]

I was wondering what a typical function generator does when changing the aplitude or zero offset while the output is active?

It definitely depends on a software used in that particular function generator. One may do it in a smooth way, other may not (may have some step).

[juppees]

Hi,

thank you for your reply!

So, if anybody has tested her/his particular devices' behavior, I would be deeply greatful for the information on the model type That would really make my search easier.

Thanks,
Josef

[Labrat101]

Hi I am Not really familiar with what you are doing. 
But when you say it needs +- 10 volt signal . Does it have to be a sine wave?
It might well except a ramp up or ramp down, which would give a smooth transition. 
In Intel we had some activators also on some delicate equipment.
The signal input may exceptable as any analog.  ie -+ 10 volt slow linear ramp .

[juppees]

Hi,

the test has to be performed using a sine signal - the transition can be arbitrary, as long as it has no step change.

Thanks,
Josef

Some further information on the test:
I am doing fatigue fracture mechanic testing on metal specimens. Therefore a specimen is clamped into a load frame, that is equipped with a hydraulic cylinder. A controller adjusts a hydraulic valve so that the piston performs a sinusoidal movement with a (in my case) fixed force amplitude and zero offset at a given frequency. The aim of these tests is to monitor the crack growth induced by cyclic loading. Since the crack tip load at a certain force increases with crack length, the piston force has to be adjusted permanently in order to maintain a constant (or otherwise defined) crack tip load amplitude. In case the signal jumps, the cylinder has to perform a step change causing a quite high strain rate in the specimen, what might affect the results.

[Vovk_Z]

I think we may find enough info about transients of some exact generator from reviews on Youtube. 

[CaptDon]

O.K., got some questions here. First, is there some sort of LVDT Linear Variable Delta Transformer or perhaps a linear position indicating potentiometer that closes a feedback loop so that +1 volt exactly equals +1mm (or what ever the scale is)? Next question, do you have a load cell in line with the hydraulic actuator to give a readback of how much force is being applied (pushing or pulling) with a safety limit set on force in case your device being tested hits a hard physical limit? All of these SHOULD be included in your setup. I have an extensive background in this area having worked for Lord Corp and testing every kind of rubber compound and damper your could imagine. We used servo controls from Moog-FCS and our available hydraulic line pressure was 3000psi. The Moog unit could generate the test signals with automatic ramps applied if needed. It could generate sines, sine sweeps, noise (white, pink,brown, etc.), and any combination of triangle or ramp. The Moog-FCS units were expensive but if you are working at university level this is the very least caliber of equipment you should be conducting your tests with. I probably wouldn't even accept the results of your testing as valid with lesser equipment. Cheers mate!!

[rstofer]

I don't know if an Arbitrary Waveform Generator, like the Siglent SDG2082X (which I have) is useful in this application of not.  You can create arbitrary waveforms using any method you like (for me it's probably Fortran) and feed the values to the AWG

Here are all the manuals for the SDG2000X series (among others).  The programming manual goes into details
https://www.siglenteu.com/resources/documents/waveform-generators/#sdg2000x-series

I'm not sure how you program the points of inflection when the amplitude or offset is changed.  Technically, I would expect to see a discontinuity at the point of change.

From the datasheet:

Innovative TrueArb technology, based on a point-by-point architecture, supports any 8 pts~8 Mpts Arb waveform with a sampling rate in range of
1 μSa/s~75 MSa/s

So, you have 8 Mpts to get the job done.

[bdunham7]

the test has to be performed using a sine signal - the transition can be arbitrary, as long as it has no step change.

Every AWG I've seen simply shuts off abruptly and starts over whenever you change any of the parameters.  If too high of a dV/dt will damage your machine, you either need to implement an analog filter or use some other method.

[ledtester]

the transition can be arbitrary, as long as it has no step change.

How about returning the control signal to 0 between changes?

Something like:

- Put a potentiometer on the output to control attenuation of the output from 0% to 100%
- Have a program ready which will configure the waveform via network/USB/RS232 commands when you press certain keys on your keyboard.
- Before changing the control signal use the pot to kill the signal and afterwards turn the pot the other way to raise the signal back to its normal level.

[Circlotron]

Some function generators are pretty rude if you have two sine waves locked together in phase and frequency and you start altering the frequency. -> https://www.eevblog.com/forum/testgear/rigol-dg1022z-rough-frequency-change/

[juppees]

Hi,
first off all, thank you for your replies!

O.K., got some questions here. First, is there some sort of LVDT Linear Variable Delta Transformer or perhaps a linear position indicating potentiometer that closes a feedback loop so that +1 volt exactly equals +1mm (or what ever the scale is)? Next question, do you have a load cell in line with the hydraulic actuator to give a readback of how much force is being applied (pushing or pulling) with a safety limit set on force in case your device being tested hits a hard physical limit? All of these SHOULD be included in your setup. I have an extensive background in this area having worked for Lord Corp and testing every kind of rubber compound and damper your could imagine. We used servo controls from Moog-FCS and our available hydraulic line pressure was 3000psi. The Moog unit could generate the test signals with automatic ramps applied if needed. It could generate sines, sine sweeps, noise (white, pink,brown, etc.), and any combination of triangle or ramp. The Moog-FCS units were expensive but if you are working at university level this is the very least caliber of equipment you should be conducting your tests with. I probably wouldn't even accept the results of your testing as valid with lesser equipment. Cheers mate!!

The testing machine for which we want to implement the system at first is an Instron VHF series machine, equipped with an Instron 8800 classic controller. The piston stroke is actually measured by an LVDT, the force by a differential pressure transducer and a piezoelectric force transducer - all periodically calibrated by an accredited laboratory. The external input signal in parsed by the controller and taken as the set point for the pid controller - therefore all security measures like user set or inbuilt limits, specimen break detection, over/underpressure dectection... are still active. The Instron controller on its own is for sure capable to perform the things I am trying to do - but since we do not have the software and the support is discontinued due to the systems age of 21 years, using an external signal generator is propably the easiest solution...

So, you have 8 Mpts to get the job done.

Unfortunately, I need a few times more. If I sample a sine at lets say 250 pts / cycle, I only have memory for 32000 cycles - of about 5...8e7 cycles for a regular test.

If too high of a dV/dt will damage your machine, you either need to implement an analog filter or use some other method.

I actually do not fear a machine damage, but a step change causes a fairly high strain rate and therefore might the test.

How about returning the control signal to 0 between changes?

Since 0 V equals zero force in my case, I would definitely interfere my test (crack closure due to local plastification) and therefore, this is not an option.


@PreDre, bdunham7 and Circlotron
From your replies, I took the message, that a function generator is not an appropriate device for such an application. I think I will look for an PC controlled analog output and generate the signal on my own. Then I have all the flexibility I need.

Once more, thank you all for your replies! 

Best wishes,
Josef