Using a signal generator to simulate the LVDT secondary coil output to the AD698, the AD698 output jumps.

Hi AD698 ,

Thank you for your question. I am reviewing the information you provided and may come back with follow-up questions. I appreciate your patience as I work on your query. 

Regards,
Gilbeys

Ok, thanks for your feedback! 

Hello,

AD698 said:

Whether the signal generator can not be used directly for simulation, or the circuit needs to be modified, thank you for your help in advance!

the AD698 uses the ACOMP signal to synchonuously demodulate the AIN signal. That works, when ACOMP and AIN are synchronized. Synchronized means: if you monitor ACOMP and AIN with a two channel oscilloscope and trigger on ACOMP, also AIN has to stand still on the scope.

In your setup you use different signal sources to drive ACOMP (driven by EXC) and AIN (driven by signal generator). If you do not actively ensure the synchronization of both sources, the output of the AD698 will be unstable.

The easiest way to run this circuit successfull would be: use a real LVDT indstead of the signal generator. The output voltage of the LVDT is always in synch to its input.

If you absolutely need a simulation with signal generators: use two signal generators, which can be synchronized (or a two-channel generator, where both outputs are synched). Take one of them to drive ACOMP with a constant sinewave. Take the other one to drive AIN with a sinewave of identical frequency, but varying amplitude/phase.

best regards

Achim

Hello,

Achim said:

use two signal generators, which can be synchronized (or a two-channel generator, where both outputs are synched). Take one of them to drive ACOMP with a constant sinewave.

thinking about this a second time, this advice may have been uneccesarily complicated. ACOMP is a comparator input. It will probably not need to be driven with a sinewave, a digital signal should work as well. And most signal generators have a digital synch output which may be used to drive ACOMP.

So you may try the follwing:

- connect -ACOMP to a fixed DC voltage like e.g. 1V as reference level of the comparator

- connect +ACOMP to the digital synch output of your signal generator

- connect AIN to the analog sine output of your signal generator

This way your simulation should probably work with a single signal generator.

best regards

Achim

Thank you for your solution and suggestions. 

Now I want to verify the function and schematic of the AD698,

so, do you think the following methods are feasible?

I follow the schematic, using signal1 driven by EXC to drive ACOMP.

In addition, I use signal 1 to trigger the 33500B signal generator and

send out signal 2(sinewave), the frequency of signal 2 is the same as

that of signal 1, so as to ensure that signal 1 and signal 2 are in phase,

then I use the signal 2 to drive AIN, but the effect is not ideal.

Hello,

so you want to use EXC to drive ACOMP. And you want to trigger your signal generator with this signal to generate the sinewave for AIN?

I'm afraid but I don't think that will work satisfactory. The problem is: you can never set the frequencies of EXC and of the 33500B to really identical values (i.e. synchronize them). Both sources have individual deviations and drifts. The triggering will initially define the phase relation. But this defined phase relationship will drift away over time and the synchronous demodulation of the AD698 will sooner or later deliver unstable results. Do the test I suggested above with the dual channel oscilloscope. Can you really make both signals stand still by triggering only on one of them? If yes, then the signals are in synch. But this synchronization will be lost over time due to termperature drifts etc.

The signals on ACOMP and AIN must not only start with a defined phase, they must also maintain the defined phase relationship over time. This can only be achieved, if both signals are derived from the same source. They must be kept at the really identical frequency over time (like in a phase locked loop). It's not enough, if the frequencies are nominally identical, and a start trigger for the second signal source will not be sufficient.

If you can disconnect ACOMP from EXC, the simplest approach should be as described yesterday: drive both from the same source (33500B). For the ACOMP the digital synch output should be sufficient, for AIN you can use the analog sine signal of the 33500B.

If you cannot disconnet ACOMP from EXC (e.g. due to limited access to the signals on your board), you must derive AIN from EXC. The easiest way to do this is to use a real LVDT (turning your simulation into a real experiment). You may replace the LVDT by other transformers or build something with a PLL.

A kind of "handmade PLL" would be to try to keep the signals in synch based on the oscilloscope measurement suggested above. The frequency resolution of the 33500B is quite fine, so you could try to manually adjust the 33500B frequency so that if follows the EXC frequency exactly over time (controled with the parallel oscilloscope measurement).

best regards

Achim

Hello Achim,

Thank you for your detailed explanation. 

Yes, I want to do it this way. I connect the EXC signal of the AD698 to ACOMP through a phase compensation circuit, just like in the schematic. Then I connect the EXC signal to the trigger input of the 33500B and monitor it with an oscilloscope. After triggering, the 33500B outputs a sine wave with the same frequency as EXC to AIN. I use another channel of the oscilloscope to monitor the AIN signal. By adjusting the phase, I aim to ensure that the two channels of the oscilloscope are in phase synchronization. However, I found that this is not easy to achieve as the waveforms on the two channels are hardly synchronized.

Thank you for your guidance.

Best regards,

Shan Yueyue

Hi,

Good day. I agree with Achim. If the two input signals ACOMP and AIN are not driven in sync with each other, this would induce instability in the circuit which makes your simulations inaccurate at some point.

Regards,
Gilbeys