Hello everyone,
I recently finished my studies as a industrial systems engineer and I'm now working for a research lab. My studies were mostly oriented on power electronics and automation, so I'm a beginner in the field of analog circuits with op-amps. Nevertheless, this kind of job has been assigned to me and I struggle a bit to complete it.
So, my job is to design an interface between a control board (B-Board PRO from Imperix (imperix.com)), and external sensors such as Voltage and current sensors (DIN 800V and DIN 50A from Imperix).
1. Sensors provide a differential pair of signal, that need to be filtered.
2. One of the task is to design a security system based on the sensors measurement, where the value of the signals will be compared with limits, to trigger a fault bus.

A circuit was recommended in the datasheet of the B-Board PRO, so I tried to reproduce it. In the attached file, you have the full circuit on LTSpice.
This circuit was then tested in real life but nothing is working. The opamps are heating a lot, oscillating like crazy. I only get noise of the output of my circuit. I'm not good at describing the symptoms of my circuit because I don't really know what is suppose to be normal or not, I'm really a beginner in this.
If anyone is willing to help me a bit, I would gladly accept it. I can also provide additional information because I'm pretty sure some things are missing in my description.
Thanks in advance for your answer.
Best regards,
Maxime Rouiller

PS: I hope I posted the thread in the good section, I'm new to the forum.

 

hi Maxime .
Welcome to AAC.
Do you have a LTspice asc file you could post.?
E

 

Welcome to AAC.

I’ve moved your thread from the general forum to the more specific Analog and Mixed Signal Design forum.

Nice to have you join us, I expect you will get some good help.

 

hi Maxime .
Welcome to AAC.
Do you have a LTspice asc file you could post.?
E

Sure, here it is.

 

hi M,
Your OPA supplies have 10R impedance to main the +/-13V supplies and there is no power rail decoupling, the
OPA's will oscillate and get hot.
E

 

hi M,
May I ask why are you injecting Noise into the power lines.
E
BTW: I have cleaned up my version of your asc file.

 

For reference, attached Is the B-Board PRO datasheet from which the circuit came. [original source]

 

May I ask why are you injecting Noise into the power lines.

In the actual circuit, I will use a AC/DC power supply that provide +/- 12V. (RT-65B datasheet in attachment) This power supply has a very noisy voltage level that I smooth with capacitors. I injected noise to have a accurate representation of a power supply with a noise voltage level around 12V.

 

hi M.
So does the circuit board that houses the OPA's have capacitive decoupling at each of the OPA's locations on the PCB, if yes what are the decoupling capacitor values.?

Checking the circuit using LTS suggests not of the OPA are overheating.
All in the lowish milliwatts region.
E

 

No, unfortunately I forgot to add decoupling capacitors to each opamp. I only placed capacitors between 12V and GND and -12V and GND at the entrance of my board. (their values are 100uF (it was the only caps I could get quickly)). So do you suggest that my opamp are heating because they need decoupling cap as well?

 

hi M,
Use say a 10uF and 10n on the power rails close as possible to each OPA power pin.
Let us know how it goes.
E
BTW: this Noise circuit for LTS is neater,

@Rouillermax Added: PDF

 

Once again thanks for you help and suggestions
Unfortunately, I could access a GND trace easily due to the design of my board (my fault). What I did was putting a cap between the 12V and -12V pins, as close as possible.
I'm not sure this setup is doing anything since it is only rejecting common mode noise, but it's better than nothing right?
One other thing I noticed while measuring if this cap was the solution, is that the 1.7V reference that I input on the - is clearly not 1.7V. For this level I used a voltage regulator (LM4120), with a cap of 22 nF (like the typical application). For a reason I don't understand, the voltage level is 8.8V instead of the one desired.
I'm starting to think my opamps are damaged and that is maybe why I get all these issues...

 

hi M,
Are you saying you could not add the decoupling capacitors for each OPA supply line.?

Where is the 1.7Vref on the circuit diagram.?

Do you have a complete circuit diagram you could post.?
E

 

I only could attach capacitors between the +12V and -12V pins of the opamps, but not between +12V and GND and -12V and GND.
The 1.7V ref is simulated by the 1.7V voltage source. Nothing changed from the schematic I posted.
(attached is the datasheet of the voltage regulator. It is 1.8V but should not change the global behavior )

 

hi M.
All I see for the 1.7Vref is a LTSpice voltage source, not the components that would provide a 1.7Vref.

I would like to see the actual circuit diagram and if possible a clear photograph of the PCB.

It is not possible to make a meaningful reply by looking at the LTSpice circuit.

E

 

The component I used is not available in LTspice. That is why I used a voltage source. But here is the circuit I have where V_0 is the 1.7V, and I also used 22 nF cap:

So, this piece of circuit replace the 1.7 V voltage source.
Then, I also attached the PDF of my PCB (interface_box.pdf)

 

hi M,
IMHO the PCB Power tracks, +/12V and Ground/0v are very poorly laid out, and I would predict they will cause you power line noise problems.
What IC type is the 1.7Vref.
E

 

Yes, I agree that the power line are really not well laid out... I'm not experienced in that matter and it was my first real PCB.
The IC is the one I mentioned earlier (LM4120) and its datasheet is in my previous post.

 

hi M,
Sorry, I missed your 1.7V PDF.

What do you plan to do regarding the PCB layout.?
E

 

In my opinion, it would be a good thing to re-do the layout properly. But unfortunately, the objective of my manager is apparently to make this one work, so it doesn't cost too much to rebuild everything.