current measurement

[charon1985]

Hi, i have built an induction heating circuit and i am trying to measure the current that goes into the step down transformer. After the transformer is the series LC circuit. I have 3 current transformers, one built by me and two built by factory. After the current transformer i have rectifying circuit, and after that i have an opamp, next will be the atmega controller to measure the current...

Now i have the following problem, at the beginning and end of the dead time, i have some oscillations spikes. Since the oscillations are not that big the rectifier circuit and the opamp are not affected, but the current measurement will suffer. I am thinking of taking 2-3 measurements and make the  mean. All 3 CT have same oscilations in the output.

What is generating the spikes and how can i remove them (even after rectification if else is not possible or the opamp)?

Thank you.

[rx8pilot]

Schematic?
Is this a PCB layout? Deadbug?

[charon1985]

Rectification, opamp and the atmega controller are on breadboard, the rest of the circuit is on PCB.

If deadbug refers to the construction method, then is not a deadbug construction.

[rx8pilot]

You are providing very general information and hoping for some information about the details of the behavior.

What you are seeing could be a design problem, a construction problem, or a probing problem. It could be all of those. If you provide only an overview of the setup - you are not likely to get an actionable answer.

[Andy Watson]

Are these spikes real? Or are they artefacts being induced in the ground lead of your oscilloscope probe?

[charon1985]

Providing to many details about a problem will give a "chance" for some people to comment on the problem providing wrong paths for solving. Usually the people who know the solving for a problem will recognize the problem even if the situation is briefly described.

The problem that i am facing, i have seen it on similar other projects, but the builders simply ignore it...

Andy Watson, it is possible to be artifacts because i have had similar problems. I am new to high frequency signals and oscilloscopes. For example on the same signal if the ground lead (the black wire with alligator clip) is not kept close to the probing pen, or the physical distance between the probing points (ground and probe tip) is large,  the rectified and smoothed signal looks like a small sinusoidal signal imposed over a DC component.
I think that the best way would be to measure the signal (with the microcontroler) and see what values i get. Would you sugest anything else to do?

[Kilrah]

Providing to many details about a problem will give a "chance" for some people to comment on the problem providing wrong paths for solving. Usually the people who know the solving for a problem will recognize the problem even if the situation is briefly described.

That's valid when what you see can only come from one or 2 pretty obvious causes. When there are way too many different things that can cause it like here not providing enough details won't get you anywhere... and not accepting it will get you ignored.

[bktemp]

The spikes are generated by the fast switching of the power transistors.
You can either ignore them, or you need to provide much more details (at least schematic and pictures of your actual circuit).

[rstofer]

The first image looks a lot like the switching noise I see on one of my op amp circuits and AFAICT, it's caused by the switching PS.
I am using a DC wall wart followed by a dual-output switcher to create +-15V.

ETA:  You can always put a low pass filter on the ADC input.  In any event, the ADC won't be measuring anything in the MHz range.

[charon1985]

Problem solved.

I hate measurement errors.

[rx8pilot]

This is the normal solution. I make my own with different lengths.

If you posted a pic of your work and test method earlier - a dozen people would have pointed out that detail right away. Power electronics are notoriously difficult to measure accurately and it is how I make a living right now. Don't be afraid to show your work if you want information. You may get some conflicting opinions, but that is way better than dead silence.

[Andy Watson]

This is the normal solution. I make my own with different lengths.

Thank you for the picture - the last time I suggested this solution I was mocked! I also suggested using a paper-clip as the source material - it gives it a bit of spring and the extra resistance is insignificant. Also, if you cut the end obliquely you can leave a sharp point for making the contact.

[w2aew]

I make my own short ground contacts like that a lot.  You can see it here (at 6:45) in my video about the affects of ground lead lengths on signal integrity:

[w2aew]

...and, if it is a spot you're going to probe often, or need to do it hands free, then make yourself a high performance probe "socket" like this:

[alsetalokin4017]

@Charon1985:

Your spikes are apparently _still there_! Note that in your first post you have the scope's vertical scale set to 200mV/div, and in the shot where you say "problem solved" you have the vertical scale set to 500 mV/div. The spikes on the sinusoid seem to be just about the same actual voltage in both cases.  (The second shot is somewhat complicated by the double-image caused by incorrect triggering.)  At the higher 500 mV/div setting the spikes are just disappearing into the noise on the flat-line trace, but they are still there.

Simply taping the long ground lead to the body of the probe does not solve the problem caused by the long ground lead. To eliminate this problem you should take the grabber clip off the probe tip and use one of the spring-type ground adapters as shown in images above, using the ground ring close to the sharp tip of the probe.

[mtdoc]

Some good stuff here: