[SOLVED] Inverter Charge Pump problem - noise

[evandro.bubiak]

Hello, we're using TC962 to generate -15V from a +15V positive regulator.
The negative voltage is used to supply 2 general purpose opamp and a differential output amplifier (THS4151).
The TC962 current output is about 23mA.
The problem is that we are observing a little noise coupled ond our output that is caused by the TC962.
Attached there is a picture showing the noise coupled on circuit output. There is 2 graphs, one is with a digital filter showing only the noise.
There is another picture showing the circuit. We tried to replace C26 for a 100u and no luck. The others IC's are using a 100nF parallel with the supply.
Is this noise normal or it can be eliminated? How?

 

[D.A.(Tony)Stewart]

Welcome to the real world.

At this point you can lookup the specs for ESR of your electrolytic caps and see how that compares to the output impedance of the port driving that capacitor.
(Both)

The IC spec says Low Output Impedance 28 OhmTyp
TC962 can source as much as 80 mA but I do not recall a graph of ripple vs current.

You can lower the ESR of the Caps,
define your noise spec mV-pp, see if signal supply noise rejection is adequate, and consider LC Filter if you need more ripple reduction.

Also be sure your probe ground leads use 1~2" clips or use coax.
Be careful of layout EMI... Isolation, CM Filtering, Shielding, ground, supply ESR, ESL


B.T.W. so called General Purpose 10uF caps have ESR around 30 Ohms in SMD and 15 Ohms in Radial
and so-called low ESR Caps will be < 1 Ohm for this value in 25V

 

[evandro.bubiak]

Welcome to the real world.

At this point you can lookup the specs for ESR of your electrolytic caps and see how that compares to the output impedance of the port driving that capacitor.
(Both)

The IC spec says Low Output Impedance 28 OhmTyp
TC962 can source as much as 80 mA but I do not recall a graph of ripple vs current.

You can lower the ESR of the Caps,
define your noise spec mV-pp, see if signal supply noise rejection is adequate, and consider LC Filter if you need more ripple reduction.

Also be sure your probe ground leads use 1~2" clips or use coax.
Be careful of layout EMI... Isolation, CM Filtering, Shielding, ground, supply ESR, ESL


B.T.W. so called General Purpose 10uF caps have ESR around 30 Ohms in SMD and 15 Ohms in Radial
and so-called low ESR Caps will be < 1 Ohm for this value in 25V

OK and thanks!!! I just forgot about the ESR on my MLCC
I Will try this and the RL filter and PI filter. I am already taking care of EMI using a good ground plane and using a SMA connector with a coax cable to measure the output signal.
In this case, for a RL or PI filter do I need to use shielded inductors or it wont make difference?
I will post the results here.

 

[D.A.(Tony)Stewart]

So what are your Cap ESR values?

Shielding depends on noise current nearby orientation. if at right angles no problem

 

[evandro.bubiak]

SunnySkyguy I was looking for the ESR of my MLCC (part number C3216X6S1E106M) but there is no information about the value on the datasheet, so aI tried to measure it using a 1Khz 1us pulse and a scope described here https://geoffg.net/Measuring_ESR.html and got a 0.05 ohms (don't know if it is correct).

I tried using a 100uH smd 1206 inductor to form a LC on the output of the inverter (10uF + 100uH) but got jus a little atenuation of the ripple (15mv Pk before using the inductor and 10 mv using the inductor).

Is there any pratical differente on puting the inductor first or the cap first on the LC filter?

I did not undestood your answer on where using the Shielding inductor, could you detail it a little bit more please?

Do think that trying to put a big cap, like 220 uF, could work for reducing this ripple? I dont know what to do now...

Thanks for all...