Noise problem in PCB caused by grounds

[satishreddy]

Hello,

we have designed a board using aduc7026 microcontroller but we are facing noise problem as we have different grounds for Vref(gnd) , vcc(gnd),avdd(gnd),24gn connected separately to connector.I think noise problem is because grounds so please help me to come out of this problem.

Waiting for the reply
Regards
satish

 

[honey 77]

Re: Noise problem in pcb

Add capacitor (0.1uF) between VCC and GND for decoupleing.

 

[satishreddy]

Re: Noise problem in pcb

Add capacitor (0.1uF) between VCC and GND for decoupleing.

Thank u for the reply.

we tried with the capacitor(0.1uf) between vcc and ground but the problem is not cleared.so please give me any other suggestions to come out of this problem.

Thanks & regards
satish

 

[jiripolivka]

Re: Noise problem in pcb

As you do not specify what noise you observe, try to find he noise source first. All microcontrollers use clock oscillators ad those generate harmonics. As you mentioned that you have separate grounds, then use 10 nF + 100 uF in parallel between those grounds, to achieve they all are "AC coupled" to kill the noise.

 

[FvM]

Re: Noise problem in pcb

Please refer to the datasheet Grounding and board layout recommendations. ADI clearly requires analog and digital ground to be connected directly at the chip or to use a common ground plane.

 

[satishreddy]

Re: Noise problem in pcb

As you do not specify what noise you observe, try to find he noise source first. All microcontrollers use clock oscillators ad those generate harmonics. As you mentioned that you have separate grounds, then use 10 nF + 100 uF in parallel between those grounds, to achieve they all are "AC coupled" to kill the noise.

noise is generated from power supply.......

Thanks & Regards
satish

 

[tpetar]

Re: Noise problem in pcb

In applications in high noise electrical environments, it is best to run the microcontroller near the high end of the allowable VCC range.

 

[satishreddy]

Re: Noise problem in pcb

too much ground noise is observed.........

 

[milind.a.kulkarni]

Re: Noise problem in pcb

try to put the inductor in series with your power lines ....also use very high value of capacitor like 1000uf after the inductor ....can be made like LC or CLC filters...this may help but the problem with high value of inductance is it will feed the power little longer than expected.... also one can also connect the AGND and GND(digital ) with the inductor to avoid noise translation from one GND to another GND .... I want to share one more thing here too...As I understand the problem you are measuring the noise on ground using CRO some time CRO also can give wrong reading ....In order to conform this claim you need to give same power source to OPAMP and input to OPAMP as GND of the board ....OPAMP must amplify the noise on Ground to give you the better idea....about the noise level on ground.....

Good Luck

Good Luck

 

[jiripolivka]

Re: Noise problem in pcb

try to put the inductor in series with your power lines ....also use very high value of capacitor like 1000uf after the inductor ....can be made like LC or CLC filters...this may help but the problem with high value of inductance is it will feed the power little longer than expected.... also one can also connect the AGND and GND(digital ) with the inductor to avoid noise translation from one GND to another GND .... I want to share one more thing here too...As I understand the problem you are measuring the noise on ground using CRO some time CRO also can give wrong reading ....In order to conform this claim you need to give same power source to OPAMP and input to OPAMP as GND of the board ....OPAMP must amplify the noise on Ground to give you the better idea....about the noise level on ground.....

Good Luck

Good Luck

If your noise comes from your power supply, throw it away! Use only a good power supply, linear if possible. Why use a bad power supply and then fighting interference?!

 

[mvs sarma]

Re: Noise problem in pcb

noise is generated from power supply.......

Thanks & Regards
satish

what type psu did you use?
If linear, add diodes of 0.01uF across each diode. if smps, add emi filter at the power input and take output thro' inductors both on Vdd and gnd wires. for distribution of power , use star distribution from the place where the power has an electrolytic cap and a disc cap.
you might also have distributed filters close to the chips pins. I hope it should solve the whole issue.

 

[satishreddy]

Re: Noise problem in pcb

try to put the inductor in series with your power lines ....also use very high value of capacitor like 1000uf after the inductor ....can be made like LC or CLC filters...this may help but the problem with high value of inductance is it will feed the power little longer than expected.... also one can also connect the AGND and GND(digital ) with the inductor to avoid noise translation from one GND to another GND .... I want to share one more thing here too...As I understand the problem you are measuring the noise on ground using CRO some time CRO also can give wrong reading ....In order to conform this claim you need to give same power source to OPAMP and input to OPAMP as GND of the board ....OPAMP must amplify the noise on Ground to give you the better idea....about the noise level on ground.....

Good Luck

Good Luck

Now the problem is almost solved as we have put ceramic capacitor between vcc and gnd of power supply.But the noise is around 100mv so this as to be reduced to 30mv as per the data sheet of tps5450.

Thanks & Regards
satish

 

[tpetar]

Re: Noise problem in pcb

Now the problem is almost solved as we have put ceramic capacitor between vcc and gnd of power supply.But the noise is around 100mv so this as to be reduced to 30mv as per the data sheet of tps5450.

Thanks & Regards
satish

:shock: What power supply you use ? Replace that. You have very bad power supply if this can called that.

Often is practice to use 100nF near as possible to +/- pins of IC/uC.

 

[FvM]

Re: Noise problem in pcb

Now the problem is almost solved as we have put ceramic capacitor between vcc and gnd of power supply.But the noise is around 100mv so this as to be reduced to 30mv as per the data sheet of tps5450.

It's instructive to take a closer look at the TPS5450 datasheet. The 30 mV ripple voltage number is a design parameter of an application design, the datasheet discusses the capacitor requirements to achieve it, particularly ESR-wise. PCB layout isn't considered in this regard.

So if you observe a higher voltage ripple, there are two likely explanations:
- Your design parameters are not meeting the requirements, e.g. capacitor ESR
- The voltage ripple is increased by "parasitic" PCB layout elements

In any case, I doubt that the output voltage of a switched mode converter either if the ripple voltage is 30 or 100 mV should directly supply an analog micro processor. Additional LC filters are strongly recommended.

As already mentioned, you can make your life much easier by using a linear regulator, which seems particularly reasonable for analog circuits of rather low power consumption.

 

[satishreddy]

Re: Noise problem in pcb

Getting noise when reading adc.when adc is not read the is noise is less around 50mv while reading it is producing spikes when observed on cro...please help me to come out of this problem...
Regards
satish

 

[milind.a.kulkarni]

Re: Noise problem in pcb

What is the sampling period of ADC....try to put 10nf or less capacitor on the Analog pin ....Bcz I think while ADC is sampling the voltage is creating the switching transients....Also try to put the inductance in series of power feed line with appropriate rating .... One more aspect is you need to check the decoupling capacitor of IC and there position near to pins....

Good Luck