Forums

56F805EVM ADC problem

Started by millindala March 12, 2003
I am working with two 56F805EVMs - one is Rev. A and the other Prague-
E version. I have many problems with their ADC's, the worst being in
Rev. A. Suppose a constant voltage of +2.5 V is fed to the ADC in any
of these DSPs, there is a wide disparity among different samples.
Furthermore, this disparity gets worser when a motor drive is driven
by the DSP. Earlier, I tried improving the grounding and placing
common mode chokes on the motor drive cables. I cannot avoid
proximity to the motor drive as the DSP is used to generate the PWM
pulses for it.

When I checked in the archives of this mailing list, I found similar
ADC problems reported (ref. Jacob Christ, 8 Nov 2002) with the first
batch of 56F827EVM, where a few decoupling capacitors on the EVM were
suggested as a remedy. Is there a similar solution available for
fixing the ADC problems in Rev. A and Prague-E versions of 56F805EVM.
I don't find any fixes for the ADC in the Chip Errata documentation.

Thanks,
mahesh--



I have used EVM to test my applications. I never
expect EVM can do a good job. Its ADC inputs are noisy
because it was designed for evaluation purposes. If
you want to get noise free ADC input. I suggest you
make a your own PCB and carefully layout your analog
traces and ground traces. Connect analog grounds
direct to power supply input pins. You also can reduce
ADC clock frequency to allow longer signal settling
time.

Charlie W.
--- millindala <> wrote:
> I am working with two 56F805EVMs - one is Rev. A and
> the other Prague-
> E version. I have many problems with their ADC's,
> the worst being in
> Rev. A. Suppose a constant voltage of +2.5 V is fed
> to the ADC in any
> of these DSPs, there is a wide disparity among
> different samples.
> Furthermore, this disparity gets worser when a motor
> drive is driven
> by the DSP. Earlier, I tried improving the grounding
> and placing
> common mode chokes on the motor drive cables. I
> cannot avoid
> proximity to the motor drive as the DSP is used to
> generate the PWM
> pulses for it.
>
> When I checked in the archives of this mailing list,
> I found similar
> ADC problems reported (ref. Jacob Christ, 8 Nov
> 2002) with the first
> batch of 56F827EVM, where a few decoupling
> capacitors on the EVM were
> suggested as a remedy. Is there a similar solution
> available for
> fixing the ADC problems in Rev. A and Prague-E
> versions of 56F805EVM.
> I don't find any fixes for the ADC in the Chip
> Errata documentation.
>
> Thanks,
> mahesh-- >


__________________________________________________