LEDs and PWM...

[Norse]

Hi,
I have a tiny puzzle for you.
My isopod V1 has been watching a steady analog signal and converted that to a 50/50 pwm dutycycle. It has been rock steady thanks to the postings about averaging the samples.
Now the odd thing happens, I turn off the Isopod LEDs and the motor starts to creep. I can see the slight change in the dutycycle on the scope.... I turn on the LEDs and the creeping motion stops... I do this a few times with the same result...

I was thinking of using the LEDs as indicators but if it causes a creep in my "neutral" position then I have a precision problem.... I have plenty of powersupply headroom for the pod, 800mA.
And, of course, the motor is driven by a your great little PWM H-Bridge nmih-0050 which is attached to a separate power-source...

I do not have to use the LED's so it is perhaps a tiny but odd problem...

Any thoughts?

Cheers,

Per.

[RMDumse]

Yes, that is a bit odd. But, let's see if we can sort it out. Since you are reading an analog signal, and the LED is a digital signal, there should be no cross over. But apparently there is. So where could that be coming from? Perhaps it is through the Vref to the A/D. The current through the LED is not insignificant, about 10 to 15 mills each if I recall. This additional current might be biasing the refernce down a bit, or the gnd plane up a bit.

So as a way to devine if this is the case... does it happen for all LED's or will only one cause it to creep?

Can you put a high precision meter across the Vref to see if it dips at all when the LED's are ON/OFF?

[Norse]

Yep, it does drop a bit.
All LED's off, Vref = 3.23V
All LEDs on, Vref = 3.22V
This was measured between VSSA and Vref on J8....

I also measured between J8 Vssa and the 3.3 voltage regulator which showed the same voltages. Across the 3.3V regulator I measure a steady 3.27V.


I guess it calls for an external Vref if I decide to keep the LEDs as indicators...

Cheers,

Per.

[nmitech]

If it is a short creep, try to tackle a 1uF tantalum cap across VREF & VSSA. If it is a constant creep, use a bigger cap on the 3.3V regulator output. Just a thought, in case if you want to try.

[RMDumse]

We should probably check on V2 as well. We moved the A/D connections on V2 to be closer to the processor, and hopefully less susceptible to digital interference. It may show an entirely different response.

Per, you've got a V2 there, don't you? Would it be possible for you to check under the same circumstances?

[Norse]

Hi Randy,
The V2 shows no change in Vref.
It stays at 3.16V regardless of the LED status... So that is good.
It was measured at the J3 Pins.
The 3.3V regulator shows 3.24V.

The V2 is therefore indeed more stable...

Oh well, I don't want to rewire my interfaces so I will have to live with it until I change it to the Plugapod footprint for the final experiments.

Cheers,

Per.