We are still fighting with the electric noise. As we said in previous blog entries this problem affects the focus signal.
We are using the PHR-803T electronics and the astigmatic method to focus de laser on the PCB. To obtain the S-Curve we calc and amplify the A,B,C,D signals obtained from the pickup’s photodiode array and generate an analog signal (FE or Focus Error) which we read from an Arduino port. In a similar way DVD or Blu-Ray players focus the laser beam on the disc surface but using more complex circuitry because they have to do it in real time as the music or movie is playing. We only focus the laser beam once when begin to print the PCB so our circuit is more simple.
We developed all this method and electronics last year and were working fine in MKI, but in MKII we did the mistake to mix the digital and analog circuitry on the same PCB, so the digital circuitry induce noise on the sensible analog signals and eclipsed the S-Curve.
We think the solution will be to isolated digital and analog circuits on different PCB areas and use different GND and VCC lines connected in a star configuration to a common point. Also to reduce PWM traces length ad minimum, isolate connector pins and decouple everything with some tantalum capacitors which are more suitable to the range of frequencies we are treating. We also included a choke to see if it filter well the power suply noisy frecuencies.
So we redesigned the circuit and the board and here it is: (courtesy of ZofzPCB 3D Gerber viewer)
The smaller area is the digital circuit: the noisy neighbourhood.