Java 2 Ada

LPS Issues

Version 1 by Stephane Carrez

This page keeps track of the issues found during the conception or building of LPS.


  • Must have a GNU Ada compiler for ATmega16 (GCC 4.1 is the target)


2006-12-22: The creation of several single sided boards failed. The clear reason is unknown. I have used different insolation time and the boards were not good enough. I suspect this is because the printed board were old (bought in 2001 or so) and the UV light has less effect on them.
2007-07-28: I've succeeded in creating 4 single sided boards (LPS supply, positive rectifiers, positive outputs and positive voltage regulator). Before doing the final boards I checked on small scratch boards. This time I've used a transparent slide for the insolation (instead of a laser print on 80g paper). The printed board failed when the insolation time was 2mn and when I used only one slide. I tested with 2 slides, it was better. With 3 transparent slides that were above each other, the insolation succeeded. The insolation time was 1mn30, revelation 1mn.
2007-11-04: With the emergence of small Ethernet controllers I'm wondering if I should switch to Ethernet instead of an RS2323 isolated solution. Ehternet is isolated by default so this would solve that issue. Second, an Ethernet device is easier to connect compared to an RS232 device. However, the software that I'll have to run within LPS will probably be more complex. The Ethernet solution is also a little bit more expensive (arround 30 euros).
2007-11-04: The LPS power supply board is working. I made a big mistake in the schema that I didn't caught during review. The voltage doubler had the diodes connected to the ground where they should be connected to the alternative inputs. Without load I get arround 57 volt/-57 volt after doubling. The +24/-24 power supply is working but I broke the +24 power under my tests. Since this power supply is not well protected I have to be careful about it (the +24 output delivers +57V; the -24 V can sustain 100mA).
2008-12-22: One year without working on LPS (due to lack of time). I'm simplifying the project and I'm configuring an analog version of LPS (ie, no controller) so that I have a working power supply that I can use until I design the digital part. The ATX power supply is gone due to lack of space within the box.