I have been designing small psu units and got tired of resistor loads. Since I had a nice cpu heatsink I decided to grab some random components at the lab and make a programmable current load.
The schematic I am using, is something like this:
Of course I could not make something that looks nice. To be honest I actually made something that looks pretty damn bad. But it works.
Measuring current and output voltage. Everything seems nominal.
The powerbox is unable to supply all the current I want it to. I must start thinking about something more beefy.
Almost like a bomb, is it not?
And here is something pretty. Everything before was very damn ugly :)
I got some pcb-s made that actually worked. One is kind of ok but other is preforming exceptionally well.
First one for the blog is my long winded project of UAV Ground station. It begins with RC transmitter and receiver builds. The transmitter pcb is ready but reciever still needs 5V stepping regulator. I will see when I will get to that.
Transmitter is designed to be really simple module. It has AtMega processor and an Xbee modem. Xbee is choosed to be XBee 868LP just so that it would have maximum compliance with RF rules and at the same time to ensure maximum transfer distance.
The bandwidth of the module allows me to have 50Hz update rate with 10 channels of control and also I can squeeze a little bit telemetry about receiver side in to it. What I want to see is battery voltage, current and maybe signal quality information. (I could use the low side quality info, but upload will have considerably more data flow than receiving side.
The other PCB is just a module of a bigger project.
I was needing a 12V and 1A supply from highly variable input voltages. I decided to use a sepic converter. It would not be neccessary, but my voltages are from 5V up to 50V. Making 12V from that area is not something one has to do every day.
The solution for my problem is here:
And if anyone questions then yes, I have re-soldered few parts more than once. At first I did not get the current out I designed the system for. The problem was in feedback decoupling. But before I found that, I tested and measured multiple possibilities. After I fixed the feedback noise problem the system became stable and I got 12V out with low side input and also with high one. Also the system is rather resistant to fast changing input voltages and varying load.
The other module for this system will be 30V and 1A output from same input voltages. That is a bit more touchy design so I will see how it works out.
This is a lesson for everyone. Double check your footprints!
Also you can see here an example of pcb that actually has been designed keeping cnc milling in mind. I got lucky and had it produced in a fab. Vias are huge, because they are there with soldering trough hole in mind. Also that is the reason for so wide signal tracks.
Today, when I was soldering my new 4×20 hitachi LCD it felt like something is amiss. After thinking about it I decided, that it was the light that my lab has. It is not nearly enough for soldering although it is more than needed for reading and paperwork. I decided to make one little fast fix for the lab.
this fix took about 20min and some random spare parts. I have no driver for it. It is running from my Current controlled lab supply but for now, it is the most awesome upgrade to my lab in past few months.