30 March 2011

Molded MiniDIN8 cable for FT817 accessories

I have just secured a small lot of MiniDIN8 molded cables, 2m long, terminated in a DB9 connector.

These will be offered as an option to fellow builders of my FT817 accessories.

28 March 2011

FT-817 Interactive Frequency Reader - current consumption

Measured on the FT-817ND 12Vdc input, the I.F.R. draws:
  • nominally about 16 mA
  • about 110 mA when communicating over the CAT interface
These measurements were taken without LCD backlight.

Since the circuit uses a 78L05 regulator, rated at 100mA continous, I would power the LCD backlight directly from the 12Vdc line, or a dedicated battery.

25 March 2011

Fiddling with a MC145106 PLL

While re-arranging the shack, I have come across a 10-11m RTX with MC145106 PLL inside. I had used it on 10m, so I decided to check its electrolytic capacitors since it was built more than 20 years ago. While peeking inside I traced damages and mods I had done and fixed, and that was about 20 years ago too!

I had always wanted to "upgrade" the RTX by feeding the PLL divider with a binary counter, in place of the rotary switch + band switches. Now that I can "master" (aehm...) microcontrollers, adding such a control will give even more flexibility.

The plan is to remove the two 4008 adders between the rotary switch and the PLL, and replace them with DIP sockets. Then build a daughter board with a ATtiny2313 and two or more buttons to get the tuning done.

If this works, meaning that I have understood how was designed the PLL circuitry, an external display could be added too, memories, scanning, restricting to 10m and so on.

The RTX will keep the orignal 10 kHz channelization, unless I manage to make the PLL chip do 5 kHz increments. But that would require close work on the PLL.

First I will check the RTX power ups and is still generating some RF :-)

The picture shows PCB traces between 2x4008 and the PLL 145106.

For the records, this RTX is called Pearce Simpson Super Cheetah but it has 6 "bands". It has a circuit very similar to Superstar 3900, which I am using as a reference for this mod.

14 March 2011

150th blog banner picture

On March 17th 2011 Italy will celebrate its 150th birthday. The only colored waves I could think of are made of photons, so let me introduce you to the new blog banner:

Celebrations will last 9 months, so if you are planning a Spring/Summer trip to Italy make sure to check out the program!

07 March 2011

Measure CFL color temperature with a picture

Say you are lighting a room with two pendants. One day you replace one pendant's CFL bulb and find out that, despite being both "warm light", their color rendering is different. You may notice this effect looking at their light spot on the ceiling, or observing usual room objects under the two bulbs.

That was my case, with a brand new "warm light" 2700K CFL looking colder than a Philips CFL. There my curiosity stroke: how to find out the actual color temperature of CFL bulbs?

Having played with digital pictures I remembered that it is possible to control while balance (that usually sits in "AWB" mode) in post-production, that is with a picture editing software. One condition: the camera must be able to save pictures in RAW format, not JPEG. Most recent point-and-shoot cameras should be able to save uncompressed pictures ("RAW").

The white balance is controlled by adjusting a slider to the light source's actual color temperature: colors will look natural only when picture settings match light source's value.

So, here's the procedure I tried:
  • take a picture of the object/room lighted with the suspect bulb, it MUST be in RAW
  • colors might look wrong, it doesn't matter
  • download the picture to your computer and open it up with your camera producer's RAW processing software (Canon's is called Digital Photo Professional)
  • edit the RAW picture settings and find out how to control the white balance
  • select "Color temperature" and move the slider until picture colors match real ones: that WB value is your actual CFL's color temperature
Canon's slider goes from 2500 to 10000 Kelvin in 100 K steps. Probably you will not be able to find the exact value, but a non-trained human eye can recognize a range of +-/ 200 K around the actual, true setting.

Few words of caution. Wait for the CFL to warm up before taking the picture. Mind that your computer screen might modify colors too. More?