Editor’s Note: This is part of a several part project by Vincent Portet. If you would like to see a specific chapter, refer to the table of contents below.
How to Build a Color Mechanical Television, Part 1
Color Mechanical Television, Part 2: Choosing Components
Color Mechanical Television, Part 3: Circuit Construction
Color Mechanical Television, Part 4: Color Systems
Color Mechanical Television, Part 5: Software Tools
Color Mechanical Television, Part 6: Tuning Procedure and Calibration
Proposed Tuning ProcedureThese guidelines are not strict procedure steps but are rather intended to explain a possible method that I consider to be simple and quite logical. It is probable that sometimes you will have to step back and forth in this procedure for finer adjustments during tests. You will need an oscilloscope. Even a basic, old, and analog oscilloscope will do.
First set the potentiometers in the following positions:
- P1 set at mid-range
- P2a set to maximum resistance (if the gamma correction circuit is used)
- P2b set to maximum resistance (if the gamma correction circuit is used)
- P2c set to maximum resistance (if the gamma correction circuit is used)
- P3 set at mid-range
- P4 set to zero resistance
- P5 set at maximum resistance
- P6 set at maximum resistance
- P7 set at maximum resistance
- P8 set at mid-range
- User-controller speed potentiometer set at mid-range
- P9 set to maximum resistance (if the color decoder is used)
- P10 set at mid-range (if the color decoder is used)
- P11 set at zero resistance (if the color decoder is used)
- P12 set at zero resistance (if the color decoder is used)
First Tuning PhaseStart playing the color bar pattern and adjust the computer volume so that the composite signal (green or luminance) is about 1V peak-peak at the maximum levels.
Then check the signal on “SYNC OUT” in the synchronization extractor circuit while viewing an entire video frame. While turning P1, look for the two limits where the extracted synchronization signals becomes degraded and then replace the cursor at the middle of these two limits.
Once the synchronization correctly operates, the stroboscope LED must flash at periodic intervals (12.5Hz, the video frame refresh rate equal to the disc rotation speed in RPS).
Then check the back-light (video controlled) LED anode supply voltage (in the low loss regulator circuit), while loading this output by about 100 ohms, just to drain some current from it. This voltage appears like a DC voltage plus a superimposed ripple, which is not an issue. Considering the maximum expected VF for the white or blue LED (about 3.5V), the common anode voltage should be set above VF + 2.4V ~ 6V to have the LED current driver(s) working properly. So adjust P3 to have the lower ripple peak at 6V.
Setting the Maximum Green LED Current
It’s time now to set the maximum green LED current by using P4. Replacing the LED by a 4.7-ohm resistor to reduce transistor heating, set P4 so that the maximum peak voltage on the 6.8-ohm resistor (BD135 emitter) is: [target_current] x 6.8.
For the green LED (part of the RGB LED), I recommend not to exceed 240mA per LED, while the proposed white LED (for the black and white version) can handle much more. If wanted, you will not reach the maximum rated LED current of 900mA with this circuit, unless you change the 6.8-ohm emitter resistor for a smaller value and increase the 47-ohm resistor in series with P4. You will also need a heat-sink for the BD135.
Adjusting and Testing the Optical Sensor
Once the video circuits are set, it’s time to adjust and test the optical sensor. It is mandatory to have it mechanically installed in front of the fully-equipped disc, as in the definitive setup.
To have the disc spinning, it’s possible to move it by the hand or to power the motor with an external power supply (for instance a battery, even if the disc spins slowly). If the “counting track” sectors are correctly made (dark black and bright white) and if the distance to sensor is correctly adjusted and if the sensor is correctly protected from strong ambient light, then there must be a position for P5 for which the 2N2907 emitter “toggles” at each counting sector all along the disc rotation angle. This should happen flawlessly all over the disc rotation. By adjusting P5, look for the limits where operation is degraded then set it back in the middle.
Checking the NE555, Adjusting P6 and P7
Now it’s time to check that the NE555 correctly generates on pin 3 a square wave synchronized with Hsync pulses. No particular adjustments should be required if using 30 or 60 lines per second.
In the motor power driving circuit, keep the user-controlled potentiometer at mid-range and, while pushing the “Sync Phase” button to disconnected “CTRL”, set P8 so that its cursor voltage equals the motor voltage required to reach 12.5rps (from your previous motor evaluation tests), plus 5%.
Release the pushbutton. The P8 cursor voltage will then show some superimposed positive and negative pulses (unsynchronized) when rotating the disc. Perform a first adjustment of P6 and P7 to have these both pulse amplitudes at about 20% of the DC voltage.
Second Tuning Phase
We are now ready for a tuning in real conditions. After having switched the system off, connect the motor and the white / green LED to their respective drivers. Then turn the power supply back on while keeping the “SYNC PHASE” button pressed.
The motor driver has a “soft start” feature which allow limiting the motor starting current. However, after speed stabilization (a few seconds) it has the ability to relaunch and brake disc rotation with low impedance against inertia to keep the best possible stability. When the free run speed is stable, keep on holding the button pressed while adjusting the user-controlled potentiometer so that the stroboscope shows that the disc is spinning slightly too fast (apparent speed at about 0.25 RPS). If this cannot be obtained then fine adjust P8 to reach this result. Then release the button…
If you are lucky, the stroboscope shows that the disc rotation is synchronized! This will look like apparent immobility. Then you will just have to play with the button to get the angle mark (as shown in the Youtube video) at its appropriate location. A stable picture must then be seen on the back-light illuminated disc sector!
If you are not lucky, you’ll see the disc apparently coming back and forth under the stroboscope light, or even totally out of sync. This means that you have to fine tune P6 and P7, or even P8.
A synchronized disc showing an excessive angle jitter (apparent oscillation under the stroboscope light) is the result of a too strong inertia against motor torque capabilities. Try anyway to change the settings of P6 and P7 to change the error gain and see if satisfying results can be obtained.
A disc totally out of sync can result from...
- P6-adjusted control pulses too weak versus P7-adjusted control pulses
- P7-adjusted control pulses too weak versus P6-adjusted control pulses
- P8 set too far from the appropriate free-run speed
Color Decoder Tuning (If Used)
The color decoder can now be connected and adjusted if you planned to use it. In that case, you should be impatient to add blue and red to your green and unclear picture!
First, check the signal on pin 6 of the LF351N (front amplifier). You should see the same as on the color carrier channel from the computer, plus a DC component. Adjust P9 to get a peak-to-peak maximum amplitude (occurring during the biggest color bursts) at 6V.
Then check the signal on pin 7 of the LM311 (right part). Adjust P10 to obtain a square wave at a duty cycle of 50% and perfectly synchronized with the color carrier. (If the signal is stable, then it is synchronized).
Then check the voltage on one of the two 6.8-ohm resistors (LED drivers, bottom right) while slightly rotating the corresponding potentiometer (P11 or P12). Increase the level until the maximum voltage (during the biggest color bursts) reaches 1,7V. Process the same way for the second LED driver.
If everything happens as expected, the picture on the disc must now be in full color!
Gamma Correction Tuning
If this small circuit is inserted before the white LED driver (black and white version), you now need to worry about this kind of exponential function called gamma correction. The “scientific” way for tuning requires the creation of a test pattern where video lines are linear increases. Both transistors are intended for changing director coefficients within the ramp-up while the diode network sets the upper clamping. Then the three P2x potentiometers can be adjusted until the gamma curve is seen at the output, shown by the oscilloscope.
The more empiric way is the watching of real videos while adjusting the three potentiometers until it appears as pleasant, crisp and clear… ;-)
Using Disc Error Correction – Calibrating ToolIn a case where the line offsets (caused by disc’s holes position errors) make the result a bit disappointing, it's possible to apply a software correction at WAV file generation time, as explained in Part 4. The program “shiftune” (available only for Linux at this time) makes this far easier than manually editing “shiftdef.dat”, leading to many iterations.
Launch the corresponding “shiftune-x0” corresponding to your resolution. The possible arguments are:
- No argument: Classical scanning, left to right, top to bottom
- Argument “-i”: Scanning from left to right but from bottom to top
- Argument “-n”: If the PC audio output inverts video polarity
- Argument “-ni”: Activates simultaneously “-i” and “-n”
FINAL NOTE:This set of guidelines does not deal with the common—but necessary—checks that any careful experimenter has to perform for any kind of electronic hardware.
Here are just a few examples of checks you should run to avoid damaging the most critical and expensive components before starting the tuning procedure:
- Ensure that LED drivers cannot damage the LED if transistors are not mounted correctly (first tests with dummy loads).
- Ensure than VDD is well regulated at 5V in the color decoder before inserting the micro-controller on its socket.
- Ensure that transistor bias points do not show aberrant behaviors (pin-out errors, damage…) and that they are not overheating.
Thank you for following along with this project!