DEVMS Test Plan

All testing here is based on rev 61 or higher.

Item	Results	Fixed	Details
Interface min limits	Rollover gets stuck	Fixed in rev 62	Using the buttons, see what happens when you go below zero for each displayed mode
Interface max limits	Rolls over to zero		Using the buttons, see what happens when trying to go beyond the highest value
Verify saw waveform	Not biased to zero	Fixed in rev 63-64	Make sure the sawtooth waveform counts incrementally. Make sure the min and max values are +DEV to -DEV.
Verify triangle waveform	Good		Make sure the triangle waveform counts up and back down. Make sure the min and max values are +DEV to -DEV.
Verify sin waveform	Good		Make sure the sin waveform looks like a sin waveform. Make sure the min and max values are +DEV to -DEV.
Verify sqr waveform	Not biased to zero	Fixed in rev 63-64	Make sure the square waveform looks like a square wave. Make sure the min and max values are +DEV to -DEV.
Verify rand waveform	Good		Make sure the random waveform verifies randomly between +(1/2)DEV and -(1/2)DEV.
Verify rtri waveform	A spike occurs	Still broken	Make sure the waveform is a triangle wave and that it starts at different locations in the waveform
Verify rsin waveform	Good		Make sure the waveform is a sin wave and that is start at different locations in the waveform
Performance	Slow	Improved ¹⁾	Reduce the time period down to zero, then vary the waveforms and listen for an unintentionally lengthened period.
Debug mode	Numbers did not show	Fixed in rev 63	Press F3 and see if the unit switches to debug mode. Press F3 again and make sure it switches back. Also, make sure the display now updates every cycle instead of every MAX_STEPS cycle
Verify gain control	Good		Make sure gain increases/decreases the output volume level
Verify volume control	Good		Make sure volume increases/decreases the output volume level
Verify bass control	Good		Make sure bass increases/decreases the output bass level
Verify treble control	Good		Make sure treble increases/decreases the output treble level
Verify period control	Good		Make sure increasing/decreasing the period causes the changing in waveforms to get slower/faster. A period of ~1000 should change step level about every 1 sec
Verify random period	Good		Make sure increasing/decreasing the random period causes random changes to the period
Verify deviation control	Good		Make sure increasing/decreasing the deviation causes the magnitude of distortion changes to get larger
Verify reset	Good		Turn on debug mode, configure any distortion waveform, press the play/pause button and verify all values return to their defaults

Performance

The system only has a 20MHz cpu and 2M of ram.
After adding the sin function I reached another limit of 200KB for the firmware size. Removing the original recorder.c code from Rockbox freed up the needed space.
Upon completing the sin() function, I could see noticeable slowness in completing each cycle when using a sin() mode.
These items were performed to make things faster
- Bit shifts were using in place of / 2
- char and short were used in place of integers
- The display is now only updated every MAX_STEPS ²⁾ cycle instead of every cycle ³⁾
- The method used to control period was redesigned
- Some waveform optimization was done
At best, these changes only gave a maximum distortion waveform frequency of 10HZ ⁴⁾ ⁵⁾

¹⁾ See performance section notes

²⁾ MAX_STEPS = 16 as of rev 64

³⁾ Except in debug mode

⁴⁾ This has yet to be measured

⁵⁾ This will vary depending on which waveform is used