Daqarta for DOS Contents
SIGNAL GENERATION CONTROLS:
YAMAHA OPL2 SYNTHESIZER:This THUNDER.ADC driver makes use of the built-in Yamaha OPL2 music synthesizer chip to generate signals. This chip has been held in low regard as music synthesizers, especially when compared to the newer "wavetable" synthesis, because the sounds they can produce are "too limited". But they generate low to mid-frequency audio sine waves at least as well as a typical laboratory function generator (about 1% distortion or -40 dB), suitable for many experiments. And although not really designed to do so, they can act as very good sources of pulsed or continuous random noise... which is actually very hard to obtain otherwise, and which can't be made at all with a wavetable!
OUTPUT LOADING REQUIREMENT:The Thunder Board output is AC coupled, which means that it has a series capacitor to block internal DC voltages. The output must drive a resistive load in order to allow the capacitor to charge properly.
FEEDTHROUGH:There is one problem when the Thunder Board is used as a signal generator: A small portion of the input signal leaks through to the output, inverted in polarity. This only happens while data is actually being collected, not in Pause mode. The feedthrough is worst for input frequencies near 300 Hz. The normal worst-case scenario would be when the input is near the maximum level before the AGC threshold (1 mV RMS), at which point the output will contain 13 mV of input signal if the thumbwheel is at maximum. This is 49 dB below the maximum output level (3.5 V RMS).
The feedthrough rises with input signal level. At the maximum input that the AGC can handle (25 mV RMS), the output will contain 43 mV of input signal, or 38 dB below maximum output.
The feedthrough goes down proportional to the output level when the thumbwheel control is used, but is independent of the Level control setting. So you should try to keep Level as high as possible, and use an external attenuator to control the actual stimulus level, if feedthrough is a problem.
Feedthrough will probably not be a problem for most purposes, however. If the response you are monitoring is an image of the stimulus, you may not even know that this is happening since it will add an insignificant amount to the output. For example, if the stimulus is a 300 Hz sine wave and at maximum output of 3.5 V the response is 1 mV, this just means the stimulus goes down (due to the inverted polarity of the leakage) to 3.49 V... a decrease of only 0.03 dB.
If the response is different from the stimulus, such as a neural discharge in response to a tone burst, there may be more cause for concern. In this example, the subject may be able to detect a click from the fed-through neural spike more easily than the intended stimulus.
Off:The synthesizer output is disabled. This reduces the signal about 68 dB below the Level setting. Set Level to Off also if you need even lower output.
On:The synthesizer output is continuously ON, at the Frequency and Level settings selected. The remaining options, which control event timing, are ignored.
Burst:Tone bursts are created with the selected parameters. One burst is output per sweep in Sequential mode only (RTime key option off). This option is not available in RTime mode, and will disappear from the menu. If you switch to RTime while Burst is active, the Off output option will be selected instead. When you turn RTime off the Burst mode will again appear and become active.
The level of the completely "off" portion of the burst is about 68 dB below that of the completely "on" portion.
While in Burst mode, the acquisition sweep is synchronized to the burst, regardless of Trigger Control Menu settings for Trigger Mode, Source, Slope, or Level. However, Trigger Delay and Cycle controls behave as usual.
Gap:The converse of Burst mode, Gap keeps the tone on continuously EXCEPT for the specified duration. Note that Rise and Fall still apply in the normal sense: The tone is on, then at the start of the sweep the Fall begins, and after the selected Duration the Rise begins to return the tone to its original level until the next sweep.
As with Burst, this is only available in Sequential mode and vanishes in RTime mode. However, unlike Burst, if Gap is active when you switch to RTime the output will go to On.
Also as with Burst, the level of the completely "off" portion of the gap is about 68 dB below that of the completely "on" portion.
Frequency:Sets the tone frequency in Hertz, from 2 to 24011 Hz. The synthesizer is not capable of 1 Hz resolution at higher frequencies, so you will notice coarser resolution there.
The synthesizer output frequency response is given here, normalized to the maximum output level at 500 Hz. Measured values are arranged in sequence, NOT TO SCALE:
dB Frequency dB 0 500 0 -0.1 113 1100 -0.1 -0.5 53 2300 -0.5 -1 36 3300 -1 -2 24 4800 -2 -3 19 6200 -3 -4 16 7500 -4 -6 11 9900 -6 -9 8 13500 -9 -12 6 17400 -12This is the output of the sythesizer alone, as measured by independent means. It does NOT include the input response of the Thunder Board.
Output Modulation Artifacts:The OPL2 synthesizer runs at its own fixed output sample rate or 49.7 kHz. A consequence of this is sampling "modulation" that is evident at frequencies above a few thousand Hertz. This is not just an artifact of the difference between the synthesizer sample rate and the acquisition rate. It is due to interactions between the synthesizer sample rate and the output signal frequency, and is "really there"... you can see it with an analog oscilloscope.
This may be of no consequence if you are only interested in the spectral content of the signal, say to determine threshold frquency responses of your subject or system under test. You can check the spectrum for yourself and see if it meets your needs.
Noise Output:You can convert from tone to Noise mode by attempting to set a Frequency of 0. The noise has a wide bandwidth and does not have any apparent repeat pattern. The non-repeating aspect is very important for perceptual experiments, since many other noise generation techniques have obvious repetition patterns that may compromise results. (You can easily hear noise patterns lasting several seconds before repeating... they sound like ocean waves rolling in to the shore.)
All the other settings apply to Noise as they apply to Frequency: Rise, Duration, Fall, and Level can be used to create noise bursts or gaps. The Level in noise mode has been set so that the RMS value of the noise is the same as that of a pure tone of 500 Hz at the same Level. The peak noise levels will be slightly higher.
Note, however, that this use of noise generation is an undocumented function of the OPL2: It may not work properly on your particular board.
In particular, on some boards the noise is accompanied by a positive or negative DC offset voltage. This DC voltage will be blocked by the board's output coupling capacitor and will thus generally cause no problems for continuous noise, but for bursts or gaps there will be a transient "thump" that may not be acceptable.
Some boards may show this DC offset problem during a given session, but if you Quit Daqarta and restart, the problem may vanish. The presence or absence of the DC seems to be "set" during board initialization, and remains that way for the remainder of the session. If you need to give noise bursts, you may thus want to check the output at the start of each session.
Rise:The Rise value is the time in milliseconds for the burst to go from 10% to 90% of its final value. Unlike standard laboratory practice, however, the shape of the rise is not a sinusoid but an exponential. This is a legacy of the musical roots of the OPL2 synthesizer chip. The beginning portion of the burst thus stays at lower levels longer than an equivalent sinusoidal rise, and at longer rise times there can be a substantial "tail" (or should that be "nose"?) that precedes the 10% start of the rise. This pre-tail is not counted in the rise-time value, but it nevertheless delays the onset of the burst, so you can't determine the 90% time by calculations based on the rise-time alone.
The rise is composed of many small steps in level that normally happen too fast to notice. But as the rise becomes slower, say 16 msec or longer, each step takes long enough that the overall effect becomes more like a staircase than a continuous curve. Even though the steps are small in amplitude, you should check to see if there are any audible or measureable consequences for your particular experiment.
Duration:This is the duration in input samples (not time units) from the start of the Rise portion to the start of the Fall portion of a burst. For a gap, it is the number of samples from the start of the Fall to the start of the Rise. Duration does not take into account the length of the Rise or Fall portions, so if you set the duration too short for a long rise time, the burst may never reach the 100% "on" portion. Similarly, a short gap duration with a long Fall time may never get all the way off.
Fall:The Fall value is the time in milliseconds for a burst to go from 90% to 10% of its final value. This is an exponential and not a sinusoidal decay, so there can be a long tail after the 10% level. Unlike Rise, however, the level steps that are used to generate this are finer and less noticeable until much longer Fall times. As a result of the musical origins of the OPL2 synthesizer, the range of Fall times is about 6 times longer than the Rise times to better approximate conventional musical sounds such as plucked strings.
Level:The setting here is in dB relative to the volume control thumbwheel setting on the back of the board. All values other than zero are thus negative, but you don't need to include the minus sign when entering them directly... Daqarta will supply it automatically since there can be no values above zero.
When adjusting the Level with the cursor keys, the up-arrow gives more output, which means the dB values become smaller (less negative).
The thumbwheel setting should be made so that when Level is set to 0 (maximum output), the board output will produce some maximal calibrated output from your system. You should then take steps to lock the control into this position to prevent an accidental change of the system calibration. A piece of tape covering the thumbwheel is probably the simplest solution.
Note that with the thumbwheel at maximum and the Level set to 0 dB, there may be a very slight clipping on the positive and negative peaks of the output. Setting the thumbwheel back just a bit to about 3.5 V RMS will eliminate this distortion.
The OPL2 synthesizer chip used by the Thunder Board has a very good level control circuit, with 0.75 dB steps from 0 to 47.25 dB. Since this means you can't always get even dB values, Daqarta will pick the closest setting.
However, due to the output input feedthrough problem, it is best to leave the Level set to 0 dB and use an external attenuator, if at all possible. The Level control can be used for fine adjustment if your external attenuator has only coarse steps, but attenuation provided internally does not reduce the interference pulse, while external attenuation does.
Also, even if there were no feedthrough, the 47.25 dB range is probably too small for many experiments. Consider that the human ear can perceive sounds over a range of 120 dB or more, and you can appreciate the need for more control. However, unless you already have an attenuator or want to build one, you would probably be better off upgrading to a new sound board like the Sound Blaster 16. This not only has 16-bit input digitizing ability, but sports more input sources, TWO independent output channels, no feedthrough, and a huge attenuator range of 135 dB.
TRADEMARKS:Thunder Board is a trademark of Media Vision, Inc. Sound Blaster and Sound Blaster 16 are trademarks of Creative Technology Ltd.
THUNDER.ADC REVISION HISTORY:
THUNDER.ADC INDEX:NOTE: The actual Daqarta THUNDER.ADC index allows you to move through it by hitting the first letter of any entry. We apologize for this "dumbed-down" Web substitue: You may scroll through it as usual, select a letter from the following line, or simply use your browser's "Find" function to search this page.
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
AC Coupling, Input AC Coupling, Output Address Parameter A: AGC Threshold Table Artifacts, OPL2 Output Modulation Attenuation, Output AUTOEXEC.BAT, SET BLASTER Line Automatic Gain Control (AGC)
BLASTER Environment String Burst Mode, Output
Calibration, Input Capacitor, Output Coupling Compression, Input Response Configuration Parameters Control, Thumbwheel
DSP Version Duration, Output
F: Parameter (Help File Omit) Fall, Output Feedthrough Frequency, Output Frequency Response, Input Frequency Response, Measuring Input Frequency Response, Output
Gain Calibration Parameter G: Gap Mode, Output
Help File Parameter F:
Impedance, Input Input Control (Menu) Input Source IRQ Parameter I:
Level, Output LPT Printer Port I/O Parameter L:
Modulation Artifacts, OPL2 Output
OPL2 Synthesizer Chip Output Level Output Loading Requirement Output Mode (Off, On, Burst, Gap)
Response, Input Frequency Revision History Rise, Output
Sensitivity and AGC SET BLASTER Line Signal Generation Controls Synthesizer Chip, OPL2 Source, Input
Thumbwheel Control Trademarks
Questions? Comments? Contact us!We respond to ALL inquiries, typically within 24 hrs.
25 Years of Innovative Instrumentation
© Copyright 1999 - 2006 by Interstellar Research