The plain circuit diagram clearly means that the software must do all the work.
Such a simple arrangement must have a few drawbacks: - The sampling rate is not impressive at 50k samples per second, assuming 8 bit samples and one channel
- The samples may not be entirely equally spaced. It may be that samples are missed altogether when Windows is busy doing something else
The site proposes that you do not use a general purpose opamp for the comparator: An FET input type opamp will certainly present the required high input impedance, but it may not like to operate near the negative supply rail. Surprisingly, if you do not trim the output offset voltage, you will sooner or later run into trouble, even though only comparator operation is requested. Resistors of 50k between pin 1 and 4 (and also pin 4 and 5) seem to cure the problem (for a pinout such as the one corresponding to the TL071), but beware that some opamps' output may not swing low enough to switch modern 3.3 Volt ports to the logic zero state.
The day that you would make an ADC converter from individual components is long gone, and prices have dropped considerably, so there is no reason to favour tracking over successive approximation converters. Here is the starting source code and program file, anyway. High frequency response will be much attenuated with a tracking converter. Strictly speaking, you should only accept a sample after an edge is detected in the comparator output. If you do not want noise in the least significant bit, you must only accept a valid reading after a positive going edge, only.
A PIC microcontroller such as the PIC18F252 can be interfaced to the parallel port, and features a much faster converter on-chip, but would not be quite as inexpensive. A very similar arrangement was used (a very long time ago), to look at speech waveforms, and also the waveshape of musical instruments, though the circuit could be put to other uses, especially if high conversion rates are not needed.
By the way, some people claim that you can use old parallel ports for input after setting outputs to a logic one and getting external circuits to pull the pins low at will. The site measured less than 100 Ohms of effective resistance between the power supply and port output, even for a ten year old port. I cannot see you pulling that low for any practical amount of time without damaging both the port and probably the external device, too. Will you please double check that the output of the comparator is connected to pin 10 (Acknowledge) of the Centronics port. The site has not verified that the comparator output is protected against overload when sinking excessive current.
| The opponent is a fool. What is the one who loses to the fool? |