MAX232

RS232 signals can be sent at voltages ranging from ±3 to ±25 Volts. This wide range of possibilities is dangerous when trying to interface with transistor circuits because the small end of that range will not drive transistors into saturation, and the large end will create surges through the transistors that is likely to break them. This problem is not unique to our purposes, and has thankfully been solved with an integrated circuit called a MAX232. This IC outputs a 0 to 5 volt square wave given any RS232 input, and outputs a ±7.5 Volt square wave given a 0 to 5 input wave. Based on our testing, the chip is very robust and reliable. We plan to use it to convert the RS232 signal into TTL before the current amplifier which drives the LED, and again after we filter and amplify the output of the detector.

The output signal of the MXS's X-ray detector looks like the figure below, taken from an oscilloscope. Its typical amplitude is 100-400 mV.

Our objective is to clean up this signal so it looks like a square wave with pulses of the same width as those we see from the output of the MXS. In other words, we should end up with a signal that looks just like the one we used to drive the LED. For illustrative purposes, if we get the signal to look similar to the waveform shown below we have met our goals.

In order to transform the MXS output signal into a square wave, we need circuits that amplify the signal and recognize whether the signal is high or low. Our approach is to use cascading inverting amplifiers, a peak-hold detector, and a Schmitt trigger.

While designing our amplifier circuit, we had to make sure that it would function as expected even at high frequencies. Most of the op-amps we had around the lab don't handle high bandwidths well, so we ordered some op-amps (AD848, datasheet here ) with a high gain-bandwidth product of 175 MHz, which means that our amplifier should be able to amplify signals with frequencies in the MHz range as long as the gain is relatively low. More specifically, we decided to support a maximum frequency of 65 MHz, so we made a series of cascading amplifiers in which the gain of each amplifying stage is 2.7. Our circuit schematic is pictured below.

The peak-hold detector is the core of our output circuitry. It allows us to transform the noisy pulses from the MXS into a comparatively clean waveform produced by tracking and holding the peaks of each pulse. Essentially, as soon as it detects an "on" pulse the circuit will follow that increase in voltage. As is the case with the MXS, even if there is noise in the signal that makes it move between many different voltage levelss, the peak-hold detector can "forget" about the small changes in voltage that occur throughout the "on" pulse. Only when the signal has been off for a specific minimum amount of time (determined by the resistor and capacitor values) does the peak-hold detector output a low "off" voltage. The basic topology of the circuit is shown below, with exact values and part numbers to be added soon.

We put the Schmitt Trigger after the low-pass filter output in order to make the output look like a clear digital signal (that is, either high or low). A Schmitt trigger has two threshold levels; as long as the signal is below the lower threshold, the Schmitt trigger will output 0V, but once the signal is above the higher threshold, the Schmitt trigger will output a 5V signal. While the signal is between the two thresholds, the output of the Schmitt trigger will not change. We are using a Schmitt trigger made by Texas Instruments, part number SN74HC14N. Its datasheet can be found here .

Our Schmitt trigger requires a signal input voltage between 0 V and 5 V, so we devised a way to shift our input waveform to meet that requirement. Shown below is a circuit diagram which takes the output of an amplifier (ranging from -5 V to +5 V) and uses a voltage divider to convert it to a signal between 0 V and 5 V.