Easy sampler for old oscilloscopes

Easy sampler for old oscilloscopes
Convert your old scope into a high bandwidth RF scope
by sv3ora

Do you have an old vintage oscilloscope that you want to use for fun? You can't afford a modern high bandwidth oscilloscope? At any case, the very simple device presented in this article will help you "see" this high frequency signal that you were missing with your old oscilloscope.

The device is called a sampler and more specifically an analogue sampler. To my knowledge, the technique of analogue sampling, was first used in the 50s and 60s from Tektronix. Tek was producing quite a big series of expensive analogue sampling plugins for their oscilloscopes. Those master pieces of engineering, could display waveforms of frequencies of several GHz on the low bandwidth analogue oscilloscopes of the era. How was this possible?

The magic is in the technique. You see, in an ordinary analogue oscilloscope, the front end amplifiers have to amplify the actual waveform without distortion or gain loss. The CRT must also be of high speed, so that it can trace this high frequency waveform accurately. Those things become very difficult as you approach frequencies of a few hundrends of MHz. Tektronix had developed special distributed plates CRTs for this purpose, which improved things much and they could reach frequencies of 100-450MHz at maximum. This was way better but still limiting.

Therefore Tektronix introduced in their products the technique of analogue sampling, which in it's simplest form (there are a few variations) works like this: Imagine a repetitive waveform, like a sinewave carrier. Instead of having the amplifiers and the CRT to "follow" the actual waveform, you sample one point on the first cycle of the sinewave and you store it's voltage in a storage device, such as a capacitor. After a few cycles (or actually hundrends or thousand cycles) you sample a second point. Of course, you have taken the first sample from the first cycle, whereas the second sample is taken on another cycle. As long as your waveform is repetitive (like a sinewave carrier), this does not matter, because each cycle is usually the same as the rest of the cycles. Essentially, you reconstruct a lower frequency waveform onto the CRT, out of multiple samples of different cycles of a high frequency waveform. Obviously this technique works for repetitive waveforms only.

The circuit above is a bare bones practical sampler and it is as simple as it gets. The 74AC74 requires a variable clock frequency (or sinewave oscillator) and drives the diode bridge with fast transitions differential pulses. The oscillator must have a high enough level for the chip to be triggered. During each differential pulse, the diodes sample a point on the input waveform and store its voltage on the 100pF capacitor. The 100nF at the output removes any DC component produced by this simple sampler. The input of the sampler is set to 50R and the output is connected to your low bandwidth oscilloscope.

So what's the catch? This very simple sampler has drawbacks compared to the more elaborate and expensive samplers. It requires a variable frequency clock. Each time you change the input frequency, the clock has to be retuned to display the waveform correctly onto the screen. Also each time you change the clock frequency the displayed waveform stretches or shrinks and it's amplitude varies way lot. Also, there is no way you can callibrate this thing. Despite these disadvantages, you can still make accurate measurements using the comparison technique described below, so it is still a valuable instrument.

The comparison technique is really simple. SW2 chooses between the RF input and a calibrated signal generator input using equal lengths and types of coaxial cables (and connectors). You switch SW2 back and forth and set the generator frequency and amplitude to match those of the actual input waveform on the CRT. When matched, both waveforms will be triggered at the same point and then you can read the frequency and amplitude from the generator. This would be the frequency and amplitude of your actual input signal too.

Note that the cables do not need to be expensive high frequency cables and SW2 doesn't need to be a special switch. With this technique, any losses in the switch or cables do not matter, because these losses are applied to both the input waveform and the signal generator waveform to be compared.

The picture above shows a 100MHz signal from a low distortion sinewave oscillator, displayed on a 4MHz bandwidth Tektronix 310A oscilloscope. The clock frequency (sinewave) was around 45MHz. You can perfectly use other lower clock frequencies too, but note that as you approach very low clock frequencies (in the KHz region) compared to the waveform you want to sample, you get aliasing effects. As you can see, this 100MHz signal triggers perfectly and it is sharp. Amazing, isn't it? With this little circuit, your vintage low bandwidth oscilloscope, can get a new life!

Back to main site