GR 1568-A Wave Analyser
This is the first item of GR equipment that I have written about. I have been resisting (and continue to resist) GR gear that in it’s way, is superb and accurate, but fussy. The mechanical construction tends to be tricky and every item I have is hard to work on. I think that the electronics must have been designed by mathematicians, not engineers. This is not helped by skimpy, confusing manuals that I need a magnifying glass to read. Given GR’s reputation, there is no excuse for such poor manuals that contain difficult to read schematics with obscure descriptions and instructions that contain many errors. It is almost as though the construction and the manuals conspire to protect the secrets within!
The 1568-A usually comes with an XLR input connector at the top left (the third pin was used to provide a microphone energisation potential), this one has an input level control. (I have marked a full clockwise arrow to indicate calibrate next to this control.) I replaced the God-forsaken 1/4″ jack input with a RCA connector. My experience of the reliability of jacks is dismal.
The GR 1568-A is a wave analyser that operates over the range 10Hz to 20KHz. According to the manual it was “inexpensive”, I take that with a grain of salt. The filter is of the constant percentage bandwidth type where the width of the response peak measured between -3dB points is always 1% of the centre frequency resulting in more than 75dB attenuation an octave away from the center frequency. Again, according to the manual, it is particularly suited to analyse closely spaced components common to the vibration spectra produced by various types of machines (presumably including record cutting and transcription machines).
The analyser consists of three main sections, an input amplifier, the variable peak filter and an output amplifier / meter driver. There is also a precision step attenuator that consists of a maximum input control and a meter range control.
The filter consists of a synchronous cascade of two resonant (second order) filters, the first filter having high Q and the second low Q. This filter system is continuously tuneable from 10Hz to 20KHz, the synchronisation being accomplished by two precision wirewound frequency potentiometers that are geared together.
Well, in a word, tricky, this is my first experience of such a machine and I have found it extremely challenging. In principle, the first thing to do is to set it in calibrate mode at the centre of the frequency range of interest. (The manual does not specify whether centre refers to the log centre of the linear centre. For example for a range of of 1KHz to 10KHz, the log centre is 3.16KHz and the linear centre is 5.5KHz. This seems like pedantry I suppose and maybe if I understood the mathematics properly I would know.) The calibrator is interesting: When in calibrate mode, the output of the filter amplifier is fed back to the input. The forward gain is set using the input and meter range step attenuator by placing white dots on each control over each other at 12 o’ clock. The loop gain is controlled by the cal control and is to be adjusted such that the gain of the analyser is equal to the loss in the feedback network such that the system oscillates, the resulting signal being displayed on the meter. The cal control is operated so that the meter needle lies in the calibrate region on the meter face.
The system then may be adjusted to indicate the fundamental frequency of the signal to be analysed at some indicated reference level, either voltage or 0dB. Then sweep the frequency upwards, noting the frequency and amplitude of the harmonics. It is nice to set it in automatic so that the frequency range indexes automatically. The meter is scaled 1-10 and 1-3 also +2dB to -15dB. If the amplitude of the signal falls too low to be resolved accurately on the meter, the meter range may be stepped downwards in 10dB increments or corresponding voltage increments, thereby increasing the sensitivity by a known amount. The overall range is 100µV to 300V full scale. The analysing range of the system is given as 80dB so for example, if the fundamental is 1Vrms, then the smallest measurable harmonic would be 100µVrms in amplitude. I would think that battery operation is necessary to realise the highest sensitivity of the unit.
The intention of the automatic mode is interesting: That is the frequency control may be rotated by a drive from a GR1521 chart recorder so that as the pen moves across the paper, the frequency sweeps upwards, the range automatically indexing with each rotation. Type GR 1521 is fitted with a drive sprocket. The knob on the 1568 is replaced with a corresponding sprocket and placed on top of the 1521. The drive is hooked up with a chain and set to a slight backlash with an idler sprocket. The days before computer based analysis! It would be fun to see this set-up working, preferably demonstrated by a competent operator, not me!
CONDITION ON RECEIPT:
Completely dead. I expected that the two 9.6V Ni-cad batteries would be defunct however, it turned out that the meter coil was open-circuit. The wire was so fine that I could barely see it using a magnifying glass, and I am quite good at finding broken ends of fine windings and repairing them. It was a Jewell taut-band 100µA device. I scoured the panel meter offerings on Ebay and eventually took a chance on a nice looking 100µA taut band meter, it that looked as though I might be able to transplant the movement from this meter into the GR meter case. I got lucky, it worked out beautifully though the needle is a little short. I find it does not hinder reading the meter significantly and the linearity of the scaling matches the GR scale well. The picture of the analyser above shows it with the rebuilt meter. Here is the original movement:
Here is the transplant:
Another issue was a failed zener in the high Q section power supply, easily replaced. Later, one section of the 4 section precision wirewound frequency pot in the high Q section went open. I got lucky here in two respects, 1, it was the top section than is relatively easy to open and 2, all it took to repair it was to re-flow the resistance wire connection at one end. Phew! Here is the open pot:
Here is a view showing showing the high Q and low Q pots that are geared together. Without doubt I got lucky with the location and nature of the pot fault! NOTE, I took GREAT CARE to place index marks on the potentiometer shaft, rotor and case before disassembly!
Here is an overall view. Partially visible on the bottom right is the potentiometer rotor. The high Q board is on the left, the low Q to the right of it. Underneath the filter boards is the frequency multiplier switch that is rotated via bevel gears (visible in the picture behind the rotary switch) either directly from the front panel or by a motor when the unit is operating in automatic mode. This compartment is completely shielded, closure being completed by a plate over the top. To the right of the filter compartment is the PSU board and above that is the meter drive board. The capacitors are for the 10Hz range with the attenuator between. The two frequency potentiometer shafts are visible at the bottom, the printed switch cuts the filter path when the dial is crossing between ranges and the micro switch on the right indexes the frequency range motor when the unit is in automatic mode or when operating manually and crossing over up to the next range. This mode is not available when the unit is powered from the batteries.