The H.P. 606 and the Hickok 288X
By Bill Jones
This paper is not meant to suggest that the HP 606 and the Hickok 288x are in any way equivalent generators. The HP was a modern laboratory generator in its day. It is mostly a tube type generator - it used miniature tubes. It was never meant for radio service. It was expensive and its features were not necessary for any radio repair shop. Today however the HP is available at a moderate cost and offers those interested in radio repair and Amateur Radio work a bargain generator. I find it quite useful and even a joy to use because of its known frequency and signal level outputs. It is simple and straight forward to use.
The Hickok generator was really meant for radio repair and in its time it was a desirable generator for the radio shop and not an inexpensive generator. The Hickok is today and perhaps has always been a misunderstood generator because of its complex design. It was cleverly designed. It is difficult to use all the features of the generator without a good understanding of the generator. Without the manual the unit is somewhat of a mystery. The unit used octal tubes and the circuitry is simple compared to the HP 606 but it still is a proper generator for radio repair. The Hickok had a crystal calibrator and a frequency range greater than the HP. It also has an FM output. The 288X had a voltmeter for monitoring the receiver output. Hickok made a similar generator called the 277X without the audio meter, but otherwise I believe they had the same outputs.
The HP 606A signal generator was available in the late 50’s. Many consider it to be the ultimate in LC oscillators because of its outstanding performance. It is however, difficult to lift because of it’s nearly fifty pounds. The generator will output a three volt rms signal into a fifty ohm load. The frequency range is from 50 kc to 65 mc with overlapping coverage. The output stays constant to within one dB as the frequency is varied. The output is continuously variable from 3 volts rms to 0.1 microvolt. The frequency is accurate to within one percent. The generator was made for years as a laboratory generator. The generator is now popular for radio repair and Ham use. The generator went through the usual improvements with some changes in circuitry. Some additional options were added, but its frequency range and output were never changed.
The Hickok 288X was designed for radio repair and has about the same frontal area as the HP, but the Hickok is only about seven inches deep and weighs only about thirty pounds. To move the HP requires both hands. But the Hickok can be easily moved with one hand. The Hickok has a wide frequency range that covers 100 kc to 110 mc. It provides a wide range of frequency modulation. There is an audio output that is variable from zero to 15 kc. It has a 400 cycle signal for internal modulation, and it has an input that allows an inputting of a modulation signal. The generator also has a voltmeter for monitoring the output of a radio for alignment purposes. The R.F. output varies with frequency and is relatively small as compared to the HP generator. The 288X has an output in the millivolt range, but it is adequate for radio repair. I worked in radio repair for about ten years and did use the 288x during that time. I worked as an Electronics Engineer for many years and used the HP in the laboratory until synthesizers became available. I presently have both generators for use on my work bench.
The piezo-electric crystal was well known before WW1 and it is has been suggested that Sonar was in an experimental phase at this time. If so, the crystal was perhaps used as the first oscillator source. I assume that it would probably have been for high audio frequencies. The LC (tuned circuit) oscillator was likely the original means (not considering the spark-gap) for generating radio frequency signals. The LC oscillator was used during WW1. The invention of some of the well-known types of oscillators is recorded around 1918 (the Hartley Oscillator, for example).
In the beginning, vacuum tube amplifiers would often have problems with undesired or spurious oscillations. These oscillations were usually the result of the triode inter-electrode capacitance and the very nature of the triode will cause these capacitances to be very high compared to the screen grid tube. The screen grid tube was very helpful in preventing these spurious oscillations.
The oscillator, however, turned out to be basic to,much of the electronic technology that we have today. We could not have cell phones, television, computers or radio without the oscillator - not that those are necessarily good things. While a TRF receiver does not need an oscillator, the transmitter does.
In today’s terms the LC oscillator is considered to have poor stability and accuracy mainly because of the stability and accuracy available from the crystal and the frequency synthesizer. The synthesizer is stable and accurate because it obtains the frequency output from a single stable source such as the crystal. While accuracy and stability are related terms, they have different definitions. About the best stability an LC oscillator has is around one cycle per one hundred thousand cycles. It is difficult to obtain this stability (one part in ten to the fifth) without an oven for the tuned circuit. For a variable frequency oscillator, the stability and accuracy is a problem generally accepted. Accepted because the LC oscillator is so convenient since it provides a wide range of frequencies, and it is small and also inexpensive. It was all that was used for many years. If stability and accuracy was really required the quartz crystal was available.
Most of our LC signals generators are fairly unstable and inaccurate. I am thinking of the generators that we might use in radio service. If you would set the dial on many of these generators for a certain frequency, it would not be a surprise if it were found to be in error by five percent or more. After all, these radio service generators were seldom if ever calibrated except for one time – perhaps at the factory. In my work in the laboratory generators were calibrated yearly.
Some of the “good” radio service signal generators, the Hickok 288X for example, added a crystal oscillator for calibration purposes and this allowed an improvement in accuracy. Hickok advertised the 288X as a “Universal Crystal Controlled Signal Generator”. Actually, the crystal did not control anything. The term “Universal Crystal Controlled” was misleading and mainly used for marketing. The crystal was used only for calibration purposes, and the crystal was not very accurate or stable by today’s standards. The 288X used a single crystal for the 100KC and the one megahertz outputs. There was no adjustment on the crystal and both frequencies were usually in error, but it was a crystal oscillator. It was a great addition – the idea of an internal calibration was great. The origin and first use of the idea is unknown but it is a desirable addition. HP also provided 100kc and 1 mc. The HP had two crystals and each crystal frequency was adjustable. The crystal output was mixed with the main signal and an output was available for headphones. This mixer provided a beat note that could be heard when the main signal was a multiple of the crystal frequency.
The Hickok was considered a good signal generator for radio service. It would, however, not be noted for its stability compared to the HP oscillator. The output of the 288X was far from sinusoidal – the output was rich in harmonics. The actual output voltage was small and undefined, and the output varied widely over its frequency range. The 288X is a versatile and useful generator, and does many things – poorly. But it is still a good generator for radio service.
The 288X has a wide frequency modulation range that is variable. It has two ranges from 0 to 30 kc and 0 to 450 kc. In addition the generator provides a zero to 15kc signal output which is difficult to obtain without reading the manual. There is also a 400 cycle output plus an output for oscilloscope sweep alignment, but this output is not linear. The 288X uses four oscillators to obtain all its outputs. To provide all these outputs it uses only six tubes - two tubes are dual triodes with one tube being the rectifier. Again - it requires a through reading of the manual if you wish to really understand how to properly use the generator. The manual does provide a good operational table that indicates how to set the knobs for desired signal outputs – be sure to read the footnotes.
Again, the attenuator on the 288X is not the same for all the Hickok’s. I have had two of the generators, and one generator had a forty dB first step on the attenuator and the other had ten dB for its first step. Not that this matters much because the output of the generator is small and not well known. It has an output control that varies the signal level. I usually leave the signal level at its maximum and just use the output control knob for signal level adjustment, but the attenuator is there if needed.
The 288X went through several versions with some tube changes and minor circuit changes. There were versions that allowed an additional frequency range at additional cost. The oscillator provided reasonable frequency stability after warm-up because the oscillator has a double enclosure on the main chassis, and the cabinet then enclosed the main chassis. The generator has a decibel / voltmeter which allow monitoring the radio output during alignment.
Crystal oscillators can be about ten times better than the LC oscillator. The integrated circuit made the frequency synthesizer possible, but a stable source is still necessary for the synthesizer. The quartz crystal was the original stable source. Today, there are oscillator sources much more stable than the quartz crystal. The synthesizer is as accurate and stable as its source oscillator. Tube synthesizers were used in the 1950’s but they were to my knowledge experimental and difficult to use. They were much better than the LC oscillator but usually not as good as a crystal.
The crystal itself is very complex and there are special crystal types (cuts), and desirable crystal frequencies for maximum stability. To obtain the best stability from a crystal will require an oven and a vacuum enclosed crystal. The crystal frequency will change over time and this is called ageing. The ageing can be reduced somewhat by keeping the crystal drive low – that is to keep the crystal signal amplitude at a low level. The circuit should have an amplitude gain control circuit, and the oven itself should be analog instead of the on-off type of oven.
HP designed the 606A for laboratory work but the synthesizer made the LC oscillator obsolete –at least for the laboratory. The frequency accuracy and stability available today from synthesizers was unthinkable when the 606A and Hickok were designed. The synthesizer is very expensive compared to the LC oscillator. The synthesizer is difficult to repair and totally unnecessary for radio repair requirements.
The HP is mostly all tube (19 tubes) and a bit difficult to repair. I say –mostly all tube - because it does use silicon rectifiers in the power supply and some of the later generators use transistors in the power supplies. The power supply is complex because it uses several regulated voltages and also has a DC supply for the tubes.
I used an HP 606A for my engineering design work for several years before synthesizers became available. After retirement, I searched and found a 606 for my personal use and have never had a problem with it. It is one of my favorite pieces of test equipment. My HP is actually a 606B but has all the specifications of the 606A with an addition that allows the frequency to be changed a small amount with an input voltage.
The 606A provides means to check the sensitivity of a receiver since the output of the generator has a known amplitude accuracy and is variable down to less than one microvolt. It allows you to check the gain of a stage because of its accurate ten db per step attenuator with its infinitely variable output adjustment. The frequency accuracy is quite desirable for radio alignment. The signal leakage on the HP is undetectable by normal means.
The pass band of an I.F. amplifier is designed to be flat and have a certain bandwidth only at the design frequency. It will not be flat, and have the correct bandwidth at any other frequency. While this is not a problem, (who cares or who knows) for most inexpensive radios, however, there are some who want the pass band to be flat as designed by the manufacturer- Ham radio operators, for example.
It will generally be found that in most radios a good “flat” pass band is only obtained with a sweep generator even if you do have the correct center frequency. Without the correct frequency it is difficult to obtain a flat pass band. The flat pass band is important in some applications such as data transfer. The flat pass band is normally ignored by most radio service personnel. I like the Hickok because it provides sweep for the AM radio as well as the FM radio.
The internal structure of the HP generator is fairly massive and necessary because a small change in the structure dimension will result in frequency drift. The unit is designed around a thick cast aluminum base. The oscillator coils (inductors) are placed on a large turret that is a round plate about eight inches in diameter. The inductor wires are attached to contacts on the turret. The frequency selector control turns the turret to move the inductor contacts around to the oscillator contacts. The turret arrangement keeps all the wires in the oscillator circuit very short and thus improves both the stability and the frequency range available. There are actually two turrets – one for the oscillator and one for the output amplifier circuit.
The mechanical design is remarkable. The band-switch is unusual in that it disconnects itself from the turret after a band change. This is an important feature because when the band switch contacts are moved even slightly; there will be a small change in frequency. The HP has a frequency range for each band of about 3.5 to 1. This infers a variable capacitance range of over ten to one. This variation is difficult to obtain when band switching is involved.
The 606A has a sinusoidal R.F. output, which is not usual for the inexpensive signal generator but desirable and even necessary for some work. It also provides a sinusoidal audio modulating oscillator. The modulation is variable from zero to 100 percent. Two audio frequencies are provided; 400 hundred cycles and 1000 cycles. The oscillator that generates the modulation signals comes from the original HP type circuit that made the HP Company famous. The audio oscillator was the first product available from the company and the oscillator is still popular today.
There are two meters on the front panel of the HP generator. One meter reads the percent modulation. It is variable from zero to one hundred percent. External modulation is also allowed from a front panel connector. A second meter on the front panel monitors the R.F. output. The HP has an excellent attenuator and provides an output change in ten dB steps.
If I were only interested in radio repair I would be willing to rely only on the Hickok generator, but I have interests that make the HP useful for other work. The HP is difficult to find as compared to the Hickok but the HP can often be found in working condition. The 288X is easily found on the internet but it is usually found in an unknown operating condition – that is to say - it probably is not working. The output cable is normally in bad condition, and not so easy to replace. I would look mainly for a unit in good physical condition. They are not difficult to repair. I put a BNC connector on my Hickok and it makes it easy to use a coax cable with clip leads or a cable for the output. In any case a manual for the HP or Hickok is desirable if not necessary if one wants to use and repair their own equipment. Pictures of the two generators are shown below.
Questions or Comments? Please e-mail me at whnj
@att.net Thanks, Bill