Really Expensive Stuff

 Some of these items cost a tremendous amount of money when they were new but today they are hardly worth their weight in scrap metal.

If they stop working you'd have to pay to have them carried away!

 Hewlett Packard HP8551B Spectrum Analyser


 HP8551B with the 5th harmonic of a 500MHz signal displayed at 2.5GHz

The top unit is valve-less, using transistors

The bottom unit is a "plumber's delight" handling the 10MHz to 42GHz specification

I suspect the lower limit may be a lot less as the dials imply that it beats a 2GHz oscillator with a VFO, tuning down to 2GHz

My Marconi sig genny has a similar arrangement which lets you generate zero Hz (is that meaningful?) when the dial is set to "0"!

 Hewlett Packard 608E VHF Oscillator

 Below, my Type 608E VHF generator, which isn't much use in a domestic repair workshop as it do not include FM. I suppose they would be more useful for overhauling aircraft equipment as they still use VHF-AM.

This monster, weighing in at around 100lbs covers 10-480MHz.

These equipments were made in the days before clever methods of frequency stabilisation had been invented and relied on mechanical ruggedness (=weight) to ensure that the output frequency was stable. Another requirement of these things, with their very accurate output attenuators (the dial on the bottom right), was not to leak any RF which might upset measurements.

Even with professional test equipment, RF leakage can pose problems. I was once servicing a marine VHF transceiver and was checking the distress frequency calibration, when the local Coast Guard popped up and complained about my test signal on his receiver. This despite the use of screened leads and a dummy load.

Click to see the HP608E circuit diagrams

And click to see the full manual (13MB)



Above, you can see a row of 8 rectifier diodes, and below on the left a pair of disimilar 6080 double-triode valves which have a very low value of anode resistance so waste little when used for say regulating HT.


 Below, the metal cover conceals a turret tuner.

 And with the cover detached you can see a good example of precision mechanical engineering. Two sets of coils can be rotated by the front panel bandswitch knob. I didn't remove the top cover, but this conceals three triode RF valves. These being a 4042, 4043 and an EC8010. Clipped to the chassis behind the scale is a special tool for extracting the 4042 and 4043 valves. Another valve is V9, a 6U8 which from the presence of the adjacent 5Mc/s crystal is an oscillator.



 Not only a set of valves, but lots of transistors. I counted the following 13 valves. V2 and V21 12AU7; V9 6U8; V10, V11 12AT7; V12, V14 6AU7; V13, V16 6080; V15 5651 (=86V stabiliser); V18 V19 6AH6; V20 6AL5. Inside the RF box are V6 RF oscillator and V8 RF amplifier disk seal triodes types 4042 and 4043 plus V22 RF buffer EC8010. Stop looking at this point because numbers V1/V3/V4/V5/V7 and V17 are not allocated.



 Above... I didn't think it was advisable to mount valves upside-down, and what's that bulb-like object.. a barettor?

Below, lurking in a corner of the chassis I spotted this 2N1544 Germanium PNP 106 watt power transistor, and also what appears to be a mixer diode inside the RF box. In the centre is a picture of a 4042 triode (similar to the 4043)



 Cossor Model 1052 Oscillograph



 To me, it's just an oscilloscope, but the label tells me it's a portable oscillograph. I read that the latter can make a permanent record and if so it must be via a camera fitted at the front. This example was made in 1952 and of course uses valves, 14 of them. These are two 7Y4, eight 6AM6, two 12AT7 and one each of 6F33 and 6AL5. Notable are the full wave rectifiers; the 7Y3 was a pre-war American valve which was made by Cossor after 1945, and the 6F33 which is able to be cut off by some 10 volts negative applied to its suppressor grid. The CRT is a type 89D. Pictures below...





 The labelling of the front panel controls is not particularly forthcoming when it comes to details, although it's clearly a twin beam scope, given Y1 and Y2, but timebase-wise is anyone's guess. The implication is that if the Y bandwidth is no better than 2Mc/s, and if that refers to the 3dB point, then the timebase must range up to say a few microsecond per cm.

I've now found an instruction book so can reveal some details:-

Time Base Ranges:


 Velocity Anti-Clockwise

 Velocity Clockwise






















 External Scan



The Y amplifiers are identical and have a 3 position switch controlling gain and bandwidth.


 Max sens.


 B/width from

 B/width to

 Undistorted deflection






 Full screen




 15c/s -250Kc/s









Y inputs are 1Mohm and 50pF


 And at the other end of the scale the cheapest piece of test equipment ever?

A bulb tester that came with some very early reflector bulbs


Return to entrance