The Tektronix P6042 is a current probe measuring currents from DC to 50MHz in a range from ca. 5mA to 10A. It appeared around 1967. Despite its age it is still a very competitive device and a valuable tool.
Differing from its successors the probe head is hardwired to the amplifier through its cable. The amplifier is specialized and compensated exactly to its probe head. Epicenter of the whole device is the probe head, which is made from a split high-permeability core into which a Hall-element is indented. Hall-elements sense DC- and low-frequency AC-currents. Also 25 windings of wire are looped around the core to capture AC-currents with higher frequencies. Both branches are fitted together in the amplifier circuitry. To capture small AC-currents a high-permeability core is essential.
To prevent the core from saturation at already small DC levels the built-in Hall-Element senses the magnetic flux. A amplifier translates the sensor voltage into a proportional current and feeds it into the coil so that the generated magnetic flux just cancels the DC signal flux. The AC-signal current is modulated onto the DC-current.
The core halves have to be very precisely manufactured and aligned in their position to guarantee sufficient sensitivity and linearity. The core halves of the P6042 are fine lapped precision parts with additional 2-layer Mu-metal screening. When the moveable slider of the top half of the core almost closes the probe you have to push the button even further. A small guide clicks into a slot fixed to the lower half and a small metal ball presses down on a fixture plate that firmly presses both core halves together for the smallest airgap. Only then releases a electronic switch the probe head, resp. the amplifier output and the red lamp "Probe unlocked" is going out.
The lower fixed half also contains the Hall-element. Both structures are encapsulated in small Mu-metal screening troughs and potted with epoxy. Unfortunately they are quite sensitive mechanically. In many probes offered as defect or malfunctioning the cores are damaged ... and they are not serviceable. Replacement cores are almost unobtanium and if you find one be sure that the individual-to-each-core tuning resistors are part of the deal.
The amplifier casing also holds a row of very special and tightly tolerated and screened parts. But most of it can still be serviced or replaced by new subassembalies. Typical for Tektronix devices of these years is the enornous effort to which they went. All transistors are held by dedicated sockets, pairs are tightly matched and thermally coupled, the copper traces are massively gold plated. The outer casing is made from a outer frame and a braced drawer fixed to each other by just one screw. Even the 330µF power supply caps still measured more than 400µF and less than 0.3Ohm ESR!!
At first switch-on things looked quite ok, but the DC-Offset could not be dialed smoothly. Too coarse and erratic. So back to study the service manual and to check testpoints. Partly the values were a bit off and needed a new calibration, but overall everything seemed in good condition .... apart from the offset value of the Hall-amplifier. This special device (named M18) in a 12-legged OPAmp-like casing fortunately proved perfectly ok. So troubleshooting went on towards the probe head as it appeared that the Hall-element might show a defect. Bingo! when moving the head around the jumps in offset occured again. Now the cause was cleared to be a cable break resulting in a shortcut. As it happens all so often the location was found close to the entry to the strain relief boot. The cable itself is also a special thing where probabely no new replacement exists, but fortunately it is quite long and I just cut off approximately 20cm of length.
The current probe is basically functional again and may require only adjustement and calibration.
When it appeard in 1967 the probe costed 625$ which equals about 4.000$ nowadays. Modern DC+AC current probes from A-brands cost similar, base on the same principle and are hardly any better. What those lack though is the easy serviceability and they probabely won´t last half as long.
Further inspection revealed that the Degauss circuit didn´t work. I could trace it down to electrolytic (C65) which was totally flat. This cap generates the decaying amplitude of the Degauss-oscillator.
Since it was down the collector voltage of the oscillator transistors was permanently turned off. Snipping one pin and temporarily connecting a 100µF cap instantly produced nice oscillations but only with a decay of ~100ms instead of the specced 200ms. Either there is some leakage still to find or the 100µF caps TEK used were rather of 200µF in value.
Beeing a lazy cat I prefer the latter guess. Anyway it made it clear that even when the supply caps measure fine I will replace all lytics with new ones.
Also I found out that the Degauss switch and the current/div switch need some cleaning and I will replace the DC-Offset and current/div balance potis ... probabely with multi-ganged types like Vishays 534 series or similar.
In short ,,, a complete overhaul is ahead.
meanwhile I not only hacked and simulated all subassemblies into LTSpice, but also found in the depths of LinearTechnology´s App-notes a replacement circuit for the special M18 OPamp designed by Jim Williams and simulated it.
The output amplifier prooved to be a basically linear amplifier with very high gain. Astonishing how Tek managed to achieve 60dB of gain at 50MHz bandwidth with essentially only four transistors. Understandable that this required multiple tunable compensation networks and very tightly matched and screened transistors.
Possibly this could emerge in a future project ... to build the electronics completely new.
With new modern OPAmps it should be possible to build the amplifiers with increased bandwidth and improved behaviour as well as reducing the compensation effort. The Probehead itself seems to offer potential for 100MHz bandwidth.
This redesign could also tackle issues like the distinct temperature sensitivity.