For a couple of decates I worked in the microwave semiconductor business and a key part of that was making DC measurements. This requires forcing either voltage or current and measuring the resulting current or voltage respectivly. A lot can be learned from this type of testing.
The test equipment can be very simple. For example I made a simple 1 mA current source which can be used to measure the forward current of a diode. The Fluke 87V DMM uses a 1 mA current for the Diode test. Also made the Transistor Tester from the GE Transistor Manual "Safe and Simple Transistor Tester", in addition to a number of Semiconductor analyzers that work on three terminal devices.
Hewlett-Packard [ (Wiki), Agilent (Wiki), Keysight (Wiki)] made a number of test instruments that were used to characterize microwave semiconductors when I was working at Aertech.
Keithley
Prior to the introduction of HP Rocky Mountain Basic (RMB) in the early 1970s, all testing was done manually using analog equipment. I got my MSEE in 1965 so worked for a number of years using manual analog equipment. HP introduced the HP 8410 Microwave Network Analyzer in 1967. This was a manually operated analog instrument (Keysight). It was a huge improvement over using a Slotted Line (Wiki) to measure the VSWR of a microwave component.
By the late 1960s Aertech was making their own Tunnel Diodes. These were made one at a time by hand. There was no such thing as a wafer. The etching was done using an electric toothbrush with a special diode holder that allowed rinsing the diode in water and then, using the Tek 575 curve tracer seeing the peak current. The operator would go back and forth between the etch and water until the Ipk was near the desired value. After manual measurements of capacitance the cutoff frequency could be calculated. When introduced in 1968 the HP 9100 allowed doing the calculation much faster.
HP introduced the 2114, 2115 & 2116 Digital Computers in 1968 (HP Memory Project). A couple of years later they had RMB which was both an operating system and programming language aimed at controlling test equipment. This was before HP-IB so there was an interface card in the computer for each test instrument. In the case of the 5100 Synthesizer that means over a hundred wires.
Tektronix 575 & 567 Curve Tracer
The 575 transistor curve tracer (Wiki) came out in 1957 (TekWiki) and is a CRT (Wiki) based manual analog tester that displays the I-V curve (Wiki) of a device on the screen. I had no experience with the Tek 570 vacuum tube curve (TekWiki) tracer that came out in 1955. During my college years (1961 - 1965) earning a BSEE & MSEE from San Jose State (Wiki) some of the classes taught vacuum tubes and some taught transistor circuit design. As part of my home lap I bought a new Tek 515 (TekWiki) Oscilloscope from the Tek field office next door to Ford Aerospace in Palo Alto (Wiki). I got a tour of the Tek facility and the most interesting part was the "car wash" booth. It turns out that the 515 and pretty much all the Tek products at this time (late 1950s, early 1960s) were designed to be pressure washed. This was important in the days of vacuum tubes since a small leakage current would cause a tube to change it's operating point. So keeping the inside clean was important. In order to use a soap and water wash the Tek equipment was designed without using paper or other materials that would absorb water. For example instead of the common phenolic terminal strips that were very common then, Tek used a ceramic terminal strip. (See photo on TekWiki).
HP 4145A Semiconductor Analyzer
This is a digital version of the Tek 567 curve tracer. It has a CRT but the display is digitally driven, not analog. It was featured in the October 1982 issue of the HP Journal. This was the first time I'd heard of a Source Measurement Unit (SMU). The operating system for the 4145 needs to be loaded from an HP formatted floppy disk along with whatever program you want to run. The HP 1983, 1984, 1985 catalog shows the 4145A (pg 112) but does not have a price. By the 1986 HP Catalog page 359 the price was listed as $34,100 for the basic instrument with a list of accessories that could double that price. I expect this price is why I did not buy one of these. Inflation between 1986 and 2020 would make that about $79,759 or two to three times the cost of a new Cadillac.
HP 4142A Modular DC Source Monitor
The 1993 HP Catalog lists the 4142B (page 522) for $12,100 for the mainframe and $4,000 to $5,000 for each SMU plug-in. I used a number of these. In order to have Kelvin connections to the Device Under Test (DUT) they use Triaxial connectors (Wiki). So either you buy special cables or use the Pomona adapters to standard BNC cables. Note there are two flavors of triaxial connector, one that has 2 lugs and one that has 3 lugs. The photo on the Wiki page shows a 2-lug version.
Keithley 225 Current Source
This is a manual analog current source. It has not provision for remote programming of any kind. Note that HP power supplies had the capability of remote programming by means of a voltage input or resistance for many years prior to having a digital programming capability.
Range 0.1 uA to 0.1 Amp. 3 digits of precision, with up to 100 Volts compliance. 0.5% of reading, 0.05% of full scale accuracy.
Internally a Kelvin-Varley Divider (Wiki) is used on the three decade switches.
Keithley 220 Current Source
This is a modern current source that can be fully controlled by a computer. For example see the system for testing Tunnel Diodes. It is much more economical to use separate current and voltage sourced than a combined SMU.
I was watching a YouTube video about the Weston Sensitrol Relay and noticed the Keithley 2450 SMU. A check on eBay showed these selling used for over $5,000. A current source, like the Keithley 225 that sold for about $200 would make the same measurement when coupled with a voltmeter. This was the motivation to make this web page.
Fig 1
Fig 2
PRC68, Alphanumeric Index of Web pages, Contact, Products for Sale
Page Created 2020 May 15
