The latest addition to my lab equipment is a Keithley 230 programmable voltage source which can output up to ±100V. Unlike many of the power supplies and voltage sources I have, the Keithley 230 is capable of operating in all four quadrants, meaning it can both source and sink current in either of the output polarities. Although the current capability is limited to ±100mA.

In most of its output ranges, the voltage accuracy is rated at 0.05% with a temperature coefficient of 0.005% which is quite impressive for a device in this class. The minimum voltage steps are determined by the range of operation and the resolution of the 12-bit DAC used.

Voltage sources like this one are commonly used in automated testing environments. As you can see on the rear panel, the 230 has trigger input/output BNC connectors and digital I/O connections. Along with the GPIB connection, operator can program the 230 to perform sequences of tasks automatically.

Upon removing the cover, you can see the typical clean Keithley layout inside. From the date code printed on many components, I suspect that this unit was built sometime in late 1987 or early 1988.

While I was doing some testing on the outputs from different ranges using my Keithley 196, which is a 6½ digit bench multimeter, I noticed that some of the ranges were slightly out of calibration. But considering the 30+ years of service life of this unit, it remains quite impressive that the calibration only drifted ever so slightly. And each of the adjustment points is clearly marked, making re-calibration an easy task with the right equipment.

Towards the upper right corner is the GPIB/digital IO daughter board. An enlarged picture of the board is shown below.

The GPIB controller chip used here is a TMS9114A.

Underneath the metal shielding cover is the sensitive portion of the circuitry which is responsible for generating and amplifying the voltage output by the DAC to obtain the correct output voltage.

The two large transistors seen below (MJ15011, MJ15012) are for the final output stage after the the bootstrapped op amp circuitry to drive the output between the ±100V range.

The picture below shows the closeup of what is underneath the shielding cover. At the heart of the Keithley 230 is an Analog ADDAC80N 12-bit digital to analog converter. The temperature coefficient (tempco) of the voltage reference internal to the ADDAC80N largely determines the overall temperature stability of the voltage source.

Note the damaged component towards the bottom left of the picture below. It is a 100 nF bypass capacitor for the DAC on the +15V rail. It did not cause any noticeable issue during my testing and after I dissembled the boards I replaced it with a new one (see in picture above).

The remaining board located at the bottom of the case contains the power supply and the digital portions of the circuitry. In the picture below, you can also see the vertically mounted 7-segment display board.

And here is a closeup of the microprocessor portion. The microprocessor used is a classic Motorola 6800 chip MC6802P. I have dumped the firmware binaries (version B13) and they can be downloaded towards the end of this post.

Here are a couple of pictures showing the linear regulator on the bottom board next to the power transformer. Note the optoisolators in the picture. These are used to isolate the control signals.

The Keithley 230 programmable voltage source has 100 onboard memory locations and each location can be programmed independently to store unique values of voltage/current and duration. With this capability, we can generate very precise arbitrary waveforms at low repetition rate.

The pictures below illustrate using the Keithley 230 to generate a pulse train and a staircase signal respectively.

A video of the teardown and experiments can be found below:

Keithley 230 Firmware (B13) download

Be Sociable, Share!

6 Thoughts on “Teardown of a Keithley 230 Programmable Four-Quadrant Voltage Source”

  • Happy New Year!!

    Inspired by your video on the Keithley 230, I found one on eBay for my lab. For the price I paid, my expectations were not high, however what I received was as close to “like new” as any eBay used I have bought before. Except for the 10V range, the unit was in spec. I have several other Keithley meters and have had no trouble calibrating them to spec, so I thought the 230 would be easy as well. The calibration instructions are simple if read carefully and followed exactly. Like “word for word” exactly. (where’s the fun in doing that?)

    The only remaining problem was that the 10V range “+19VDC” calibration point could not be reached due to the trim pot hitting the stop.The “-19VD” was reachable however. If I input “+19.000,” my calibrated 6.5 digit bench meter reads “+19.031671” which is a tad over 10mV higher than the spec limit. I checked the 5V supply (5.01 \ish) and the 15volt rails: -15.0027 and +15.257. The regulators are 7815’s and 15.257 is within their expected range from the data sheet. The interior of the case was “lab-clean” and there were no obviously bad parts. And, no I did not poke my greasy fingers on the precision resistors or even remove the analog board. I did remove the EMF shield to inspect, but it was in place for calibration.

    Were you able to get your 230 calibrated to specification? Do you have any advice? My first instinct would be to replace the trim pot on the +10V op amp–based on some experience with vintage gear mechanical components.

    Any ideas?


  • Hello Gary and Kerry,
    Donal here. I have a few schematic drawings that might help. First, many thanks to Kerry for the pioneering work. Next, what’s the best place to send the documents? They will be PDF files..


    • If you send the PDFs to me (kerryw at I can post them here. Or if you can provide me with the links I can link them in as well. Thanks in advance!

Leave a Reply

Your email address will not be published. Required fields are marked *