Keithley 619 Electrometer

I picked this thing up for cheap on ebay market as defective. It was stuck in its boot cycle. Before we delve in: some explanation

Volt meter Operation

What’s an electrometer?

In short a voltmeter with an extremely high input impedance, so high it is almost unmeasurable. By employing a few tricks this machine can also be made into a current and an ohm meter.

As a voltmeter the input stage is simply switched as a unity gain opamp followed by an A/D convertor. The only special thing is the actual input stage itself. As opposed to being constructed with a normal opamp the electrometer uses paired jfets to obtain a very high input impedance.

In Ampere meter configuration the opamp is switched into differential mode. The output is connected to a buffer (green) and sent through a ranging resistor (red). A traditional ampere meter sense a current through a resistor.  Assuming a sensing resistor of 1 ohm is used , a current of 1 ampere would drop 1 volt across this sensing resistor.This drop is called the burden voltage.

The electrometer does not have a burden voltage. The incoming current flows through the sensing resistor into the buffer stage. This buffer stage absorbs the current. The electrometer steers the buffer in such a way that the voltage at the + terminal is equal to ground.

Amp meter Operation

A word of caution is in order as the correct operation of the machine depends on the ranging resistor and the driving capabilites of this buffer. Certain electrometers, like this machine, have a buffer that can create very dangerous voltage. 

The model 619 uses a +260 volts (+Vh) and a -260 volts (-VH) rail. In open loop this can cause all kinds of nasty side effects from test subject destruction to operator electrocution. Therefore these machines have a safety system built in. 

The compliance voltage detector will limit the buffer driver voltage to a user settable limit. by default after power up the compliance falls to 5 volts. Only when you really need the high voltage mode should you enable it.

Complete Block Diagram

The image on the left shows the entire construction. 

The input opamp is clearly visible. The buffer is shown as two transistors feeding + Vh and -Vh into the feedback path. The feedback resistances close the loop. 

Additional switches set loop gains and allow the machine to ‘null’ all offsets by digitizing them and the mathematically substracting them during measurement. 

Only the 2 ampere range uses the traditional ampere meter construction using a 0.1 ohm sensing resistor.

in Ohms meter mode a 1 volt reference , in combination with the feedback resistances, makes a current source. 

Eye Candy

IMG 3932

Front Panel

The inside modular buildup. From left to right : two input channels, the A/D analog section , A/D Digital section, Isolator , CPU board and GPIB board

IMG 3935

The GPIB board uses a Motorola 68488 GPIB chip in a nice side brazed ceramic dil package.

IMG 3938

The main CPU is built around a 6802 processor, two 2Kbyte Eproms and 2Kbyte of Ram organised as twin banks of two1024 nibble rams (2114 type)

IMG 3939

The front panel interface drives the LED displays and scans the keyboard.

IMG 3943

The A/D convertor digital sections is basically a huge counter that performs the timing related functions of the dual slope convertor. It uses its own crystal as maste clock reference.

IMG 3945

The analog section of the A/D convertor holds the input amplfier, comparator and reference.

IMG 3947

The venerable LM199 temperature controlled buried zener is is used to create the reference voltage. The AD542 opamp in metal can is the input comparator. Due to the very high impedances parts of this system is built on teflon stand-offs. Cleanliness is key here. 

Any flux residue or other contaminant will cause errors. 

IMG 3948

This is one half of an input stage and contains the high voltage generator (transformer on the right) and the driver amplifier (heatsinked transistors)

IMG 3951

The second board houses the input switching and ranging as well as the input amplifier.

Under the metal can we find again the teflon standoff construction. The metal can AD515 is the electrometer opamp.

The back of the machines shows the two electrometer modules. These function independently from each other and are isolated.

The machine can do differential and offset measurements between channels.


At powerup the machine runs a selftest , illuminates all segments, then shos the firmware revision and then should begin measuring. Mine didn’t. It was stuck showing the firmware division. The service manual did not mention this kind of behavior and was of little help.

So the base check is : are all voltages present. I quickly found that the 5 volt analog rail was collapsed to 1 volt.

I started by pulling out all modules and only leaving the CPU and front panel board attached to see if this would restore the rail. Maybe there was a short in one of the analog boards. Checking the analog 5 volt rail now showedover 13 volts. That’s not good. This means the regulator is completely shot. If it were a mere short the regulator should make a nice clean 5 volts. Replacing the regulator made a nice clean 5 volts rail.

The machine now gave an isolation error. This indicates there is a problem commuicating through the isolation logic that splits the common circuitry off the input channels. Well, that was progress. And, as the modules were not installed yet this looked normal. 

A quick powercycle and module installation later the machine booted correctly and proceeded to measuring. I verified correct operation by comparing voltage readings against a calibrated 34401 on different ranges and the machine was spot on. Current ranges were tested as well and all was well within spec.


Service Manual

© Vincent Himpe 2013