US2431335A - Direct-current amplifier - Google Patents

Description

Nov. 25, 1947. R. v. LANGMUIR DIRECT CURRENT AMPLIFIER Filed July l5, 1945 Patented Nov. 25, 1947 DIRE CT-CURRENT AMPLIFIER Robert V. Langmuir, Schenectady, N. Y., assigner to Consolidated Engineering Corporation, Pasadena, Calif., a corporation of California Application July 15, 1943, Serial No. 494,808

1 Claim. l

This invention is concerned with D. C. amplifiers for small currents and especially with D. C. amplifiers included in the output circuit of a mass spectrometer.

A mass spectrometer is an apparatus for sorting ions into a plurality of beams according to their mass-to-charge ratio, i. e., their specific mass. These different ion beams are directed at an ion collector, so that the currents they bear are discharged. The several ion currents thus produced represent the mass spectrum of the material being analyzed in the spectrometer, for example, a gaseous mixture of different molecular constituents-say a mixture of hydrocarbons.

The ion currents thus produced are Very small and even when discharged through a large resistance of the order of 1012 ohms produce but a very low potential. They must be amplified greatly for purposes of recording and interpretation, for example by means of an ampliiier connected between the collector and a recording galvanometer. In many types of mass spectrometers the beams are unidirectional and produce direct currents. The amplification of these small direct currents Without distortion and while maintaining a high ratio of signal-to-noise is desirable.

There are several sources of noise in a D. C. amplifier, the term noise being employed to describe the randomly uctuating signal that appears in an amplifier due to changes in elements thereof. Such noise is present even when no signal is being amplified and is generally of the same magnitude regardless of the strength of the signal. It may be due to the production of gaseous positive ions Within the envelope of a vacuum tube due to the impression of potentials high enough to ionize residual gas therein, to thermal agitation of atoms and electrons within the input resistance, to statistical variations in the rate of electron emission from the cathodes of the vacuum tubes in the amplifier, or to the flicker effect which is apparently due to changes in the surface conditions of the emitting cathodes. These sources of noise are discussed at length by McDonald (The thermionic amplification of direct currents, Physics, Aug. 1936) who states that heater tubes are too irregular in their action to be of practical value in D. C. ampliers.

As a result of my investigations, I have discovered that, under certain conditions described in detail hereinafter, a heater type tube (specifically a type 38 tube, i. e. a pentode type of tube provided with a cathode [including heater filament] a control or signal grid, a screen grid, a

suppressor grid and an anode or plate and having a gain, as connected, of unity) may be employed without difficulty and in fact with advantage as the rst amplication tube in an inverse feedback operated D. C. amplifier, and that such an amplifier may be employed in place of an electrometer tube circuit having a large input resistance connected in the grid circuit, which has been proposed heretofore for amplication of small currents While maintaining a large signal-to-noise ratio, the term electrometer tube being employed to describe a tetrode type provided with lament, space charge grid, control or signa1 grid and anode or plata and having a gain, as connected, in substantial excess of unity, say in the neighborhood of 50. Such electrometer tube circuits usually incorporate the following features:

1. They utilize electrode potentials below the ionization potential of any residual gases Within the electrometer tube.

2. The electrometer tube itself, the input resistance, and the lead to the ion collector are mounted in a highly evacuated chamber (pressure equals 1 to 10 millimicrons Hg).

3. Temperature compensating and battery voltage fluctuation compensating circuits are used to maintain uniform amplification factor.

The practice of the present invention does not require the use of electrometer tubes or of the voltage and temperature compensating circuits associated therewith. Instead, my invention contemplates (in a D. C. amplifier for small currents and including a circuit containing a plurality of amplifier tubes connected in tandem and a current indicator in the output of the circuit), the combination which comprises an input heater-type amplifier tube connected adjacent the input of the circuit and provided with a cathode, an anode, a signal `grid and a highly evacuated envelope within which the anode, Vcathode and grid are disposed, means for imposing a negative bias on the signal grid, means establishing at the anode a positive potential, a negative feedback connected between the output and the input of the circuit, a grid resistor connected between the grid of the input tube and ground and adapted for connection to the source of small current, and a high resistance di-electric enclosing the grid resistor. A second highly evacuated envelope encloses the envelope of the input amplifier tube and in my preferred conductively connected to ground.

Preferably, the input amplifier tube has an input signa1 voltage amplification factor that is greater than unity. Moreover, it is desirable to establish at the anode a positive potential below that at which gas present in the envelope may be ionized by electrons emitted from the cathode.

Another aspect of my invention contemplates a mass spectrometer provided with a D. C'. ampliposing a negative bias on the grid, means for establishing at the anode a positive potential, a negative feedback connection between the output and input of the circuit, an ion collector in the mass spectrometer, a grid resistor connected between the grid and the cathode of the input tube and to the ion collector and the aforementioned second highly evacuated envelope.

In its preferred aspect, `the amplier of the present invention is a negative-feedback-operated D. C. amplifier utilizing a type 38 tube in its input or first stage, this tube being operated at the optimum heater lament current, (about ture or voltage fluctuation compensation circuits. Moreover, noises which originate either in the plate circuit of the type 38 tube or in any of the subsequent circuits yof the amplifier are negligible compared to the noise in the grid circuit. Hence the signal-to.,noise ratio is maintained at a Yhigh level.

These and other features of my invention will be more thoroughly understood inthe light of the following detailed description taken in conjunction with the accompanying two figures which are wiring diagrams of amplifiers incorporating the principles ofV my invention.

The circuits illustrated inFigs. 1 and 2 are the same except for differences in the manner of supplying potentials to the electrodes of the various amplifying tubes. In the apparatus of Fig. 1, the potential supplyY circuits are of a con-r ventional type, but those of Fig. 2 utilize potential dividers between successive stages, these dividers' being connected across a pair of bati tential Vwith respect to cathode shouldrbe about 5 volts.

Under these conditions the grid current is duce no appreciable effect on the input voltage.; There would be little advantage in making the Y 4 about 2 1O12 amps., and the effective internal resistance of the tube between the grid and cathode is about l012 ohms. With a plate resistance of 0.5 megohm, the gain of the tube is about 100.

At the input ofthe amplifier, there is a large input resistance R1 connected between the signal grid and ground,A A small `current such as that detected by an ion collector of a massV spectrometer passes through this resistance, thereby producing a voltage to be amplified.

With the tube operated as described hereinl above. the largest input resistanceY which it is possible to use is about 101 ohms, but this resistance is small compared to the grid-to-cathode Y resistance, so that iiuctuations in the latter proresistance larger than about 5 l01 ohms in the present circuit, because the use of larger resistances will not improve the signal-to-noise ratio. This arises because thermal noise increases as the square root of R1 while the shot noise and the signal increase in direct proportion to the input resistance. Becafuse of this, an input resistance is used which has a value such that the noise due to the shot effect is somewhat larger than the thermal noise.

A resistance R2 is connected between ground and the cathodes of the input and'output tubes., T1 and Ta. This resistance feeds back part of the output signal to the input circuit, thereby effecting stabilization of the amplifier according to well-known principles of negative feedback.

When the product a is very large compared to unity, where n represents thetotal amplification of the tubes T1, T2 and T3 without feedback and represents the feedback ratio, the current amplification factor of the. amplifier may be shown to be equal to: Y

tery. The values of the resistances R3, R4 andV R5 are so proportioned with respect to the in*- f ternal resistance of the ouput tube Ta that the movements of these. sliding l contacts. change thej V biases on the tubesTzy and T3 andV thereby con-v trol the amount of current appearing in thelout,

junction J2 between resistors R4 and rR5 is'at about Ythe same potential as the junction J1 be- Y tween `the anode of tube T3 and the resistance R3'. vA current measuring orindicating device (not shown) may be connected between the junctions J1, J2`. may comprise four galvanometer circuits of different sensitivity Vsuch as those shown asy ele-` merit 45 in Fie: 1, 0f: the Hoskins-Langmuir patent application, Serial No. 437,922, filed April 6, 1942,`

now Patent 2,380,439, datedv July 3,1, 1945;

vIn, one of the leads to the output circuit, I

Y utilize a monitor meter which` is placed Von theA panel ofthe mass spectrometer so that visualof'ionA indications arej given of the intensities beams striking theion' collector.

The signal grids of tubes T2 and T3 are-connectedl to. slide wireV contactsv S1, S2 of potentiometers included in `rthe plate circuit' of Vthe preceding tubes y'i1/and T2IfeSpeCtively. '.The,

Such a current measuring device Y put circuit and affecting the galvanometers, thereby affecting the Zero output level, i. e., the galvanometer deflection corresponding to no input signal. The output current is more sensitive to changes in bias of the signal grid of tube T2 than it is for tube T3. For this reason, the sliding contact S1 connected to the tube T2 is called the coarse zero control, while that (S2) connected to the grid of tube T3 is called the fine zero control.

In order to shield the amplifier from electrostatic pickup, the evacuated envelope which includes the rst tube T1 and the input resistance R1 is preferably made of metal.

In practice, the bias battery in the input grid circuit, the battery which supplies potentials to the screen, the lament supply for the input tube, and the feedback resistor and the plate resistance of the first tube are all mounted on a panel external to the evacuated chamber, and are connected to the appropriate part of the amplifier tubes through a shielded conduit.

In general, it can be shown that if the various stages of a D. C. amplifier are such that:

(where en vis the output voltage of the nth stage and en 1 its input voltage, and kn is the output voltage of this stage when there is no input to the D. C. amplifier), the instability in the output voltage, den due to some changes, clk, in the lcs of the individual stages is:

am de lil lll/2 MIIJZILS The small changes dk may be due to changes in plate supply, screen voltage, etc. If the dlcn are all approximately the same and if the 'rLs (amplification of factors) are all equal to or more, the only important term is the first. To keep this term small it is important to have a large gain in the first stage of the amplifier. A in Equation 2 is the operating gain of the D. C. amplifier, which is approximately l/.

From the foregoing, it will be seen that increasing the gain of the later stages to get more overall negative feedback will not reduce the zero instability due to unstable ampliiier elements. This can only be done by increasing the gain of the rst tube. But; there is no advantage in increasing this too much, for other instabilities such as changes in the heater voltage, which the amplifier does not distinguish from an input, become of greater importance. Assuming a large ri, the only advantage in increasing it is the improvement in linearity. The linearity of the present 3 tube amplier, which has a ,a of about 35 db., is better than 0.1%.

I claim:

In a D. C. amplifier for small currents, including a circuit containing a plurality of amplifier tubes Iconnected. in tandem and a current indicator in the output of the circuit, the combination which comprises an input heater type amplifier tube connected adjacent the input of the circuit and provided with a cathode, an anode, a signal grid, a screen grid, a suppressor grid, and a highly evacuated envelope within which the anode, cathode, and grids are disposed, means imposing a negative bias on the signal grid, means establishing at the anode a positive potential, a negative feedback connection between the input and the output of the circuit, a grid resistor connected between the signal grid and ground and adapted for connection to the source of small current, a second highly evacuated envelope enclosingthe envelope of the tube as well as the grid resistor and the connection between the grid resistor and the tube, at least one of the two envelopes being metallic and connected to ground.

ROBERT V. LANGMUIR.

REFERENCES CITED The following references are of record in the file of this patent:

UNITED STATES PATENTS OTHER REFERENCES Pub., Review of Scientific Instruments, vol. 7, Jan. 1936, page 47. (Copy in Div. 36.)

Radio Engineering, a pub. by F. E. Terman, McGraw-Hill, 1937, page 42418 only.

Images