Current mirror
In electronics, the current mirror is an elementary transistor circuit with which it is possible to derive a further current from an existing reference current. The current mirror enables currents to be copied and scaled, and thus represents a current-controlled current source .
motivation
The current mirror circuit is primarily used as a sub- circuit in analog integrated circuits , such as differential amplifiers , operational amplifiers and comparators , for example for setting operating points in amplifier stages. It is also possible to shift the DC voltage level of analog signals.
MOSFET current mirrors are also used as electronic loads, with the output resistance of the current mirror serving as the load resistance of an amplifier.
In circuit technology , a replica of an existing current is often required. With current mirrors it is possible to generate an identical or a fixed ratio to an existing current. For the latter variant, transistors are used that are identical or shaped in terms of their design, but are not of the same area - so the ratio of the currents can be determined via the area ratios of the active areas, gate area or emitter area, solely via their areas.
If the transistors of a current mirror are closely spaced (same temperature) and have been created by the same manufacturing process (i.e. on a chip ), precise current mirrors can be built; this is not possible with discrete circuits .
The properties of operational amplifiers can only be achieved through the internal use of current mirrors. In conjunction with NPN and PNP transistors, this means that circuits can also be implemented that are impossible with electron tubes .
Functional principle of the current mirror
Current mirrors are controllable current sources with the highest possible output resistance. They have a low input resistance at the feed point of the reference current and become voltage-controlled current sources through an upstream resistor.
Circuit with bipolar transistors
Bipolar transistors , if they are not operated in a saturated state, behave like a current source on the output side; they are therefore well suited for current mirrors.
A simple current mirror consists of two transistors, as shown in the picture on the right. The transistor Q 1 has its collector and base terminals connected to one another. If an input current I e flows through the transistor Q 1 , a base-emitter voltage U BE is established . The base connections of the two transistors are connected so that the same voltage U BE is applied to both transistors .
A base-emitter voltage arises at transistor Q 1 , which is linked to the input current solely via the transistor properties and the temperature.
If the properties of the transistors Q 1 and Q 2 and their temperatures are the same, the dependence of the collector current on the base-emitter voltage is the same and the same collector current flows through Q 2 as through Q 1 . A current mirror is therefore obtained.
Clarification
The base-emitter voltage U BE is approximately linked to the input current I e , the temperature voltage U T and the reverse current I S via the following relationship
(Part of the input current I e flows as a base current into the transistors Q 1 and Q 2 ; this part has been neglected here.)
The output current I a through the transistor Q 2 depends on U BE , the following also applies:
Current mirrors are therefore often built from similar transistors located on a chip . They are then created through the same manufacturing process and have the same temperature properties.
If the base currents are taken into account, this results to be exact
Thus, in the case of a simple current mirror with bipolar transistors, there is an error that decreases as the current gain factor increases. Another error arises from the early effect , which causes the output current to depend on the voltage by modulating the virtual base width. The associated parameter Early voltage shows sample variations , should be large for good current constancy and increases with the physical base width and thus usually with decreasing current gain. Transistors with a high current gain therefore reduce the error due to the base current, but deliver a less high output resistance - the output current changes more with the output voltage.
In integrated circuits, the use of multi-emitter transistors enables the current to be multiplied or reduced by connecting the multiple emitter connections in parallel. If, for example, Q 2 is implemented as a multiemitter transistor with three emitters in the above circuit , the collector current of Q 2 is greater by a factor of 3 than the collector current of Q 1 . If, on the other hand, Q 1 is designed as a triple multiemitter transistor, the collector current of Q 2 is reduced to 1/3.
Circuit with MOSFETs
If the simple current mirror is constructed with identical MOSFETs, the relationship between the input current I e and the gate-source voltage U GS is obtained when both transistors are in saturation:
The following applies to the output current I a :
(k n represents a component-specific quantity. In the picture on the right, it is an n-channel MOSFET, hence k n ). Thus we get for the ratio of the two currents:
Since all parameters such as W, L and U th are the same for identical transistors and U GS is the same due to the interconnection, these values can be shortened and obtained .
The voltage U th is the threshold voltage at the gate from which the channel begins to conduct.
A deviation from this ideal behavior arises due to the channel length modulation of the transistors, which, however, has been neglected here. Errors result from the temperature dependence, among other things, on U th ; therefore and because of the required uniformity of the layer thicknesses and thus the component parameters, such current mirrors are preferably constructed monolithically (on a chip ).
The disadvantage of the MOSFET current mirror compared to the bipolar current mirror is the voltage dependency of the controlled output current.
Examples
The table shows typical examples of current mirrors in bipolar transistor technology and the equivalent MOSFET circuits. The resistors in the bipolar current mirrors are optional, they are unusual in MOSFET technology.
Essential criteria when selecting a current mirror circuit are the minimum supply voltage, the output resistance and the accuracy requirements for the mirrored current.
| Simple current mirror | 3-transistor current mirror (supported current mirror) |
cascode -Stromspiegel | Wilson current mirror (regulated current mirror) |
|
|---|---|---|---|---|
| Bipolar |
|
|
|
|
| MOSFET |
.svg)
|
.svg)
|
.svg)
|
- Widlar current mirror
- The Widlar current mirror , named after its developer Robert Widlar , is a variation of the simple current mirror and represents a particularly simple current mirror with only two transistors. The current ratio is set via the emitter multiplicity (gate multiplicity). Alternatively, the current ratio for bipolar transistors can also be set via a negative feedback resistor in the emitter branch, as shown in the adjacent figure, if the accuracy requirements are not too high. The Widlar current mirror can be obtained from the simple current mirror by short-circuiting the emitter negative feedback resistor shown on the left in the reference branch of the simple current mirror. Due to the strong current dependence of the transformation ratio, the Widlar current mirror is usually only suitable for constant currents.
- 3 transistor current mirror
- The transistor in the middle reduces the error caused by the base current of a bipolar transistor. If you add a resistor in the middle of the bipolar version, which leads from the common base to ground, this current mirror becomes considerably faster. This circuit is mainly used in frequency-dependent emitter circuits.
- Cascode current mirror
- With a simple current mirror, the output current depends on the output voltage due to the finite output resistance of the transistors. This effect can be reduced by cascoding the transistor on the output side. A transistor is also inserted on the input side to set the operating point.
- Wilson current mirror
- Further variants are the Wilson current mirror and the extended Wilson current mirror . The latter comprises four transistors and offers improved linearity compared to the Widlar current mirror.
In order to meet the accuracy requirements, current mirrors are built as symmetrically as possible, usually in quad layout or common centroid layout . In MOSFET current mirrors only gates of the same length and width are used, the mirror ratio is therefore only determined by the number of transistors. In order to reduce the effects of structuring or deposition in the edge areas of the electrically active areas during production, the current mirrors are surrounded by so-called dummy gates (a method commonly used in microelectronics). For applications with high accuracy requirements, the voltage drop across the drain-source path must be as identical as possible and the gate connection must be kept currentless, otherwise the gate-source voltage of the various transistors would be different, which would lead to large errors in the mirror ratio.
See also
literature
- Ulrich Tietze, Christoph Schenk: Semiconductor circuit technology . 12th edition. Springer Verlag, Berlin 2002, ISBN 3-540-42849-6 .
Individual evidence
- ^ Hans R. Camenzind : Designing Analog Chips . 2nd Edition. Virtualbookworm.com Publishing, College Station, Texas 2005, ISBN 1-58939-718-5 ( Online - Chapter 3 Current Mirrors : Pages 3-5).
- ↑ Patent US3320439 : Low-value current source for integrated circuits. Applied on May 26, 1965 , published May 16, 1967 , inventor: Robert J. Widlar.