BULLETIN OF THE POLISH ACADEMY OF SCIENCES TECHNICAL SCIENCES Vol. 52, No. 3, 2004 Multiple-input floating-gate MOS transistor in analogue electronics circuit L. TOPÓR-KAMIŃSKI1*andP.HOLAJN2** 1 Institute of Measurements and Automatic Control in Electrical Engineering, Silesian University of Technology, 10 Akademicka St., 44–100 Gliwice, Poland 2 Institute of Theoretical and Industrial Electrotechnics, Silesian University of Technology, 10 Akademicka St., 44–100 Gliwice, Poland Abstract. Conceptions of analogue electronics circuit based on a multiple-input floating gate field-effect transistor MOS (MIFGMOS) have been presented. The simple add and differential voltage amplifiers with one and two MIFGMOS transistors and multiple-input operational amplifiers with their application have been proposed. One of them was used for the realisation of a controlled floating resistor. Results of circuit simulations in SPICE programme using the simple substitute macromodel of MIFGMOS transistor have been shown. Keywords: floating-gate MOSFET, analogue amplifier, voltage controlled resistor, multiple-input operational amplifier. 1. Introduction ating region is utilised in electronic circuits with very low supply voltage, even below one-volt [5]. Floating gate field-effect transistors have been applied in a digital circuit until quite recently. At present appear the propositions of application of FGMOS transistors in 2. Properties of FGMOS transistor analogue circuits [1–4] especially non-linear circuit’s in Floating-gate MOS transistor is generated by forming an the form of multiple-terminal electronic devices, called additional conductive layer, between control terminal G then multiple-terminal MOS transistor with floating-gate and channel DS isolated from the environment, called (MIFGMOS). They have the same basic properties as floating gate. For transistors equipped with larger quan- equivalent ordinary MOS transistors but widened by cer- tity of control terminals Gi than one (MIFGMOS), they tain additional features. The most significant of them, are contacting with that gate through the capacities Ci there is the ability of summing gate controlling input sig- created between them. Floating gate is contacting with nals as well as the possibility of reduction of threshold the channel CH through the capacity of oxide layer C0X U value voltage TH. In some technological solution, the and with source, drains and bulks through the capacities value of charge collected in floating gate can be addition- CFS, CFD and CFB. Values of those capacities depend ally controlled. These transistors can operate as normal on the area of input gates Gi, floating gate FG and chan- MOS as saturated or non-saturated within the region of nel as well as on thickness of oxide layerbetween then, strong inversion or atypically within the region of weak that is to say, on the shape of structure of the whole inversion called sub-threshold region. That second oper- MIFGMOS transistor made on semiconductor (Fig. 1a). Fig. 1. Layout of k-input FGMOS (a), equivalent circuit model (b), and symbol (c) * e-mail: [email protected] ** e-mail: [email protected] 251 L.Topór-Kamiński,P.Holajn Potentials of transistor terminals UGi, UD, US, UB, 3. Simple voltage amplifiers with MIFG- floating gate potential UFG and charge QFG collected MOS transistor on that gate as well as channel surface potential ΨS, exert the influence on the phenomenon’s taking place in Using MIFGMOS transistor with large number of input the channel and thus on drain current ID. Depending terminals Giin above-mentioned amplifier, simple sum- on transistor design, operating conditions and values for ming voltage amplifieras in Fig. 3 is obtained [4, 7]. individual potentials and capacities, one can state in many cases that floating gate potential is only a function of potential UGi of input gate terminals summed with weights depending on CGi and CT values, that means k UFG = wiUGi (1) i=1 where: CGi wi = (2) CT where CT is the sum of all capacities connected with floating gate. In the region of strong inversion and saturation of MIFGMS transistor with the channel of “n” type and underassumption that UB = 0, drain current for that transistor is described by the following relation: k 2 IDFG = β wiUGi − UTH (3) i=1 in which β = µC0X W/L is a transconductance parameter and UTH — threshold voltage. Drain current for that Fig. 3. Reversing/summing voltage amplifier transistor within non-saturated region has the following with one MIFGMOS transistor value: k IDFG = β wiUGi − US − UTH (UD − US) For operation within the saturation region that am- i=1 plifieris described by the following dependence: 1 2 n n − (UD − US) . (4) I w 2 U 1 2 − j U U − k U . 0 = w β w j = 00 j j (5) The most important properties of MIFGMOS transistor 0 j=1 0 j=1 are the following: additional input control signals UGi and possibility of threshold voltage UTH reduction. The And, at the same time following condition must be ful- circuits containing MIFGMOS transistors can be, basing filled: on relation’s (3) and (4) simulated using macromodel n [3, 6] (Fig. 2) composed of MOS transistor as well as wj Uj > (1 − w0)U0. (6) additional capacities of gates CGi and controlled sources j=1 w U i i. In order to obtain non-reversing voltage amplifier, two MIFGMOS transistors (in Fig. 4) should be applied. When the both transistors FGMOS1 and FGMOS2 will operate within the saturated region, the following equa- tion will be obtained: √ 2 I2 w2 I1 w2 w1 U0 = − + U1 w02 β2 w01 β1 w02 w01 = U02 + k12U1. (7) It describes this circuit when input U1 fulfils the following condition: w01 1 2I1 1 − w02 2I2 Fig. 2. Equivalent macromodel of k-input FGMOSFET U1bounD = − <U1 w1 w01 β1 w2 β2 252 Bull. Pol. Ac.: Tech. 52(3) 2004 Multiple-input floating-gate MOS transistor... 1 − w01 2I1 < = U1bounG. (8) w1 β1 where U1bounD, U1bounG — boundary values. If in the circuit (in figure 4) MIFGMOS transistors with next additional input terminals are applied sum- ming/subtracting voltage amplifiers shown in Fig. 5, is obtained. Forthe case when the both FGMOS1 and FGMOS2 transistors will operate within the saturated region, am- plifieris described by the following equation: √ 2 I2 wP I1 wP w1 U0 = − + U1+ w02 β2 w01 β1 w01 w02 Fig. 6. Multiple-input OA with the pair of FGMOS transistorsg wP w2 w3 w4 + U2 − U3 − U4. (9) workin in the saturation region w01 w02 w02 w02 Making an assumption that transistors are identical, have equal current supply and weight value, that is: wa1 = ... = wan = wb1 = ... = wb1 = w. (10) then it is described by relation: w n n U0 = A Uak − Ubk . (11) w0 k=1 k=1 In order that transistors work in the saturation region each of them must fulfil the inequality: w I w U − w 0 2 0 U k (1 0) w β + w TH (12) Fig. 4. Simple non-reversing voltage amplifier The simple practical realisation of such an amplifier being an integrated circuit realised in the CMOS technique has been shown in Fig. 7. Fig. 5. Summing/subtracting voltage amplifier Fig. 7. MIOA simple practical realisation in the CMOS technique with a use of two FGMOS transistors 4. Multiple-input OA with FGMOS tran- sistors working in saturation region 5. Multiple-input OA with FGMOS tran- sistors working in non-saturation region Making connection of differential OA with the amplifica- tion A to the pair of FGMOS transistors (Fig. 3), working If differential current amplifier with voltage output and in the saturation region it was obtained multiple-input transconductance Rm has been connected to the pairof voltage OA circuit shown in Fig. 6. FGMOS transistors from Fig. 3 working in non-saturation Bull. Pol. Ac.: Tech. 52(3) 2004 253 L.Topór-Kamiński,P.Holajn region in the way presented in Fig. 8, then multiple-input compensation voltage should have value: voltage amplifieris obtained. [8] |Uk| = UTH/wk. (15) 6. MIFGMOS transistor as floating linear controlled resistor For floating gate MOS transistor with four control ter- minals, operating within the non-saturated region, the dependence (4) describing it, is rearranged to the follow- ing form: IDS =(w1UG1 + w2UG2 + w3UG3 − US + wT UT + Fig. 8. Multiple-input OA with a pair of FGMOS transistors working β in non-saturation region 1 2 + UTH)(UD − US) − (UD − US) . (16) 2 Forthe same weight value of all input terminals it is w U U U U U U described by the dependence: Assuming: T T = TH, G1 = C , G2 = D, UG3 = US, w2 = w3 =0.5andw1 = wC , the following n n relation is obtained: U0 = wβERm Uak − Ubk . (13) U − U k=1 k=1 D S 1 RDS = = . (17) In order that transistors work in the triode region each of IDS βwC UC them must fulfil the inequality: That relation describes a theoretical model of floating n linear resistor, adjustable by input voltage UC (in Fig. 6). 1 ID UTH Uk UG = + . (14) The model of controlled floating resistor with one w 2β 2w k=1 MIFGMOS transistor, shown in Fig. 10, cannot be prac- In orderto obtain lineartransition function also for tically realised, since the sum of all input weights of that voltage lowerthen limit voltage UG, it is possible to transistor is always less them unity therefore whichever apply two pairs of complementary FGMOS transistors two of then cannot be equal to 0.5 at the same time. So, powered by E voltage with opposite sign and with com- in order to realise it practically, additional amplifiers com- pensated threshold voltage UTH. The possible structure pensating lowered values for weight w2 and w3 should be of practical integrated circuit of MIOA realised in CMOS inserted between terminals D and S and input gates G2 technique has been presented in Fig.
Details
-
File Typepdf
-
Upload Time-
-
Content LanguagesEnglish
-
Upload UserAnonymous/Not logged-in
-
File Pages6 Page
-
File Size-