single-chip frequency converter

8/8/2019 Single-Chip Frequency Converter

1/4

C O N V ER T E R S A R E B A S I C

F carrier of a signaldio receiver.Frequency conversion, or hetero-

1 .from m ixing, their

9-IvlHz input and a 2-MHz7 and 11 MHz.Building a frequency converter is

w than it's ever been becausemixer in an 8-pin DIP, as

. The NE602 was originally de-

descriptionThe NE602 uses a double-balanceduencies, not that of theF input or LO. You can thus connec t

the output of an NE602 directly to areceiver without overloading it . Witha conventional mixer, you'd have toadd a tuned LC circuit to eliminate theLO output. The NE602 LO is alsowell isolated from its RF input; youcan thus connect a receiving antennadirectly to the RF input terminals ofthe IC without worrying a bout radiat-ing the LO signal back out through theantenna. This is important in direct-conversion receivers, where the LOfrequency is so close to that of theinput, that the two ca n't be isolated bya tuned LC circuit.

PRODUCES outputs at the sum and dif-ference of input and LO frequencies. In thecase of the NE602, since it's double-bal-anced, both the input and LO signals areabsent from the output.

The combination of the differentialamplifier and mixer in the NE602 isknown as a Gilbert cell. The m ixer hason-board voltage regulat ion, anddraws 2.5-3 mA at 4.5-8 volts. Forbest performance, bypass the powersupply with a 0.04-pF capacitor asclose to the 1C as possible. T he ab so-lute maximum supply voltage is 9.0volts, but a 9-volt battery often ex-ceeds that, and 9-volt wall transfor-mers o ften deliver as much as 11 volts.For safety, use 1000-ohm dropp ing re-sistor Rl as shown in Fig. 3; using aZener diode, you can use automotivepower supplies up to 18 volts.The RF input and mixer output canbe either single-ended o r balanced asshown in Figs. 4 and 5. Using a bal-anced input reduces harmonics , whilea balanced output gives better sup-pression of the input RF and LO sig-nals. However, even in the simplests i n gl e - e nd e d c o n f i g u r a t i o n , t h eNE602 gives much better perfor-mance than the one-transistor mixercommonly found in receivers.The input and output impedancesof the NE602 are about 1.5K at lowfrequencies, and decrease with in-creasing frequency. The input signal %should be weak to prevent harmonic s; ?the third-order intercept point is for a r;;-15 dBm input, but the recommended 8


2/4

level is -25 dB m or below. That corre- The input signal is amplified prior dled on the chip. That makes i t easy tosponds to 68.87 millivo lts into 1.5K if to mixing; the voltage gain is about build many different oscillator typesyou use direct coup ling, or 1 2.82 mil- 10. Thus, a receiving converter built with few external components.livolts into 52 ohm s if you use im ped- with the NE602 can increase a re- Figure 6 shows some of the mainance matching. The NE602 works ceiver's sensitivity. The NE 602 LO is versions; there are many others. Thewell with microvolt-level signals from a transistor with connections to its NE602 can be used as an oscillatoranten nas. base and emitter, with biasing han- without the mixer. One way is to sa m-~ l ehe L O output at pin 7 : better way

amplifier (2647 and mixers Q2-Q3an d Q4-Q5 s called a Gilbert cell.

. . . .current limiter (Rl) and integrator (Cl) , as well as for isolation. In (a), + 4.5-8.0 volts DC isthe n ormal operating range of the NE602. n (b),R1 drops voltage , and is use d since a +9-volt battery can go higher, and a +9-v olt w all supply can produc e up to 11 volts. In (c) ,an+8-18-volt DC supply is regulated using 8.2-volt Zener Dl .

(a) being for no impedance matching, (b ) or inductive matching (c) for capacitivematching. By contrast, (4 s for a balan ced input with reduced second harmonic.

is to unb alance the mixer and u se it toamplify the LO signal, as shown inFig. 7 . The unbalance is created by a10K resistor froin one input pin toground, w hich changes the bias volt-age slightly. The output level of suchan oscillator is abou t 0 . 5 VAC P-P.Basic crystal oscillatorMany f requency conver ters a recrystal-controlled; Fig. 8 shows themost basic version. The low side ofXTAL l and C 2 can be returned eitherto ground or to VCc; the latter is 11101-ecompact, because pins 6 an d 7 areadjacent to Vo (pin 8 ).The values ofC1 and C2 are imp ortant. If C1 is toolarge, or C2 is too small, there's toomuch feedback and the osc i l la torwaveforrn is distorted, with a strongthird harmonic. If C l is too small orC2 is too large, oscillation doesn'toccur.Some suggested values for C1 andC2 are shown i n Fig. 6 along withformulae for calculat ing them. Athigh frequencies, C1 can be some-what than the value shown becausestray capacitance does some of thework. The values shown are for thebest sinusoid. If you want to be surethat a relatively inactive crystal willoscillate and don't mind harmonics,make C1 three times larger. The thirdharmonic from such a circuit could beused for VHF. There's also a lowerfrequency limit; the unmodified cir-cuit will oscillate with a 45 5-kH z ce-ramic resonator, but not a 100-kHzcrystal. Adding 22K from pin 7 toground will increase the oscillatorgain, and improve your chances w ithlow-frequency crystals.Precise frequency controlA crystal won't necessarily oscil-late at its exact rated frequency. Thereare two kinds, series- and parallel-resonant. They're electrically identi-cal, the only difference being that se-ries-resonant crystals are cut to anexact frequency, whereaa p arallel-res-onant crystals are cut slightly longer,so as to r e sonat e i n depe nden t lyslightly below their rated frequency.For tha t reason, para l le l - resonant


3/4

1il1 ; ~ ~ERSA) ~ r $ [ i ca b c

FIG. 5.-OUTPUT CONFIGURATIONS. Here. (a ) is the simolest sinale-ended aooroachwithout impedance matching, (b) is a single-ended a ppr oa ih for a ti n e d LC circ'u'it load,and (c) s for a balanced approach for better supp ression of input and LO signals.

FIG. 6.-BASIC NE602 OSCILLATOR CIRCUITS; (a) is Colpitts crystal-controlled, (b) isColpitts LC-tank-controlled, (c) is Hartley LC-tank-con trolled, and (4 s controlled b y anexternal os cillator. Many other configurations are poss ible.

URATION for an NE602. To make an LOsignal appear at OUTA (pin 4) an d O UT B (pin5), IN A (pin 1) is grounded through R1.

f MHz I ~1=100p~i f l C2=1000pF/ f

10 1 32 I i n n

frequency of oscillation to their ratedvalue. In Fig. 8 , C1 is this capacitor,but it's usually larger than 32 pF andhas less effect than the one depictedhere.Th us , at 10 M Hz , parallel-resonantcrystals oscillate about 100 parts pe rtnillion (p pm ) below their rated fre-quency, w hile series crystals resonateabout 300 ppm above. A parallel-res-onant crystal can be pulled up to itsrated frequency using a small variablecapacito r in series with it, as in Fig. 9 ,letting you adjust the oscillator as de-sired. However, even without this ca-pacitor, the frequency error won't bemore than 300 ppm (0.03%).Overton e crystal oscillatorAbove 20 MHz, crystals oscillatein overtone mode, and the oscillator

F I G . 8 . -F U N D A M E N T A L C O L P I T T SCRYSTAL OSCILLATOR. Note that thejuncture of XTALl and C2 can go to eitherground or Vccneeds a tuned LC circuit to select thedesired harm onic. For example. a 27-M Hz third-harmon ic crystal can reso- .nate at 9 MHz (fundamental) or 45MHz (f if th harmonic) . The NE602

FIG.9.-A VARIATION ON FIG. 8, includingC3 to adjust the frequency of XTALI (par-allel-resonant), bringing it up to its ratedvalue.

data sheet recommends a modifiedColpitts oscillator for overtone crys-tals, but the Butler oscillator in Fig.10 gives much better results. Its crys-tal is series-resonant, and L, nd C1are tuned to the crystal frequency.This circuit is reliable to at least 60MHz. Just adjust L, and C1 until os-cillation occurs. By adjusting thistuned LC circuit, you can trim thefrequency by about 50 ppm; for great-er variation, use a parallel-resonantcrystal in series with a variable capac-itor for adjustments.

OSCILLATOR. with C1 as trimmer. Here:L1=1300 pH , and both L1 and C1 have to betuned to the frequen cy of XTAL1.Frequency doublerFigure 11 shows a crys ta l -con-trolled frequency doubler with notuned LC circcits. That circuit isuseful in the 20-40 M Hz rang e, butthe same method could be used withovertone crystal oscillators for evenhigher output frequencies.

The doubling is achieved by feed-ing the LO from pin 7 into the mixer.The output is 2 x f (where f is theoscilla tor fundamental f requency) ,while the difference frequency is zero(o r DC) , disappearing due to capacl-tive coupling.The output still contains some ener-gy at the LO frequency and isn't pure,gut is good enough for hobbyist pur-poses. A tuned LC circuit can easilyprovide pure output. O f the oscillatorsshown here, this is the only one thatcan't be used with the mixer, becauseone mixer input is occupied (althou ghyou could feed a signal to the othermixer input).Figure 12 shows a Colpitts LC os-cillator using coils and capacitors.Here, L, forms a resonant circuit withC1 and C2 in serie s, plus C4 in paral-lel. Also , C3 blocks DC from pin 6 toV, or ground; it has little effect onthe resonant frequency. F igure 12 alsogives formulas for component values.At very high frequencies, a 22K re- nsistor from pin 7 to ground (no t V,,) rwill change the bias point and in- L;COcrease gain. o


4/4

PRODUCES a sine wave at twice the fre-quency OF XTAL1. Note that output istaken only from OUT a (pin 5), while OU T A(pin 4) is left open.

FIG. 12.-COLPITTS LC OSCILLATOR.Here: L1=7 pHif, Cl=C2-C3=2400 pFif,where f is in MHz.

VERTS LONGWAVE signals from 350-500kHz up to 4.35-4.5 MHz, enabling them tobe received via a shortwave receiverplugged into J1.Longwave receiver converter

Figure 13 shows a frequency con-verter front end for a shortwave re-ceiver to receive longwave signals(3%-500 kHz). It mixes the incom-ing signal w ith the 4-MH z signal fromthe LO. For example, 400 kHz incom -ing produces 4.4 and 3.6 MHz. Theshortwave receiver will receive thesignal if tuned to either frequency.The input has a tuned LC circuit toprevent spurious response.If the receiver is set to 4.4 MHz,then without the tuned LC circuity o u 'd l i s t en to 4 0 0 k H zan d 8 .4 MH z,because each gives a 4.4-MHz outputwhen mixed with the LO. The tuned

LC circuit at the input selects one andrejects the other. This circuit was at-tached to a shortwave receiver, andimmediately received several long-wave navigational beacons in nearbystates. A long wire antenna works,but loops pick up less noise becausethey are directional.Direct-conversion receiver

A frequency converter can shift fre-quencies up or dow n. However, if youshift an R F signal down to au dio, youget an audio signal. This is calleddirect-conversion reception, and candemodulate Single-Sideband (SSB)and Mo rse c od e C ontinuou s- Wave(CW) transmissions. It demodulatesAM, but there's a whine if the tuningisn't perfect.Figure 14 shows such a direct-con-version receiver for the 40 meter ban d(7.5 MHz), that was able to receiveseveral amateur radio stations using a3-foot whip antenna. The designcould be refined; tuning would beeasier with a variable capacitor in-stead of an adjustable coil.

CEIVER for the 40-meter (7-MHz) amateurradio band, where CW is directly down-converted to audio.

ConclusionThere are basically three RF circuittypes-amplifiers, osc illa tors , an dfrequency converters. The NECi02makes frequency conv ersion easierthan ever. Both it and related IC's w illeventually become basic buildingblocks of RF design . This article hasonly scratched the surface of the pos-sibilities for.the NE602. In an IF sec-t i o n , i t m a k e s a g o o d p r o d u c tdetector. By mixing audio with RF, itcan act as an AM o r DSB m odulator.By mixing audio with audio, the

NE602 can be the heart of an ultra-sonic l istener (by down-convertinghigh-frequency audio) or a speechscramb ler to add security to telephoneconversatins.You can get NE602's at $2. 00 ea ch,plus $2.5 0 per order postage and han-dling per order, from the Small PartsCenter, 6818 Meese Drive, Lansing,MI 48911, (517) 882-6447; there 7s,n ominimum order . They ' re a re a l soavailable from Arrow Electronics,Schweber Elec t ronics , and manyother Signet ics d i s t r ibutors , wi th$25.00 typical minimum orders. Besure to specify whether you want theNE602N (8-pin DIP) or NE602D(surface mount package).You also may prefer to order theNE 602A , which will be replacing theNE602 imminently; it has somewhatimproved intercept chara cteristics, re-sulting in less harmonic generationand intermodulation distortion. Tospecify the desired package type,you 'd re fer to the NE602AN orNE602AD. We would like to thankPhi l Anzalone, Ali Fotowat , andCraig Hirtz of Signetics for their in-valuable assistance in preparing thisarticle. R-E

POWER LINE GLITCHEScontinued from page 42

13-16. Figure 13 show s a norma lpower sinusoid, Fig. 14 shows har-monic distortion, Fig. 15 shows abrief glitch, and Fig. 16 shows a briefpower outage.ConclusionThose graphs shown in Figs. 8-16show only a few of many possibledisturbances. Pow er glitches are com-mon and readily identified. Most areeasily fixed, the culprit often beingpoor wiring, bad grounding, or loadswi tching-al l can be cor re c tedcheaply. The most com mon, practicalcounterm easure is to install a separatepower line from the circuit-breakerbox involved to the device being inter-fered with, like a PC.Power monitors make identifica-tion easy, but they're generally tooexpensive, and would be needed tooinfrequently, to warrant purchase bythe average hobby ist. They can , how-ever, be rented for short periods, on anas-needed basis, letting you derive thebenefits of their technology withoutmaking a major investment. R.E

single-chip frequency converter

Documents