Nederlands Radiogenootschap

Transcriptie

1 Tijdschrift van het Nederlands Radiogenootschap DEEL 26 No A survey of physics and technology of high frequency transistors by W. E dlinger * ) Paper given at the conference of the Nederlands Radiogenootschap in The Hague, on February 17th, Introduction A rem ark ab le tren d of the tra n sisto r developm ent and p ro d u ctio n of th e la s t y e a rs h as b een th e successful im p rovem en t o f tra n s is to r p erfo rm an ce a t high freq u en cies. W T ils t during 5 y e a rs, in th e e a rly tra n s is to r d ay s, it seem ed im possible to o p e ra te tra n s is to rs a t frequen cies sufficiently h ig h er th a n 1 M c/s, th e ra p id im provem ent of th e la s t y e a rs h a s lead to w a rd s o p e ra tin g freq u en cies of sev e ra l 100 M c /s. A s a m a tte r of fa c t, it h as been believed fo r some tim e, th a t th e tra n s is to r w o u ld be a useful device fo r low fre q u en c y ap p licatio n s only, th e h ig h er freq u en cies being a dom ain of th erm io nic v alv es. T he in v en tio n of th e so-called 'd rift tra n s is to r', i.e. tra n s is to rs w ith a n in te rn a l a c c e le ra tin g field, b ro u g h t an a b ru p t change to th is belief. L a te s t com m unications b y se v e ra l la b o ra to rie s re p o rt a b o u t o p e ra tin g freq u en cies as high a t 1000 M c /s an d no th e o re tic a l lim itatio n w ith re s p e c t to frequen cy resp o n se h as been found up to now. A sim ilar developm ent is going on in the field of sm all signal sw itching, and also the high frequency po w er field is subje c te d to m any re c e n t im p rovem en ts. I t is th e aim of th is p a p e r, how ever, to give a survey prim arily of the sm all signal am pli- N.V. P h ilips G loeilam penfabrieken, Sem iconductor boratories, Nijmegen, The N etherlands. D evelopm ent La-

2 108 W. Edlinger fication field, as th is field seem s to be m ost stab ilized w ith resp ect to technologies, a t the p resen t m om ent. I t is obvious from the above m entioned facts, th a t high fre quency tra n s is to rs a re c h a ra c te riz e d b y th e p resen ce of an in te rn a l field. O n th e o th e r h an d, low fre q u en c3r tra n s is to rs do n o t have an y significant in te rn a l field, a n d a re b a se d on the c a rrie r m ovem ent in the base by m eans of diffusion only. So th is p a p e r m ight also be called 'su rv e y of d rift tra n s is to rs ' w hich m eans a d e sc rip tio n w ith re sp e c t to th e design, w h e re a s the nam e high frequency tra n sisto rs ' is re la te d to the app lication. T he su rv ey to be given co n sid ers m ainly germ anium tra n s ito rs, w hich a re alm o st ex te n siv ely used fo r sm all signal am plification. O n ly m inor tech n o lo g ical m odifications a re n e c e ssa ry fo r silicon tra n s is to rs, w hich a re, h o w ev er, so m e w h a t slo w e r in frequen cy response due to the low er m obility of charge carriers in silicon. 2. Physics of high frequency transistors T he b asic id ea of am plification b y m eans of tra n s is to rs is sim ilar of th a t of a th erm io nic valv e. I t is th e co n tro l of a c u rre n t in a high-ohm ic circu it b y m eans of a sm all p o w e r applied to a control electrode. W h ils t in the case of a th e r m ionic v alv e, th e e le c tro n c u rre n t is tra n s p o rte d in high vacuum due to th e p resen ce of an electric field, an d w ill be influenced b y th e p o te n tia l of a so-called grid, th e c u rre n t tr a n s p o r t is com pletely w ithin a solid sta te m aterial in the case of tra n s isto rs. T he m ost com m on o p e ra tio n m ode of tra n s is to rs is the co n tro lled injection of m in o rity ch arg e c a rrie rs b y m eans of a /-^ -ju n c tio n an d th e ir v o ltag e in d e p e n d e n t collection by an o u tp u t electro d e, th e so-called collecto r. B esid es the fact, th a t th e tra n s is to r is a solid s ta te device, th e re a re tw o m ajo r differences w ith re s p e c t to th e th erm io n ic v alv e: F o r th e co n tro l of th e o u tp u t c u rre n t, th e tra n s is to r needs a c e rta in p o w er, i.e. th e in p u t im pedance is low, w h e re a s th e therm io nic valv e n eeds only a v o ltag e. T he o th e r difference is th e m echanism of th e tr a n s p o r t of th e charg e c a rrie rs w ith in th e device: w ith in a tra n s is to r, th is tr a n s p o r t can be done by diffusion due to a c o n c e n tra tio n g ra d ie n t, w h e re a s in a valve, th e c u rre n t is a lw a y s a p ro p e r field c u rre n t. T he first designs of tra n s is to rs h av e exclusively u sed th e diffusion tr a n s p o r t m echanism. T his diffusion, h o w ev er, is a very slow process and it tak es, therefore, a long tim e

3 High frequency transistors 109 b efo re th e ch arg e c a rrie rs h av e m oved from th e in p u t to th e o u tp u t electro d e. T his fa c t n eeds n o t to be a re stric tio n in re s p e c t to frequency. H o w e v e r, th e n e c e ssa ry co n c e n tra tio n g ra d ie n t h as to be e stab lish e d an d v a rie d in th e rh y th m of the in p u t signal an d th is m eans a v a ria b le ch arg e in th e b ase o r a large capacitive current. O ne speaks ab o u t a diffusion cap a city of a tra n sisto r because this capacity arises from the diffusion m echanism. T here are, of course, o th er capacities p resent, the m ost imp o r ta n t of w hich is th e co lle cto r cap acity, w hich p ro v id es a feedback from ou tp u t to input. T he m o st im p o rta n t of th e in h e re n t re sista n c e s is th e b ase re sista n c e, w hich is th e ohm ic re sista n c e b e tw e e n th e activ e tra n sisto r region and the base connection of the device. L e t us now co n sid er a classic allo y tr a n s is to r w ith a b ase la y e r of c o n sta n t re s is tiv ity : In o rd e r to o b ta in a good p e rfo r m ance a t high frequencies, th e cap acities n eed to be as sm all as possible, an d th is m eans a v e ry sm all b ase th ick n ess (fo r in sta n c e 5 /*) an d a sm all co lle cto r a re a (fo r in stan ce 0 = I 50 /u). A v e ry th in b ase la y e r, h o w ev er, is in c o n tra d ic tio n w ith th e w ish fo r a low base re sista n c e r I f w e w o u ld in crease th e co n d u ctiv ity of th is b a se la y e r, w e w o u ld also in crease th e collector capacity and decrease the b reak d o w n voltages. In th ese co n sid era tio n s w e h av e even n eg lected th e fa c t th a t the collector-base junction has a sm all b u t n o t negligible thickness, w hich depends of the collector voltage and the conductiv ity oi the base, and w hich m ust not touch the em itter. A s show n here, the design of an alloy tra n sisto r is a comprom ise w ith resp ect to operating voltage and operating fre quency, n o t to fo rg e t th e p ra c tic a l difficulties in o b tain in g v e ry sm all b ase w id th s b y th e alloying tech n iq u e. A n optim ization w ith re s p e c t to fre q u en c y le a d s to a device w hich is ab le to o p erate up to approx. 20 M c/s. In 1953, it w as proposed by K röm er and E a r ly *2), indep e n d e n tly from each o th e r, to p ro v id e tra n s is to rs w ith a b ase, th e co n d u ctiv ity of w hich being a functio n of th e p lace. W h ils t the m ain idea of K röm er has been to reduce the diffusion cap acity, E a rly thou ght prim arily of a reduction in collector capacity. The idea, how ever tu rn ed out to be so universal, as to *) Kröm er, N aturw issenschaften 40, 1953, p ) E arly, Bell STJ , p. 787.

4 110 W. Edlinger d ecre ase n o t only th e diffusion an d co llecto r cap acity, b u t also th e b a se re sista n c e, a n d to in crease th e co lle cto r b re a k d o w n voltage, all of them in a ra th e r unlim ited m anner. The p ractical w a y is to m ake the conductivity n ea r the colle c to r v ery high a n d n e a r th e e m itte r v e ry low. In doing so, one h as o b ta in e d a v a ria b le co n c e n tra tio n of fre e c a rrie rs an d th u s an in te rn a l e le c tric a l field, th e s tre n g th of it depending on th e difference b e tw e e n co n d u ctiv ity n e a r e m itte r to conductivitj" n e a r th e co llecto r, an d th e b a se w id th. T his in te rn a l field can be in th e o rd e r of 400 to 2000 V /cm and th e actio n of it is to tr a n s p o r t th e m in o rity c a rrie rs m erely b y th e field a n d n o t b y diffusion a n y m ore. T he c u rre n t th ro u g h th e b ase is now a field or d rift cu rren t and th ere the nam e 'd rift tra n s isto r originates from. T his in te rn a l field sp eed s up th e m ovem ent of th e c a rrie rs in th e ir w a y from e m itte r to co llecto r. I t is ev id ent, h o w ev er, th a t the requirem ent for v ery sm all dim ensions in the tra n s versal direction rem ains unchanged. T he p ra c tic a l re a liz a tio n of th ese id eas is easily p o ssib le an d w ill be d iscu ssed in a la te r c h a p te r. A t th is m om ent, w e w ill tr e a t som e p h y sical p ro p e rtie s of th ese high fre q u en c y o r d rift tra n sisto rs, prim arily w ith resp ect to th e ir application. 2.1 Power gain and stability T he p rim a r in te n tio n fo r th e cu sto m er of a tr a n s is to r is to o b ta in am plification. T he p o w e r gain d ep en d s, of course, of th e circu it used, th e frequen cy, th e b a n d w id th, th e s ta b ility etc. T he cu sto m er is ab le to design a circu it a ro u n d th e tr a n s is to r if he k n o w s th e fo u rp o le p a ra m e te rs of th e tra n s is to r. T he d esig n er of th e tra n s is to r, h o w ev er, n eed s an u n d e rsta n d in g of how these fourpole p aram eters depend from the physical stru c tu re, i.e. he w a n ts to know the relatio n sh ip b etw een the dim ensions, th e co n d u ctiv ity of th e b a se region of th e device, an d th e fo u rp o le p a ra m e te rs. A n useful ste p to w a rd s th is re la tio n sh ip is th e eq u iv alen t c ircu it of th e tra n s is to r. T his e q u iv a le n t circu it is a circu it co n sistin g of re s isto rs, cap a c ito rs, c u rre n t a n d v o ltag e sources etc. w hich b eh av es like th e tra n s is to r. T he com ponents of th e e q u iv a le n t circu it should be in d e p e n d e n t of fre q u en c y. M a n y e q u iv a le n t circu its can be e sta b lish e d b y m easu rem en t of th e e le c tric a l b e h a v io u r of a tr a n s is to r a g a in st frequen cy. The m ost useful equivalent circuit, how ever, is one w hich show s

5 High frequency transistors 111 d ire c tly som e of th e phen o m ena w ith in th e tra n s is to r. So the d esig n er p re fe rs to see in th e eq u iv alen t circu it th e b ase re s is tan c e, th e co lle cto r cap acity, th e diffusion c a p a c ity etc., w hich he can, in tu rn, c o rre la te w ith dim ensions, doping levels an d so on. A v e ry sim ple eq u iv alen t circu it h as b een given b y Jochem s, M em elin k an d T um m ers 3) w hich holds fo r m any high frequen cy tra n sisto rs up to reasonable frequencies. B y m eans of th e eq u iv alen t circuit, th e e le c tric a l en g in eer is able to calculate input and o u tp u t im pedances, fourpole p a ra m e te rs an d th u s th e am plification, sta b ility, an d b a n d w id th of his circuit including the tra n sisto r. F o r th e d e sc rip tio n of a p a rtic u la r tra n s is to r, th e p o w e r gain u n d e r w e ll defined conditions is o ften used. I t is th e so-called u n ila te riz e d p o w e r gain, an d th a t m eans th e p o w e r gain w h en th e in p u t an d o u tp u t is m atched to optim um am plification an d th e in te rn a l feed b a c k of th e tra n s is to r is n eu tra lized. In such a case, as it can be show n b y m eans of th e eq u iv ale n t circuit, the p ow er gain is approxim ately given by P.G. const. f ( Cd + Q Cc r b r0 w here f m eans the operating frequency, C d the diffusion cap a city, C e the em itter capacity, C c the collector capacity, r b the b a se re sista n c e, r0 = ---- k T th e ju n ctio n re sista n c e, Ie being th e q I e e m itte r b iasin g c u rre n t. W h e th e r n e u tra liz a tio n is used, is a q u estio n of econom y. In m ost ra d io receiv ers, th e n e u tra liz a tio n is to o expensive in m ass p ro d u ctio n an d th en, th e p o w e r gain w hich can be o b ta in e d u n d e r s ta b le conditions, is d ifferent. A t I F freq u en cies, it is m ostly lo w e r th a n th e u n ila te riz e d one. In m any cases, fo r th e fro n t end of a receiv er, use is m ade of th e in te rn a l p o sitiv e fe e d b a c k of th e tra n s is to r in g ro u n d ed b a se connection. In an y p ra c tic a l case it is im p o rta n t fo r th e d esig n er, to m ake th e cap acities an d th e b ase re sista n c e as sm all as possible, fo r a high am plification of the device. 2.2 Noise figure The usefulness of an am plifying device is n o t only given by 3) Jochems, M em elink and Tummers, Proc. IR E, 46, 1958, p

6 112 W. Edlinger th e p o ssib ility of a high p o w e r gain, b u t also, if n o t p rim a rily, b y its noise figure. N o ise is p ro d u ced in a tra n s is to r in th e base resistance, a t the p~n- junctions and in the base as d istrib u tio n noise. V a n d e r Z i e l 4) h as given, am ong o th e rs, a good su rv ey, in p a rtic u la r a b o u t th e high fre q u e n c y noise. O n e can in tro d u ce noise c u rre n t an d noise v o ltag e g e n e ra to rs to a noisefree tra n s is to r fo u rp o le, o r in to th e e q u iv a le n t circuit. O n e can also ex p re ss th e to ta l noise figure in term s of b asic tra n s is to r p ro p e rtie s, such as cap acities an d b ase re sista n c e. T he e x a c t fo rm u la is r a th e r com plex a n d d ep en d s am ong o th e rs from th e source im pedance of th e circuit. A ro u g h a p p ro x im atio n can be given fo r high frequencies in the follow ing form N F = i + C f.rqn (Cd + c C being a c o n sta n t. F o r m edium frequen cies th e noise is a p p ro x im a te ly in v e rsely p ro p o rtio n a l to th e b a se re sista n c e r/,. T he se form u lae a re found to be q u ite c o rre c t in p ra c tise and so one can develop a tra n sisto r to a certain noise re quirem ent. I f one c a lc u la te s th e optim um source im p edance fo r minimum noise an d fo r m axim um am plification, one finds th a t th e re is little difference. T his is quite d iffe re n t from th e therm io nic v alv e, w h e re one h as, som etim es, a c o n sid e ra b le m issm atch fo r eith er one. T he noise figure is co n sid ere d b y m any p eo p le to be th e m ost im p o rta n t p ro p e rty of a device like a tra n s is to r. A s a m a tte r of fact, it tu rn s out, th a t a tra n s is to r w ith a sufficiently lo w noise figure h as a lw a y s a p o w e r gain of g re a te r th a n 10 db, so th a t th e noise figure of th e follow ing stag e can, in g en eral, be n eg lected. 2.3 Envelopes B esid es good e le c tric a l p ro p e rtie s, a tr a n s is to r n eed s to be such th a t it can be easily handled, and soldered into the circuit. I t is also quite w ell know n, th a t th e tr a n s is to r su rface is sensitive to am b ien t conditio n s, esp ecially m o istu re. T he tra n sisto r envelope has the double ta sk to provide a herm etic 4) van der Ziel, Proc. IR E, 46, 1958, p

7 High frequency transistors 113 enclo su re of th e tra n s is to r an d to m ake it e a sie r to h an d le fo r the set-m aker. O f course, the envelope should not d e te rio ra te th e in trin sic electric tra n s is to r p ro p e rtie s. T he pro b lem s a rise especially w ith re s p e c t to c a p a c ity a t th e o u tp u t an d w ith re s p e c t to le a d self in d u ctan ce a t th e in p u t. T h ere is a g e n e ra l tre n d to go to sm aller dim ensions fo r all electric com ponents. T his fa c ilita te s red u cin g th e self in d u ctan ces of the leads, b u t m akes the capacities of the envelopes larg er. E sp e c ia lly in th e high frequen cy tra n s is to r field, a c e rta in sta n d a rd iz a tio n h as a lre a d y b een o b ta in e d b y m o st of th e tr a n s is to r m a n u fa c tu re rs. T his sta n d a rd iz a tio n gives a s ta n d a rd lead arran g em en t w hich fits in the com m ercially available a u to m atic insertion m achines. A V hether th e re w ill be a n y coaxial envelopes in th e fu tu re, d ep en d s v e ry m uch from th e in trin sic tra n s is to r p ro p e rtie s w hich can easily be o b tain ed. In g en eral, a coaxial p ack ag e is expensive as such an d fo r th e re a liz a tio n of th e circu it into w hich to fit. 3. Technology of high frequency transistors T here are several m ethods for m aking high frequency germ anium tra n s is to rs. A ll of them h av e one com mon p ro p e rty, th a t is th e non c o n sta n t doping level in th e b a se, w hich can e a sily be realized b y diffusion of d o n o r elem ents into th e solid s ta te germ anium a t te m p e ra tu re s n e a r th e m elting poin t. B y m eans of diffusion tech niques, one o b ta in s a d o n o r c o n c e n tra tio n w hich is h ig h est a t th e su rface a n d lo w e r inside th e c ry s ta l, the d istrib u tio n being a p p ro x im a te ly a co m p lem en tary e r ro r function. T he se diffusion tech n iq u es a re easy to p erfo rm on germ anium an d a re of ex cellen t re p ro d u c ib ility. W e can distinguish b e tw e e n pre diffused Impunfy 3 m ajo r ty p e s of tra n s is to rs, each c o rre la te d w ith one o r m ore m anufacturers. C onstruction of D rift tran sistor (R C A ). 3.1 D rift transistors, type RCA The designation d rift tra n s is to r applies, as a m a tte r of fact, to all high frequency tra n s istors. A s the construction to

8 114 W. Edlinger be d e sc rib e d h a d b een th e first high frequen cy tra n s is to r on th e m a rk e t, it h as becom e th is nam e a n d still k eeps it, an d new designs have chosen o th er nam es. T his device consists of a classic allo y tra n s is to r, into th e w a fer of w hich a v ariab le donor concentration has been diffu sed b efo re th e alloying p ro cess (see figure 1). E m itte r an d co llecto r a re a llo y e d from d iffe re n t sides into th e w a fe r an d the base con tact consists of an an n u lar tab. T hese tra n s is to rs a re n o w a d a y s m a rk e te d b y R C A, T elefunken, B rim ar and others. Philco an d th e ir licencees S p rag u e, C B S, a n d o th e rs, h av e im p ro v ed th is c o n stru c tio n b y th e ir sp ecial chem ical etching an d d ep o sitio n m eth od w hich p erm its to m ake th e germ anium w a fe r very thin a t the place w here em itter and collector are to be electro p lated (see figure 2). prediffused impurity D rift transistor (Philco) T hese c o n stru c tio n s h av e th e a d v a n ta g e, th a t th e y d iffer from o rd in a ry allo y tra n s is to rs only b y th e use of a d iffere n t w a fe r. T hus, th ese tra n s is to rs could be p ro d u c e d w ith th e sam e jigs a n d m ach in ery as th e allo y ty p e s. T he m ajo r d isa d v a n ta g e, h o w ev er, is th e fa c t th a t one h a s to c o n tro l th e b a se w id th an d even a ll p e n e tra tio n d e p th s w ith refe re n c e to th e p red iffu sed la y e r, to extrem e close to le ra n c e s. O n ly P hilco h as succeeded to come in to th e o p e ra tin g region of se v e ra l h u n d re d M c/s, b ecau se of th e ir special etching te c h nique. T he Philco co n stru ctio n, h o w ev er, h as still th e fam e of being less re lia b le, w ith re s p e c t to shock a n d surge c u rre n ts, th a n th e o th e rs, b ecau se of th e lo cally v e ry th in c ry s ta l. T h ese P hilco tra n s is to rs n eed fo r th e ir p ro d u c tio n a series of r a th e r so p h istic a te d a n d expensive m achin ery, b u t th e y a re allo y ed to re ally close tolerances. A ll of th ese 'd rift tra n s is to rs ' h av e th e a d v a n ta g e a g a in st th e su b se q u e n t ty p e s, th a t th e ir co lle cto r series re sista n c e is negligible, w hich is especially useful in sw itching and large signal applications. 3.2 Mesa transistors, type Bell T hese tra n s is to rs a re p ro d u ced b y W e s te r n E le c tric, T ex as In strum ents, M o to ro la, Siem ens, C S F and m any others.

9 High frequency transistors 115 The idea of this construction is (see figure 3) th a t all m anufa c tu rin g o p e ra tio n s come from one side of th e w a fe r. F irs t o f all, a n-type skin is diffused into the / - ty p e w afer. A fte rw a rd s, pre diffused N - layer collector p emitter CAIJ base CAu-Sb) mesa collector tab Fig. 3 M esa transistor (Bell). 3 an e m itte r is e v a p o ra te d and allo y ed (o r diffused) from th e sam e side of th e w a fe r. A lso th e b ase c o n ta c t is o b ta in e d b y e v a p o ra tio n. In o rd e r to m ake th e co llecto r c a p a c ity as sm all as p o ssib le, th e tra n s is to r is m asked, and a table -like constru c tio n is etch ed. F rom th is a p p e a ra n c e, th e nam e 'm esa- tra n s is to r' h as b een d eriv ed, as th e y a p p e a r v e ry m uch like th e 'm esa-m o u n tain s' in th e S o u th -W e s t of th e U.S.A. T he c o n ta c t w ires a re g e n e ra lly a p p lied by th erm o co m p ressio n bonding, an o p e ra tio n w hich is r a th e r d elicate a n d is c a rrie d o u t a t ap p ro x im a te ly 350 b y lo cally ap p ly in g high p re ssu re b etw een the ev ap o rated electrodes and a gold w ire. T he ease of fa b ric a tio n of th is ty p e of tra n s is to r is still su b ject of m any discussions. T he m ain difficulty seem s th e v ery th in re c ry sta lliz e d e m itte r region a n d its e le c tric a l b eh av io u r, w hich lim its th e choice of th e b ase co n d u ctiv ity. A n o th e r difficulty is th e fa c t th a t th e b a se -to -c o lle c to r ju n ctio n is only I yu o r so b e n e a th th e su rface a t w hich one ap p lies, d u rin g th erm o co m p ressio n bonding, lo cally v e ry high p re ssu re s. In th eo ry, h o w ev er, th is s tru c tu re should be unlim ited w ith re sp e c t to size, b ecau se ex trem ely sm all geo m etries can be o b ta in e d b y evaporation techniques. T he h e a t tr a n s p o r t a w a y from th e co lle cto r ju n ctio n is m ore fa v o u ra b le in th is case th a n fo r th e d rift tra n s is to rs. T h e device h as got, h o w ev er, a r a th e r larg e co lle cto r series re sista n c e. 3.3 Post alloy diffusion transistor T his c o n stru c tio n h as first b een in tro d u c e d b y P h ilip s 3). I t is a com binatio n of se v e ra l id eas. A ll o p e ra tio n s a re c a rrie d o u t from one side like in th e m esa tra n s is to r. D o ts a re used, h o w ev er, w ith a co n sid era b le p e n e tra tio n in to th e c ry sta l, from w hich th e b ase la y e r is diffused d u rin g th e allo y diffusion cycle (see figure 4). O n e d o t serves as em itter, the o th er one

10 116 W. Edling er as base connection. The a d v a n ta g e s of th is co n stru ctio n a re m ainly in th e field of e a sy p ro d u ctio n. D iffusion an d th e c o n ta c ts a re m ade in one step, th e w ire a tta c h m e n t is easy, b ecau se of th e re la tiv e ly larg e Post Alloy Diffused transistor (Philips) d o ts, an d the b a se -to -c o lle c to r ju n ctio n is fa r a w a y from th e su rface. A m esa s tru c tu re is also etch ed in o rd e r to d ecre ase th e co llecto r a re a. T he d isa d v a n ta g e of th e high co llecto r series resistan ce is also p resen t in this stru ctu re. 4. Practical results W e have considered the 3 m ajor technologies for high fre quency tra n s is to rs a n d w ill discuss now th e re su lts o b ta in e d in the sm all signal am plification field. O ne typical application is the interm ediate frequency am p lifier a t 10.7 M c/s. B y m eans of tra n s is to rs of e ith e r of the 3 co n stru ctio n s, gains p e r sta g e in th e o rd e r of 24 db a re e a sily o b ta in a b le w ith th e b a n d w id th an d sta b ility re q u ire d. H isto ric a lly, th e first a p p lic a tio n w a s th e s h o rt w ave am plifier, o sc illa to r an d m ixer. A lso this field can be co v ered w ith ty p es of all th e d e sc rib e d technologies. F o r th e o sc illa to r, th e p h a se sh ift should be lo w a t th e h ig h est frequen cy re q u ire d in o rd e r to g et good a n d c o n sta n t fe e d b a c k conditions o v er th e w hole frequen cy ran g e. T he noise figure of tra n s is to riz e d s h o rt w a v e receivers is in the o rd e r of 4 db. F o r 100 M c/s am plification, o sc illa to r an d m ixer ap p licatio n s, the existing tra n sisto rs according to all of the th ree technologies y ie ld a p p ro x im a te ly 22 db p o w e r gain an d 10 db noise figure fo r a tw o sta g e fro n t end. T he tra n s is to rs acco rd in g to th e 'd r if t tra n s is to r' co n stru ctio n of R C A fo r o p e ra tio n a t 100 M c /s a re m ade accord in g to th e Philco etch ing p ro cess r a th e r th a n th e o rig in al R C A c o n stru c tio n, in o rd e r to realize th e n e c e ssa ry sm all ju n ctio n d ista n c e w ith a high y ield. Som e su p p lie rs give sp ecial g u a ra n tie s fo r th e p h a se sh ift of th e ir tra n sisto rs a t these frequencies, in o rd er to m ake a self oscillating m ixer possible. F o r telev isio n tu n e rs up to an o p e ra tin g fre q u e n c y of 240 M c/s, th ere are also some tra n sisto rs on the m arket, again

11 High frequency transistors 117 accord in g to all of th e th re e tech nologies m entioned. N o ise figures in the o rd er of 6 db can be realized. O th e r functio ns in telev isio n receiv ers can be co v ered b y com m ercially a v a ila b le high frequen cy tra n s is to rs w ith o u t an y difficulty: video am plification, jungle, tim e bases etc. L a te ly, some p a p e rs h av e a p p e a re d 5) d escrib in g tra n s is to rs su itab le fo r use a t 1000 M c /s, m ade acco rd in g to th e d escrib ed tech nologies an d h av ing a p p ro x im a te ly 10 db p o w e r gain an d 10 db noise figure a t 1000 M c/s. T he co n stru ctio n of tra n s is to rs fo r v e ry high frequen cies is m ainly a m a tte r of h an d lin g th e v e ry sm all dim ensions re q u ire d. T he e m itte r d o t d ia m e te r needs to be a p p ro x im a te ly 150 /x fo r 100 M c/s o p e ra tin g frequen cy, 75 /a fo r 200 M c /s an d 35 /ll fo r 1000 M c /s resp. fo r d rift o r allo y diffused tra n s is to rs. F o r m esa tra n s is to rs, it tu rn s out th a t one n eeds so m e w h a t sm aller dim ensions fo r th e sam e p erfo rm an ce, fo r in stan ce a rectan g u lar em itter of 25 x 50 for 200 M c /s operation. The m entioned figures show th a t the sm all signal, high fre q u en cy tra n s is to r can com pete successfully w ith th e electro n ic valves w ith resp ect to am plification and noise. T here are, how ever, tw o p ro p e rtie s w h e re th e tra n s is to r m ight be so m ew h at w o rse th a n th e electro n ic v alv e, i.e. th e c a p a b ility a f h an d lin g large signals of the antenna, and d isto rtio n or cross m odulation. In th e field of larg e sig n al high fre q u e n c y tra n s is to rs, the situ a tio n is still changing ra p id ly an d th is is e a sy to u n d e rsta n d : H ig h frequen cy m eans sm all dim ensions an d high p o w e r m eans la rg e ones. To give a ro u g h id e a of w h a t h as b een o b ta in e d in th is field, w e m ight s a y th a t a t 100 M c /s tra n s is to rs m ight d eliv er an o u tp u t p o w e r of a p p ro x im a te ly 1 A V att w ith an efficiency of 50 /0 an d an am plification of 10 db. Silicon is su p e rio r to germ anium b ecau se of th e p o ssib ility of h ig h er ju n c tio n te m p e ra tu re s w ith o u t d e g e n e ra tio n of p erfo rm an ce. T he tra n s m itte r o u tp u t ap p lic a tio n s fo r low frequen cies, w h e re p o w e rs a re re q u ire d in th e o rd e r of k ilo w a tts, w ill p ro b a b ly n o t be realized by tra n sisto rs w ithin the n e a r future. F o r th e sak e of com pleten ess it n eeds to be m entioned, th a t in m ost of th e ab o v e m en tioned fields th e d ev elo p m en t is n o t finished y e t a n d m an y ty p e s a re n o t a v a ila b le in m ass q u antities for a cheap price. s) Saari et al, 1960 International Solid-State C ircuit Conference.

12 118 W. Edlinger 5. Conclusion C oncluding, it m ight be said, th a t all high frequency tra n s is to rs co n tain an in te rn a l field in th e b a se and differ in th is re sp e c t from th e lo w freq u en cy tra n s is to rs. T h re e d iffere n t technologies a re u sed fo r th e p ro d u c tio n of th e se tra n s is to rs in th e case of germ anium, all of them lead in g to a p p ro x im a te ly the sam e perform ance. T he a v a ila b le high frequen cy tra n s is to rs a re m ain ly in te n d e d fo r am plification of sm all signals an d th e ir o p e ra tin g frequen cies a re still in creasin g. A t th is m om ent, freq u en cies up to 1000 M c /s seem to be p o ssib le. T he p ro p e rtie s of th e se high fre q u e n c y tra n s is to rs co m p are w ell w ith those of electronic valves. The ad v an tag es of tra n s isto rs like sm all d issip atio n, sm all volum e etc. h o ld also fo r high fre q u e n c y ap p licatio n s. T he m ain p ro b lem s fo r th e use of tra n s is to rs seem th e sm all p erm issib le v o ltag e a t th e in p u t an d in some cases the d isto rtio n a t high frequencies. T he high fre q u e n c y p erfo rm an c e of a v a ila b le tra n s is to rs is lim ited b y th e tech n o lo g ical difficulties in h an d lin g th e n e c e ssa ry very sm all dim ensions. In th e h ig h er h o w e r field, th e d ev elo p m en t is still p ro ceed in g v e ry f a s t an d n ew tech nologies a re lik ely to be in tro d u c e d fo r these applications, m ainly using silicon. Manuscript received March 27, 1961.

13 Deel 26 - N o De werking en eigenschappen van het vaste-stof thyratron door O. W. M em elink *) Voordracht gehouden voor het Nederlands Radiogenootschap op 17 februari Summary An interesting new sem iconductor device is the controlled rectifier or P N P N - sw itch, w hich consists of a silicon cry stal in w hich four regions o f different co n d u ctiv ity -/ - or ^-type-have been incorporated. The currentvoltage characteristics of the controlled rectifier are very sim ilar to those of a conventional gas-filled th y ratro n. The device is able to block potentials up to 500 V olts in either direction. A pplication of an electric pulse to a control electrode brings the device into a state of conduction in w hich it conducts large currents w ith a small voltage drop. In the present article the operation of the controlled rectifier is explained and a few exam ples of its applicability as an electronic sw itch are m entioned. The controlled rectifier show s m uch prom ise for the use in stabilized pow er supplies, regulation and D C -A C converters. 1. Inleiding D e P N P en N P N la g e n tra n s isto re n zijn th a n s volkom en ingeburgerd in de electrotechniek. B etrekkelijk nieuw, doch fund am en teel op dezelfde h a lfg e le id e r eigenschappen b e ru ste n d is h e t v a s te -s to f th y ra tro n, ook w e l genoem d: P N P N -sc h a k e la a r of gestuurde gelijkrichter **). Z o als bekend m ag w orden v eron dersteld, vertonen de conv e n tio n e le " tra n s is to re n, w a t hun u itg a n g sk a ra k te ristie k e n b e tr e f t g ro te o v ereen k o m st m et v acu u m p en to d es. In h e t volgende zal blijken, d a t de stro o m sp a n n in g sk a ra k te ristie k e n v an de gestu u rd e g elijk ric h ter veel lijken op die van h e t gasgevulde th y ra tro n, v a n d a a r de n aam v a s te -s to f th y ra tro n. *) N atuurkundig Laboratorium N.V. Philips, Eindhoven. **) In de A ngelsaksische v ak literatu u r bekend onder de nam en: trinistor, thyristor, of controlled rectifier.

14 120 O. W, Memelink In d it a rtik e l zullen wij de w erk in g van de g e stu u rd e gelijkrich ter bespreken en enige toepassingsm ogelijkheden noem en. 2. De werking van de gestuurde gelijkrichter In fig. 1 h eb b en w e schem atisch de o p b o u w van de g e stu u rd e gelijkrichter of v aste -sto f th y ra tro n getekend. Zij b e s ta a t uit een " s ta p e lin g v an v ie r lag en v an v ersch illen d geleid in g sty p e in een silicium één k ris ta l. D rie v an deze lagen zijn voorzien van co n tacten, z o d a t zij electrisch a a n g e slo te n k u n n en w o rd e n. In analogie m et h et gasgevulde th y r a tr o n w o rd e n deze c o n ta c te n : anode-, k a th o d e - en s tu u rc o n ta c t genoem d. In fig. 2 zijn de stro o m -sp a n n in g sk a ra k - teristiek en van de gestuurde gelijkric h te r w eerg eg ev en v oor v ersch illen d e w a a rd e n van de stuurstroom Is. Indien de stuurstroo m nul is b lokk e e rt de A V W /^ -stru ctu u r stro o m in beid e rich tin g en to t spanningen van enkele h o n d erd en V o lts. O m d it n a d e r de gestuurde gelijkrichter. te k u n n en v e rk la re n m erken w e op, d a t de gestuurde gelijkrichter opgevat De stroom -spannings karakteristiekenschaar van de gestuurde gelijkrichter.

15 De werking en eigenschappen van het vaste-stof thyratron 121 kan w orden als een serieschakeling van 3 /W -d io d es, fig. 3. W a n n e e r Vak n e g a tie f is s ta a t de diode D a in de v o o rw a a rtsric h tin g, en d ra a g t d a n p ra c tisc h geen spanning. D e diodes 1 en 3 s ta a n in de sp erric h tin g, w a a rb ij sp eciaal de o v erg an g 3, V k Fig. 3 Een eenvoudig vervangingsschem a van de gestuurde gelijkrich ter voor het geval, d a t Vak < 0. overeenkom end m et diode D 3 zo gedim ensioneerd is, d a t zij een hoge sp e rsp a n n in g k a n d rag en. O v ersch rijd t de spanning tussen anode en k a th o d e de d o o rslag sp an n in g v an diode D 3, d an neem t de stroom snel to e op an alo g e wijze als in een norm ale h alfg e leid er diode, die in h e t doorslaggebied bedreven w o rd t. W o r d t Vak p o sitie f te rw ijl Is = O b lijft * dan zullen de spanningen o v er D 1, D2 en D 3 v an te k e n om keren, D 1 en D3 kom en in de v o o rw a a rts- en D 2 kom t in de sp erric h tin g. H e t is th a n s o v erg an g 2 die de aan g eleg d e spanning Vak d ra a g t. O v e rs c h rijd t Vak een w a a rd e v an enkele h o n d erd en V o lts, d an neem t I a w e e r toe te rw ijl nu bij deze p o la risa tie (in tegenstellin g m et het voorgaande geval Vak O ), de spanning Vaky via een gebied van n eg atiev e d iffe re n tia a lw e e rsta n d, te ru g v a lt op een lage w aard e. D e o o rzaak van de o v erg an g van een to e s ta n d van geringe geleiding n a a r een to e s ta n d v an g ro te geleiding m oet gezocht w o rd en in h e t gecom bineerde tra n sisto re ffe c t van de P N P en de N P N com binatie binnen de P N P N -stru c tu u r W a n n e e r Vak P O, w e rk e n de b u ite n ste /h V -o v erg an g en 1 en 3 (fig. 1) ieder als em itter van resp. het N P N en P N P gedeelte en de m iddelste overgang 2 functioneert als gem eenschappelijke collector. D e stroom, die de m iddelste P N overgang p a sse e rt is gelijk aan I a = I ( i + O -N P N I k + CL P N P I a ( I ) cinpn Ik is de g ecollecteerd e stroom v an h e t N P N g ed eelte en apnp Ia de g eco llecteerd e stro o m v an h e t P N P g ed eelte. Id is de stroom o v er de overg an g 2, die ook zou vloeien, indien e r geen tra n sisto re ffe c t aan w ezig w a s. M.a.w. Id k an o p g ev at w o rd e n als de stro om b eh o ren d bij diode D 2, w a a rv a n de g ro o tte b e p a a ld w o rd t d o o r de sp annin g sv al o v er deze diode. W a n neer Vak P O is dus ld de sperstroom van. In fig. 4

16 122 O. W. Memelink h e b b e n wij h e t g ed ra g v an de g e stu u rd e g elijk ric h te r w eerg eg ev en in een eenvoudig v ervangingsschem a. D e stro om a^rpn^k + o-p n p Ia is, zoals te doen gebruikelijk is, voorgesteld door een stroom bron p arallel aan diode D 2. Indien %o3 I s = O is Ik = I a = l d + ( (XPNP + O-NPn ) Ia (II) D e stro o m v ersterk in g sfacto ren vlnpn en ölpnp zijn niet consta n t, doch afhankelijk van de w a a rd e I a. H ie ra a n liggen v ersch illen d e p hysisch e m echanism en ten g ro n d slag, w a a r wij n ie t n a d e r op in zullen g aan. Bij kleine an o d estro o m I a is olnpn + aptfp <C<C I en zal I a ongeveer gelijk zijn a a n de d io d e sp e rstro o m /^. Bij o p v o eren v an Vak zal ld op den d u u r g ro te r w o rd e n d o o r een toenem ende lek van diode D 2. D it b etek en t, d a t ook I a en d a a rm e e clnpn + <*pnp toen em en en de stro o m b ro n in fig. 4 een ste e d s g ro te re Fig. 4 E en eenvoudig v erv angingsschem a voor het geval, d at Vak 0. fra c tie v an de to ta le stro om o v er de m id d elste o v erg an g zal voeren. D it effect w o rd t zo ste rk, d a t op een gegeven m om ent de d io d estro o m ld w e e r afneem t, d o o rd a t de stro o m b ro n v an lig. 4 p ra c - tisch de gehele anodestroom voert. A an gezien de diodestroom ld d irect gekoppeld is m et de spanning over de overgang 2 b e te k e n t een d ale n v an ld een d ale n v an Vak D e to ta le a n o d estro o m in d it gebied n eem t ste e d s to e en w e k u n nen dus sp re k e n v an een n eg atiev e d iffe re n tia a l w e e rs ta n d. W a n n e e r de an o d estro o m zo g ro o t is g ew o rd en, d a t up/pn + &PNP = volgt uit (II), d a t ld O, m.a.w. de sp an ning o v er de m iddelste o v erg an g is ju is t nul. D it g e b e u rt m e e sta l bij stro m en v an enkele tie n ta lle n m illiam pères. D e sp an n in g Vak o v er de g e stu u rd e g elijk ric h ter is nu ongeveer I V e n b e s ta a t u it de som v an de v o o rw a a rtssp a n n in g e n v an de o v erg an g en 1 en 3. V e rd e r o p v o eren v an I a h e e ft to t gevolg, d a t clnpn + clpnp 1 w o rd t. O p d a t vergelijking ( II) nog blijve gelden m oet de overgang 2 zich nu in de v o o rw a a rts ric h tin g p o la rise re n, m.a.w. de diode D 2 m oet een teg e n stro o m lev eren, zo g ro o t, d a t w e e r a a n (II) v o ld a a n w o rd t. In " tr a n s is to r ta a l zou m en h ie r sp rek en van een verzadigen (Eng. "bottom ing ) van de m iddelste over-

17 De werking en eigenschappen van het vaste-stof thyratron 123 gang. D e to ta le spanning o v er de g e stu u rd e g elijk rich ter is nu gelijk a a n de v o o rw a a rtssp a n n in g van de o v erg an g 1, die ca. 0,6 V o lt b e d ra a g t plus de spanningsval over de tw ee b asisgebieden. D e spanningen over de diodes D 2 en D 3 heffen e lk a a r nagenoeg op, d a a r b eid en in de v o o rw a a rtsric h tin g zijn g ep o larise erd. In de nu b esch rev en to e s ta n d lijk t de g e stu u rd e g elijk rich ter veel op een h alfg e leid er k ra c h td io d e, d.w.z. zij g eleid t een g ro te stro o m te n k o ste v an een kleine sp an n in g sv al. E v e n a ls bij de iw -d io d e b e ru st de v o o rw a a rts geleiding v an een g e stu u rd e g elijk ric h ter op reco m b in atie v an g eïn jecteerd e g a te n en electro n en, w a a rb ij in h e t gev al v an de g e stu u rd e g elijk ric h ter de reco m b in atie in de twee b asisg eb ied en p la a ts v in d t. H e t tra n s p o r t v an la d in g d ra g e rs d o o r de b asisg eb ied en v e rg t tijd en d it h e e ft to t gevolg, d a t bij snel in- en u itsc h a k e le n tra a g h e id se ffe c te n een ro l spelen. O m h e t b eeld eenvoudig te houden, h eb b en wij v o o rlo p ig aangenom en, d a t de stu u rstro o m Is nul w a s. V o e re n wij een zek ere stu u rstro o m toe (fïg. 2), d an b lijk t de g e stu u rd e g elijk ric h te r re e d s bij lag e re v o o rw a a rtssp a n n in g e n in te sch ak elen. D e o o rzaak h ie rv a n is, d a t tengevolge v an h e t to ed ien en v a n basisstroom aan h et N P N gedeelte van de gestuurde gelijkric h te r de stro o m I a stijg t, a^pn en ölpnp toenem en, en h e t sch ak elen bij lag e re spanning g eb eu rt. Bij vold oende stu u rstro o m re d u c e e rt de v o o rw a a rts -k a ra k te ris tie k to t de v o o rw a a rts -k a - ra k te ris tie k van een P N g elijk ric h ter (fig. 2). In h e t alg em een is d it re e d s h e t gev al v o o r een stu u rstro o m I s = 5 m^ bij een spanningsval tussen stu u relectro d e en k ath o d e van 1 k 2 V. T enslotte nog enige opm erkingen over h et gedrag van de gestuurde gelijkrichter bij hogere tem p eratu ren. D e silicium gestu u rd e g elijk rich ter is b ru ik b a a r to t 125 C, in som m ige gevallen zelfs to t 150 C. N atu u rlijk liggen de lekstrom en bij deze tem p e ra tu re n h o g er d an bij k a m e rte m p e ra tu u r. D e m axim ale blokk erin g ssp an n in g en in v o o rw a a rts en te g e n rich tin g v e ra n d e re n d a a re n te g e n n ie t n o e m en sw a ard. O o k b lijk t de g elijk richter m ak k elijk er in te sch ak elen, d.w.z. m in der stu u rstro o m nodig te hebben bij hogere tem p eratu u r. D e v o o rw a a rtssp a n n in g in geleidende to e s ta n d en de schak eltijd en w a a ro v e r wij nog zullen sp re k e n v e ra n d e re n n ie t veel.

18 124 O. W. Memelink 3. Het gebruik van de gestuurde gelijkrichter 3.1 Het inschakelen O m h e t m om ent v an in sch ak elen zo goed m ogelijk v a s t te leggen, is h e t g e w e n st een stap - of pulsvorm ige stro om a a n h e t s tu u rc o n ta c t to e te voeren. T u ssen h e t m om ent v an stu u r- stro o m to enam e en a n o d estro o m toen am e lig t een tijd sd u u r van ongeveer 1 /jl sec. D it is de tijd die nodig is om de diode D x (fig. 3) te " o p e n e n. D o o r Is g ro te r te m ak en w o rd t de vertra g in g stijd b e k o rt. N e e m t de an o d estro o m een m aal m e rk b a a r to e, d an is in o n g ev eer 1 /cl sec. de w a a rd e b e re ik t, die b e p a a ld w o rd t d o o r de u itw en d ig e sp annin g sb ro n en b elastin g. D eze stijgtijd w o rd t b e p a a ld d o o r de looptijd v an g a te n en electro n en in de tw ee basisgebieden. N a a s t de v ersch illen d e m eth o d es, v o o r h e t v erk rijg en v an o n tste e k p u lse n, die ook in de th y ra tro n te c h n ie k to e p a ssin g vinden, lenen - w egens de geringe b en odigde stu u rsp an n in g en - tra n sisto rsc h a k e lin g e n zich b ijzo n d er goed v o o r de o n tste k in g van g estu u rd e g elijk richters. D ik w ijls o n tw e rp t m en d a n een bistabiele schakeling, die spontaan oscilleert, of uitw endig a a n g e sto te n w o rd t. In d it v e rb a n d b ie d t een germ anium tunneldiode, gecom bineerd m et een tra n sisto r in teressan te m ogelijkheden. O o k k an een klein ty p e /VVT^iV-schakelaar als b ista b ie le inrichting geb ru ikt w orden. 3.2 Het uitschakelen van de gestuurde gelijkrichter E en g e stu u rd e g elijk richter, die een stro o m g ro te r d an O, I A g eleid t is alleen u it te sch ak elen d o o r de a n o d e stro o m zoveel te verm inderen, d at ajvpn + <*pnp I w o rd t, zo d at de ano destro o m zich n ie t la n g e r k a n h a n d h av en. In de p ra k tijk w o rd t dikw ijls h e t circu it m et de g e stu u rd e g elijk ric h ter d u sd anig o n tw o rp e n, d a t de an o d estro o m sp o n ta a n v an rich tin g o m k eert na een periode van geleiding. O o k kan de gestuurde gelijkrich ter uitgeschakeld w orden door een tegenstroom van voldoende g ro o tte toe te voeren. D it k a n b ijv o o rb eeld g eb eu ren d o o r op h e t g ew en ste m om ent een g eladen c o n d e n sa to r electrisc h m et de g elijk ric h ter te v erb in d en. D e aan leg v an een teg ensp an n in g h e e ft b o v en d ien nog een gunstige invloed op de u itsch a k eltijd. Z o a ls w e re e d s e e rd e r b e sp ra k e n v in d t de v o o r- w a a rtsg e le id in g in g estu u rd e g elijk rich ters p la a ts d o o r recom - binatie van geïnjecteerde lad in g d rag ers in de tw ee basis gebieden.

19 De werking en eigenschappen van het vaste-stof thyratron 125 D it b e te k e n t, d a t hoe g ro te r de v o o rw a a rtsstro o m is, hoe g ro te r de e x tra c o n c e n tra tie s van la d in g d ra g e rs in de b asisg eb ied en zullen zijn. W o r d t de stro om o n d e rb ro k e n, d a n zal enige tijd nodig zijn v o o rd a t alle e x tra la d in g d ra g e rs d o o r reco m b in atie v e rd w e n e n zijn. G e d u re n d e deze tijd, die te vergelijken is m et de deïonisatietijd van een gasgevuld th y ra tro n, kan de gestu u rd e g elijk rich ter geen v o o rw a a rtssp a n n in g e n g ro te r dan enkele V o lts b lo k k eren, aangezien zij d o o r de aan w ezig h eid v an de e x tra la d in g d ra g e rs onm iddelijk w e e r in sch ak elt. W o r d t in p la a ts v an de a n o d estro o m to t nul te re d u c e re n een tegenspanning aan g eleg d, d an zal een g ed eelte v an de re s tla d in g afgevoerd w orden door de tegenstroom. W a n n e e r deze tegenstro o m van dezelfde g ro o tte -o rd e is als de v o o rw a a rtsstro o m k a n de u itsch a k e ltijd een fa c to r 2 k 3 b e k o rt w o rd en. D e uitsch ak eltijd zo n d er to ep assin g v an teg enstro o m b e d ra a g t 10 k 20 // sec. 3.3 Toepassingen vayi de gestuurde gelijkrichter In bijna al die gebieden w a a r h e t th y ra tro n em plooi vond, zal de gestuurde gelijkrichter dezelfde ta a k kunnen verrichten. Zij h e e ft d a a rb ij de volgende v o o rd elen : 1. klein van afm etingen, v e rg e lijk b a a r m et die van een k ra c h t- diode en robuu st. 2. geringe eigen d issip a tie bij 50 A is de sp an n in g sv al n iet g ro ter dan 2 V. 3. geringe ontstekingsspanning 1 ^ 2 V. 4. sc h a k e lt snel in en uit. D e m axim aal te blokkeren spanningen liggen voorlopig b e ned en de 1000 V, w a t in h e t algem een la g e r is, d an v o o r een gas-gevuld th y ra tro n. T o ep assin g b eg in t de g e stu u rd e g elijk richter te vinden in de volgende v ier geb ied en : 1. G estab iliseerd e voedingsap paratuur. 2. R egelcircuits voor electrische m otoren, verlichting, etc. 3. O m zetters van gelijkspanning in w isselspanning m et fre quenties van 50 H z to t 20 kh z.

20 126 O. W. Memelink 4. D a a r w a a r k o rte p u lsen v an g ro te stro o m en lage re p e titie - frequentie nodig zijn: autom obielontsteking, puntlassen, etc. In de drie eerstgenoem de gebieden is h et afgegeven v erm ogen v an de sch ak elin g afh an k elijk v an h e t ty p e g e stu u rd e g elijk richter. V o o r een 10 A 300 V ty p e k a n m en re k e n e n op een afg ifte van 1 i 2 k V A p e r g e stu u rd e g elijk ric h ter m et een eigen d issip a tie v an 10 k 20 w. Manuscript ontvangen 21 juni 1961.

21 GEOGRAPHICAL LIST OF TV-TAPE INSTALLATIO NS Country Location Customer Equipment Austria 625 lines Vienna Oesterreichischer 1 Ampex Rundfunk 2 RCA Belgium lines Brussels INR 2 Ampex * Interswitch 405- Denm ark 625 lines Copenhagen Danmarks Radio 4 Ampex E gypt 625 lines Cairo UAR 4 RCA Damascus UAR 2 RCA Finland 625 lines Helsinki Yleisradio 2 Ampex France 819 lines Paris R T F 6 Ampex Interswitch RCA Research Paris Intercontinental Television 2 Ampex Mobile 525 lines Germany 625 lines Baden-Baden Südwestfunk 5 Ampex 1 mobile Berlin Sender Freies Berlin 3 Ampex CCC Film 1 RCA mobile Berliner Tages- Zeitungen Ampex Bremen Radio Bremen 1 Ampex Cologne W estdeutscher RF 2 Ampex 4 RCA 2 mobile Frankfurt Hessischer RF 3 Ampex Freies Fernsehen 5 Ampex 3 mobile Hamburg Norddeutscher RF 4 Ampex Ludwigshafen BASF Ampex Tape production Leverkusen Agfa Ampex Tape production Munich IRT 1 Ampex Research 127 RIVA-Film 2 Ampex 1 RCA Bayerischer RF 2 Ampex Saarbrücken Europäischer RF Ampex to be installed Saarländischer RF 1 Ampex Holland 625 lines Bussum N TS 2 Ampex to be installed Italy 625 lines Rome RAI 10 Ampex Interswitch Luxembourg 625/819 lines Luxembourg T élé-luxembourg Ampex Interswitch

22 128 Country Location Customer Equipment N orw ay 625 lines Oslo Norsk Rikskring- 2 RCA kasting Spain 625 lines M adrid Televisione Espanola 2 Ampex Sweden 625 lines Stockholm Sveriges Radio 3 Ampex Switzerland 625 lines Zürich Schweizerisches Fernsehen Ampex Geneva Télévision Suisse 1 Ampex U nited Kingdom 405 lines Belfast Ulster T V Ampex Bristol T W W 1 Ampex mobile, Interswitch Cardiff T W W Ampex Glasgow Scottish T V 2 Ampex Hayes, Middlesex E.M.I. 1 Ampex Tape production London area ABC 'Television 2 Ampex 1 mobile, Interswitch RCA to be installed Alpha T V 6 Ampex Interswitch Assoc. Rediffusion 6 Ampex 1 mobile, Interswich Newcaste Tyne-Tees T V 3 Ampex 1 mobile Norwich Anglia T V 2 Ampex Plymouth W estward T V 1 RCA Southampton Southern T V 3 Ampex 1 mobile to be installed Assoc. Television 3 Ampex 1 mobile Interswitch BBC 14 Ampex 3 mobile, Interswitch to be installed Granada T V 2 Ampex Ministry of Aviation Ampex closed circuit Zonal Films Ampex Tape production Manchester ABC 'Television 1 Ampex Granada T V 9 Ampex 1 mobile, multistandard Eire Dublin Radio Eirann 2 Ampex 1 mobile, Interswitch to be installed Total number of V T R s in Europe 152 *TM Ampex Corp.

23 M AP OF TV-TAPE FACILITIES IN EUROPE 129 DRIEDIM ENSIONAAL W AARNEM EN PER RADAR In juli werd door de Amerikaanse Luchtmacht een contract getekend voor de ontwikkeling door International Telephone & Telegraph van een uniek driedimensionaal werkend systeem, waarmede geheel nieuwe mogelijkheden voor luchtvaart en ruimtevaart kunnen worden onderzocht. IT T kreeg de opdracht doordat reeds geruime tijd gewerkt werd aan een volumetrisch 3D-systeem, dat van alle kanten en óók van boven kan worden bekeken en w aarvoor géén speciale stereoscopische brillen nodig zijn. Het systeem is speciaal bedoeld voor de U.S. Air Force hoewel ook andere toepassingsmogelijkheden denkbaar zijn die de toepassing in bemande vaartuigen bestudeert. V erw acht wordt, dat grote mogelijkheden aanwezig zullen zijn voor de luchtverkeersbeveiliging, het opsporen en geleiden van raketten en andere ruimtevaartuigen, duikbootbestrijding en vele aanverwante gebieden. Het systeem ( ± 1/30 m3 inhoud) is onlangs voor hoge autoriteiten in W ashington gedemonstreerd. Binnen in een doorzichtige cilindrische plastic huls draait een lichtgevend scherm om een verticale as bij een constante snelheid waardoor het scherm onzichtbaar lijkt. W anneer lichtpuntjes vliegtuigen of andere soorten vaartuigen op het draaiende scherm worden geprojecteerd, schijnen zij in de cilinder te zweven. De besturing van de lichtpunten vindt plaats door 3D-Radar of informatieverwerkende systemen. Een fijn lijnenspel kan eveneens in het systeem worden geïntroduceerd w aardoor de plaats van de objecten nauwkeurig kan worden bepaald. Ook is het mogelijk elk lichtpuntje in een andere kleur te projecteren.