The development of the multigap resistive plate chamber

UCLEAR PHYSIC~

PROCEEDINGS SUPPLEMENTS ELSEVIER

Nuclear Physics B (Proc. Suppl.) 6IB (1998) 250-257

The development of the multigap resistive plate chamber M.C.S. Williams

INFN Bologna, Italy A new variety of resistive plate chambers has been in existence for the last year. In this paper we examine some basic problems with resistive plate chambers. We then discuss the solution offered by this new design. We will conclude with the current status of the development of the multigap RPC. 1. I N T R O D U C T I O N Resistive plate chambers (RPC) have been in use in various e x p e r i m e n t s since the 1980's. In essence they are very simple devices. They consist of a gas volume bounded by two parallel plates. These plates are m a d e of r e s i s t i v e m a t e r i a l , with electrodes covering the outer surfaces. By applying a high voltage to these electrodes, one generates a strong electric field across this gas volume. Through-going charged particles create clusters of positive ions and electrons in the gas. If the electric field is sufficiently strong, these electrons avalanche as t h e y move towards the anode. This avalanche induces a signal on the external electrodes. At sufficiently high electric fields, the avalanche grows large enough to initiate a streamer; in this case the high resistivity of the plates keeps this discharge localised. Thus there are two operation modes of the RPC; either 'avalanche mode' or 'streamer mode'. In all Letters of Intent for the LHC, these devices are the prime candidate for the muon trigger. The proposed type of RPC has a 2 m m g a s gap b o u n d e d by 2 m m thick (linseed-oil treated) Bakelite plates (we will refer to this as the conventional RPC). There have been many questions raised concerning the performance of this type of RPC - and w h e t h e r it is suitable for the hostile environment of the LHC. The most basic question concerns the rate capability of the RPC. At the time of the original Letters of Intent for the LHC, the conventional RPCs were operated only in 'streamer mode'. They were believed to have a rate capability of 100 Hz/cm2 [1]; although o t h e r s m e a s u r e d a m u c h lower r a t e capability of between 1 and 5 Hz/cm2 [2,3]. However, it soon became clear that the RPCs 0920-5632/98/$19,00 © 1998 Elsevier Science B.V. All rights reserved. PII S0920-5632(97)00570-7

would have to w i t h s t a n d a much higher b a c k g r o u n d count r a t e t h a n originally expected due to neutrons and gammas. The r e s u l t was t h a t in t h e following LHC Technical Proposals [4,5], it was suggested t h a t the conventional RPC were operated in qow gas gain mode' (now normally referred to as avalanche mode) in order to extend the rate capability of the RPC. This change in operation mode and the need for high rate capability have been the drivingforces of the R&D effort for RPCs during the last years. 2. AVALANCH~ MODE The conventional RPC was originally conceived as a low rate device that could be constructed at low cost. It was operated in s t r e a m e r mode; the s t r e a m e r produces a large signal with a fast rise time. The resistive plates were needed to keep the streamer discharge localised, so that only a small area of the device was affected by each streamer. They could be operated at full efficiency well above th_e 'knee' of the efficiency plateau; in fact, the time resolution improved with the degree of over-voltage. This over-voltage also allowed a rather loose tolerance on the size of the gas gap; t h u s cheap mass production was attainable. The problem with the streamer mode is t h a t a large amount of charge is produced inside the gas gap, which has to be discharged through these resistive plates. This is the cause of the rate problem. The solutions are: (a) to find a very low resistivity plate material; or (b) to reduce the amount of charge. However, if we reduce the plate resistivity below 109 f~.cm, we can enter a mode of continuous discharge. Thus, to achieve high rate capability, we are forced to reduce the production of charge by o p e r a t i n g the chambers in avalanche mode. Avalanche

M.CS. Williams~Nuclear Physics B (Proc. Suppl.) 61B (1998)250 257

mode is the same mode of operation of parallel plate chambers with metallic plates; as is well known, it is difficultto reach high efficiency with these devices for m i n i m u m ionising particles; also they need to be constructed with a gap tolerance of 5 micron with carefully smoothed edges of the plates to prevent discharges around the border of the chamber. To find a way to construct low cost RPCs and to operate in this inherently difficult avalanche mode is the fundamental problem for the design of the RPC. The gas-filled gap of a parallel plate chamber is both the medium for production of primary ionisation and the growth of the avalanche. With Argon based gases (in common usage in gaseous detectors) there are 3 clusters of primary ionisation.per m.m for a through-going m i n i m u m lOmslng particle. Thus, there is a 4% probability that there is no ionisation within the nearest m m to the cathode. Thus, to operate a 2 m m gap R P C at high efficiency (-100%), the gas gain has to be set at a value high enough so that a single electron avalanching over the remaining I m m can produce a detectable signal. The fast signal is only 1/aD of the total (-7%), so a gain of 106 per m m will produce a 10 fC signal from a single electron avalanching over I m m . A gain of 106/mm is a gain of 1012 over 2 m m . This is shown

251

schematically in fig. 1. Since avalanches of 108 are a t the threshold for s t r e a m e r formation - one can easily see that it will be impossible to operate the conventional RPC at high efficiency with an Argon-based gas working in avalanche mode without having a large fraction of streamers. Since one needs very sensitive amplifiers for avalanche mode operation, streamers show up as very large clusters of hit strips (not such an ideal characteristic for a trigger device). There are two possible solutions to overcome this problem. One solution is to use a dense and strongly quenching gas in the 2 m m gap chamber. A dense gas increases the number of clusters of primary ionisation, while strongly quenching gases can inhibit the formation of streamers. Another solution is to increase the size of the gap; for exsmple, if we consider an 8 mm gas gap, again one needs the 1 mm closest to the cathode for the creation of clusters of primary ionisation; however in this case the 10 minimum gain is across the remaining 7 ram, which leads to 7 106 gain across the full 8 ram. We have tested these two solutions using a gas containing 85% of freon 13B1 (CCI3Br) in a 2 mm gap RPC compared with an 8 mm gap chamber filled with an Argon based gas mixture. The results have been previously presented[6]. The main findings of this comparison were that the 8 mm gap chamber has a better rate capability, lower power dissipation in the gas and could be constructed with a more relaxed gap tolerance. However the 2 mm gap chamber has a better time resolution. 3. BASIC PROBLEMS FOR AVALANCHE OPERATION OF RPCs

Figure 1. Schematic r e p r e s e n t a t i o n of problem encountered if conventional RPC is operated in 'avalanche mode'. A gain of 106 is needed across 1 m m to obtain high efficiency. This results in a gain of 1012 across the 2 mm gas gap.

If the gas gap of an RPC is increased to enhance the streamer-free range of operating voltage, the time resolution is degraded. The reason for the time j i t t e r is t h a t the production of p r i m a r y ionisation is a statistical process. On an individual event basis one does not know where the clusters of ionisation will occur, or the exact number of electrons in a cluster. We set the gas gain such that detectable avalanches are produced from any ionisation cluster within the closest mm to cathode. Since the drift velocity of electrons at this field is -10 ns/mm, it is easy to conclude t h a t a time resolution of this order, -10 ns, is attainable. In general, as one increases the gap size, one works at

252

M.C..SI Williams~Nuclear Physics B (Proc, Suppl.) 61B (1998) 250-257

lower electric fields (therefore the drift velocity is slower). In addition the change in gain with distance is lower, so detectab_le avalanches can be initiated further from the cathode. These two effects increase the time tter of the avalanche signal. However the asic problem is t h a t a finite distance is n e e d e d for t h e p r o d u c t i o n of p r i m a r y ionisation. There is another basic problem for the RPC. This is t h a t the number of electrons, N, created in an avalanche is given by the formula: N=NoeaD; where a is the Townsend coefficient, D is the distance the avalanche avalanches and No is the initial n u m b e r of electrons. However this is just the average number; there are fluctuations on this size which are exponential in nature. We have studied this effect by simulating the charge spectrum obtained from various RPCs[7]. Our finding is t h a t in a single gap of an RPC, the avalanches from i n d i v i d u a l clusters of ionisation 'merge'; thus one is dominated by fluctuations. These fluctuations have a very bad effect on the charge spectrum; the most probable value is peaked at zero and there is a very long tail of large sized avalanches. This is t h e worse possible distribution of pulse sizes for a detector, as one needs to work at the lowest possible threshold but also contend with very large avalanches (which also can provoke the formation of a streamer). A possible mechanism of this 'merging' is as follows: close to the anode the avalanches are large; thus an avalanche can be affected by the tail of positive ions from a preceding avalanche. Either there can be recombination, or a reduction in growth of the avalanche due to the reduced electricfield. Thus the two basicproblems of operating R P C s in avalanche mode are: (a) that there is onlX one 'merged' avalanche in a single gap RPC; the size of this avalanche is dominated by avalanche fluctuations; (b) one needs a finite distance (close to the cathode) in order that clusters of ionisation are produced; the statistical n a t u r e of ionisation within this distance generates a time jitter.

~

regions of 0.5 ram. Operating these devices at the appropiate voltage for full efficiency, one finds t h a t the maximum gain, across the (a) M O N O - G A P

RPC

9ram

I m m closest to j Cathode J ~'i

Anode

Cathode

(c) MULTI-GAP RPC \

3mm

4. S O L U T I O N S In the top two diagrams of fig. 2 we show a schematic cross section of (a) 9 m m monogap RPC a n d (b) double gap (2 x 4.5 mm) RPC. For the double gap, the I mm region n e e d e d for creation of p r i m a r y ionisation clusters can be divided into two

Anode

|

Figure 2. Schematic representation of a monogap RPC (top); a double gap RPC (middle) and a multigap RPC (bottom).

M.C.S. Williams~Nuclear Physics B (Proc. Suppl.) 61B (1998) 250 257

9 m m gap of the monogap RPC, is the same as the m a x i m u m gain across the 4.5 m m in the double gap case. Thus, in the case of the double gap, the characteristic distance which controls the time jitter is reduced by a factor 2. In addition we now have two independent gaps, t h u s two i n d e p e n d e n t avalanches. Therefore there will be an averaging of the avalanche fluctuations. Thus the double gap RPC is a big improvement compared to the mono-gap concerning the two f u n d a m e n t a l problems of t h e RPC discussed in the previous section. However this solution has its drawbacks. The gas gap is half the size of the monogap chamber; therefore the tolerance on the gap size needs to be a factor of two smaller. The chamber can be read-out with a single p l a n e of read-out strips, however they are sandwiched between the two RPC planes. This c o n s t r a i n s the segmentation of the read-out. In the bottom graphic of fig. 2, we show (c) a multigap RPC. The construction is similar to the 9 m m monogap shown in fig 2(a), except t h a t two extra resistive plates have been added, dividing the single gap into 3 sub-gaps of 3 m m each. The electrodes, to apply the voltage and to pick-up the induced signals, are mounted on the external resistive plates. No electrical connections are made to the internal plates; they are allowed to take the voltage given by electrostatics. Each subgap is independent. The resistive plates act

253

as dielectrics for the fast induced signal from the avalanches, t h u s the pick-up electrodes sum over avalanches in all three sub-gaps. The intermediate plates have electrons fed into them from one sub-gap, and positive ions from the other sub-gap; thus the net charge is zero. If the voltage does vary on one of the plates, the gains of the sub-gaps change. This results in a net flow of charge into the intermediate plate which adjusts the voltage such t h a t the gas gain in all sub-gaps is equal. We have built many multigap chambers and have measured efficiency plateau curves many times in a pulsed-beam environm e n t at various fluxes. In all our tests we have never observed any sign of instability. For the multigap RPC, the i m m sensitive region of the monogap chamber is divided into 3 regions of 0.3 mm, thus bringing about a corresponding improvement of the time resolution. There are now three independent avalanches, t h u s the Quilt-in' averaging m e c h a n i s m will b r i n g a b o u t a big improvement of the pulse height spectrum. In fig. 4 we show the s i m u l a t e d charge spectrum for a 9 mm monogap RPC and a 3x3 m m multigap; in both cases the average charge was set to 2 pC. In figure 4 we show the measured charge spectrum from a 3x3 mm multigap compared to simulation; in this ease the average charge is 0.5 pC. 0

1

2

3

4

5

6

*C

5pC 400

[9 m m MONOGAP [ 100

200300~

~ average = 2pC --

Data

....-.-...-...

i~!~i;:':~!~!; S i m u l a t i o n ~

0

"''!

25o

....

"% 13 3

mm

MULTIGAP

'

100150

~ r ~ _

I '

"

50. u~ ....... 1....... '2....... "3....... 4 ....... Fast induced signal [pC] Figure 3. Comparison of simulated charge spectra for a 9 m m monogap and 3x3 m m multigap RPC. In both cases the average charge is set at 2 pC.

25

50

75

loo

125 150 175 2 0 0 225 2 5 0 A DC c o u n t s

Figure 4. Comparison of 3x3 mm multigap data with simulation.

254

G. H I G H L I G H T S MULTIGAP RPC

M.C.S. Williams~Nuclear Physics B (Proc. Suppl.) 61B (1998) 250-257

O F I t & D W I T H THE.

At the start of 1996, t h e multigap was a n e w c o m e r in the f a m i l y of resistive plate chambers. The goal of the R&D was to gain experience with this design in order to fully u n d e r s t a n d the operation. To facilitate this u n d e r s t a n d i n g we have built and tested 3 variants of this chamber - (a) 2 gaps of 4 m m each (2*4 ram) (b) 3 gaps of 3 m m (3*3 mm) and (c) 4 gaps of 2 m m (4*2 mm). All three chambers were extensively tested in the T9 test b e a m at CERN, where m a n y parameters were m e a s u r e d as a function o f voltage and rate. One satisfying outcome was the very good a g r e e m e n t between the simulated and the m e a s u r e d charge spectra shown in figure 4. This has been fully discussed[7]; the main f'mding is t h a t the avalanches 'merge' in an individual gap - and t h u s single gap devices are d o m i n a t e d by avalanche fluctuations. This accurate simulation can be used as a tool; we h a v e u s e d it to i n v e s t i g a t e constructional aspects of the chamber, such as the effect of spacers and the effect of gap tolerance. 5.1 R e s u l t s f r o m t h e t e s t b e a m Efficiency plateau curves for various rates for the three types of chamber are shown in fig. 5. We also show the dark current, which is very low and decreases with the n u m b e r of aps. We have noticed previously t h a t the ark c u r r e n t starts to rapidly increase with over-voltage; the cause is the production of a large n u m b e r of'random' streamers; low dark c u r r e n t is an indication of low n u m b e r of 'background' streamers. One can also see t h a t there is a step decrease of efficiency at elevated voltages and at high rates. This is caused by 'dead areas' in the RPC created by s t r e a m e r s (initiated by previous throu6hgoing particles). Of course this is a regmn w h e r e one does not w a n t to operate the chamber; however t h e r e is a s t r e a m e r free plateau of -500 V for the 3*3 m m multigap and -400 V for the 4*2 m m multigap. One would expect s m a l l e r gaps to g e n e r a t e avalanches with a higher density of positive ions; t h u s this earlier initiation of streamers is expected. In fig. 6 we show the time spectra of these three chambers. We are using a fast current amplifier followed by a time-over-threshold discriminator. We can t h u s m e a s u r e the time of the leading edge, the trailing edge

~

2x4 m m m u l t i g a p 100

-

....

; ....

,-~-~- . . . .

6o ............. i = i

*

I m 20 Hz/¢m 2 I ] • 140Hgcm 2 j 0 650Hz/©m2

I

............ ............. i....

..]

6

.~ 40 ............~..0".....~ ............ i ......... "'"i''?" ~20 11.5

12.5

4 2

= .

~

13.5

HV (kV) 3x3mm multigap

[ D 75 Hz/cm 2 ] @ 1.3 kHz/cm 2

I I

80

............

..........

........

0.8

..............................................,............

~., 60

r

0.6 "~

40

= ........... i.............~-............~-.............,............ 0.4 E~ 20 ............*.......................................'............ 0.2 ': '~! L! -! j . ~ N=J ' ' 0 ~14.5 15.5 16.5 .

100

.

.

.

.

6o

(D 4 0

r=.l

.

.

.

.

.

.

,IJ@ r" 2.5 ]d4,/c 4 x 2 m m mHV u l t(kV) igap 700HrYcm2 Hz/cmm2 I .• . . . . . . . . . . . . ----- . ~ . ] o

80 ~

.

........

..........' " '

.........

•

i

.........

"~i

o.s

I 0.6

....

20 .

. . . . . . . . . 15 16 HV (kV)

0.2 17

Figure 5. Efficiency versus high voltage for 3 versions of the multigap RPC for various fluxes (flood illumination). The dark current corresponds to a chamber of 24x24 cm2 active area. The gas mixture used was 80% Argon, 15% C02, 4% iso-C4H10, 0.5% C4F10 . and form an average of the two on an event by event basis. This average is the first order slewing correction, and gives an indication of what we m a y expect from a constant fraction (or similar) discriminator. The binning is 1 ns per bin for the leading and trailing edge h i s t o g r a m s , a n d 0.5 ns per bin for t h e average. One can m a k e varmus observations. All time spectra shift to earlier times as the

M.C.S. Williams~Nuclear Physics B (Proc. Suppl.) 6IB (1998) 250-257

255

I 2*4ram multigap I • !300 ~i

Leading edge "300Trailing timing ~tI edge timing

:

"~

9ns

o

5o.

• t

/t300

I iJ~

loo

~[

8ns

15o

Average timing

~ 100

5.5ns - f

.~

s;

[ 3*3 mm multigap ] I Leading edge ,f400 I Trailing I A [ [~ ] edge ] timing |[~]0 ~ timing ] :I200

-

0 ..... ,~-i;o-{k;'l

'1 4" 2mm b t 400

~ Leading edge

, [

II

'I

![300

timing

~t

!

25

I Average [} timing

50

multigapl [ Trailing"-edge I A I timing

~ns

75

i

|

II-~f

-f 1-

25

50

:oo:

Average

]

=m,=gi 3.Sns

75

I

i00 ]

Tim e [ns] Figure 6. Time spectra for 3 differentversions of the multigap RPC. A time-over-threshold discriminator is employed and the time of the leading edge, the trailingedge and the average on an event-by-eventbasis is shown. The histograms of the leading and trailingedge have I ns/bin;

the 'average' histogram has 0.5 ns/bin.

gap width decreases; this decrease corresponds to a drii~speed for the avalanching electrons of 10 ns/mm. The time resolution improves with an increasing number of subgaps as expected. Another important aspect of RPC time resolutionis the time walk with voltage and rate. W e are going to have to vary the voltage to follow atmospheric conditions; additionally,differentregions of the RPC may be on different parts of the

plateau curve due to variation in gap dimensions or due to a difference in rate. The mean of the time spectra is shown in fig. 7 for the 4*2 mm multigap (similar curves are obtainable from the other two chambers that were tested). It is clear that as the voltage is increased the avalanche signal becomes larger; thus the leading edge

becomes earlier and the trailingedge later; however the average (i.e.slewing corrected)

M.C.S. Williams~Nuclear Physics B (Proc. Suppl.) 61B (1998) 250 257

256

O l e a d i n g edge u t r a i l i n g edge{

4*2 mm multigap 70 .... ~ ....~ .... ,

]

~ average - - -

!

. . . .

I - -

-

!

-

65 ..........•...........:......................T.........."..........".......... 60 il rat~ 100 Hz)cm 2' i i •

~

.

.

i

E

55 ...........r..........~ .........•...................~ ........-r .........

so .......... i........... i. . . . . . . . . . . 45

".... . . . . . . . . . . . . . . . . . . . .

" .......... " ..........

......... ~ | ii

15.2 70H P.

......... : Iw * ," ' ,.,

, ' "i ,

15.4 ........ ..................

, , .:

|" " ' ' " i

|" " '

15.6 HV (kV) !

........ ..%L.m....

'"

|" " '

15.8

I00" ~ ~

........ ...................

~ G a s g din 3"2 1071 3 m m ] equivalent toapplied

~

95-

.......... e==~

4 6 0 . H . V . = 15.7kV : ....................................... • "..............................i .................. : ........

................................. ..................................

~ 90-

SI~TI

8s-

ON l ~

I Are.

80-

[

100 1, 000 10,000 Rate [Hz/cm 2] (flood i l l u m i n a t i o n ) F i g u r e 7. M e a n of time spectra for the l e a d i n g e d g e , trailing edge a n d the average versus high voltage and rate. spectrum has a time walk of only 1 ns over a 600 Volt variation. A similar behaviour can be observed for the variation with rate. It is clear t h a t the front-end electronics needs to have some slewing correction built-in if we are going to fully exploit the time resolution of the RPC system. 5.2 U s i n g t h e s i m u l a t i o n a s a t o o l One obvious advantage of the multigap is the installation of the spacers (necessary to fix t h e size of t h e sub-gap). These are staggered from sub-gap to sub-gap such t h a t a through-going particle only traverses a t most a single spacer. Thus, in the other two sub-gaps (in the case of a 3x3 mm multigap RPC) avalanches are produced and give a signal. Additionally since the space h a s a dielectric constant greater t h a n the gas filled gap, the e l e c t r o s t a t i c coupling b e t w e e n t h e a v a l a n c h e s a n d t h e pick-up s t r i p s will increase. There is yet another effect; if there

Region without

75

o

ct dr

by spacer

]

20

40 60 80 100 Threshold [fC] Figure 8. Effect of spacers in a m u l t i g a p RPC. L e a k a g e c u r r e n t across the spacer leads to a higher field in the two remaining sub-gaps. This effect has not been simulated. is a leakage c u r r e n t across the spacer, the field a r o u n d the spacer will be reduced; h o w e v e r t h e r e will be a c o r r e s p o n d i n g increase of electric field in the other two gaps. We can compute the effect of having 3 sub-gaps a n d also the case of 2 sub-gaps (with increased electrostatic coupling). This is shown in fig. 8, where the efficiency versus threshold is shown for these two cases. I f one worked at a 15 fC threshold and covered 20% of the active a r e a with spacers, the global efficiency would fall from 99.5% to 98.7%. To c o m p u t e t h e e f f e c t of l e a k a g e current is not possible; however we scanned over a spacer bar with a very defined beam. We observed a small rise in efficiency leading us to believe t h a t the leakage current helps by increasing the electric field in the two subgaps t h a t are without spacers at t h a t point. A n o t h e r i m p o r t a n t a d v a n t a g e of the multigap is t h a t the electric field is defined

M.C.S. Williams~Nuclear Physics B (Proc. Suppl.) 61B (1998) 250 257

b y t h e voltage applied across the three gaps. The field in the individual sub-gaps is the same for each. We can compute the effect of a change in the total gap dimension. For example if we consider the case of a 3x3 mm multigap, there are 9 mm of gas. If we increase this gap to 9.3 mm keeping the applied voltage the same, there will be a lowering of the electric field; thus a lower gas gain. However the avalanche has now an increased distance to develop in, which can ~artially compensate the drop in gas gain. In g. 9 we show in the upper plot our measurement of the Townsend coefficient a, versus applied voltage for the 3x3 mm multigap. In the lower plot we show efficiency versus threshold for 4 cases. The top curve is the 3x3 m m multigap with an applied voltage of 15.9 kV. The lowest curve corresponds to an electric field obtained by increasing the gas gap to 9.3 mm, but keeping the individualsub-gaps to 3 mm in size. The two i n t e r m e d i a t e curves correspond to two cases: either the 300 ttm is applied to only one sub-gap, or each sub-gap is increased by 100 ttm. One can see that 60.

,lO

8 la

~'~ 55

i07~ 50

.-- Gas gsin across 3 mn~

15.2

15.6

16.0 Voltage [kV]

16.4 a - ~7

I

I

.-, 9 0 "

8o •~ 70"

~ 6o" l agaps°fl 13~aP sd~N

a=5325 I 3 m m [ I 3.1mm l [ 2 g a p s o f 3 D m m 50" SIMULATION [ 1gap of3~3 mm |

4 0

-

0

.

.

I

20

.

.

.

iI

40

.

.

.

iI

6)

.

.

.

Thresho Id[fC]

iI

80

i

i

i

I

100

Figure 9. Simulating the increase of the total gap width from 9.0 to 9.3 mm. See text for details.

257

there is a compensation effect; the reduction in field caused by the increase of gap size is partially compensated by the additional distance that the avalanche has to develop in. This compensation is about a factor 2. Whether there is a compensation effect or not depends on the variation of gas gain with electric field, and the variation of field with t~p dimension. Since the voltage applied to e multigap is across the full 9 ram, the change in field with gap size is much smaller than for conventional narrow gap RPCs. Thus, we can calculate the tolerance we need for the construction of the multigap RPC; this is (500V/15500V)*9000 lam *2 = 580 ~tm (i.e. a = 580/'/12 = 170 tim). 6. CONCLUSIONS We h a v e d i s c u s s e d the p r o b l e m s associated with operating resistive plate chambers in avalanche mode. We show that the m u l t i g a p design alleviates t h e s e problems. In addition the multigap RPC satisfies the requirement for a muon trigger device for the LHC collider. We also have had two 1 x 1.5 m2 3x3 mm multigap chambers built by Pol-hi-tech[8]. It is essential that we understand the problems of mass production; this is the start of such investigations. References 1. M. Bertino et al., Nucl. Instr. and Meth. A283(1989)654. 2. M. Iori and F. Massa, Nucl. Instr. and Meth. A306(1991)159. 3. I. Crotty et al., Nucl. Instr. and Meth. A 329(1993)133. 4. The ATLAS T e c h n i c a l Proposal, CERN/LHCC/94-43, LHCC/P2 (1994). 5. The CMS T e c h n i c a l P r o p o s a l , CERN/LHCC 94-38, LHCC/P1 (1994). 6. E. Cerron Zeballos et al., Nucl. Instr. and Meth. A374(1996)132. 7. Avalanche Fluctuations within the Multiap Resistive Plate Chamber. E. Cerroneballos, I. C r o t t y , D. H a t z i f o t i a d o u , J. Lamas-Valverde, R.J. Veenhof, M.C.S. Williams and A. Zichichij Accepted for publication in Nucl. Inst. and Meth. 8. Pol.hi.tech. s.r.l., S.P. Turanense Km 44,400, 67061 Carsoli (AQ), Italy

~