CROCS Electronics
Introduction
As a quick and dirty way collecting available documentation and drawings, this is initially a collection of various e-emails plus various drawings and circuit diagrams.
Overview
The overall scheme of a central Control Station with up to 4 satellite Detector Stations is shown in Fig1. The plan is to use a multi-core cable both to provide electrical power to the Detector Stations and to carry signals to and from the Control Station.In this scheme the power is derived from a single 24 Volt mains adapter.
Just what is needed in each of the Detector Stations is shown in Fig2: this shows in detail how the power rails needed by the Silicon PhotMultiplier (SiPM) and the various bits of electronics are generated, and the other functional blocks. These comprise:
1)The Signal Chain circuits that amplify and condition the tiny pulses from the SiPM and send them along the cable to the Control Station.
2)The Bias Control circuits that allow the SiPM operating conditions to be finely adjusted to have adequate gain with stability.
3)A communication path that allows the Control Station to both to adjust the SiPM bias (and possibly other features of the electronics) and to monitor what is going on: temperature, voltage, current for example).
For preliminary work much of this circuitry is redundant and far too elaborate. I suggest that for the early stages, and to keep the project within the scope of a school environment, the remote control aspects be dropped. Fig3 shows a reduced circuit with a manual SiPM bias control that should meet this need. It is this that a school wishing to investigate the operation of a scintillator and SiPM might build.
The e-mails
These give some explanation of the system and the thinking behind it that should prove useful. Note that the ideas evolve, and some of the details expounded at the start have been changed: the circuits given above should be taken as the latest recommendation.
1)%%%%%% MJG to the CROCS Community on 28/01/2011 %%%%%%
IhavebeengivingsomethoughttotheconnectionbetweentheControlStationandtheDetectorStationsofa"CROCS"CosmicRaysetup.Itwouldmakesensefor ustoadoptthesamegeneralscheme,soIwilltrytogiveanideaofthesort ofthingI'mproposing.(Butnotethatnoneofthisisfixed-andwillbemodified inresponsetofurtherthoughtandexperience-includinginputfromyou).
EachDetectorStationholdsascintillatortileandSiliconPhotoMultiplier(SiPM). Thelatterneedsastable,lownoiseandadjustablebiasvoltageofupto35V(luckily ofverylow current). Thereisalsoapre-amplifierthatneedsaround10Vatmaybe10mA. Therewillalsohavetobesomeformofdrivercircuittoallowthesignalstobesent backalongthepathtotheControlStationinacleanway:thiswillneedalowersupply voltage,typically3.3Vor5V.Thecircuitrytoallowtheadjustmentofthebiasvoltage and other communication functions will probably also run from 3.3 or 5V. There may be otherfunctionswewouldliketodowithintheDetectorStation-e.g.signal conditioning,orvariousmonitoringtasks-andweshouldchooseaconnectionscheme thatleavessomeroomforfutureextensions.
Thisallboilsdowntothefollowingrequirementsfortheinterconnection:
1)electricalpowerdelivery
2)gettingthesignalpulsesbacktotheControlStation
3)controlofDetector(primarilytheSiPMbiascontrol)
4)gettingmonitoringdataback-I'mplanningtomakethecontrolpathof3)bi-directionaltoachievethis.
5)theseparationoftheunitscouldbe>10metres,andthecablingpathwill be longerthanthis.
StevehasproposedtheuseoftheubiquitousRJ45connectorsandCAT5Ecablingasa cheapandreadilyavailablestandard.Thishas(Ibelieve)4twistedpairsofrather thinwire.Ifweallocate1pairtosignalreturnandanothertothecontrol communicationpath,thisleaves4wirestodeliverallthepowerneedsofthedetector station.It'sworthspendingabitoftimeplanninghowthisistobedone.
If we standardise on delivering a single supply, and generate the other voltages we needfromthis"Bulk"supply,wewillonlyneed1pairofwires(assumingtheycan handlethecurrent): if wehave2pairsfree,wecouldparallelthemuptoreduce theresistance.
We could derive the other voltages from the bulk supply using "Linear Voltage Regulators"-buttherearetwoproblemswiththis:
1)thebulksupplyvoltagehastobehigherthanallofthederivedsupplies.
2)lotsofpowerwillbewasted(discuss)
LuckilytherearerelativelycheapandsmalldevicescalledDC-DC("DC-to-DC")convertersthatcanefficientlychangepowersuppliedatonevoltagetoahigher orlowervoltage(functionallyasortofDCtransformer).(Howdothesework?)Theyallowustogreatlyimprovedeliveryofthepowerneededoverthe long,thinwires.Saythetotalpowerneededis10Watts,andtheresistanceofthe wiresis3Ohms.Ifwedeliverthatpowerat5Volts,that'sacurrentof2Amps. Thatmeanswelose6Valongthewires.Soweneedtodrive11VfromtheControl Stationtogetthe5Vatthefarend.Andinadditiontothe10Wdeliveredtothe detector,wearewasting2Ax6V=12Winthewires.Noneofthisisgoodnews.
Butifwechoosetodelivertherequiredpowerat20V,thecurrentdropsto 10W/20V=0.5Amps,andthedropalongthewires will be 3 x 0.5 = 1.5 Volt; thepowerlostinthewiresis0.5Ampx1.5Volts=0.75Watts,whichisfar morereasonable.
SoIhopeIhaveconvincedyouofthesenseofthisapproach.I'msorrynottohaveproducedtheodddiagram,butIhopetheargumentiscleartothosewhounderstandV=IxRandP=IxV.IshouldmentionthatthereisapotentialproblemwiththeseDC-DCconverters:theygenerateelectricalnoisethatmightupsetoursensitiveelectronics;butIbelievethiswillnotbeaproblemifwechoosegoodconvertersandtakecarewiththeiruse.
Thedevilisinthedetail:thepracticalitiesare:
1)TRACOmakealargerangeoftheseDC-DCconverters,andaselectionofthemareavailablefromFarnellandRS.WeshouldusePCB-mountingvarieties,withanoutputvoltageandwattagetosuittheparticularsupply.Someregulatetheoutputvoltagetomakeitconstantaseithertheloadcurrentortheinputvoltschange:thesearetheonesweshoulduse.
2)someacceptawiderangeofinputvolts,whileothersspecifyarestrictedrange:weshouldgofortheformeraswe'renotsurejustwhatthedropdownthewireswillbe.Iwouldliketostandardiseonadeliveryvoltageof24V(itkeepswirelossesdown,issafe,andisrelativelystandard)
3)theoutputvoltagesavailablearelimited:12V,5V,3.3Vareallavailableinarangeofpowers,but35Volts(thatweneedfortheSiPMbias)isnot.
4)Mysolutiontothisproblemistousethefactthattheoutputsareusuallyfloating,soitispossibleto"stack"theoutputofoneontopofanother.Thuswecoulduseadualoutputsupplythatprovides±15Vasa30Vsupply,andreferencethistothe+10Vsupplyusedforthepre-amplifier,givinga40VrailfortheSiPM.
5)TheSiPMbiasneedstobeadjustablewithinsafelimits.Giventheverylowcurrentneeded,wecandothisusingapotentiometerto"tap-off"therequiredfractionofthis40Vrail.Foryourlabinvestigations,thiscouldbeaconventional20-turnscrewdriver-adjustablepot.Addseriesresistorsattopandbottomofthispottorestricttheavailablevoltagetoasafeandusefulrange(suchas25-32V?LookattheSiPMdatasheet).Notetheoriginaldeviceswegotranfrom17or18V.Forremotecontrol,Ihaveyettodecideonascheme.
IfyougotothispageofthePowersolvewebsite,youwillseetherangesofPCB-mountDC-DCconvertersthatI'vebeenlookingat:
TheTSM2415Dinitially lookshopefulforthe+24Vi/p,±15Vo/psupply,butis not regulatedagainstinputvoltagefluctuations,sonotthisone.
TheTMR2423looksbetter:forinputsof18–36Vitgivesawelldefined±15V.Theavailablecurrentis65mA,whichisfarmorethanweneed.
Fromthesamerange,theTMR2421mightbeagoodchoiceforthePre-ampsupply:itgives±5Vatupto200mA:thiswouldgivethe10Vrailweneed,and,asabonus,a5VrailthatcouldbeusedfortheotherfeaturesIplanonincludinginthefullsystem,suchastheremotecontrolofthebiasvoltageItalkedaboutinrequirement3)above.
Ihaven'tasyetcheckedpriceoravailabilityofeitherofthese.
Ihopeallthisiscomprehensible,typosandlackofdiagramsnotwithstanding.Iwilltrytoputtogethersomedrawingsasthingsdevelop,andwillletyouhavethem.
Maurice
2) %%%%%% MJG to the CROCS Community on 02/02/2011 %%%%%%
I've put together a couple of drawings that will help make sense ofmy last e-mail - see the attached .pdf
The first page shows the general arrangement of the central ControlStation with the (up to) 4 Detector Stations with the power and controlinterconnections. Note that it is all powered from the 24V power adapter.This should eliminate electrical safety issues. I have suggested a unitthat can be purchased from Farnell.
The second page shows the power circuits of the Detector Stations:essentially 2 DC-DC converters and a handful of passive components.
What I imagined is that you might make up the power circuits, and usethem to drive the SiPM and its amplifier. At this stage you could omitthe RJ45 connector, and take the 24V in using a power socket that matchesthe jack on the power adapter. Or you could include the RJ45 to preparefor the next stage: if so, I have a suggested pin-out:
Pin / Signal1 / Pulse+
2 / Pulse-
3 / Control+
4 / +24V
5 / +24V
6 / Control-
7 / PowerGnd
8 / Power Gnd
*** Warning: I have changed this recommendation: see Fig2***
Note that the Amplified SiPM output ("Pulse") is carried on atwisted pair of wires as a balanced signal, as is the Controlsignal. This gives good speed and noise rejection. Doubling upon the power wires would reduce the voltage drop along the wires.
This scheme is compatible with "Power Over Ethernet", which mightpay off in some way.
I haven't tested any of this, so please shout if you see anythingthat doesn't make sense.
Maurice
[Type text]CROCS_Electronics.docx of 2018-11-02Page 1 of 4