LED Pulsers for MULAN


Collaboration Update May Power Point.

Current goal is to complete the construction of one driver for testing.

·     16 channels card completed (PCBexpress board DIP connection)

·     ecl-ttl board designed and schematic and PCB complete

·     test of the PASS cables complete

o    Long cable-ecl-TTl

§     operates at 10 ns

o    Long cable-ecl-TTl-long cable-ecl-TTL

§     operates at 20 ns



  Collaboration talk Feb 2003 for downloading (power Point)

 A few more details of the after pulsing using the above slides.

 ARCHIVE:  Summer2002: Design & Status  A prototype was made for the 2002 summer PSI run. Details and discussions etc. are archived here.  Developments as they are made can be found below.

Proposal drafted to build the system. To download the file proposal.doc.

This outlines the expected costs to produce 500 boards and 25 16 channel driver boxes.

The total cost will amount to about $50.00 per channel.


Photoelectrons. Suggest that we use 100 pes as our signal goal.


Flasher rate maximum is 100 kHz based on the recharge time.


*********** Upcoming run Sept 24-Oct 31


DrG traveled to PSI for about 2 weeks (9/21-10/6).

·      Tested all installed LEDs (all but two were working, see prelim detector spreadsheet)

·      4 old boxes were updated using new cables and patch panels.

o     2 have all 12 channels working

o     2 have 10 channels working

·      Two twelve channel drivers installed hex 9 and hex 12.

·      Flight simulator generating 12 signals on each of its two outputs.

·      FS-driver-2 hexes working through MIDAS


“Howto” for the system


Some pics from the experiment.


Our first results (note two detectors were unplugged, only one detector from each detector pair)

 Hex 9 adc spectra, resulting gains

 Hex 12 adc spectra, resulting gains



Lid Connectors ***************************************************************

Lid connectors are shipped to U of Ill.  Installation  and cable connection.



Fiber connection and Flasher ****************************************

Fibers have been completed and shipped to U of Ill 7/18/03. Final design for the internal cabling and coupling is complete. There will be a 1’ cable with two connectors on the LED flasher side. One connector will be “type RED” the other “type Green”. Pins will be eliminated in these special connectors (Cabling diagram). The fiber-LED coupler is a small cylinder of plastic. Fibers (three) are glued to the coupler. Fibers are then glued to the light guide. Shrink tubing provides a light tight path. A cable tie with a adhesive will secure the fibers/flashers to the house wall. A diagram can be found here.

The method follows approximately the original design (original fiber/coupler design e , original cabling memo).



Old Driver New Flasher ****************************************

Here are some ways to use the old driver with the new flashers.



Flasher Boards *******************************************

Boards are done. Some modifications were added to the new boards = the new circuit.


340 boards were delivered to U of Illinois. About 50 of the 400 hundred boards were not working. The reasons for the bad flashers are not yet known. Once a flasher was tested as good it seemed fairly robust. It could be handled and connected to the drivers without failure. One flasher was used to test cables and was connected – disconnected several hundred times with no noticeable effect. The maximum driving voltage is 24 Volts. A flasher was run for an hour at this voltage with no noticeable effect. The flasher was connected and disconnected a few times to verify that it was not sensitive to the connection process.


Inserting the cable on the flasher in the improper orientation may cause damage or complete malfunction of the flasher.


Several flashers were installed in houses. Connections are described above. These flashers were driven by the old driver system using a special cable. One hex house was successfully pulsed a  7 kHz rate.


A flasher pulsed at 3 kHz is producing a steady 740 mV signal with FWHM=6.7 ns using a xp2020 - 2” tube directly viewing the LED. The pulser has been running continuously for over one week.


ecl connections *******************************************

A 34-pin 16 channel (1 gnd pair) ecl ribbon cable passes triggers to the driver. The FS (flight simulator) and

ecl-TTL/nim pin outs are shown here.



Drivers *******************************************

·      2 34 pin standard (0.1x0.1) connectors for trigger inputs from flight simulator.

·      ecl translator to make ecl to TTL conversion board

·      xilinx programmable gate array (for uprgrades) (use 12V from power supply)

·      same circuit for the driver boards but we need to double the channels/board and make the corresponding changes to the connector outputs for added channel as well as the new cabling scheme (see above).

·      12 V fan (on/off switch). (use 12V from power supply or include an AC to Dc converter)


Driver box details ***************************************************************

1U = 1.75 inches

Driver Boxes are 2U (3.5 inches)

         height        3.5 inches

         width        17 inches (box width) =>19” rack (front panel width)

         length       15 inches total 14 inches box dimension 1” clearance for cables (could be less)





Summary of status 8/19/03

o       Flashers (pulser boards) complete.

o       Installation in houses

§        method complete

·        fibers

·        connectors

·        cables

§        Final installation underway at Illinois

o       LED drivers

o       mechanical layout basically decided

§        detailed quotes for work are in progress

§        boxes ordered

o       cables completed

o       power supplies

§        Use the same as summer 2002

·        Final power consumption review needs to be completed

o       TTL driver and power boards

§        overall design complete

·        fine tune the component layout and test with a prototype

§        order components

§        order boards

§        assemble

o       ecl translators

§        working model on breadboard but further testing underway

§        analyze some alternatives

finalize design


§        build boards

§        order components

§        assemble

o    Internal cables need to be made









·      Chris Church –mechanical layout of the driver boxes (LED pulse width)

·      Eric J Stofferahn – Pulse lengthening (Fiber couplers)

·      Mike Clemens- gate array (Fiber testing) 

·      Simon Hale – Continue work on the ECL-TTL translator

·      Dan (D'Antuono,Matthew) – Circuit board layouts and manufacture

·      (graduated 2003) Charlie  Arnold – coupler, gate array (graduated: left mid July 2003)

·      (graduated 2003) Andrew Werner- gate array (graduated: left mid July 2003)



****************  Below is a list of things that were under study in order to reach the final design.  The list may not be up-to-date.

DRG’s TO DO LIST (below)

In order to complete the task of designing and building the pulsers the following needs to be done:

1.   Decide on the logical function of the LED boards.

a.    Testing DAQ

b.   Testing wiring

c.    Will they test for systematic timing errors as a function of where in the measurement period they occur?

LED will test the readout wiring, DAQ, Histogramming …. 

The option of using them during the run will be maintained.  The time of the pulse must be well know for systematic studies.

2.   Decide on the function of the drivers

a.    The pulse may need to be widened to drive the pulser (PECL should fire the comparator but he simulator will produce very short signals.)

b.   Based on the input there may need to be a second pulse very soon afterwards.

c.    probably want to go to a PGA (programmable gate array)

                                                   i.     buy the Zylex package and test it out.

3.   Find vendor and price for

a.    boards

b.   stuffing


4.   Determine source or a correction for afterpulsing

Solution: 1  kΩ in parallel with the diode.

5.   Determine the best rise time and minimum pulse shape attainable.

a.    We need to check correlations between amplitude and rate. (should be small)

b.   Check the correlation between trigger pulse width and amplitude (should be zero)

Solution: Diode can produce a 7 ns FWHM pmt response.

Test results that show the diode and the driver pulse were summarized at the February 2003 MULAN meeting. Conclusion was that a pulser could be built that operated at 300 kHz with 7 ns FWHM pmt response.

6.   Set specs for the LED pulser.

a.    Given the limitation in the number of independent trigger pulses (Flight simulator generating 170 independent triggers) and the interest in developing a double pulse (two LED/ detector) for some pmts. There may be a need to develop some multiplexing. Using the 10 pin connectors perhaps three (B+, trigger+, trigger-) would be assigned for one board and three more would be assigned for a different board. The LED pulser boards perhaps could be jumpered to choose either set of the three inputs. A single pulser driver channel could route these signals on a 10 pin cable. The cable could be teed at the detector end and control two pulsers.

Assume that pulser driver should be designed with option of upgrading to two LED board/channel.

b.   Perhaps a optical fiber can be used to deliver the pulse. This would greatly simplify installation.

First test indicate that a reasonable light flash could be achieved with a simple cap and short silica fiber.


Companies dealt with so far:

Gorilla circuits (made the first boards)





Gorilla recommends a company to do the stuffing (they have been the contact).

Orion Manufacturing in San Jose,CA 408-955-9001


for the 550 boards we will be building or you there will be no extra charge for not having parts on the reel.

Dino Paquiz

Orion Manufacturing

Account Manager

408.955.9001 telephone

408.386.1247 mobile





wrong ORION**************************************************


Address: 690 Cone Park Ct., Suite A, Merritt Island, FL 32952
Tel: (321) 452-1670 Fax: (321) 452-1678 Email:



Lauss suggested for stuffing:

Digicom Electronics, Richmond, Ca, Contact person: Mo O'Hady

Tel:  510 - 527 8084

Fax:  510 - 527-8187











7.   Find CAD program to do the work.

ORCAD university package for $2500 is reasonable and compatible with BU and Berkely

8.   Layout boards

a.    LED

                                                   i.     need to add the second trigger and B+ option

                                                ii.     need to add the resistors.

b.   Driver

c.    NIM/ECL

The flight simulator will deliver PECL. The comparator should work fine with this. (Pulse may need to be widened)

9.   Test out power supplies and order.

Rick ran the supplies and found no excessive heating. He thinks that the power supplies are okay that we perhaps made errors that caused the problems at PSI.  Keep the same supplies.

10.                   List and order electronic components

a.    see list

b.   Maxim 961 eua on tape

                                                   i.     quote to maxim minimum order seems to be 2500 (emailed for lower quantity)

                                                ii.     I can buy 500 not on tape perhaps I can put them on tape (Neward does this for large orders)

                                              iii.     Newark quote for 500 # 1508533   800-463-9275

                                               iv.     stknet.com

Ordered Max chips

Impossible to buy 500 on tape the minimum quantity is 2500. However the parts can still be purchased in bulk and mounted.  The option of having a second vendor put the parts on tape is not a reasonable solution (manufacturer’s statement) because the alignment is critical and the reliability of the tape packaging is often not good enough.  Buy the bulk chips and pay a bit extra in mounting cost.

11.                   Build prototype and test.

12.                   Redesign boxes, mount connections with improvements

Soccer Ball
truncated Icosahedron
20 hexagon

12 pentagons.





scints per house

total scintillators


Lid connectors










hex houses
















beam pentagons
















total scintillators







total connectors

















total pulsers











channels /box

total channels




driver boxes








extra drivers








total drivers








total channels








Rack 19” mountable (17” width usable). The absolute minimum is 1u=1.75”. We choose 2u vertical height 3.5.

1.   10 pin connector

2.   2 pots, 2 test points, 2 triggers (rework)

3.   2 simulator input 32 pin twisted pair (32 triggers/box)

4.   1 100 MHz clock

5.   gnd

6.   possible computer connection