High Voltage Board - Prototyping

Original Design

The High Voltage board is used to power the Photo Multiplier Tubes (PMTs) which are used in the simulation of data taking in order to test the front end electronics. As the MobiDick was developed there were additional patches applied to the HV board. Most noticeably the Optocoupler and the DC DC converter which altered the incoming +24V to +5V. The LED lights were mounted on the surface of the board. This was later changed to include small pins to allow extension wires to move the LEDs off the board for better displaying purposes.


The image to the right shows where the optocoupler was added onto the board. This was done in order to isolate the HV board from the main motherboard. There was an issue when the HV board was powered on that caused numerous components to burn out on the motherboard. The optocoupler isolates the HV board by converting the electrical signal to optical and then back to electrical. This optocoupler is setup to work in one direction only. This means that the HV board is truly isolated from the main motherboard and the problem was fixed. The DC DC converter is not shown in the image because it is located on the actual wire of the power connector.

Upgrade Design

The first step in the upgrade of this board required the capturing of the PDF schematics into a digital format. The program used is called Orcad Capture. This program sets up the logical properties between all the components on the PCB board. Once the logical connections were made the part footprints were then defined. These footprints specify the physical dimensions of the component such as the package outline, number of pins or the size of the mounting pads. This must be done for every component that is used on the board.


The image on the bottom shows a portion of the schematic that was captured from the PDF files. This is the optocoupler which is now included into the actual design. The solid red lines are the physical copper traces in the board. The DC DC converter as well as the correct pins for the LED lights were also added to the schematics which meant that they could be included into the board layout. The diagram at the top is a close up of the physical layout of the optocoupler.


The diagram on the far right is the bottom view of the entire board. The traces in this design were all routed "by hand". This is the final design of the board and it has only recently been printed.

Production of the Board

The HV board has been produced. It recently arrived which means the component placement can begin. This will require soldering each part individually onto the board by hand. We will omit the Ultra Volt Converter for the first stage. Once this step has been completed we will begin testing for errors in the circuitry. This is the reason we will not include the Ultra-Volt since it can be dangerous.


If no errors are found then the we will mount the remaining components and begin the process of testing the board in operation. This will require access to the ATLAS detector where we will connect the MobiDICK4 (containing the new board) to the draws and perform the required tests.

Mounting Components

The smallest components are generally mounted first since they do not cause any obstruction. The picture on the bottom left gives an illustration of the size of the components that we are mounting. They are fairly small but large in terms of PCB manufacturing standards. Since they are small by our standards and especially since the mounting is being done by hand it requires patience and a steady hand to align and solder the components correctly.


The picture on the top right shows Robert Reed (PhD Student) working on the soldering process using tweezers to hold the components in place while he applies the heat required to melt the tin.


The picture on the far right is the board with almost all components mounted. There were some faults that were found in the design which will be patched up in this board to allow testing but will be fixed in the next version of the board. The large gap that is missing on the board is the Ultra Volt component. As mentioned before this will remain off the board until we know there are no faults that have not been fixed.


The board was tested for faults and a few were found and fixed. Once powered on the +24V rail LED lit up which indicated that it was functional. A signal of +3V was then sent through the input to simulate the power on signal. This resulted in the +5V LED powering up indicating that the VCC rail was also working correctly. The input for the Ultravolt was tested and all voltages were correct. The Ultravolt was then connected and the board was powered on with the input signal on. The bright blue LED turned on indicating that the high voltage was present. The voltage was measured to be slightly lower than expected so the input to the Ultravolt was adjusted, using the potentiometer, to increase the voltage. The completed board and some testing is shown in the pictures below.

Wits HEP Group, School of Physics,
University of the Witwatersrand,
1 Jan Smuts Ave, Johannesburg, South Africa
Telephone: +27 11 717 6848
Fax: +27 11 717 6879