Part VI
Building your own

Introduction
Building your own photon counting system has several advantages. First is the satisfaction of creating a sensitive instrument capable of detecting individual photons. You will also have a first hand knowledge of the instrument so repairs can be easily made. Most units seldom need repairs, however. Lastly, by building your own photometer system you can save considerable money.

Figure 8 shows a block diagram of a basic astronomical photon counting system.

Figure 8
Astronomical Photon Counting System Block Diagram

Figures 9 and 10 show close-ups of the photometry equipment built and used at the Hopkins phoenix Observatory.

Figure 9
Telescope and Photometer Head

Figure 10
10 MHz Pulse Counter and -1,000 VDC Power Supply

Photometer Head
Building your own photometer head can be rewarding. Someone once said you will build at least three before getting one that you are happy with. Hopefully the following information will make it easier for you.

Figure 11 shows a basic photometer head. Many variations of this basic design can be made.

Figure 11
Basic Photometer Head

The basic requirement for a photometer head is that the photomultiplier tube must be in the light path of the object being measured. Because of the photocathode surface variations, a Fabry Lens is used to spread the light on the photocathode and minimize the effects. A diaphragm is used to stop down a star and minimize background light from the sky and other stars. The diaphragm must have several openings ranging from a large opening for initially centering the star to the smallest, usually 60 arc seconds. To center the star in the diaphragm, an illuminated dual-cross hair post-view eyepiece and first surface flip mirror are required. A flip prism can be used in place of the flip mirror. A pre-viewer eyepiece and flip mirror is optional, but makes finding the star initially easier. A good finder scope on the main telescope can serve this purpose better than the pre-viewer, however. For faint objects beyond the capability of the finder scope, the pre-viewer allows use of the main telescope's aperture for finding the object. Lastly, for filter photometry a filter must be in the optical path just before the photocathode. Since various filters are normally used, a means of easily changing the filters must be made.

How the items are packaged is limited only by your imagination.Large telescopes can allow generous use of volume and weight of the head. Telescopes on the order of the popular 8" SCTs must have a more compact and light weight design. To create such a light weight and compact photometer head more clever designs are needed.

The addition of first surface mirrors can be used to fold the light path to the photomultiplier tube. Prisms and first surface mirrors can be used to change the light path for the post viewer. Only first surface mirrors, never prisms, should be used in the light path between the star and detector. Aluminum is the choice of materials for the housing as it can provide a strong light weight and electromagnetic shielded enclosure. All aluminum should be black anodized and/or painted flat black inside and out.

In the early 1980s I designed and tested several photometer heads. I settled on a very compact and light weight design. The design uses a 4.7" X 3.7" X 2.1" die case aluminum box (BUD #CU-234). While the box is close to ideal, the sloping sides make working with it difficult. I had a machinist friend do the milling and drilling work for me. For those skilled with access to a milling machine, the whole head could be milled out of a solid block of aluminum.

Figure 12 shows a diagram of the compact photometer head designed and used at the Hopkins Phoenix Observatory. The photometer head is about the size and weight of a typically 35 mm camera.

Figure 12
Photometer Head Top View Drawing

Figure 13 shows a photograph of the inside of the photometer head. The case is aluminum that has been black anodized. There is a lip around the cover which provides an excellent light seal. A small screw at the middle left side of the case is used to align the fixed mirror for the photomultiplier tube (the tube assembly is removed during this alignment). When done, the screw is fixed in place. The filter cage and diaphragm use screw type detents to key positions. The rods for the flip mirror and fixed mirror are 1/8" brass brazing rods painted black. Note the filter cage is in the dark slide position.

Figure 13
Photometer Head Inside

The Fabry lens is mounted on a bracket on the cover. See Figure 14. When the cover is in place the lens is in position between the filter cage and fixed mirror.

Figure 14
Fabry Lens Mounting

Figure 15 shows a close-up of the UBV photomultiplier tube photometer head used at the Hopkins Phoenix Observatory. It's shown mounted on a C-8 telescope. The white box on the back of the photometer head holds 2-AA batteries for the post-viewer illuminated reticle eyepiece. The knob on the white box varies the brightness of the reticle illumination. The top knob on the black case is for the flip mirror. The bottom knob on the black case is the 4-position filter selector (UBV and Dark). The two RG-58U high voltage and signal cables can be seen out the bottom (opposite side from the knobs).

Figure 15
UBV Photometer Head

Photomultiplier Tube
The photomultiplier tube detector is the key to the system. The tube produces small current pulses when illuminated by photons. The most popular low cost side-window photomultiplier tube is the 1P21. There have been several manufactures of these, including RCA and Hamamatsu.

The 931A (usually the lowest cost tube) is identical to the 1P21 except the 1P21 is a selected 931A. EMI Gencom, Inc. made an equivalent side-window photomultiplier tube, the EMI 9781. The EMI Gencom, Starlight-1 Photometer uses an EMI 9924A tube. The EMI 9924A is an on-end window type tube which can complicate a compact design.

Note:
Many 931A tubes are just as good and some are better than some 1P21s. To play safe, however, it is best to spend a few extra dollars and specify a 1P21. If you can afford it, buy several tubes and experiment to select the best tube. Remember, sometimes tubes may exhibit high dark current when first turned on, but after sitting in the dark for several hours with power applied, the dark current may go down to an acceptable level and may even remain there when first turned on next time, assuming no exposure to bright light in between. Tubes that at first may seem less than acceptable may prove fine. Once you have found a tube, it can last you a life-time.

Other side-window photomultiplier tubes with different spectral responses can be used or special projects. A 1P28, for example, extends the sensitivity further into the ultraviolet region, but at the cost of much higher dark current. For UBV work, where matching to the standard UBV system is desired, a 1P21/931A is recommended

Figure 16 shows a typical spectral response curve of the 1P21 photomultiplier tube.

Figure 16
1P21 Spectral Response

The 1P21 and equivalent photomultiplier tubes require a power supply capable of producing at least -1000 VDC @ 1 - 2 mA. It has been found, for analog operation (DC amplifier), voltages from -600 VDC up to -1000 VDC can be used. Once a voltage is determined, it should always be the same for future observations. If changed, the system zero points will change. Never change the high voltage in the middle of an observation set except to correct a drifting voltage. The different sensitivities will make the data inaccurate. The higher the voltage, the higher the gain of the tube (also the higher the dark counts). For photon counting a constant setting of -950 VDC seems best. At 50 degrees F and -950 VDC typical dark current/count is around 50 counts per second.

The typical gain of a 1P21 tube is over 10^6. For DC systems the output current is in nanoamperes. For photon counting the pulses are in the low microvolts with low nanosecond pulse widths and sub-nanosecond rise and fall times. The voltage and times are with the signal terminated in 50 ohms. High impedance terminations greater than 50 ohms will lengthen the pulse width. That is undesirable as it will increase the dead time.

Figure 17 shows a picture of a 1P21/931A photomultiplier tube along with a drawing of dimensions.

Note:
The tube's photocathode is aligned with the tube socket key (between pins 1 and 11).

Figure 17
1P21 Photomultiplier Tube

For new tubes see Hamamatsu at
http://usa.hamamatsu.com/

For surplus tubes see:
http://www.elexs.com/7Photo.htm http://www.tubesandmore.com/
http://www.vacuumtubes.net/ http://www.vacuumtubes.com/
http://www.vacuumtubesinc.com/
http://www.tubeworld.com/tubes.html
http://www.surplusshed.com/new.html

Voltage Divider
One of the secrets of a good photometer head, with a photomultiplier tube, is the voltage divider design. Figure 18 shows a voltage divider configuration that performs well with a 1P21. It can be used with either photon counting or analog systems.

Figure 18
Voltage Divider Schematic

If only analog systems are used, the pulse forming capacitors (lower four 100 WVDC capacitors) can be eliminated. They are very important for photon counting, however. The 0.001 uF 1000 WVDC capacitor at the high voltage input is important and must be there. Note that the voltage rating of the pulse forming capacitors is only 100 WVDC. This is because the voltage between stages is less than 100 volts. The smaller capacitors with lower working voltages can therefore be used.

The voltage divider provides a graduated voltage to all the electrodes in the photomultiplier tube. When constructing the voltage divider it is important to make all leads as short as possible. Low nanoseconds relates to frequencies greater than 100 MHz, so high frequency RF techniques should be employed. A special low leakage 11 pin photomultiplier tube socket can be obtained from Hamamatsu Corp.

Socket
A high quality low leakage tube socket should be used. The 1P21 uses an 11 pin socket. Figure 19 shows a diagram of the tube socket with dimensions.

Figure 19
1P21 Tube Socket

Note:
The tube's photocathode is aligned with the tube socket key (between pins 1 and 11).

These sockets an be obtained from Hamamatsu Corp.
They are: Side-Window PMT Tube Socket Part number E9678-11A
Hamamatsu Corp. 420 South Avenue Middlesex, NJ 08846 (202) 469-6640

Magnetic Shield
The gain of the photomultiplier tube is very sensitive to magnetic fields. It is even affected by the Earth's magnetic field. Just re-orienting the photometer head in the Earth's magnetic field (while the photometer is moved from one position to another) produces a significantly different gain. Stray fields from power lines, motors, relays, etc. can all have sever effects on the gain. A simple solution is to use a magnetic shield. Magnetic shields have been specifically designed for photomultiplier tubes and can be purchased from most photomultiplier tube manufacturers.

Figure 20 shows the mechanical details of a magnetic shield for a 1P21. These shields are important and should not be left out. Make sure the opening in the shield is aligned with the photomultiplier tube's photocathode (same side as the socket key for the tube). The small holes at the bottom of the shield is used to lock the shield to the photomultiplier tube so it doesn't turn. I tapped them and used a small 6-32 nylon screw for this.

Figure 20
1P21 Photomultiplier Tube Magnetic Shield

Magnetic shields can be obtained from Hamamatsu Corp.
They are: Side-Window PMT Magnetic Shield Part number E989
Hamamatsu Corp. 420 South Avenue Middlesex, NJ 08846 (202) 469-6640

Mirrors
The fixed and flip mirrors should be first surface mirrors. These can be attached to small brass plates soldered to rods using a small amount of RTV adhesive.

Pre-Viewer (Optional)
The pre-viewer uses a normal eyepiece typically 25 mm or longer focal length. It's best to focus the telescope with the post-viewer first. Then slide the pre-view eyepiece in or out to focus it. Pre-viewers are good for faint star work where the stars cannot be seen in the finder. They add complexity and weight to the photometer head, however.

Post-Viewer
To adjust the position of the star in the diaphragm, a post-viewer is needed. Usually a 12 mm illuminated dual-cross hair reticule eyepiece is used. The telescope's optics should provide a focus of the star at the diaphragm. To provide a focus on the diaphragm hole a Relay Lens is needed. To set the focus of the post-viewer eyepiece, with power off and the back cover of the head off (make sure the filter cage is in the dark slide position), illuminate the diaphragm and focus on the edge of one of the holes. After finding a star and centering it, focus the telescope so the star is in focus in the post-viewer eyepiece. This will put the star in focus at the diaphragm.

Relay Lens
By making a relay lens holder that fits inside the bottom of the eyepiece, the exact focus of the Relay Lens is not critical. The eyepiece can be moved in and out of the photometer head as well as the relay lens holder moved in and out of the eyepiece. Once adjusted, the relay lens can be held in place with a bit of silicon adhesive. The eyepice can be locked down with a set screw. Figure 21 shows a layout for a relay lens holder for a 1 1/4" eyepiece barrel.

Figure 21
Relay Lens Holder

Edmund Scientific relay lens part numbers 95034 and 94749.
Edmund Scientific Co.
101 East Gloucester Pike Barrington, NJ 08007-1380
(609) 547-3488/573-6250

Diaphragm
The purpose of the diaphragm is to mask out all but a small part of the sky around the star. This has two functions. First, it cuts down the sky or background readings considerably. Second, it allows a single star to be selected for measurement out of a field of close stars. Typical diaphragm sizes are: maximum opening, 120, 90 , and 60 arc seconds. Some people, with very accurate mounts and well sheltered from the wind, use 30 and even 15 arc second diaphragms. Usually the larger the diaphragm you can use the better. This reduces the time spent with lost data if the star moves out of the diaphragm. Figure 22 shows a diaphragm plate used with the Hopkins Phoenix Observatory photometer head. The plate has been machined from a 1/16" aluminum plate and black anodized.

Figure 22
Diaphragm Plate

Making small diaphragms holes can be a problem. A number 80 drill (0.0135") will make about a 35 arc second hole when used with an 8-inch f/10 telescope. Longer focal lengths for the same size hole will produce a smaller angular opening. To calculate the angular size, use the following:

D(arc seconds)= 206,265 X (d/f)

Where:
D= the hole in arc seconds
d= the hole diameter in inches
f = the telescope's focal length in inches.

For:
d= 0.0135" and f= 80"
D= 34.81 arc seconds

Fabry Lens
Because photocathodes are not completely uniform and to obtain the most consistent readings, it is important to keep the same area on the photocathode illuminated. Atmospheric turbulence, tracking errors, etc. cause the image to drift. To help with this a Fabry lens is used. The Fabry lens causes the star's image to be spread out on the photocathode. The spread out image then remains stationary even when the star's image drifts slightly in the diaphragm. With the Fabry lens in place, the star can vary in its position in the diaphragm and the illumination of the photocathode will remain constant. To further increase the accuracy of the readings, always place the star in the same position in the diaphragm, e.g., in the corner of the cross hairs or in the exact center of the cross hairs. Find a convenient location in the eyepiece close to the center of the diaphragm. The photocathode of the 1P21 tube is rectangular with its width a bit over 0.25". The ideal spot size on the photocathode is about 0.125" in diameter. This is one reason why filters larger than 0.5" square for a side-window tube are a waste of space and money.

Figure 23 shows how to calculate a Fabry lens.

Edmund Scientific Part Number 94749.
Edmund Scientific Co. 101 East Gloucester Pike Barrington, NJ 08007-1380 (609)
547-3488/573-6250

Figure 23
Fabry Lens Calculation

Where:
D= Objective Diameter
d= Fabry Lens Diameter
F= Objective Focal Length
F'= Fabry Focal Length
W= Photocathode Dot Size
L= Fabry Lens to Diaphragm Distance

S= F + L
S'= Fabry Lens to Photocathode Distance

Since F is nearly equal to S, S/S' can be replaced by F'/F

Therefore:

F'= FW/D (F/D= telescope's focal ratio f)

The Fabry Lens diameter must be large enough to capture all the light diverging from the diaphragm.

Thus: d= L/f + d'

for a spot size W= 0.125" and f= 10

F'= 1.25" and d= 0.1* L + d'

where d'= largest diaphragm to be used for observations

As can be seen the values for L and d can be adjusted. Also, since most photometers have a large diaphragm for initial star locating, the size of that opening should usually not be used for d' but instead use the largest size to be used for photometry measurements.

Note:
It is not necessary to put the diaphragm at one of the Fabry Lens Focal points, because it images the objective (not the star) onto the photocathode.

Filters
For standard UBV or UBVRI photometry, special filters must be used. The following filters are UBV filters used with 1P21 and 931A type photomultiplier tubes.

V - GG-495 (2 mm thick)
B - BG-12 (1 mm) + GG-385 (2 mm)
U - UG-5 (2 mm) or UG-2 (2 mm)

They are available from Schott Glass Technologies, Inc.
SCHOTT GLASS TECHNOLOGIES, INC.
400 York Avenue Duryae, PA 18642 (717) 457-7485 http://www.schott.com/optics_devices/english/products/filter/glass_filter.html
and
http://www.andcorp.com/Web_store/Filter_Glass/filter_glass.html
and http://www.omegafilters.com/index.php?page=omegatext/prod_astro_index&ps_session=10aadfa8008db1459b8450bb166380bb

Note:
The UG-5 is more sensitive than the UG-2 filter, has higher ultraviolet transmission, but is more expensive and has more red leak than the UG-2. The UG-5 was originally recommended by the AAVSO. The UG-2 is used for UBV photometry by Kitt Peak National Observatories. A GG-495 filter can be used with the U filter to measure the red leak. This can be important for "red" stars which have little ultraviolet light. Since the spectral response of the 1P21 tube drops to near zero in the region the red region passed by the UG-2 , normally there is little concern about red leak.

If there is sufficient interest, filters can be purchased, cut and distributed at a significant savings. These filters are supplied from Schott in 2.0 inch square glass with an average price of around $50 each (1980 price). Since only 0.5" square filters are needed, a set of 8 filters can be made from each original 2" square filter. That assumes no breakage. For more information, contact Jeff Hopkins at phxjeff@hposoft.com. Figure 24 shows typical transmission curves for these filters.

Figure 24
UBV Filters Transmission Spectral Curves

Filters for use in the Red and Infrared bands are as follows:

R - Corning 3480 + 4600 (2 mm thick)
I - Corning 2600 + 3850 (2 mm thick)

Note:
These filters will not work with the 1P21 as the spectral response of the tube drops close to zero in longer wavelength bands. A special photomultiplier that is sensitive in the infrared region is needed with these filters.

Techniques For Mounting Filters
Mounting the filters can be done in several ways. The basic criteria is that the filters must be in the light path after the point of light deflection to the post-viewer eyepiece. Since visible light will be nearly completely attenuated with the U filter, looking through it will not work. In addition, by placing the filters close to the photomultiplier, the size of the filters can be reduced. When changing the filters the positioning should be smooth and accurate as well as light tight.

Any technique that allows smooth changing and light tight operation is acceptable. There should be one position that is blocked, no filter and no opening, to provide a "dark" position for when the photometer head is not in use. This will protect the photomultiplier tube from stray light. Whatever technique is used the filter mount should be black anodized and/or painted flat black on all sides.

Slides
The filters can be mounted on a slide. It is very important that the slides be light tight. The Optec photometers use this method. The original AAVSO Manual for Astronomical Photoelectric Photometry describes a photometer head using a slide for the filters.

Wheels
Filter wheels are very popular with CCD cameras,. However, a filter wheel used for the photomultiplier tube system must be positioned after the post-viewer eyepiece. This can complicate the implementation. The operation must also be light tight. Customized units can be made, however.

Cages
A filter cage that just fits around the magnetic shield can provide a very compact and light tight arrangement for the filters. The finished unit should be black anodized and can be machined out of solid aluminum. Small, 1/2" filters can be attached with a dab of silicon adhesive. Detents at the top can provide precise stops for the filters. See Figure 25 for one of the filter cages used at the Hopkins Phoenix Observatory.

Figure 25
Filter Cage

Cooling
Cooling a photomultiplier tube will increase its sensitivity and reduce noise. People have devised many different means to cool them, e.g., dry ice, cold water and liquid nitrogen. Cooling complicates the design considerably and the increased bulk makes most cooled systems unsuitable for small telescopes. For bright star work, cooling is not needed. Even fainter star work does not require cooling if the ambient temperature is not above 50 degrees F and/or a large telescope is being used.

High Voltage Power Supply

Introduction
A photoelectric photometry system, using a photomultiplier tube, requires a high voltage power supply. One can be constructed easily and inexpensively. Typically the tube uses -500 VDC to -1000 VDC at around 1 - 2 mA. The current requirement is a function of the tube's voltage divider. While it was not too long ago that a high voltage power supply would have required the use of vacuum tubes, the following circuit provides an easy to build and inexpensive solid state high voltage power supply. The basic requirements for the high voltage power supply are an adjustable negative high voltage of - 500 VDC to -1000 VDC with at least 2 mA of current available and regulation to at least 0.1%. The following is a description of a circuit that meets these requirements.

Note:
Because temperature can affect the power supplies regulation, it is best to turn the power supply on and let it stabilize for at least 30 minutes in addition to monitoring the voltage and adjusting any variation during the photometry measurements.

There are three parts to the high voltage power supply, the low voltage section, the oscillator section, and the high-voltage section. Figure 26 shows a schematic of the high voltage power supply.

Figure 26
High Voltage Power Supply Schematic

Low Voltage Section
The low voltage section consists of a bridge rectifier (BR1), a 1000 uF 35 WVDC filter capacitor (C1), an adjustable three terminal voltage regulator LM317 (VR1), a 120 ohm resistor (R1), 1k ohm potentiometer (R2), and capacitor C2. Either 12 VAC or 12 VDC can be used. If AC is used, BR1 provides rectification. If DC is used BR1 corrects for the proper polarity (either input terminal can be positive). C1 provides filtering of the input voltage. VR1 allows adjustment of the output voltage from about 3 VDC to 10 VDC. R1, in conjunction with R2, provides a feedback voltage for the regulator. By adjusting R2 the output voltage can be varied. The regulator provides better than 0.1% regulation. This is line regulation. There is no load regulation and it is not necessary because the photomultiplier tube's voltage divider provides a constant load (about 1 milliampere) several orders of magnitude greater than the variations due to the photomultiplier tube current (typically less than a microampere). C2 provides transient protection for the regulator. The voltage adjust potentiometer R2 can be removed from the board and a panel mounted multi-turn potentiometer used. Just connect the three wires from where R2 was to the panel mounted potentiometer.

Oscillator Section
The oscillator section consists of Timer (8 pin Dip) integrated circuit U1 (CA555), drive transistor Q1 (2N2219), timing components resistor R1 (1k ohm) and potentiometer R2 (100k ohm multi-turn) and capacitor C1 (0.001 uF), bypass capacitors C2 and C3 (0.1 uF), and current limiting resistor R3 (330 ohm). U1 provides a 30 kHz pulse whose amplitude is proportional to the supply voltage. The potentiometer R2 is adjusted for an output frequency from U1 of around 30 kHz. The exact frequency is dependent on the transformer. The frequency is adjusted for a peak in the output high voltage with a corresponding dip in the input supply current. This is a point of resonance for the transformer and provides the best operation point. Once set, it should not need further adjustments. The high voltage output is then proportional to the amplitude of the primary pulse. Adjusting the supply voltage changes the amplitude of the primary pulse and thus the output high voltage. U1 is not capable of driving the transformer directly so drive transistor Q1 is used. The resistor R3 provides a current limited pulse to the base of Q1. The supply voltage is connected to one side of the primary winding of transformer T1. When Q1 is biased on, due to a pulse on its base, it goes into saturation and effectively grounds the other primary lead of T1. This causes a high current pulse to flow in the transformer's primary circuit. The primary of T1 should be 4 turns of number 26 enamel coated wire. The primary should be wound first with an insulating layer of mylar on top.

High Voltage Section
The high voltage section consists of the secondary winding of transformer T1, four high voltage diodes forming a bridge rectifier CR1 - CR4 (1N4007), three filter capacitors C1, C2, and C3 (0.001 uF, 1000 VDC), voltage divider network consisting of resistor R1 (10 megohm) and potentiometer R2 (100k ohm multi-turn), and an optional digital panel meter (X-34). The secondary of T1 consists of three layers of number 36 enamel coated wire (with mylar tape insulation between each layer and each layer having 150 turns) totaling about 450 turns. The exact number is not important. It is better to have a few extra than not enough, however. Diodes CR1 - CR4 make a bridge rectifier and capacitors C1 - C3 provide the filtering.

Low Voltage Power Supply

Introduction
The pulse conditioning circuit requires low voltage power of +5 VDC @ 50 mA, -5 VDC @ 50 mA, and +12 VDC @ 20 mA. The 10 MHz Pulse Counter requires +5 VDC @ 350 mA. The high voltage power supply requires 12 VDC/AC at about 500 mA. To supply these power requirements the following Low Voltage Power Supply can be used.

Circuit
Figure 27 shows a schematic of a low voltage power supply that can be used to power the Pulse Conditioner, High Voltage Power Supply, and 10 MHz Pulse Counter. Three terminal regulators are used to provide regulated +5 VDC, -5 VDC, and + 12 VDC at up to 1 Amp each. A -12 VDC circuit is included but not needed for the above requirements.

Figure 27
Low Voltage Power Supply Schematic

It is very important to use heat sinks with these regulators. This may be done by adding special heat sinks to each regulator or using the enclosure they are mounted in as the heat sink. If the enclosure is used, be sure to use heat sink grease and insulating parts (e.g., a mica insulator and insulating feed-throughs and washers) to electrically isolate the regulator from the enclosure. If individual heat sinks are used the insulating parts are not required but the grease is. Parts can be purchased from most electronics distributors and as shown on the schematic. Typically the regulators are under $2 each and the transformer around $10.

Data Output

Analog
While the output of a photomultiplier tube is a series of current pulses, they can be integrated or averaged with a small capacitor to produce a current that is proportional to the input light. A current amplifier with several switchable stages of amplification is used.

Display Meter
A simple analog microampere meter can be used to read the current from a DC Amplifier. Getting precise reading is difficult, however. A digital meter would definely be a better choice.

Chart Recorder
A step up from the meter is a chart recorder. This was popular for a while. It not only gave a permanent record of the data, it also allowed a bit more accuracy than the meter. On the downside, you could easily go through a lot of chart paper and keeping the data would take non-insignificant amount of space.

Voltage-to-Frequency Converter/Counter
A trick to make reading the data more precise, is to convert the current to a voltage and use a voltage-to-frequency converter, e.g., the Analog Devices AD537. This allows a greater range and more precise readings. A frequency, as with photon counting, can be used to display the frequency or counts. This is not the same as photon counting , however, and does not provide the dynamic range of photon counting.

Pulse Counting (Photon Counting)
As noted above the output of the photomultiplier tube is a series of current pulses. Instead of averaging those pulses and reading the current, the individual pulses can be amplified and counted. This allows more dynamic range, sensitivity and perhaps easiest to use astronomical photoelectric photometry system. The photomultiplier tube photon-counting system provides at least 2 magnitudes more sensitivity than a solid state system plus a dynamic range of over 10^7. In addition, the linearity is excellent. With the large dynamic range, no scale or gain switching is needed. This allows very accurate work to be done using a comparison and program star that differ greatly in brightness. Data reduction is also easier as there are no gain factors to worry about and reading are precise to 6 or 8 decimal positions. Correction for the system dead time must be made, however.

The basic requirements for a photon-counting system are a detector (e.g., a photomultiplier tube) that produces current pulses in proportion to incident photons, some sort of pulse conditioner, low voltage power supply, and a pulse or frequency counter. The output pulses from a photomultiplier tube are negative going with an amplitude typically in the microvolt region. The pulse width is typically less than 10 nanoseconds with rise and fall times in the sub-nanosecond range. The voltage and times are with the signal terminated in 50 ohms. Higher resistance termination will result in wider pulses and correspondingly larger dead time. The pulse conditioner must take these fast low-level pulses and convert them into something that can trigger a frequency counter or a computer counter circuit.

A counter capable of at least 10 MHz should be used. Bright star work can result in counts into the millions for a ten-second integration. Counts above that should be avoided. With larger telescopes and brighter stars, the photomultiplier tube may become saturated. When saturated the tube loses linearity and should be avoided. It may be necessary to stop down the telescope to make sure the photomultiplier tube operates well below saturation. A counter with eight seven segment LED displays is ideal because it allows displaying up to 99,999,999 counts per integration period and is easily visible in the dark. Also a 10 MHz counter is capable of triggering on 10 MHz or 100 nanosecond wide pulses so again dead time will be minimized. If a computer interface is used, the counter part should have the same minimum specifications as the 10 MHz counter. Figure 28 shows a block diagram of a typical photon counting system.

Figure 28
Photon Counting Block Diagram

Pulse Conditioner

LeCroy MVL 100
Originally LeCROY made the MVL100 (Amplifier/Comparator) monolithic integrated circuit that did most of the functions of a pulse conditioner (see Figure 29).

Figure 29
LeCroy MVL 100 Pulse Conditioner

New Pulse Conditioner
The MVL100 worked well and only cost about $25. LeCROY no longer makes these devices and replacements are hybrids costing in the $300 to $500 range. That is the reason for the development of the pulse conditioner shown in Figure 30.

Figure 30
New Pulse Conditioner Schematic

Circuit Description

Video Amplifier
Integrated circuit U1 (14 pin DIP) is a high-speed video amplifier (LM733CN) made by National Semiconductor. It has a bandwidth of 120 MHz with selectable gains of 10, 100, and 400. The device operates from + 5 VDC and - 5 VDC with a total current requirement of less than 30 mA.

As used in this circuit, the device is programmed for a gain of 100 (specifications indicate the gain will be between 80 and 120). This is accomplished by connecting pin 3 to pin 12. The output pulse from the photomultiplier tube is terminated in R1 (51 ohm). R2 provides protection from reflections of the high speed signal. The "IN 2" (pin 1) input is used as the signal input and causes the output to be inverted (positive going). The "IN 1" (pin 14) is connected to ground.

Power is applied to pins 10 (+5 VDC) and 5 (-5 VDC). The 0.1 uF bypass capacitors are very important and must be used. They must also be close to power pins 5 and 14. The output of U1 is differential. The "OUT 2" (pin 7) is terminated, through a 0.1 uF capacitor, with a 1 k ohm resistor. The "OUT 1" provides an amplified positive going pulse to the low-pass filter. The purpose of the low-pass filter is to reduce high frequency harmonics in the pulse.

Voltage Comparator
Integrated circuit U2 (14 pin DIP) is a high-speed voltage comparator (LM710CN) made by National Semiconductor. Because not all pulses coming from the photomultiplier tube are due to photons (some are produced by thermal emission) it is desirable to select only the ones due to photons. The pulse produced by a photon goes through all the stages of amplification and attains the highest gain or amplitude.

Some thermal electrons also may go through all stages but many are emitted from lower potential stages and produce pulses of lesser amplitude. By using a comparator with an adjustable threshold voltage, the lower-level pulses can be eliminated. The other thermal electrons going through all stages produce the measured dark counts. U2 operates from +12 VDC (pin 11) and -5 VDC (pin 6) supplies.

Total current for each voltage is less than 10 mA. Again, the 0.1 uF bypass capacitors are very important. The pulse from the low-pass filter is routed to the "+" input (pin 3) of U2. Because of the high input impedance of U2 it is necessary to have resistor R4 (1 k ohms) terminate the signal. The "-" input to U2 (pin 4) is used for the comparison voltage input. Any input pulses to pin 3 that equal or exceed the voltage on pin 4 will cause the output of U2 to go low. The threshold voltage is produced from the + 5 VDC supply and voltage divider R7 (100k ohms) and multi-turn potentiometer R6 (1k ohm). Typical threshold voltage on pin 4 of U2 is 20 mV. This is a good starting point for determining a more precise threshold voltage. The output of U2 is on pin 9.

Monostable Multivibrator
Despite its awesome name, the monostable multivibrator is a very simple and easy-to-use device. Integrated circuit U3 (14 pin DIP) is a monostable multivibrator, or one-shot, (74121) made by various manufactures. The purpose of U3 is to form a constant pulse width for each input pulse and produce two TTL level outputs. The output pulse width is set by resistor R8 (2k ohms) and capacitor C2 (150 pF) and is about 150 nanoseconds. The device is not re-triggerable, meaning if a second trigger is applied before the output pulse is finished, it will be ignored.

With the input to U3 connected to pin 4 (A2) and pins 3 (A1) and 5 (B) tied to Vcc, the device will produce an output pulse when the input is a negative going transition. Only +5 VDC (pin 14) is required for U3 and the supply current is less than 40 mA. The outputs are complementary. Pin 6 (Q) produces a positive going pulse while pin 1 (Q NOT) produces a negative going pulse. Either of the pulses can be used with a counter because counters count either positive or negative transitions and both outputs produce one of each per pulse.

Note:
If there is sufficient interest an order of printed wiring boards can be made for both the pulse conditioner and high voltage power supply. A basic set of parts could also be included. For more information, contact Jeff Hopkins at phxjeff@hposoft.com.

Frequency/Pulse Counter

Introduction
Most any electronic counter will work. It should be able to count at a minimum of 10 MHz, however. Input should allow TTL pulses to be counted. The display should have a minimum of 8 LED displays. Surplus counters can sometimes be purchased for $100 or less.

Check:
http://www.bgmicro.com/

If you wish, you can build your own counter and save money.

Figure 31 shows a basic counter schematic of an easy-to-build (about $100) Intersil ICM7226AEV/Kit 10 MHz Universal (Pulse) Counter kit. In addition to buying the kit, it is suggested that eight right angle sockets be purchased so as to allow the display to be positioned at right angles to the printed wiring board. This makes mounting the board and display much easier. While Intersil no longer makes this, searches may turn up surplus kits.

Figure 31
10 MHz Pulse Counter

 

Present Page Version as of 23 March 2004

phxjeff@hposoft.com
www.hposoft.com