Jump to content

  •  

* * * * *

Photoelectric Photometry of Variable Stars - Past and Present


Discuss this article in our forums

Photoelectric Photometry of Variable Stars - Past and Present

By Russell Milton


Someone on the Classic Telescopes forum asked me (see reply #3750) to share what I called "another story" of my experience with photoelectric photometry (PEP) of variable stars. I hope this article will encourage others to make such contributions to scientific studies of variable stars. So what follows will largely be what I remember about my experience as a photoelectric photometrist back 1985-90. Many of the concepts and techniques are still valid today. You'll find technology has improved greatly since 1990. In particular the use of digital imaging cameras provide for measurements of a number of stars within the camera's field of view. Historically in the 1980s it was a laborious process, measuring one star at a time. There may be differences in achievable accuracy comparing 1985 versus 2020 methodology.

 

I had started off as a member of the American Association of Variable Star Observers (AAVSO), making visual observations of the Spring variable R Leonis. This was great fun, gathering useful measures of that long-period variable star as it changed in brightness. Such estimates are accurate to perhaps a couple tenths of a magnitude. But being a science sort of guy, I looked into the higher accuracy measurements afforded by photoelectric photometry or PEP. Here's the current AAVSO introduction to that method of measure:

For those wanting to pursue PEP with digital cameras, Sky and Telescope magazine in its December 2020 issue has some further guidance in an article entitled Measuring the Stars (page 60).

 

If memory serves me correctly, my measurements were typically accurate to ±0.007 magnitude, and sometimes as low as ±0.003 magnitude. This appealed to my science-guy mentality. And measures of that quality are greatly valued by professionals. Being of an age that I call the "vacuum-tube generation", I looked into utilizing a photomultiplier tube (or PMT), which had been invented about 1934 by a Russian scientist. Its subsequent development led to the venerable RCA 1P21, which was state of the art for some years. My own photometer used a Japanese equivalent of that photomultiplier design, which was less expensive than RCA's product.

 

In general these photomultipliers work like this as measurers of light intensity. A light sensitive surface inside the photomultiplier tube is exposed the starlight being measured. This photo-cathode, on being stimulated by the incident light, ejects a number of electrons via the photoelectric effect. Albert Einstein received a Nobel prize for his discoveries in this area. These photo-electrons are exposed to an electrical field (voltage) that accelerates these to the surface of a nearly metal plate inside the tube called a dynode. This is the first stage of the 10-stage photomultiplier tube. Each electron impacting that first dynode) kicks loose an additional number of electrons to be sent to the next stage by the voltage present. This process is called secondary emission. So each of the 10 stages multiplies the electrons freed in the preceding stage. Each of these 10 stages has an electrical potential between successive dynodes of 100 volts, giving a total of 1000 volts that must be supplied to the tube.

 

So in summary, a small amount of star light produces a growing cascade of electrons which exits the PMT from its output called the anode. Such a cascade of "bazillions" of electrons, is a very tiny electrical current, that is a measure of the star's brightness. Measuring this tiny electrical current is what is involved with the sub-discipline of PEP called direct current or DC photoelectric photometry. An analog micro-amp meter is the output measuring device.

 

But the story is even more interesting than that. Light can behave either like waves or like individual particles, call photons. So if a single photon enters the photomultiplier tube, it produces as before a tiny electrical current. But this current arrives as a pulse of electrons - one pulse for each photon of light entering the PMT. It is counting these pulses that is the methodology for photon-counting photoelectric photometry. Now light, even very faint starlight, is composed of a huge number of photons. As a consequence the pulses are very numerous, and they come very fast. So the pulse counting electronics must be specially designed to accurately count all the pulses. 

 

There is another minor adjustment that needs to be made. When a photon impacting the photo-cathode has caused some electrons to be accelerated toward the first dynode, the cathode is temporarily insensitive to any more photons arriving during a short period of time called dead time. These missed photons can be accounted for by a process called dead-time correction. This adjustment, though minor, is best applied for highest accuracy measures. The method employed has been lost in the fog of my memory. So it will not be discussed further.

 

With these thoughts in mind, I'll now describe the methods used in photoelectric photometry. In 1985 PEP was quite labor-intensive and time consuming, compared to more modern methods. In my case the method used was called differential photometry, which made the process less complex. One measures the difference in brightness between the variable star and a nearby star of constant brightness called the comparison-star. Now if the comparison star is not in fact constant, then the differential measurement is not accurate. So a second nearby constant star (called the check-star) is measured at the same time. If the differential magnitude (comparison to check star) is constant, this means both comparison-star and check-star are of constant brightness. And this is what we want for our measures of the variable to be valid.

 

For a short time I used my home-made 8-inch reflector for differential DC photometry. 

 

But I soon learned that the mount's lack of slow motions controls made it very difficult to find and center variable and comparison stars. But it did produce one notable finding from that period, the time of minimum of the eclipsing binary star V1010 Ophiuchi. Here is a hand plot of that result.

 

So in 1985 I sold my cherished 1973 Celestron-5 to finance a Celestron Super C-8 Plus. This proved much easier to use, and was employed for the rest of my photometry.

 

As for PEP hardware I purchased a commercial HPO photometer head (by Hopkins-Phoenix Observatory) with a Hamamatsu 1P21 photomultiplier tube. This resided in a light-tight housing with coaxial cable connections for measurement channel & the 1000-volt supply voltage. From plans I made the 1000 volt, 1 milliamp supply, which worked quite well. Jeff Hopkins of HPO was a great help in getting all this going. Previously the measuring instrument for DC photometry was as mentioned a simple analog DC amp meter (seen in the above photo). But I had learned of the increased accuracy afforded by pulse-counting techniques. So with Jeff's help I assembled a pulse-counting electrical chassis. Here's a photo of the PEP equipment:

 

In my home made carrying case can be found the pulse counter on left, 1000-volt DC power supply on right and the photometer head in front.

 

Here's the observatory built in part to house my PEP rig.

 

For both DC and pulse counting methods the sequence of measurement is this:

  • Comparison-star
  • Variable-star
  • Comparison
  • Variable
  • Comparison
  • Variable
  • Comparison
  • Check-star
  • Comparison

 

This gives three differential measures of the variable star, which allows for a simple statistical estimate of probable error. This ± value is known as "standard deviation" in statistics. The single measure of the check star is good enough to ensure the comparison-star isn't variable.

 

Each sequence of measures shown in the above series takes about 20-minutes at the telescope. I did have software to compute the magnitude brightness of the variable. But for consistency of measures by different photometrists, all AAVSO data was processed by that organization, using its own software. So my raw measurements were sent via US mail to AAVSO on 4-inch floppy discs - remember those? The raw data was transcribed onto the disc in standard format for efficient processing by AAVSO. Here is a raw data sheet for a single data point of a variable star's light curve, as captured at the telescope. In this case the variable is Rho Persei.

 

In practice I was diligently watching the pulse counter's screen, transcribing the 7 numbers displayed onto the paper data sheet. This needed to be finished before the next group appeared 10 seconds later. Notice that a measure of the sky brightness in the vicinity of the star is taken, so that background light can be removed, leaving only the star's light to be measured. Both variable and sky brightness were taken through a small aperture in the telescope's focal plane that excludes all other light. As I recall the "dark current" (when no light was reaching the photometer) was about 250 counts in 10 seconds, depending on the ambient temperature. So in the above raw data for Rho Persei, the vast majority of the over 3 millions of pulses counted in ten seconds resulted from individual photons. These had entered the telescope's 8-inch aperture and passed through the photometer's aperture, finally energizing the photomultiplier tube's photo-cathode. Photons were indeed being counted. I found that quite astounding and still do.

 

In closing, another organization to consider along with AAVSO for collaboration with professional astronomers is IAPPP (International Amateur-Professional Photoelectric Photometry). These scientists are most grateful to obtain high quality data obtained by amateur PEP photometrists. They often include such amateurs as co-authors in the scientific papers they publish. I was privileged to be associated in this way in around half dozen or so papers published in scientific journals. So that is an added benefit from this activity.

 

As best as I remember, that is a snapshot of the state of the art in 1985. I do hope some others will be moved to take up this important work. I got out of it around 1990, because of becoming too much driven by gathering data. I had been spending many hours in variable star photometry to the detriment of other important activities. I felt at that time my love of astronomy had been skewed too far in the direction of science as an end. But perhaps in the future I'll do some limited photoelectric photometry using modern methods. These allow measuring the brightness of a whole field of stars in one exposure utilizing a digital camera - much easier than historic methods.

 

There is an interesting epilog regarding the photomultiplier tube. While the vast majority of vacuum tubes from the first half of 1900s have been replaced by solid state devices, that has not been the case with the PMT. It still finds current use in many specialize applications. Sometimes the "old dogs" do the new tricks better, if you catch the pun.  But I do hope you have enjoyed this perspective from that earlier era.

 


  • eros312, quality guy, rekokich and 3 others like this


8 Comments

Thanks again for posting your experiences with this. Very interesting!

    • Rustler46 likes this

I'm glad you enjoyed my writeup, Joe. I found it fascinating back in the day. I may get back into it in a limited way. Maybe my AT115EDT refractor with ZWO ASI290MC can be employed. I would not want to spend much money for needed software, since it would just be an occasional exercise. I do intend to keep my yearly license for SharpCap Pro. That is 10 pounds well spent. Perhaps it has a variable star utility. If not there may be free or inexpensive software that could be used. Any ideas?

 

Best Regards,

Russ

Talk about "timely." I collected a number of PMTs and bones of a Starlight-1. More curious, I'd like to also try to remake one of these. Have a few Books as well (Russell Genet's, Jeff Hopkin's)

Talk about "timely." I collected a number of PMTs and bones of a Starlight-1. More curious, I'd like to also try to remake one of these. Have a few Books as well (Russell Genet's, Jeff Hopkin's)

That sounds exciting. If I can be of any help, let me know. I'll look in my things to see if there is any written material that could be of use in your project. I have many pages of written (snail-mail) communication of that era.

 

As I recall the counter I made had some commercial sub-assemblies inside:

  • Amplifier-discriminator
  • Counter
  • LED digit display driver.

I just wired everything together, made the holes and cutouts in the aluminum project box. And it all worked without me knowing a whole lot about the specific circuit design. While I do have an electronics and physics background, I never worked in that field. In contrast Jeff was an electrical engineer with RCA. So he knew his stuff.

 

The HV power supply was home-made following instructions provided by Jeff Hopkins. It involved winding a small toroid-choke transformer on a very small bobbin with tiny gauge magnet wire. Hopkins provided a bare printed circuit board made for the purpose. I just soldered components on the board. My coil would never produce more than 750 volts, which limited now faint a star I could measure. So Jeff allowed me to trade in my home built unit for a working one he had for sale. This would give the required 1000 VDC. Jeff was most kind to me, patiently helping me get going in PEP. Sadly, someone said he has since died.

 

Keep us posted as to your progress.

 

Russ

    • Starsareus likes this
Photo
photoracer18
Nov 26 2020 04:27 PM

In my case when I was a junior in HS (1964) I was looking around for a senior science fair project in astronomy and building a working amateur stellar photometer seemed like a good one at the time. At first my advisor and I had problems because there were no PMTs out commercially and that was the only possible direction. However about that time RCA brought out the commercial version of the IP21 the 931A and my father who had retired from the USAF the year before was working for a defense contractor and he managed to get me one. We built the entire circuitry from scratch based on the factory data sheets and some other circuitry we tried earlier. I was never so happy as when we had everything completed and started getting meter readings when it was mounted on my RV-6 and had a star in the center of the field. I had solved the tracking issue by adding a slow motion control to the DEC axis. I did it as more of a feasibility study since I was not involved with the AAVSO at the time, as I was more into astrophysics at the time. I donated the complete system to the school after the state science fair in 65.

    • quality guy and Rustler46 like this

In my case when I was a junior in HS (1964) I was looking around for a senior science fair project in astronomy and building a working amateur stellar photometer seemed like a good one at the time. At first my advisor and I had problems because there were no PMTs out commercially and that was the only possible direction. However about that time RCA brought out the commercial version of the IP21 the 931A and my father who had retired from the USAF the year before was working for a defense contractor and he managed to get me one. We built the entire circuitry from scratch based on the factory data sheets and some other circuitry we tried earlier. I was never so happy as when we had everything completed and started getting meter readings when it was mounted on my RV-6 and had a star in the center of the field. I had solved the tracking issue by adding a slow motion control to the DEC axis. I did it as more of a feasibility study since I was not involved with the AAVSO at the time, as I was more into astrophysics at the time. I donated the complete system to the school after the state science fair in 65.

Thanks for your comment, photoracer18. We are about the same age. I was a senior in 1964. But it was few years later when I got into variable star observing, first visual, then DC PEP, finally with the photo-counting rig. I ended up donating all my PEP gear to Humboldt State College in California. Great fun while it lasted.

 

Best Regards,

Russ

Photo
noisejammer
Nov 29 2020 12:58 AM

An interesting recollection indeed. During my junior year, I had use of a 12" Tinsley. One of the boxes contained a PMT and pulse counter but no one - at least among the active amateurs at the time - could recall who had used it.

In the fullness of time, I joined the Center for Backyard Astrophysics and made many thousands of high cadence measurements of several cataclysmic variables. Targets were chosen so that they were near opposition allowing 800 - 900 measurements to be made per night and with sites around the world, near continuous coverage was possible.

Using my 12", I could usually do better that ±2 mmag, at least to ~15m with 28 seconds integration... the missing 2 seconds was to download the 640x480 image through the printer port. How times have changed. lol.gif
I discovered that an AO-7 (early SBIG adaptive optics unit) allowed the system to be extended down to near 17m (with up to 2 minutes integration.) On brighter stars, noise floors could approach 1 mmag.

Needless to say, the capture, calibration and data extraction from so many images took a fair bit of automation. This was aided by code include in Richard Berry & Jim Burnell's book, the Handbook of Astronomical Image Processing. Sadly, this is one of the treasures that has disappeared along with Willman-Bell.

I found a sales slip for a book I used to have Astronomical Photometry by Hendon & Kaitchuck. But I was passed on to Humboldt State University along with my photometry gear. I do still have John Percy's The study of variable stars using small telescopes. It has extensive chapters on PEP. 

 

At one point I was interested in digital capture of the photometer counts. There was some sort of digital output channel in my pulse counter. But I never felt like mastering what was needed to make that work. So writing down 7 numbers every 10 seconds along with the time became my method of operation. At the end of each month transcribing the data onto floppy discs for AAVSO processing took a bit of time. There were so many stars that the scientists would like having data on. Fascinating work! I believe for AAVSO my total was around 850 observations. But there were some other non-AAVSO programs that had my contributions. 



Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics