ATM My thoughts on spatially resolved PMT photometry: long

[Dominic-Luc Webb molmed <Dominic.Luc-Webb@molmed.ki.se>]

On Tue, 12 Dec 2000, Tim Walters wrote:

> Hi Dominic,
> I have an old PMT that I was going to use for some communications
> experiments about 30 years ago... Is there any info available on how to
> go about using it in amateur astronomy?
> 
> TIA,
> 
> Tim Walters
> Atlanta, GA

Some other people hit me with this question too, so bombs away...

Yeah, you seem to be the main "crowd" that is familair with and 
appreciates photomultiplier tubes for photometry. I am 36 years 
old and am nearly alone. Pretty much ALL the people I know using 
PMTs are considerably older than me. I have met several PMT users 
and also know the ages of some of the members in the ATM group that 
have built/used them). None of them that I know of is even close 
to my age. At times, I feel like the last of a fading breed. But 
this should not be so, and I would like to state my case along 
with some suggestions on how to go about pursuing PMT photometry.
With luck, someone below 30 (20???) will join us.

Pretty much all of the younger people have been drawn to computers 
and seek out things they can do using a computer. Seemingly 
anything that is made possible to be done on a computer gets 
prefered over the historical non-computerized methodology, even if 
its less efficient. CCDs and computers got together very early and 
younger people picked up on this addition to their computers 
quickly. Many of the people I have discussed this with admit that 
the computer came first, then the CCD that attached to the
computer, and then, only as an afterthought, they thought it would 
be "cool" to put this onto a telescope. Clearly, CCDs, by design, 
give spatial resolution. So they get a picture, and they feel 
happy (safe?), being able to see the object that has been recorded. 
CCDs also automate calibration against guide stars, etc. Fast and easy! 
While CCDs usually still need cooling systems, as do PMTs, CCDs 
don't need high voltages. To all of this I say, so what! 

A PMT could do all of this and in many cases outright outperform a 
CCD if designed to do so. Herein lies the problem. PMT astrophotometry 
systems were not designed from the onset to do spatial imaging, and 
they are still not a "cool" peripherial for a PC, but they could be. 
High voltage circuits are not too big a deal either, and you can buy 
very compact ones for pretty cheap. Actually, a simple PMT system 
without spatial resolution can be built considerably cheaper than a 
cookbook CCD. With some additionaly parts, some degree of spatial 
resolution could also be added still below the cost of a cookbook. But 
ok, by now you want to know about a real world application where 
PMT imaging has been done. Well, I confess, I have never seen this 
done with a telescope, but....

We have several confocal microscopes in my lab from Leica. We use 
these for high speed two and three dimensional imaging. They ALL use 
PMTs because CCDs are too slow. Leica currently sells the system 
and will gladly furnish information about these systems. Further, 
there is no problem measuring shorter wavelengths as with a CCD, 
right down into the UV, like 320 nm. Our newest system can sample at 
better than 1 Khz, and I do not mean pulse counting of a single point. 
I mean we sample the optical field with a resolution EQUAL TO or 
BETTER THAN a CCD at 1 Khz or faster! I have been working on getting 
some sample pictures together to demonstrate the technique. Most of 
our pictures thus far are for research and the images wouldn't make so 
much sense to the general public. I hope to have some pictures that 
most people can relate to in coming days and if anyone wants them I 
will gladly send them along. Better yet would be a web site I 
can deposit a couple of samples. The basic principle can be 
directly applied to a telescope and need not be technically 
or economically unsurmountable. I am getting darned close to 
a fully functional pulse counting prototype that thus far is well 
below 500 USD including all electronics, PMT and mechanicals, etc.
I am shooting for lower resolution like 100x100 pixels to 
demonstrate the technique. If you have access to a precision 
metal lathes, and have skills in electronics and electromechanical 
design etc, I could easily see far better prospects. I am managing 
with no special skills to boast of. In fact, most of my design 
ideas are taken (intentionally) from high school level electronics 
and mechanics books, like multivibrator attached to speaker 
attenuator coils to move a mirror. I am now certain that anyone 
wanting to build this can do so. I believe this will have applications 
in which CCDs won't due and avalanche photodiodes will be too 
expensive or not of appropriate sensitivity to required wavelengths.

I am documenting my adventures as I go and I hope to have some 
written work in the future in the form of a long article or short 
book that "focuses" (pardon the pun) on remote and robotic telescope 
with spatial imaging using a PMT. Its a one-man operation and my 
research slows this work down, though I have serious plans to finish 
this long before I retire :)

The history of the PMT goes back to the 1940's, if we start with 
the earliest PMTs that were used for astronomical photometry. I 
highly recommend reading some of the papers by Gerald Kron. I 
can furnish some of these if anyone wants them (I know they can be 
hard to find, and I have been trying to collect them). Kron has some 
papers where he clearly goes through step by step the circuitry 
and application of PMTs. The historical "classic" PMT was the 
RCA 1P21, which was replaced by the 931A and then the 931B. I don't 
know details, but it appears that the 1P21 PMT was what was used to 
define the first UBV standards. I am told by Burle (which is the 
manufacturer of the RCA 1P21, 931A and 931B) that the 1P21 is still 
produced for the sole purpose of maintaining the standard. However, 
the 931A is somewhat superior, and the 931B is even better still. 
The 931B is the current PMT of choice for this type of photometer 
(there are other types). These or some equivalent PMTs can be had 
for as little as 15 USD dollars (used) or about 90 USD from Hamamatsu. 
I regularly scout Ebay for these and hope to collect enough extras to 
offer to other people at cost.

There are not so many books I could recommend. There are 4 books 
that can be had from Willman Bell at www.willbell.com that are 
immediately relevant. All of these I have been able to find at 
technical university libraries:

1. Astronomical Photometry. Henden and Kaitchuck, 9.50" by 6.00", 
392 pages, hardbound, 2 Lbs.2 Ozs. ship wt.,$29.95 Cheers,

2. Photoelectric Photometry of Variable Stars. Hall and Genet, 
6.00" by 9.00", 240 pages, hardbound, 2 Lbs. ship wt.,$24.95

3. Software for Photometric Astronomy. Kaitchuck and Henden, 4-5.25" 
360K diskettes, 14 Ozs. ship wt.,$69.95.

4. Solar System Photometry Handbook. Genet, 9.00" by 6.00", 
214 pages, softbound published. 1983, 1 Lb. 4 Ozs. ship wt.,$17.95 

Book 2 seems to stand out as the one with the example circuits. 
Now for a warning. If you try to build your own high voltage power 
supply using the 555 based PS found in (I think) book 
number 2, you may be unsuccessful. I did not have such great luck. 
I got tremendous help from people in the group here and have even 
encountered one person on an electronics newsgroup who swears by 
these circuits and says they can work very nicely. Winding your own 
transformers, as indicated in the book is not very practical as 
hand wound ones tend to be really inefficient. For one thing, they 
can vibrate and make a lot of noise and heat, which implies they are 
not getting good coupling between primary and secondary and then 
you never get your voltage. I did end up making a couple of decent 
transformers by using something like 15 winds on primary and about 
5 thousand on the secondary, but it ended up more a matter of my 
demanding to not be defeated by this project. The advice I have 
been given and now advise to others is to try something like a flyback 
transformer from an old TV. If you want to dish out about 100 bucks, 
Hamamatsu sells a fixed voltage power supply something like 900 volts 
that is integrated onto a socket. This is not bad deal. Sockets are 
harder to find than PMTs.

In terms of counting photons (I am not much into DC mode), I 
have been working on the concept and realization of using a 
PC parallel port by amplifying/conditioning PMT output pulses 
to TTL compliant signals. For lower counts, sound cards could 
be used. I found parallel ports simpler to understand and 
implement. The books describe ICL all-in-one pulse counter ICs, but 
I never found a good way to use these all-in-one counter ICs 
for higher speed applications. Getting/saving 100K high 
or low states per second from a parallel port should not be 
difficult. I understand it is capable of 8 Mhz.

Next, there is the IAPPP, something like International Amateur-
Professional Photoelectric Photometry organization. There is 
an article about them as well in the December issue of Sky 
and Telescope, along with the two short articles about rapid 
and blue bursts I mentioned in my initial post. If you wanted to 
follow the time course of these bursts, for instance, I just don't 
see any way you could manage with a CCD. The IAPPP folks seem to have 
a lot of experience with PMTs and are worth contacting (as well as 
reading the article about them), although I think most of their 
members are now using CCDs and a couple of brave futuristic types 
have already gotten into avalanche photodiodes. IAPPP has a web 
page - www.iapppwest.org.


Dominic-Luc Webb


North 59 37' 30"
East 17 48' 10"