Thursday, March 29, 2018

A Tale of Three Batteries

I recently complete my piggyback power and USB distributor that should allow me to image using only two cables running to the mount. (I'll have more about that after it's been field tested.) I'm standardizing power connections using Anderson Powerpole connectors, and the plan is to use all my batteries connected in parallel. This means adding Powerpole connectors to all of the battery boxes, which got me thinking about the batteries themselves.

The old batteries are 6 and 7 years old, and there was little doubt they were starting to show their age. Both are the wet lead-acid type, which is a little messy and a tad bit risky to haul around. They're also heavy, each being rather large (Form 27) and weighing about 55 pounds.  I could justify replacing them with nice new AGM sealed batteries if they proved to be in poor shape, which raised the question of how to test a battery's capacity.

Both batteries have reserve capacity (RC) values of 175 and were advertised as being deep-cycle. I can't attest to their deep-cycle ability, but they served me through several Iowa Star Parties and four Nebraska Star parties.

RC is the number of minutes a battery at 80°F can provide 25A before it is fully discharged (it's open circuit voltage falls to 10.5V). RC is not the same as amp hours, and certainly not the same as usable amp hours.

To see why, let's pretend we can convert RC to Ah directly:

Amp hour capacity = amps provided • elapsed hours =  25A • (RC  / 60) =  0.417 • RC

An RC of 175 gives a capacity of 72.9 Ah. Seem easy, doesn't it? The catch is that you almost never want to drain a battery that far-it will gradually decrease the battery's capacity. If you only discharge it about 80% you probably won't notice any loss of capacity until you're well past 200 discharges. 200 discharges is a lot of nights out imaging! So that 72.9Ah is really more like 58Ah. Sorry!

(Much of the literature about deep cycle batteries is written for solar power users who discharge their batteries nightly. In their case a maximum discharge of only 50% or less is needed if the batteries are to be economical.)

But wait, it gets more complicated! Astrophotography seldom requires anything like 25A. (For example, my setup for imaging with a DSLR requires only about 3 to 4A; it jumps to about 6A when using my CCD) Batteries are more efficient at providing power when the current is lower, meaning that the "real" Ah capacity in that case is higher. My batteries also have a stated alternate RC value of 200 based on 23A draw, which gives a full drain capacity of 23A • (200 / 60) =  76.7Ah. Two amps less and you get 3.8 more Ah. Great!

We're not done yet, though. What you're powering also will come into play. As the battery is drained its voltage falls, and as that happens voltage converters and inverters may have to work harder to provide regulated power. Some devices (Kendrick dew controllers, for example) may shut off. So draining a battery by 80% may not work for you.

Possibly the "best" way to determine how a battery will perform is to test it yourself. I recently found a nice way of doing this that employs a simple AC electric clock, AC lamp, and an inverter with a low voltage alarm. An inline DC power meter can be useful, too, but it's not necessary.

I used a 40W bulb to better match my typical current demand. (The calculation is easiest if we use amps = power / volts, so in this case 40W / 12V = 3.33A.) The inverter and clock draw power, too, so  My inline meter suggested that inverter, bulb and clock used about 43 watts and would draw about 3.6A initially. My inverter is programmed to sound an alarm when the loaded voltage drops to 11V; in reality it appeared to cut off at 11.2V

As a reality check I tested my relatively new and well cared for 35Ah AGM battery. I was unable to find an RC value for it that I could trust, so I took values from its spec sheet and (with the help of a spreadsheet) interpolated a capacity of 31.65Ah at a drain of 3.6A.

The battery powered the inverter for 399 minutes, and the meter reported 28.5Ah were provided, but a closer look at current and voltage measurements suggests this should be adjusted by a factor of about 0.93--so 28.5 becomes 26.5. given the uncertainties, I'll use that as the output of the battery. 26.5 is 83% of the estimated 31.65Ah capacity.

Even though all of these numbers are fairly approximate, I think it's safe to conclude that the relatively healthy 35Ah battery was able to provide something like 80% of its capacity before the voltage fell low enough to shut down the inverter. Given that, the test can be repeated with the other batteries; if they are  in perfect shape they could be expected to each deliver about 80% of their 73Ah, which is 58Ah. This is very likely an underestimate because we drain at a much lower amperage than 25A.

Battery One, which is seven years old and was once allowed to go dead, provided power for only 213 minutes. That means it provided only about 3.6A • 213 minutes • (1 hour / 60 minutes) = 13Ah. Awful!

Battery Two, which is six years old and was better maintained, provided power for much longer and ended up delivering 26.5Ah, which is still less than half of what a new battery its size could be expected to provide. It should be replaced, too.

The options are 50Ah batteries, $100 ($2/Ah), 30 pounds, or 75Ah, $140, ($1.90/Ah). I think I'll go with the 50Ah ones! Time to place an order!

Sunday, July 30, 2017

Back from the 2017 Nebraska Star Party

Ah, to be back in the land of 10,000 lakes and high dew points!

NSP 2017 is over, and it was a memorable one. Sunday (7/23) was clear and about as good a night as I've ever enjoyed at NSP--or anywhere, for that matter. The day had been in the low 90s, but the temperature dropped quickly as the sun set. There was a strong and gusty breeze that kept the mosquitoes away. (I realize big scope owners didn't like the wind, but I was imaging with a 200mm lens wildly overmounted on a CGEM.) The transparency was exceptional overhead; the Milky Way was nothing short of spectacular.  I don't think you could read by its light as one person predicted, but it was definitely casting a subtle, diffuse shadow.

I was able to collect some light frames of my current target, the Rho Ophiuchi Cloud Complex, then switch over to the Sadr area for the Butterfly:


The Rho frames have yet to be processed. The next night I decided to spend the entire night imaging the Veil:


What I like about this 200mm lens is how flat it is at f5.6, and free of vignetting.  There are two new things about these images:
  1. Instead of stopping down the lens with its internal iris I used a step down ring. This gets rid of the diffraction spikes caused by  the blades of the iris.
  2. I used PHD and BackyardEOS to dither the images.  It seemed to work very well!
The last night was cut short by clouds, so I only got five frames of the Rho area at 135mm:


As you can see, this lens has substantial vignetting.

By the way, 2017 NSP was on the hot side.  The Monday and Tuesday highs were 106 and 107 degrees, respectively. But it was the usual "dry heat" and both days had those great strong breezes.  I spent the afternoons sitting and reading semi-comfortably in the shade of the supper area canopy. Thank goodness NSP didn't happen the previous week when it had been both hot and humid.

Now it's on to eclipse planning and back to meteor counting!

Tuesday, July 18, 2017

Back To Imaging for a Bit; The Nebraska Star Party Nears

My wife got a new desktop computer after the 4th of July, and that translated into more than a week of transition from old computer to new. Because she didn't get a new monitor our ancient flat panel monitor was pressed into service so that she could run both computers at the same time. And because it was the monitor on my meteor-detecting computer, that activity was put on hold for the duration.

Luckily we just had a very nice run of third quarter moon clear nights. I was able to get out for two out of the three nights and tested the setup intended for the Nebraska Star Party. This year I'm not chasing any astronomical league program certificates so I'm keeping it simple: Big mount, DSLR, and short lenses.

From my second night out, here's an example:


It's the Lagoon and Trifid nebulae. This is a 200mm f/5.6 image based on only 10 4-minute light frames (ISO 800) and it gets reasonably deep.  This was taken at a light pollution yellow-green transition zone site; I'd like to try this again at NSP to see how much the difference in sky brightness affects the outcome.

Another target will be the Rho Ophiuci area and the dark lanes to its east. And, if there are enough clear hours, some of the dark nebulae that dot the area.  I may even try some super wide fields!

The weather at Valentine has been on the warm side, with some daily highs in the 100 to 105 degree range. The forecast for the first Sunday is much nicer at this point--A high of only 87, and a partly cloudy night with a pleasant low of 60. Monday's high is forecast to be only 88! It doesn't get much better than this.

The rest of this week is NSP preparation. No more imaging until then!

Sunday, July 2, 2017

Meteor Detection Using Argo

I should have posted some of the screen caps to show what Argo is capable of showing when using TV stations for meteor detection. These are from a session monitoring CHBX Channel 2 Sault Ste. Marie, Ontario broadcasting at 55.24MHz.

Here's one showing a couple of typical events. At left is an epsilon reflection from a meteor, while the slanted line at right is probably an aircraft flying at flight level somewhere near the midpoint of a line between my location in Minnesota and CHBX.


The next image is of an overdense reflection from what I assume was a larger meteor.


Most meteor reflections are faint and very short in time span (less than a second). The overdense and epsilon reflections are less common, and can come in a variety of forms--here's one:


I would like to catch a nice head reflection eventually.

Because Argo shows so many faint meteors, analyzing a series of screen caps will be interesting--and a lot of work.  I probably won't do much more with this until the Perseids, which fall conveniently between the Nebraska Star Party and the eclipse.

Saturday, July 1, 2017

Meteor Detection using National Weather Service weather stations

This is more of an update on what I've been doing since I got things working using TV station video carrier frequencies. Last time I was able to get what looked like a decent diurnal cycle out of a day's observation of Channel 2 (CHBX, Sault Ste. Marie, Ontario); before that I'd tried using regional weather service stations and been stymied by what looked like aircraft-created reflections. I live about 6 miles from Minneapolis-St. Paul International Airport (A Delta airline hub) and it has a lot of traffic.

Argo did so well at getting the most out of the TV data that I thought it worth trying with the weather stations.  It did not fail me.

Here is an 8 minute plot of activity at 162.475MHz:



It's a real mess, isn't it? Most if not all of that is air traffic. To see what it's telling us, let's start with a quieter time and add a little annotation:


The curved lines are reflections from aircraft arriving or departing the airport; the horizontal lines are the carriers from several weather stations. Note that they're slightly out of tune. The audio shift between them is roughly equal to the difference in the carrier frequencies, so all four are within about 45Hz of each other, an error of one part in four million. The strongest two stations ("B" and "C") are probably Rochester MN and Spooner, WI.

Argo shows the reflections clearly. The first diagram contains not just curves like these but other lines that probably represent flyovers, diversions, and other course changes aircraft make near an airport. Unfortunately, none of them look like the things I was able to see with TV.

I'll let the system run overnight and hope that the signals don't saturate like they have in the past. Maybe I'll see some meteor signatures in the flight free early hours of morning?

If I don't, I'm going to have to question the usefulness of weather stations for meteor detection. Even with that negative result I'll probably take my setup to this year's Nebraska Star Party where the aviation activity is considerably less.

--------------------------------------------------

The Virgo Cluster mosaic is now half done:


This is all I'm going to get this year, but it's not bad for such a terrible spring.


Monday, June 26, 2017

Meteor Detection: Some Initial Results

I've analyzed the results of a 24-hour run of the detection system from a few days ago and the results are promising.

The analysis was the visual inspection of 1,542 screen caps taken at 55 second intervals


The columns represent meteor counts in a given hour; for example, column 0 show that about 150 meteors were counted between midnight and 1 A.M. Because I tried to count only things that I felt were clearly meteor signals the above diagram represents an underestimate of meteor activity. Even with that, this is a fair replication of the diurnal meteor cycle. More days of data would be better, of course.

This was performed using my attic antenna monitoring Canadian station CITO, channel 3, at 61.26MHz.

When I went to repeat this it was evident that something had changed with the signal. Meteors were now producing a strange oscillating siren signal that appeared as interwoven sine waves on the waterfall chart.  I noticed that LiveMeteors.com was showing an anomalous signal as well, and shortly after that switched frequency to 55.24MHz, that of a channel 2 station (probably CHBX in Sault Ste. Marie, Ontario).

I've now switched to that same frequency and, using Argo, am collection another day's worth of data.  Argo presents what is essentially a highly zoomed version of the SDR# waterfall, but with a couple of advantages. Argo makes clear that when I use the channel 2 frequency that I'm monitoring several stations at once--there appear to be slight offsets in the carrier frequencies. It also makes it easy to see when a signal is due to a reflection of a local channel 2 from aircraft. (I'll post some diagrams showing this soon.) The Argo strip chart allows me to do a screen cap every eight minutes, which will make counting meteors much easier.

I'll post more as I start analyzing results and gain a better understanding of what Argo is showing me.

-----------------------------------

And it's clear tonight! I should be able to get the first half of my mosaic done! Yay!!!


Tuesday, June 20, 2017

Meteor Detection Using Canadian TV Stations

Last time I wrote about the problems I was having with a meteor detection system that relies on National Weather Service weather radio stations. The problems were so substantial that I indicated I'd be looking into using Canadian TV station signals as the basis for meteor detection. I've now done that in at least a preliminary way, and the outlook is much more promising.

Good introductions into this method are found here and here. In brief, there are a handful of Canadian TV stations that continue to broadcast powerful analog video carriers at frequencies that are near optimal for meteor detection.  Because these stations are located near the U.S.-Canada border, it's possible for most of the eastern and midwestern U.S. to use them for meteors.


There are only a couple of differences between this method and the one using NWS stations:
  • The antenna. Not only must it be designed for the lower frequencies, but because you will be listening to a specific station you can use a directional antenna. One catch, though: With the switch to digital TV, lowband VHF antennas are becoming more difficult to find.
  • Sidechannels. If you've watched MeteorsLive.com, you probably have noticed that it's tuned to 61.26MHz and not the nominal 61.25MHz. This is because transmitters have added side channels at +/1 0.01MHz in order to reduce interference between stations. Why 61.26 was chosen by MeteorsLive instead of 61.24, I don't know [Edit 6/21/17--now I do...see the next post.]
 Here are the stations that can be used:

Canadian TV stations useful for meteor detection (with channel number)

The nice thing about these is that they're all far enough away that aircraft arriving at and departing from the nearby airport can't cause enhanced reflections. At the same time all of them (except Channel 2 in Sault Ste. Marie) are far enough away that aircraft flying at flight level 410 also can't cause spurious reflections.

My antenna is an old FM/TV multi-element antenna hanging in the attic of a wood frame house. It's directed about 40 degrees east of north. I'm initially monitoring the same station in Timmins, Ontario, used by MeteorsLive.com , which is about 61 degrees east of north. The misalignment probably isn't significant given the antenna design.

I've surveyed the available channels, and only 4 and 6 did not show evidence of any meteor-like signals. Channel 4 in Sydney, Nova Scotia, is beyond the 1400 mile limit and can't provide single-reflection meteors. Channel 6 in Medicine Hat, Alberta lies at a right angle to the axis of my antenna.

Radio path between my position and Channel 3, CITO-TV, near Timmins, Ontario

I'm happy to say that Channel 3 at 61.26MHz is producing a substantial number of strong pulses that resemble those suggested to be reflections from meteors. It actually seems to be working, and working well.

The question now is: Are the meteor-like signal pulses really of meteoric origin, or is some other process at work?  The only way I know to answer this is statistically. Will meteor counts match the known diurnal variation? Will meteor counts become enhanced around the times of notable showers?

(In case you're asking why this couldn't be confirmed using an all-sky camera or some other visual means, you've got a point. A light-amplifying camera set up and pointing toward the transmitter might pick up a few meteors. In my case that camera would be pointing over the Twin Cities, and the light pollution would ruin any chance of picking up anything other than a bright fireball. Worth a try, yes, but not in my plans.)

At this moment I'm running the system, doing screen caps of SDR#'s waterfall display at regular intervals. I'm going to run it for 24 hours and analyze the data to compare frequency and time of day. I'll probably have the results later this week.

I'll do the same at the time of the Perseid shower in mid August.

Stay tuned!