Ben probably took the long way around to get this done, but since it was as much a learning exercise as it was to put a telescope on the net, it matters not.

'Make install' Palantir.

Linux is different to windows. You can not poke the cd at the computer and install it and once it is installed, you cant just download a zip file and run setup.exe. If you want to follow Bens footsteps and your new to Linux, be prepared for some frustration. In fact, Ben recommends that unless you are very patient, have a good web connection and or have a friendly local Linux wiz, don't try this! (Thanks again for your help Dan and NJH!)

Ok, so you have a linux box and want to have a bash?

What flavor Linux you running? It makes a difference.
Ben tried Knoppix, Mandrake, Red Hat and 'Debian stable' and 'Debian testing' and only 2 "worked" (Mandrake and Debian testing) The Linux aspect of the project was probably the most frustrating. It was a huge amount of work to go from Dan showing Ben Palantir on a bootable Linux Distro (Dyne:bolic) to Ben getting it set up on a PC at home.

quick and dirty (pad this out more soon)

tar zxvf palantir
cd palantir
cd server
vi Makefile
change the user to one already on your box (or make the user palantir and skip this step)
make
vi telmu.c
change the baud rate (if using the picaxe 2400 is max, if you are not using the com port or the picaxe, skip this)
make install

if you get this error;

gcc -Wall -D_REENTRANT -O3 -g -c -o yuv2rgb.o yuv2rgb.c
In file included from /usr/include/sys/time.h:31,
from /usr/include/linux/videodev2.h:19,
from /usr/include/linux/videodev.h:14,
from video.h:2,
from yuv2rgb.c:26:
/usr/include/sys/select.h:78: conflicting types for `fd_set' <---
/usr/include/linux/types.h:9: previous declaration of `fd_set' <---
make: *** [yuv2rgb.o] Error 1

David has this to say from the Palantir email list (sign up from his web site)

" Looks like some v4l/v4l2 incompatibility. Other Mandrake 9.1 users
had the same problem some time ago, and this was my suggestion (yes, it is
a hack. Yes, it is ugly. Yes, it is time to put it into the FAQ...):

You might try to disable the videodev2.h file inclusion by turning the |
#if 1 | | at line 7 in file /usr/include/linux/videodev.h into | | #if 0
I do not believe this will have side effects on the compilation of |
other programs (unless - of course - they require the video4linux2 |
interface). | For safety, you can restore the '#if 1' after compiling
Palantir."

If you get this error;

/usr/include/linux/spinlock.h: In function `bit_spin_lock':
/usr/include/linux/spinlock.h:413: error: invalid type argument of `->'
/usr/include/linux/spinlock.h: In function `bit_spin_trylock':
/usr/include/linux/spinlock.h:430: error: invalid type argument of `->'
/usr/include/linux/spinlock.h:433: error: invalid type argument of `->'

<snip a whole more of the same>

Give up. Ben has no idea, he had to switch to a "proper" debian/linux install to get it to make install.

PIC or PICAXE or Opto22??

If the picaxe would do serial port and servos at the same time, it would win hands down and be a no brainer, but, it doesn't, so, its not. (Update 15th nov 04, the new PIXAXE is supposed to do this now, Ben has given up on them and sold his PICAXE kits so cant comment).

The Oatley monster servo driver circuit diagram. (630kb)
MDS servo driver information

We are using "standard" RC servo type signals to drive the mount. Oatley Electronics have a kit that allows a standard motor to be driven like a servo. As it happens Oatley also have the mount we are using. As it also happens the mount has a 10 turn pot on each axis that we use as the feedback for the servo. This system is neat as all's you have to do is tell the servo where to go and the servo chip takes care of the rest.
Its also a hassle as the mount needs to move 360° on the pan axis and 90° on the tilt. This is WAY more than a standard RC servo will turn. Mark has spent a lot of time getting the pic to drive outside of this standard RC servo range.
One possible problem with this system is that it always "hunts" around its current position. You need a pretty tight gear chain to keep this movement to a minimum. What the brush wear will be like is also yet to been seen. Ben is thinking it might be better to run the mount on stepper motors?
(Add way more about the pic here real soon)

Update Aug 04: We have totally ditched the servo concept. The constant buzz of the servo hunting not only drove Ben nuts, but it caused the telescope image to blur, it also wore out the front motor bearing.

We now hope to drive the motors with pwm from the pic and an MDS 'H-bridge'. The whole thing works fantastic on the bench and we are truly excited about getting it into the telescope mount.

Update: From bench to mount was not all that straight forward. One of the things we wanted to do was make this project "open source and repeatable". That is, we wanted it to be cheep and we wanted others to be able to do what we have done....... While Linux and thus Palantir is open source, the hardware to drive the mount is getting very hard to keep cheep and readily available. We have hit a bit of a glitch with the mount electronics from Mark. They don't work as well on Bens mount as they do on Marks bench. To prove a point and get some breathing space, Ben has put the mount onto his Opto22 gear. It has shown that the mount is fine. Will things stay this way? For the short term yes. The Opto22 gear is as easy for Ben to work with as Marks Pic stuff for him. Neither are cheep and open, but both work. The other thing Ben and Mark wanted was pretty much identical systems at each location.....time will tell how this comes about.......

The video cards.

Linux can not handle usb web cams real well (as far as Ben can tell). Add to this that the frame integrating camera only outputs comperset video and you have to get some sort of video card working with Linux. The hardest thing Ben found was the hardware. He very quickly found out that "Brooktree" capture cards were supported without having to rebuild the Linux kernel (not hard, but best avoided by newbies). Ben then found out that not all "Brooktree" cards are equal. Thankfully he was in a position where he could try 4-5 different cards and see what worked....... Here are a few photos of the current setup;

Here are the 3 cards in the HP computer;

Ben is a little disappointed that there are only 3 PCI slots in the HP, but then again, how many video feeds is enough? As a side note, Ben scored the HP for free, he has figured out why most computers are chucked out when they stop working, by luck the local computer shop owner is happy to throw them Bens way. About 1/2 hour work changing the bloated capacitors on the motherboard has the computer back up and running good as new! No, sorry, Im not going to tell you how to change them. If you have to ask, you should not be doing them.

Update: 15th Nov 2004, as soon as Gary gets some time to pull the card out of his computer, Ben will be trying an 8 input video card (it only takes up one PCI slot)!!! Yeah, that should be fun!

The Main "Telescope" (a 3.5inch (88mm)) "TV" zoom lens.

If it was not for Oatley Electronics, this project would still be on the garage floor.
What a deal, a 3.5 inch tv style zoom lens for 250 bucks. Super, I will have 3 of them (their web site said they had 4 left).......Hmmmm, seems they only have the one left. Doh! Looks like mds and Dad will have to miss out on this one. Pity..... Still, here we have it..

Sometimes (very rarely) you come a cropper with stuff from Oatley. If its a mission killer, they will take the thing back and give you your money back. In this case, there was no way no where Ben was going to get another lens like this, so, he pressed on.
The problem was that the auto iris was stuffed. No time to fault find the cct (and no real need, the camera has an electronic shutter and can cope with most light variations) Ben simply found the iris motor and ran the wires down the control cable so he can drive it from the study. True, we never know what its set at (what ever looks best must be right?), but again, as long as its adjustable and the camera takes up the slack, it "matters" not. (A fully open iris generally (on pretty much every lens) gives a poor image, fully shut and its a bit dark. Thus you get to the point of what looks right is right for your setup whether its f1.2 or f5.6)

Here is what it looks like on the inside and from the back;

The two cables coming out the back are auto iris (defunct) and zoom,focus controls.

All in all, so far, we are pretty happy with the thing. It weighs in at a not insignificant 3.5kgs. Here (250kb) is the pdf for the thing if you are really interested.

The zoom aspect of the lens just makes it too good to pass up. Ben was full on going to mount his 4'' Chinese refractor on the mount, but after playing with the 2'' zoom lens, it became really obvious that the zoom was the "killer app" in this setup and so this lens is the result of the search for aperture with zoom. The 4'' was getting tricky setting up a zoom on (not a lot of room for the zoom mechanics), so this one was a no brainer. BTW its a f1.8 for the record.

Update May 2004: Ben sold the lens to Mark.

The camera in more detail.

Why pay 900 aus dollars for a "special astronomy video camera" When you can buy THE EXACT same thing down the road for $398? That was the question Ben faced for many a month, a whole lot of web time finally lead him to believe they were the same (or near enough) and thus part with his hard earnt money.....

Here is what it looks like on the inside......

To Ben, the interesting side is the inside......

What are all the un-used connectors for? They are all over the place! One day, when this little project has been in use and doesn't require any work on it for about 20 weeks (it will take him that long to save the 400 bucks), Ben will buy another camera just to play and to see if he cant figure out what the extra connectors do......if you know, please, don't be shy, drop Ben a note at astrotvl at yahoo dot com dot au (remove anti-spam stuff), save him some money and time (aint that what the web is about?)

Here is the business end of the photon catcher and the back end;

The cut wires show Ben has been there. Extending the push buttons into the study so all functions of the camera can be adjusted from there.
This last photo very badly shows that the ccd is sitting in a fashion that its pretty easy to cool, either with direct air cooling, or with a cold finger. This is another reason Ben would like to buy another camera to play with. cooling.

Its hard to see, but trust me, there is a "big" gap between the back of the ccd and the pcb, should not be a major project to slide a cold finger in between the two...... one day... maybe....

The mount in more detail.

mds finally got himself the last two mounts that Oatley Electronics had, and just as Ben suspected, as soon as Mark got them, the mount firmware advanced in leaps and bounds.

Here is some information that mds gleaned in one quick session...... more info and photos to come......