[hpsdr] Mechanical considerations for ATLAS cards

Tue May 23 00:39:32 PDT 2006

Alex,

I can understand your interest in the South Pole project, but that's way
out of my hands. The engineering is pretty much done and the orders are
going out. I can give you a few parameters -

1) There will ultimately be ~100 boxes that contain single-board CPU's
along with 8 specially build I/O cards in each box. Each I/O card has
eight robust twisted pairs connected to them; the other ends of the
pairs are on instruments from 1.5 km to 2.5 km down into the ice.
Long-term reliability is paramount as from Feb. 15 to Oct. 15 each year
there is no physical access whatsoever to the Pole from the outside
world for parts supplies or anything else.

2) Lowest possible power is probably the next most important parameter.
All electricity is generated on site from jet fuel which must all be
flown in during the few open months of the year. Noticed the price of
fuel lately?

3) There will also be a couple of hundred conventional, more or less,
PC's for the remainder of the data acquisition and analysis system. The
most interesting data are shipped North during the < 1/2 day that
wayward geosynchronous satellites peek up over the horizon. Really, our
data comes in over satellites that are so broken they can no longer hold
their stations in orbit. The rest of the data is on tape, so not
available for a year or so.

4) Next year, we are scheduled to move into a new computer house, but
for now, the somewhat less than 10% of the installation currently in
place is in a temporary counting house consisting basically of two
freight containers with windows and heavily insulated doors. The room
has a tendency to fluctuate between snow blowing in the vents and
overheating at 90 F from all the electronics.

5) The station is at > 9000' altitude => about 1/2 atmosphere pressure,
so the fans don't work all that well.

6) When I was there over Christmas - mid-summer - it was about a 15
minute trudge from the home station to the counting house. That was in
full daylight (bl**dy weird endless daylight) at -11 F and winds of a
few mph. It is now dark all the time, -67 F with 22 mph wind and blowing
snow. There are on occasions a certain reluctance of the Winter-overs to
slog out there.

http://icecube.wisc.edu if anyone is interested.

	Chris - AE6VK

P.S. - There is a ham radio station there - KC4AAA. One of our
Winter-overs - Victoria - has her Extra Class license, so could in
principle be on HF some time.

-----Original Message-----
From: Alex [mailto:harvilchuck at yahoo.com] 
Sent: Monday, May 22, 2006 6:53 PM
To: Christopher T. Day; Lyle Johnson; hpsdr at hpsdr.org
Subject: Re: [hpsdr] Mechanical considerations for ATLAS cards

Chris,
For the station at the South Pole I will gladly help with a case for
them that is easy for them to maintain.
Please provide me with some of their requirements (space constraints,
weight constraints, power constraints, etc.)

Building a unit for the south pole station is quite cool (no pun
intended) and very interesting. I think I have a couple ideas based on
your comments so far.

Alex, N3NP

----- Original Message ----
From: Christopher T. Day <CTDay at lbl.gov>
To: Alex <harvilchuck at yahoo.com>; Lyle Johnson <kk7p at wavecable.com>;
hpsdr at hpsdr.org
Sent: Monday, May 22, 2006 3:37:43 PM
Subject: RE: [hpsdr] Mechanical considerations for ATLAS cards

Typically, exactly the opposite arguments are made for all the in-house
electronics made for every large High Energy Physics experiment I've
ever been on.

1) Having the I/O connectors opposite the backplane connector means that
all meaningful access to the card can be done from the front of the
rack. With a top-loading box, the whole box must be removed from the
rack to replace a card. This is awkward at best and dangerous if the box
is on slides and the rack can tip over.

2) Having the cables in the back means that they must be removed from
all cards, typically, to pull out the box. At the very least, all cables
must have enough slack to pull out the box.

3) At least front and back of the rack both must be readily accessible.

Currently, the WinterOvers at the South Pole are grumpy enough having to
replace boards in a back-plane design. Having to slide boxes in and out
would be completely impractical.

The PC box's connector placement is, I suspect, designed for a bottom
motherboard in order to provide lots of room for active components on
that board. We only have a few passive components on Atlas. Also
remember how old the PC design is; way back when, it took a lot of stuff
on the motherboard to make anything happen.

I don't think PC cases are particularly well designed for experimental
electronics. They're just cheap.

    Chris - AE6VK

P.S. - It's not really me that you have to convince, but the guys doing
the layouts. So far, they all seem to have connectors on the opposite
side from the DIN-41612 connector.

-----Original Message-----
From: Alex [mailto:harvilchuck at yahoo.com] 
Sent: Monday, May 22, 2006 12:17 PM
To: Lyle Johnson; hpsdr at hpsdr.org
Subject: Re: [hpsdr] Mechanical considerations for ATLAS cards

***** High Performance Software Defined Radio Discussion List *****

(a) it forces me to fabricate a new mounting plate for ATLAS that is
perpendicular to a perfectly good mounting bracket on the bottom of the
case
(b) I still need a bracket for the external connector. Now in order to
swap out the card I have to remove the entire ATLAS assembly in order to
remove a card, since the insertion direction for the card into the ATLAS
is perpendicular to the insertion direction for the external connector
bracket.

So I stand by my statement of preventing me from using any computer
case. I didn't say it was impossible to use the case.

If somebody wants to make a minimal-size HPSDR/SDR-1000 configuration
you just grab an old surplus pc case, remove a section from the front
half of the case and reattach the front plate to the back half of the
case (like chopping down a car). Or just cut out the rear expansion
slots section of the case and use it in your own from-scratch box.

Or buy one of the cheap mini cases and mount everything in it - less
time and hassle.

Plus the added card size gives more room for increased funcationality
per card and thereby reducing the number of cards in the system (since
we seem to be getting more cards than we can fit on a backplane). 

Don't get me wrong, I think it's a good thing to have more
functionality. Keep thinking of ideas for cards.

Alex, N3NP

----- Original Message ----
From: Lyle Johnson <kk7p at wavecable.com>
To: hpsdr at hpsdr.org
Sent: Monday, May 22, 2006 1:24:53 PM
Subject: Re: [hpsdr] Mechanical considerations for ATLAS cards

***** High Performance Software Defined Radio Discussion List *****

> I am saying that not relocating the connectors to the perpendicular
side PREVENTS me from using any computer case (rackmount or desktop).

Not if ATLAS is perpendicular to the bottom of the case.

73,

Lyle KK7P

_______________________________________________
HPSDR Discussion List
To post msg: hpsdr at hpsdr.org
Subscription help: http://lists.hpsdr.org/listinfo.cgi/hpsdr-hpsdr.org
HPSDR web page: http://hpsdr.org
Archives: http://lists.hpsdr.org/pipermail/hpsdr-hpsdr.org/

_______________________________________________
HPSDR Discussion List
To post msg: hpsdr at hpsdr.org
Subscription help: http://lists.hpsdr.org/listinfo.cgi/hpsdr-hpsdr.org
HPSDR web page: http://hpsdr.org
Archives: http://lists.hpsdr.org/pipermail/hpsdr-hpsdr.org/

 1148369972.0