This was brought up in the MER forums but I figured it might be useful here where discussion of strategy and planning for missions goes on.

Having visited a DSN station this year I have a new found interest in the amazing mission these sites are tasked with. Unfortunately like everything in these belt-tightening days the DSN is underfunded for what it is they need to accomplish. That said a friend pointed out to me that there is a simple formula to calculate the DSN costs for a mission:

Missions calculate DSN costs using this formula: AF = RB [AW (0.9 + FC / 10)]

Where:
AF = weighted Aperture Fee per hour of use.
RB = contact dependent hourly rate, adjusted annually ($1057/hr. for FY09).
AW = aperture weighting:
= 0.80 for 34m High-Speed Beam Waveguide (HSB) stations.
= 1.00 for all other 34m stations (i.e., 34m BWG and 34m HEF).
= 4.00 for 70m stations.
FC = number of station contacts, (contacts per calendar week).

Online tools available:
DSN Aperture Fee Tool - Excel file 706k
DSN Services Catalogue - PDF 675k

The DSN is so critical to nearly everything we discuss here. Everyone should take the time to learn more about it, and if possible try to schedule a visit or tour. The people who work there are the unsung heroes of the headline grabbing feats that NASA and JPL accomplish.
http://deepspace.jpl.nasa.gov/

In the case of Spirit and Opportunity, the DSN costs are the costs of communicating with the relay orbiter. I suspect that the relayed data may be a small portion of the total, since the orbiter is also collecting its own data.

The majority of uplink is still done directly to the HGA from the DSN however.

I would guess that the answer to how much of the MERs' extended mission costs are DSN costs incurred for uplinks and paid to the MODY /MRO project teams for their relay services lies in the detailed budget accounting. Is the detailed budget data (down to line item) for the MER extensions available publicly? That would be quite instructive, I would think.

It would also be good to see budget detail at that level for other missions, like Cassini, New Horizons and LRO (to name but a few) to see just what the percentage of operations costs commonly consist of DSN charges.

-the other Doug

You act as though the DSN is getting rich off of this. The DSN is already underfunded and oversubscribed. Your comments about looking into this, referring to DSN costs as "charges," and your earlier comments make it sound as though it is somehow unjustified.Regardless of what the cost is, unless you have a better idea of how to communicate with a distant spacecraft (and I don't just mean an idea, but something you can actually build and demonstrate to reliably work), acting like this isn't a necessity is like arguing that planetary exploration would be cheaper if we stopped spending money on launch vehicles or stopped maintaining flight teams to operate the missions. Frankly, I find this to be a disrespectful and uninformed way to describe the folks who do the Herculean task of commanding the spacecrafts and recovering data from their faint signals.

It turns out that LRO doesn't use the DSN, it uses a Ka-band system that GSFC built in White Sands, NM.

http://cio.gsfc.nasa.gov/centers/goddard/p...main_GV3_16.pdf

Some proposed lunar missions I know of planned to use Universal Space Network's system -- http://www.uspacenetwork.com/index.html -- obviously lunar missions don't need big antennas.

That's exactly what I've been thinking as well.

I think you guys are overreacting to some natural curiosity about DSN costs. As with so many other things, DSN costs could be lowered and efficiency increased; DSN isn't perfect. That said, I'd tend to believe that DSN costs are not the majority, or even a very large fraction, of ongoing mission ops costs, except maybe for extremely-extended missions like VIM.

I was set to defend the desire to know the breakdown of costs, including communication costs, of space missions. But in preparing my defense came upon this in the "Getting Unstuck in West Valley" thread:

In all that has been said in these two threads, this is really the only uncalled for accusation. I would assume that if communication fees are high, its because the costs of operating and maintaining the equipment are high. Maybe dvandorn would like to explain or retract this statement.

It would be interesting to see, per mission, how the costs divide between cost of the probe, cost of the launch, salaries to manage the probe, salaries to analyze the data, and (of course) cost of using the DSN. Among other things, that breakdown would make it easier to see what to try to optimize. Given limited budget dollars, I'd far rather see NASA (and other agencies) optimize the process rather than cut probes.

But I don't think I've ever seen this kind of cost breakdown for any probe -- ever. Does it even exist? That is, does even NASA know?

--Greg

I know it exists in the proposals (Discovery, etc.) I don't know whether or not there is a final report available. As I said, I am not criticizing wanting to know what it costs. It is the attitude of the posts that I find offensive (the earlier ones more so than the latest one).

Of course NASA knows. For every mission extension a detailed plan and budget has to be written and reviewed by HQ. I've never seen one of these publicly released, but you could always file a FOIA request if you really care (instead of relying on what was put in wikipedia, as this discussion appears to be doing

There isn't even a breakdown of how much the MER mission will cost in the 2010 NASA budget that I can see, but perhaps I missed it.

To be fair - the calculations for DSN time are from published documents from the DSN itself, not Wikipedia.

I'm a little bemused about why this is so offensive when all kinds of technical misinformation, unfounded opinions, etc, to my ear more technically grating, go uncommented upon in this forum. Seems a bit fanboyish to me.

Certainly, and we could discuss the fraction of cost that went to DSN for a given mission extension, if we knew how much that extension cost, but as far as I know we don't.

So comment.

I would agree that the defense of DSN may border on fawning, EXCEPT that the bit I quoted in my last post seems to be an outright accusation of price gouging. Why someone should want to bite the hand that feeds us so much great data from afar, I don't know. Maybe he knows something?

p.s. I thought we were all fans here.

This statement is what has offended people. The statement that DSN time was 'astronomical' and should be 'attacked'. I would like to see some numbers to establish if indeed the charges are 'astronomical' before making such a bold statement. Wouldn't you? It's quite a bold claim.

If wanting some numbers and evidence before claiming something needs to be attacked for being astronomical means I'm a fanboy - then I am a fanboy.

Let's try and make an estimation. Looking at the MER Analyst notebook - you can see MER schedules. Sol 1798B, for example.




The HGA starts at 11:09 - and including beep, it must be finished by 11:50. The same pattern is on all sols I can found where uplink was done. Let's round up to an hour.

1hr DSN pass + 45min setup and 15min setdown (that's a req in the docs) is 2 hrs of DSN time. Twice a day. 4 hrs per day.

AF = $1057 * ( 1 * (0.9 + 14/10) ) =, I think, $2431 / hr.

So - I get about $3.55M / year.

I'm not even accounting for the fact that the DSN can now operate multiple spacecraft from one dish at the same time - thus the costs can be shared across multiple missions in some way. Furthermore, for some significant swathes of MER ops - they don't uplink daily - especially during very low power periods. But let's ignore that - and stick with the $3.55M / year.

If the figure I've heard of around $20M/year for MER ops is true - then about 17% of MER costs would be required for 1 x 1hr uplink, per rover, per day.

The costing of downlink is going to be far more complex as it's tied in to Odyssey DSN fees. When Dan and I were at the DSN - it was the 70m dish that was Mars pointed (which costs 4x as much ). BUT - a 15 min UHF pass might generate 30-60 mins of MODY downlink budget. So perhaps it's fair to basically double that $3.55M / year figure for the MER share of downlink as a ballpark estimation. That basically comes out to DSN being 1/3rd of the alleged MER budget at, what I would classify, a worst case rough estimation.

Now - is that astronomical - does it need attacking? I don't think so.

Hands up if you do.

Talking about being a fan.
To me the 3 letters of DSN are as magic as the one from JPL.

When I said "attack" I meant to aggressively investigate why the costs are so high, as in "we need to attack this problem and get it solved, people." (That is a perfectly acceptable use of the word, and not intended to imply violence or disrespect against the institution being discussed.) Is it intrinsically extremely expensive to operate these radio receivers? I grant you, they have to separate out extremely faint signals from background noise that fairly swamps them. But does the equipment needed to do so cost exorbitant amounts? Or does it require millions and millions of dollars a year in maintenance? These are the questions that were on my mind.

One thing that was on my mind, as well, is that the way supply-and-demand economics works, you can charge pretty much what you want to charge if you have no competition, and for the most part the DSN has no real competition. I guess I'm just wondering, in this age when electronics are so incredibly advanced over what they were when the DSN was designed and built, are there cheaper ways to accomplish what the DSN accomplishes? Even granted the initial outlays that would be required, the fact that the DSN is so time-limited by so many demands on its time, it would be quite useful to have a second or even third set of receivers active at (roughly) every DSN location around the globe (or, as has been suggested, though I think the cost would be really outrageous, place a new DSN into LEO). If a new network can be built that can be operated much more cheaply, with the same quality of data return, as the current DSN, has anyone contemplated how to get this done? *Can* it be done?

Those were the questions that were on my mind.

I certainly didn't mean to offend anyone. I just have run across several cases where DSN costs have been discussed as being very high, tens of millions of US dollars a year to support a single planetary probe. Everyone talks about ways to lower costs so we can fly more missions, but they seem to concentrate on launch vehicles and ASRGs and such, when over the course of a long, multiply-extended mission, it would appear that DSN costs are one of the really high-ticket items in the budget. And the DSN is getting to the point where we can't fly many more missions at one time than we have going right now; we seem to always be looking at trade-offs from one mission to the next based on which mission gets the DSN time at any given day and hour. (I know we've lost some Cassini data due to DSN conflicts, that's been mentioned here before.)

IIRC, one of the main DSN dishes is offline now (repairs that will take a couple of years to complete), and it is making the juggling of requests that much more difficult. We're now in a position, it appears, that if we lost another dish in the DSN, we may have seriously degraded our ability to operate the missions currently in flight. I'd say that calls for an expansion of the capability. And if the many millions (perhaps billions) of dollars US spent in total for DSN time every year underfunds the network, then how in the world are we going to expand it?

These were my concerns. I sincerely apologize if I offended anyone, I certainly didn't mean to. I just wanted to point out some concerns over planning expanded planetary exploration when the DSN seems to be a bottleneck, both in terms of cost and in terms of capability.

Forgiven?

-the other Doug

To my way of thinking, you don't need forgiveness (frankly this forum would benefit from a little less uncritical adulation) but your idea that DSN costs are such a large fraction of total mission budgets is, so far as I can tell, wrong.

That said, DSN would benefit from some more modern infrastructure, and there could be some cost reductions after that initial investment (why do you think GSFC built their own system?) but I suspect the savings might be 2x, not 10x.

You typed your breakdown while I was typing my response, there, Doug. Those are much lower figures than I've seen on some other fora for DSN costs, and also your mentioning of ability to communicate with multiple spacecraft at one time through one dish is something that wasn't in the literature last time I looked through it (about 6 months ago).

And please, stop harping on the term "attack." It means, in this context, to aggressively seek resolution to what appears to be a valid concern. As I've explained, that's certainly how I meant it; I'd appreciate it if you can accept that.

-the other Doug

No disrespect, but I think you're wrong. This forum is an online oasis of respect and good-natured support. I'd much rather have it this way than have to wade thru page after page of angry woo-woo posts attacking NASA for hording images, hiding discoveries and telling lies. In these times of Commissions, shrinking budgets and shrinking visions, I would have thought that adulation - uncritical or not - would be welcomed by anyone involved with NASA.

It wasn't an especially considered or ambiguous statement you made toDoug, so forgive me for taking it at face value - ignoring the attack word - you still described the DSN as "the real funding-drain of any and all deep space missions".

My very rough estimation demonstrates, I hope, that this isn't true.

As for multi-spacecraft on one dish, it's called Multiple Spacecraft Per Aperture - it's all over google in various places. We mentioned it here in '06.

Thanks, Mike. I was going off of some discussions on a couple of other forums I read and sometimes post to, which were quoting DSN costs as as much as $15 to $20 million per year (if not more) for pretty much every spacecraft that uses it, and from what I was seeing from the publicly-available cost formulae at the time, these guesstimates didn't seem out of line.

If I'm wrong, I'm certainly more than willing to admit it. In the beginning here, I was just responding to concerns expressed and opinions given over the cost of yearly extended ops for the MERs, and was trying to determine where all of about $20 million a year was being spent to keep the old girls going. Not that I feel it's money poorly spent, just that, with the world economy in deep recession, I'm afraid that our wonderful planetary exploration vehicles, including the MERs, could be cut back or cut off entirely unless we figure out some way to operate them more cheaply.

For example, I recall that there was a "groundswell" campaign to collect money to keep the Viking landers going when Congress threatened to cut off their funding, and that the money collected wasn't really enough to pay for the DSN time, much less data analysis and storage. And that was nearly 30 years ago. That certainly reinforced the idea that DSN costs were a majority of the continuing operating expenses.

If Doug's numbers are right, then it would seem that we're paying something like $16 million a year for the rather limited staffing and facilities required to operate the MERs. Apologies, but that doesn't sound realistic, either.

-the other Doug

EDL! EDL! EDL!

Take a trip to JPL. If you're lucky Scott or Paolo or someone can show you around. I've seen it, and Cornell as well. It's a lot of people, a lot LESS than it was, but still a lot of people, a lot of offices, a lot of stuff going on. Remember the rows of desks during the early MER ops, Steve and Justin with the images on two monitors - that whole room is still used, day in day out. There's multiple sequencing rooms, theres instrument teams, there's the ISIL. Multiple daily teleconfs between JPL, Cornell, and host institutions of other participating scientists and engineers. The not insignificant effort in PDS releases. Training, IT support, security, oh the damn security, PAO, the servers for the jpg's. Hell - they've spent more on sand, grit, dirt, clay etc than a small building firm.

If it were a business, I can very easily imagine an 8 digit annual turnover.

For comparison - Cassini's 2 years extension is $80m per year.

Again, apologies. I'd still be very interested in seeing what the actual DSN costs are in these budgets, though. As I said to Mike, if about $20 million per year in extension funding isn't mostly DSN costs, I have to wonder how much it costs to keep 20 to 50 people employed, in office space, and with decent computer access these days...

-the other Doug

Again, sorry -- writing posts and getting interrupted at work and you're answering my questions before I can post them...

-the other Doug

What made a "DSN in space" idea attractive, at least to me, was the idea of letting probes use laser instead of radio to send back data. In theory, that gets you far higher bandwidth for less power (on the probe) and also less weight. JPL seems to have had a group working on this, but it looks like they disappeared around 2003 or so.

--Greg

At Microsoft, we used to estimate that the cost of a "head" was roughly double the salary. If those fifty people made $100,000/year each, that'd give you $10M. Throw in some high-priced consultants, and the cost could easily be mostly salaries.

--Greg

Wow! That's a lot, Greg. Of course, MS has never been known to scrimp on salaries.

I guess I've just been figuring that the mean salary of an engineer is on the order of $50K a year. Doug has noted to me personally that it's likely a lot more -- something that surprises me, but then again, I only know about salary ranges for the companies I've worked for and that friends and acquantiances have worked for. (Not to say our intrepid JPLers don't deserve every penny they get, I was just, from my own work experience, vastly underestimating the salary ranges.)

So, please -- if anyone out there thinks I'm undervaluing the men and women at JPL, APL, Cornell, etc., etc., all I can plead is that the software and hardware engineers I know who work in the Minneapolis area seem to think that $50K is a great salary. Please forgive me for assuming that salary ranges where I live, for people who do different things than supporting space probes, are at all representative of salary ranges across the country. Mea culpa, mea culpa, mea maxima culpa....

-the other Doug

This hardware engineer thinks 50K would be a cruddy salary-- and that's just a relative statement; I'm not a spendthrift saying it's unlivable; I save a lot with my salary-- no matter where you are except perhaps right out of college somewhere where you can't get to the ocean in a day's drive.

But you can see actual numbers:
http://www.glassdoor.com/GD/Salaries/compa...6850&locT=C

(A search for engineer salaries in Pasadena, CA at glassdoor.com)

Just the cost of living in Southern California would make that salary not very large at all.

Also to back up what Greg said earlier - the salary that is paid to an employee is just part of the overall cost of each individual to an organisation. Having worked for them I know Dell & Intel do the same as Microsoft - the basic rule of thumb is to double the salary when estimating the cost of warm bodies. When you employ someone you have account for all of the extras - there's obvious things like additional taxes\social insurance levies, medical, pension etc but you also have to pay for the organisational growth needed to support an extra person (their management obviously enough but also the extra HR, finance, payroll, security, maintenance and cleaning staff etc) and you have to pay for their office costs - the floor space they take up, the power you use to heat and cool it, telephony. Then you get into travel and training budgets, paying for conferences and subscriptions\fees to professional\academic bodies\journals etc.

The Glassdoor numbers Steve linked to look about right to me and I would be very surprised if the cost per head in accounting terms for any of these projects was less than $200k per annum.

Plus the unusual extra that you don't have in the typical corporate situation. The ISIL, maintaining two TB rovers, liasing with other missions for UHF relay etc, science teams (outside that number of people figure)

Just 50 people at $200k a head. Plus $5m DSn. Plus $5m for the 'only in MER land' extras. $20m.

I don't honestly think it could be done cheaper without compromising mission safety or productivity significantly.

One pertinent tweet from Scott last night



I still think MER at $20M is CHEAP.

This also helps explain why Alan Stern was so pleased that NH uses a staff of ten or less during cruise, as compared to Voyager which needed 100.

http://pluto.jhuapl.edu/overview/piPerspec...ective_1_5_2009

Perhaps we should ping Alan and ask if he'll share with us what New Horizon's actual costs are per year during cruise. That'd be a rock-bottom minimum, since there's no data analysis to speak of.

--Greg

If the total annual O&M budget of MER is only $20M to cover all the activities described plus personnel costs, that's RIDICULOUSLY cheap. I know of programs whose sustainment activities are confined to software maintenance alone for one (1) platform using two or three bodies part-time, and they spend $3-5M per annum just to do that.

One thing you can't put a price tag on is the dedication of people like Scott & Paolo & all the other MER team members. (Or maybe you can; MER would probably require 2x-3x more funding & people if the program's staff weren't truly passionate about it!)

Although in that case, the mission would have been over after the first 90 days!

--Greg :-)