Welcome to the ECE labs upgrade page!

(Maintained by John McNeill; all opinions are mine unless otherwise noted!)

New:

Summary of discussion from April 7, 2006 meeting of Lab Upgrade Subcommittee (Bitar, Jarvis, Ludwig, McNeill)

Recommendations:

Modify course scheduling to increase efficiency of lab utilization: every lab should be used every term.
Consider moving to 2 hour lab sessions with open lab time for people who can't finish in 2 hours. Would allow 4 lab sessions to be scheduled on Wednesdays and would also ease scheduling conflict crunch for non-Wednesday lab session.
Combine 3601 and 2011 into one course that is offered every term. This course would use a basic lab facility.

Need for five different lab facilities identified (see notes on courses and lab "owners"):

Lab	Courses/Terms				"Owners"
Lab	A	B	C	D	"Owners"
Basic analog	2011	2011	2011	2011	??
RF	3113				Ludwig, Makarov (see comments)
Advanced analog	2201	3204	2201	3204	Bitar, McNeill
Basic Digital	2801	2022	2801	2022	Duckworth, Jarvis
Advanced Digital	3801	3803	3801	3803	Duckworth, Jarvis

Comments:

Co-ownership allows continuity for sabbaticals, off-campus advising, etc.
RF lab could also be used by Makarov antenna course, maybe King photonics? Also could revive Comm Circuits course in this lab to cover mixers etc. which have no hardware home right now.
Ludwig and Makarov have a proposal in to NSF; fold this into big lab proposal if not funded
3801 (Advanced VHDL/Verilog course) moved into A,C terms to ease lab scheduling crunch
Already have description of Advanced Digital lab station from Prof. Duckworth
Action item: to Bitar, Jarvis, Ludwig, McNeill for equipment list for other labs

Pitch big proposal based on above to NSF for implementation starting in Fall 2007. If we don't get big money, implement over longer timetable using our own money. "Top down" philosophy: start with advanced labs, work toward labs earlier in curriculum.
Philosophy: All lab stations will use LXI bus for connectivity to PC for waveform and data capture.
Above does not address needs of power, robotics curriculum! Also missing lab experience for 2111 (Physical Principles of ECE, need coverage of resonance and second order systems)

Philosophy behind digital labs above (and friendly recommendation to CERC):

Level	Content		Comments
Level	FPGA/VHDL/Verilog	Processor	Comments
Simple	2022	2801	"New-User-Friendly"
Advanced	3801	3803	For "Power Users"

We considered splitting labs along the FPGA vs. processor axis, or the Simple vs. Advanced axis. We decided the tools used in the lab would be similar for both subdisciplines; the big difference is in course populations. Pretty much everybody should take 2022 and 2801, so these courses should have the most "user-friendly" lab equipment (very basic logic analyzer). Being more advanced courses, both 3801 and 3803 could take advantage of more advanced equipment and also might have smaller populations, so a smaller facility with fewer (but more expensive) stations would be appropriate.

Lab experiences in signals courses could mostly be done with MATLAB on student laptops? Not sure if DSP class could use cards that students take with them, or what.
Decided against a lab experience for 2112 (Fields): barely room for coverage of material that is there now; save lab experience for 3113 which is more hardware appropriate.

Summary of Lab Visits December 2005

Link to pictures

As we review the different areas of our curriculum, one area that I am especially concerned about is the quality of our students' lab experience. To that end, I plan to write a grant in the spring of 2006 to NSF, Tektronix, Agilent, and whoever else has money, to upgrade our undergraduate labs.

To get an idea of where we stand relative to other schools, Steve Bitar and I visited NU, BU, and MIT on December 15th for a whirlwind tour of these institutions' undergraduate teaching and project labs. We measured physical size of the labs and took lots of pictures. Here is a summary comparing total number of lab rooms, # of stations, total lab square footage, and for scale the approximate number of BS grads/year for each program:

School	Total # of Labs	Total # of stations	Total square feet for labs	Avg sq ft/station	BS grads/year
MIT	7	233	16725	72	200
BU	4	103	8400	82	100
NU	10	95	10250	108	100
WPI	6	107	6300	59	100

(The Details of the Lab Space Comparison are in this spreadsheet).

Some conclusions I draw from this data:

Total # of labs: Most for NU; this reflects NU's philosophy of many smaller labs, each specifically tailored to the needs of one or two courses. BU is more toward the one-size-fits-all model (and also our impression was that in general their faculty are least interested in the UG students' lab experience). MIT is Big Education.
In terms of Total # of stations, the rule of thumb appears to be you need about as many stations as you have graduates from your program each year. In terms of quantity of stations, we are OK. It is more the quality of our stations that I worry about, as you can see from the photo tour below.
Total square feet for labs: We are low both in total space devoted to labs and average square footage per station. Perhaps this is beacuse we have been devoting more lab sapce to research labs in recent years. However there are areas we are not using effectively (e.g. old fab lab on first floor).
Average square footage per station: NU's very high area is a reflection of their inclusion of good teaching resources in each lab: as you can see from the photo tour, most labs have an instructor's bench or teaching area. Perhaps not surprisingly, MIT is more efficient in using lab space, partly because their labs are just huge and the instructor's bench / teaching area is a smaller fraction of the total.

Bottom line: It's not our imagination that some of our lab space feels a little cramped. However, we're probably not going to get a lot more space anytime soon. We need to think about how to use our space effectively. I think some of our space is not used very effectively right now. MIT in particular was very good at structuring their program so that every lab was in use by a course every semester. We might want to adopt a similar philosophy when scheduling our courses, for example in the digital / CE area as mentioned in their latest update.

Lab Photo Tour

Follow this link to see photos of the NU, BU, and MIT labs. I think you will get a good sense of what these schools are like. Our impressions (more details to back these up will be forthcoming):

MIT: big labs, uses space very efficiently, advanced labs have excellent equipment, introductory labs are surprisingly bare-bones
NU: Lavish on lab space, especially in capstone and by including instructor bench and teaching resources in most labs.
BU: Faculty's #1 job is getting research funding, labs not a priority

Program comparison

A quick/rough comparison of programs in terms of faculty size, BS graduates per year, and research $/year:

School	Approx. T/TT Faculty*	BS/year	Research $/year
MIT	70**	200	$ ∞
BU	40	100	$ 9M
NU	40	100	$ 10M
WPI	20	100	$ 2.5 M

* Not including adjunct (FT or PT), PoP, emeritus, etc.
** 142 faculty in EECS department; assume ½ ECE, ½ CS

Bottom line:

Let's not compare our program to MIT.
For BU and NU: basically they have 20 additional faculty which buys them about $7M in additional research funding / year.

one way to look at it: That’s $350K (gross) per additional faculty member (assuming orignal 20 are delivering entire UG program, like we do)
Another way to look at it: Each of their faculty brings in $250K of research per year, and delivers half as much UG education as each of us do.

(Does not factor in that we advise projects too! )