Gas station without pumps

2015 November 27

Resistor assortment box

Filed under: Circuits course — gasstationwithoutpumps @ 11:36
Tags: , ,

A few years ago, I bought myself a selection of 1% resistors, like the ones I have my students get in the applied electronics course (from DIYGSM through Amazon, though I have tried other suppliers since then, since UCSC purchasing doesn’t permit ordering through Amazon, even when their prices are the best).  The resistors come in sets of 10 on paper tape (cut from the long rolls used for pick-and-place machines), rubber-banded into bundles of about 50 different sizes.  After a while the rubber-banded bundles started to disintegrate, and I’d gotten quite a collection of loose resistors that were no longer on the tapes, so finding the resistor size I wanted was often a bit of a hassle.

For a while I stored the loose resistors by poking them into a block of foam, roughly sorted by value,  but that was no longer working well—it took too long to find out whether I had a resistor of the size I needed stuck into the foam.  So a couple of weeks ago, I bought a large number of 3″×5″ plastic zippered bags and a 4″×6″ index card box, and sorted the resistors by size into individual bags:

Box with about 120 different sizes of resistors, from 0.5Ω to 5.6MΩ.

Box with about 120 different sizes of resistors, from 0.5Ω to 5.6MΩ.

I was originally planning to use white address labels to put the size values on the bags, but the address labels did not stick to the plastic.  A different sort of label (like the ones sold for marking stuff on freezer bags) might work, but I just used a permanent felt tip mark to write directly on the bags).

I still have about 100 loose resistors to file away, which I’ll probably finish doing today. It takes a while, as I try to confirm each resistance value before filing it: both reading the color code and measuring the resistance with an ohmmeter. My Fluke multimeter is broken, so I’ve been using the DT-9205A multimeter that I reviewed earlier. I found out one reason that multimeter was cheap—one of the probes fell apart within a few weeks of light usage.  I bought myself some more cheap voltmeter probes on AliExpress, which work ok (though who knows how long they will last).  The new probes have very sharp tips, which is handy for probing surface-mount boards, but a bit risky for clumsy people like me—I’ve stuck myself with them a few times by accident.

The DT-9205A meter is rather awkward for reading resistance—it often takes several seconds to settle for a larger resistor, and there is no zero-function to compensate for the resistance of the probes when measuring small resistances.   The ohmmeter is only accurate to about 2% also (a 1kΩ±1% resistor measures at 984Ω on the 2kΩ scale), which is nowhere near as good as the Fluke meter I’m used to, nor the expensive meters at work.

The box and bags is much bulkier than the original rubber-banded bundle (maybe 4 times the volume), so I’m not going to recommend this approach to my students, but I think that it will save me some trouble in future (as well as letting me know when I need to re-order a particular size resistor).


2015 November 23

Meeting for teachers of writing to engineers

Filed under: Circuits course — gasstationwithoutpumps @ 19:00
Tags: , ,

Last Spring I got a small grant from the Academic Senate to create a new “Disciplinary Communications” course for the bioengineering majors (a $7,000 “partial course relief” for 2015–16).  Most of the effort of creating the course happened last year, as we needed to offer the course in Spring 2015, but the money comes for this year.  I’m not actually taking any course relief this year, though my load is lighter than last year, since I’m not doing two overload courses this year.  The money (as all our course relief money) is being spent on hiring a lecturer—paying part of the salary of the lecturer teaching the new writing course.

But I felt that I ought to be doing something this year on improving “disciplinary communications” for bioengineers, in order to have something to report at the end of the year for the grant.  Since the new course was designed last year, the main effort this year will be on tweaking that course and other courses our students do that involve writing.  Rather than work just with the instructor of that new course, I thought it would be useful to gather all the faculty who teach writing to engineering students, to discuss (according to the message I sent out):

  • course design
  • teaching techniques
  • assignments
  • grading techniques
  • use of TAs or graders
  • creation of a “Professional Learning Community” to meet on a regular (quarterly?) basis

There was no set agenda for the meeting—just a chance to meet and talk about what we do. We had a pretty good turnout: 3 ladder-rank faculty, 4 writing instructors, and 1 staff person who teaches writing to a small group of minority students.

After self-introductions we had a wide-ranging conversation about assignments people gave, challenges they faced, approaches to making assignments work better, and so forth.  We did not talk much about TAs and graders, course design, or grading techniques, concentrating more on assignments and teaching techniques.

I’m a lousy note-taker, so I don’t have good notes of what was discussed, but I remember a few things.  I’ll present them here mainly as they apply to me, since that is what I remember best.

None of the ladder-rank faculty are teaching courses where writing is the primary content of the course, but improving student writing is a secondary goal of their courses. In my case, I’m (thankfully) not teaching either the technical writing for bioengineers course nor the senior thesis writing course this year, but I do provide a fair amount of writing feedback both in the Bioinformatics: Models and Algorithms course and in the Applied Electronics course. In the bioinformatics course, there are a couple of writing assignments, but most of the feedback is on in-program documentation. In the Applied Electronics course, there is a weekly design report due, which is centered on the graphics (block diagrams, schematics, and fits of models to measured data). Other courses include assignments to write abstracts, write proposals, write standard operating procedures, and other assignments typical of both academic and industrial writing tasks.

One aspect of teaching writing that I’ve never had much luck with is peer editing—another of the ladder-rank faculty brought this topic up as one of the challenges that help was needed on.  A couple of the writing instructors agreed that peer editing was hard, because the students had no notion of “editing” as an activity. What they suggested was having a set of specific questions for the peer editors to answer—questions relevant to the piece they were editing, like “what is the research question? Is there a summary of results? Is the approach clear?” for editing an abstract.  Without specific guidance, students tend to fall back on the if-you-can’t-say-anything-nice-don’t-say-anything meme, and provide useless “looks good to me” comments.  One technique that the faculty member who raised the issue has tried (with mixed success) is getting students to rewrite another student’s abstract in their own words.  Although this often pointed out problems in the original writing, it sometimes just reflected the inability of the editing student to write coherently.

One idea that seemed to come as a bit of surprise to some of  the writing instructors was creating the figures and figure captions of a document first, and then writing the paper around the figures.  This is a common approach in some research groups in our department, and one that some students will have to face. One of the writing instructors pointed out that the poster assignment (used in two of the courses) is good preparation for this.

We all pretty much agreed that there was no place in the writing instruction students were getting about good presentation of data and generation of figures. I mentioned that one of our junior faculty is interested in creating a course centered on scientific graphics, but it wasn’t clear whether he’d get to teach it next year or not.  I felt that students in my Applied Electronics course got a lot of instruction and got pretty good at displaying data (at least the scatter diagrams and fit models for that course), but that they really struggled with the notion of block diagrams and organizing problems into subproblems. One of the writing instructors, who saw the students mostly after they had had the applied electronics course, saw more problems with data presentation than with block diagrams.  This may be because of different expectations of the block diagram, or it may be that the data representations her students needed were not among the few types covered in Applied Electronics.

Another form of writing that a lot of students were not getting adequate feedback in was lab notebooks. Unfortunately, the different disciplines have such different expectations of the content of a lab notebook that it is hard to provide any sort of standardized assignment. A couple of the instructors who teach Writing 2 classes, mainly to STEM students, do include an observational-field-notebook assignment, which at least gets across the idea of taking notes as you go, and not trying to reconstruct what you did at the end of the day (a flaw I’ve seen in several of the Applied Electronics labs) or the end of the quarter (a flaw I’ve seen in some senior theses).

We did discuss the strategy of setting high expectations on the first assignment by giving detailed feedback on that assignment, with reduced checking on subsequent assignments.  This helps keep the grading down to an almost sane level, and the students still benefit from the practice, even if not everything they do is checked. I’ve certainly noticed on the bioinformatics assignments that by the 4th or 5th assignment I only need to spot-check the internal documentation, or check it for students who are struggling with the concepts of the assignment, as the better students generally are routinely producing decent documentation by then.

We discussed various things we could do that would be generally helpful, and I ended up with two action items:

  • Create a shared Google Drive folder where we can put assignments and examples of student work (access limited to faculty involved in the group).
  • Organize another meeting for next quarter. People were pleased enough by the meeting to want to meet again.

I don’t think that anyone will make any radical changes to how they teach as a result of the meeting, but I think that several of us came away with the nugget of an idea for a small improvement we could make. It was also very refreshing to have a discussion of teaching techniques—something we professors don’t often get a chance to engage in meaningfully.  Most attempts to foster such discussions are way too broad (like the Academic Senate teaching forums) in an attempt to include everyone, or way too bureaucratic (like the attempts of the administration to push assessing “program learning outcomes”).  Today’s informal discussion seemed to me to be focused enough to be productive, yet broad enough to involve many different courses.  I’m looking forward to doing it again next quarter.

2015 November 16

Considering splitting Applied Electronics course

Filed under: Circuits course — gasstationwithoutpumps @ 18:47
Tags: ,
I’ve been looking at changing my Applied Electronics course next year (not the version I’ll teach in Spring 2016, but for 2106–17).  One change is to move the course from “upper division” (3rd and 4th year) to “lower division” (1st and 2nd year), more accurately reflecting both the prerequisites needed and when students should take the course.
I have a few thoughts on problems in the course as it is:

  • The course is currently 7 units and a very heavy load for students. (A “unit” is supposed to be a median of 3 hours of work, including class and lab time, per week, so a 7-unit course should be 210–230 hours of work total.) The pace is fast and some students have trouble keeping up with the workload.  Moving the course to lower division will get more students earlier in their studies (a good thing), but they will be less able to handle a fast-pace course.
  • The time in lab and grading load are very high—it is difficult for me to keep up with, even spending full time on teaching the course.  This load will make it difficult to transfer the course to a different instructor, which will become necessary when I retire (and I’d like to do the switch to a different instructor gradually, so that I can train my replacement).
  • Not all the students need all the material in the course.  For the students in the biomolecular concentration, the course is required to get them some engineering design experience, in a curriculum that is overloaded with students learning about science others have done, rather than doing engineering. But half as much engineering design as in the current course may be enough for those students.
​My idea was to reduce the pace by splitting the 7-unit course into 2 4-unit courses.​  Students in two of the concentrations (biomolecular and assistive tech: cognitive/perceptual) might take only the first of the courses, while students in the other concentrations (bioelectronics and assistive tech: motor) would take both.

The advantage would be that students would have more time to digest the material and write their reports, as the 6-hour labs would be split over two weeks (three hours per week). They could get prelab homework back before the corresponding lab and have time to analyze data collected before writing up design reports. Lab time for each quarter would be 30 hours, instead of the current 60 hours, making it possible to have more lab sections, increasing the capacity of the course.  Lecture time would increase from the current 35 hours to 70 hours, which would reduce the intensity of lectures and allow more time for students to absorb the material and ask questions.  It would also allow me to drop the physics prerequisite, since I could take extra lectures as needed to cover the missing physics.  (The extra lecture time also explains the extra 1 credit over the current 7-unit course.)

In writing the book, I’ve already rearranged the material so that it could be used as a 2-quarter sequence (since I don’t know any other university in the US that routinely has 7-unit courses), so the curricular redesign needed is minimal (mainly adding more background material and slowing the pace of lectures).  The natural division occurs after Lab 7 (the low-power audio amp), with the second course having the remaining 5 labs (4 amplifier designs and the electrode measurement lab).
I’ve talked with other faculty whose teaching I respect about this possible redesign, and one of them thinks that it would help a lot in getting the students to learn the material—he teaches a course with somewhat overlapping material and is finding that students take a long time to get even the simpler concepts.  The other person I talked to was concerned that the students might not get enough engineering design if they only took the first half (a reasonable concern, since a number of the labs in the first half are more measurement than design labs). He suggested offering one (or both) 4-unit courses in the summer, which I might consider in future, but not this summer—I’ll be burned out after  teaching the intense version of Applied Electronics in the Spring.
Half the course would be a more feasible summer course than the whole thing, as it is already very compressed as a 10-week course, and the 6-week summer schedule would be crazy, so splitting the course does make summer session more feasible.
Some problems I see with the proposed redesign:
  • Needing the lab 2 quarters rather than 1 would increase the conflicts with the circuits course that uses the same lab space.  This should be easily handled by scheduling lab times and only allowing drop-ins when there are no scheduled labs.  The total number of lab hours a week are enough to handle both courses at once, as long as all labs are scheduled in advance.
  • I’m not willing to take on overload, so one of my other courses would have to be dropped from my regular schedule.
  • When I’m on sabbatical, someone will have to be found to teach either this course (no one else in our department is qualified) or the graduate courses I teach (there are some other qualified faculty for that, but whether they’d be willing is another question). I plan to take 1 quarter of sabbatical in each of 2016–17, 2017–18, and 2019–20 (or, at slightly reduced sabbatical salary, every year for the next 5 years).

2015 November 5

Book draft 2015 Nov 5

Filed under: Circuits course — gasstationwithoutpumps @ 22:39
Tags: , , , , ,

I released an updated version of the Applied Electronics for Bioengineers text today.  This draft involved several changes:

  • Added modifier for “resistor” at end of Section 5.1
  • Changed “load resistor” to “bias resistor” in microphone chapter and lab.
  • Fixed microphone schematics to use polarized microphones.
  • Figure 11.2 changed to use only one differential channel on PteroDAQ.
  • Brief explanation of RMS added to Section 3.2
  • Small fixes to Chapters 9–16 and indexing terms added.
  • Index cleaned up.
  • 60Hz FM figure added to Chapter 14
  • Updated power discussion in Sections 0.5, 12.3, 23.1
  • Updated to include Teensy 3.2
  • Major rewrite of Chapter 23 (Class D power amp)

I’m still not finished with the Class D chapter, but I managed to test today an H-bridge circuit using a 9V power supply, which could provide ±9v signals to a loudspeaker (the full 10W that the loudspeaker can take).  I did not actually drive the loudspeaker that far, but I confirmed that the H-bridge was providing the full voltage range for PWM and that I was getting clean signals at the loudspeaker for loudness I was willing to tolerate listening to.

I’m now convinced that an H-bridge design is a simpler approach to teach the students, as well as being more useful for students who go on into the “assistive technology: motor” concentration.  Modifying the H-bridge to use logic-level signals from the comparator but high voltages for the power FETs turned out to be quite simple.  I just added a small nFET and a couple of resistors to make an inverter with a small voltage swing on the output:

Q1 and the resistors R1 and R2 form an inverter for driving the pFET.  Sizing R1 and R2 determines the voltage swing on the pFET gate  (Q2) and how fast the turn on and turn off are.  Of course, when Q3 is on, there is a current through it that is wasted (not delivered to the load), but I was able to keep that down to about 15mA.

Q1 and the resistors R1 and R2 form an inverter for driving the pFET. Sizing R1 and R2 determines the voltage swing on the pFET gate (Q2) and how fast the turn on and turn off are. Of course, when Q3 is on, there is a current through it that is wasted (not delivered to the load), but I was able to keep that down to about 15mA.

2015 October 3

AOI514 nFET I-vs-V

Filed under: Circuits course,Data acquisition — gasstationwithoutpumps @ 21:03
Tags: , , ,

I wasted a lot of time today trying to get a good current-vs-voltage plot for a diode-connected AOI514 nFET, which is the one I plan to have the students use this year.  When I started on it this morning, I thought I was working on my book, but in the end I decided not to include the I-vs-V plot in the book, so most of the time ended up being wasted.

I breadboarded three different test jigs for measuring the current of the nFET:

I started with the simplest test jig, and worked my way up to higher currents.

I started with the simplest test jig, and worked my way up to higher currents.

The simplest test jig just uses the function generator to provide the power for the measurements and a Teensy board running PteroDAQ to make the measurements.

The second jig allows higher voltages on the function generator, hence somewhat higher currents (limited mainly by the 50Ω output impedance of the function generator, but also by the current limits of the function generator).

The most complicated jig uses an external power supply with the function generator controlling the current by changing the gate voltage of an extra nFET.

The 10µF capacitors for removing noise on A10-A11 are almost certainly too big, introducing bias into the measurements, but the 22nF capacitor works well.

By changing test jigs and current-sense resistors, I was able to span a wide range of currents (almost 7 decades).

By changing test jigs and current-sense resistors, I was able to span a wide range of currents (almost 7 decades).

The weird plots at high currents show the effect of temperature changes on the FET characteristics. At 3A the transistor got warm, but as the current dropped it cooled off a little, getting a bit warmer on each 5 second cycle.

At the other end, I had some difficulty measuring currents less than 1µA—current-sense resistors large enough to give sufficiently large voltages would be too high impedance to handle the noise injected by the sampling circuitry on the Teensy 3.1. I also went to 60Hz sampling, to alias out 60Hz interference from capacitive coupling to the breadboard. I still don’t trust the measurements below 500nA.

I decided that the I-vs-V curve here is too messy to put in the book, so instead of working on my book all day, I’ve wasted my time getting nothing more out of it than this blog post and a reminder that even “simple” concepts like I-vs-V plots are not so simple in the real world.

Next Page »

The Rubric Theme. Blog at


Get every new post delivered to your Inbox.

Join 325 other followers

%d bloggers like this: