Dealing with for-profit competition

In The New Yorker gets one right, “Dean Dad” praises a short article in The New Yorker about for-profit colleges;

A tip o’the cap to James Suroweicki, at The New Yorker, for encapsulating the issues around for-profit colleges clearly and well in a single page.  The piece is well worth the couple of minutes it takes, not least because Suroweicki neatly dispatches a couple of widely held, but false, assumptions.

…

The way to get the best outcome all around isn’t to ban them or to try to pass lawyer-proof regulations.  It’s to outcompete them  Flood the zone with well-funded public colleges with the staffing, the facilities, and yes, the marketing, to compete.  Force the for-profits to compete on quality.  Frankly, if they can prove they do a better job with students, I have no theological objection to them.  But the experience of the last ten years suggests that if they can only compete on quality, they’ll shrink to a much less threatening size, and students will be better off.  

For-profits met a need.  The way to beat them is to meet that need better.  Austerity in the public sector cedes the field to people with other agendas.  Beef up the publics, and the need that fed the for-profits in the first place will fade away.  They can’t lawyer their way out of that.

Suroweicki’s articleThe Rise and Fall of For-Profit Colleges has the same message, plus a bit more.  The article ends with

But if we really want more people to go to college we should put more money into community colleges and public universities, which have been starved of funding in recent years. We should also rethink our assumption that college is always the right answer, regardless of cost. Politicians love to invoke education as the solution to our economic ills. But they’re often papering over the fact that our economy just isn’t creating enough good jobs for ordinary Americans. The notion that college will transform your job prospects is, in many cases, an illusion, and for a while for-profit schools turned it into a very lucrative one.

The business model for the for-profit colleges has been to get students to take out as much debt as they can, give all the money to the college (who then transfer it to a handful of executives and investors), and deliver little or nothing useful in return, leaving the students with debts that they can’t discharge.  This was obviously a socially undesirable outcome, but legislators have been doing all they can to get rid of funding for public colleges and force them to follow the same model.  I really don’t understand politicians—do they really want the sort of world that they are building?

Futuristic Lights at MAH GLOW Festival

Futuristic Lights will have a booth at the Museum of Art and History GLOW Festival this Friday night (not Saturday, that’s Fire night):

For anyone local to Santa Cruz, CA this weekend on the 16th, Goon, Robin, Haste and Zohar will be running a Futuristic Lights booth at GLOW! If you’ve ever wanted a light show in person from any of these sponsors, now is the time to get one :-) Discounted Glove Sets + Meta Mode Kinetics will be available, so be sure to swing by and snag a set! GLOW is a digital art and fire festival featuring some of the best artists in the Santa Cruz County area. Feel free to stop by and say hi.

GLOW: A Festival of Fire & Light

Oct 16 – Oct 17 · Santa Cruz Museum of Art & History · Santa Cruz, CA

Source: Futuristic Lights

If you’ve been curious about what the Kinetics can do, the sponsored glovers at the booth should be able to give you an idea of the range of possibilities.

Mature students

In a comment on his post Growing evidence that lectures disadvantage underprivileged students, Mark Guzdial wrote,

I realize that adult learning is difficult to fit into our culture and work lives, but one could imagine a scenario where it might fit. Lifespans are much longer today. There is enough time for more than one career. Maybe we might work until (say) 45, then take a 3 year sabbatical to re-train, then launch into a second career into one’s 70’s or later. What might be learned in “second college”? How would college be different with more mature learners? What are the inherent limitations of having much older learners, and what are the inherent advantages of having learners who have 20+ years of real world experience?

Our grad program has had several “re-entry students” in their 40s and 50s who came back to college to get a PhD.  Some of them had been in industry doing computer programming, VLSI design, or engineering management for decades before getting bored with it and wanting to do something that used their skills more productively. Many of these students have done very well, both in the degree program and in their subsequent careers.  (One of the younger ones, who was only 41 or 42 when getting the PhD, is now a full professor, for example.)

I don’t see many limitations to having much older learners—there may be fewer all-night study sessions, but there will also be less need, because there will be less procrastination about deadlines.  Community colleges have been accepting older adults for decades (since the big growth of community colleges in the 1960s), and have had a lot of experience with them.  What I’ve heard is that the mature adult learners tend to be much more consistent and reliable than teens and barely post-teens who make up the undergrad population, but that many of them have lives outside of college, and can’t do more than one course at a time.

I don’t think that I’ll go back to grad school when I retire, but I am likely to take community college courses on subjects that my education is weak on (most likely hands-on skills like welding or art classes).

Heathkit moves to Santa Cruz

According to a Hackaday post, Heathkit: Live, Die, Repeat,

This morning, the president of Heathkit sent a message to the ‘Heathkit Insiders’ email group explaining the goings on and new happenings:

We’ve designed and developed a wide range of entirely new kit products. We authored the manuals for these kits, complete with the beautiful line art you rely on, preserving and respecting our iconic historic Heathkit style. We developed many new inventions and filed patents on them. We relocated Heathkit, and set up a factory, and a warehouse, and offices, in Santa Cruz, California, near Silicon Valley. We built the back office infrastructure, vendor and supply chain relationships, systems, procedures, operations methods, and well-thought-out corporate structure that a manufacturing company needs to support its customers, to allow us to scale instantly the day we resume major kit sales. All this effort enables us to introduce a fleet of new kits and helps ensure Heathkit can grow, prosper, and continue to bring you great new products for a very long time.

If this news is real, it is very exciting. Heathkit was one of my favorite companies as a child (see Thanks, Dad), and I have fond memories of building some of their kits.  I can’t think of a better place for them to be than here in Santa Cruz.

Unfortunately, the kit that they seem to basing their hopes on, the Explorer Jr AM radio kit, is rather overpriced at $150.  It will only sell to nostalgia fans, which is probably not a big enough base for reviving the company.  What they need to do is to make kits in the same price range as Velleman, but with Heathkit-quality manuals.

Crawlspace ventilation—better low-voltage handling

In Crawlspace ventilation again, I mentioned the problem that I was having with the fans whining if the power supply could not provide enough power:

With the 8Ω series resistance, the 4-fan load had a hard time getting started, and whined a bit before starting. Once one of the fans got up to speed and reduced its current draw, the others quickly came up to speed also. The regulated voltage was only 4.9V, and fluctuated a bit, rather than the constant 5.02V with smaller loads or less series resistance between the 9V power supply and the regulator.

With the 10Ω series resistance, the 4-fan load could not start at all, but just whined. The output of the regulator was only 2.2V while the fans were stalled. If a 4th fan was added while 3 fans were already running, the fans ran, but with a 4.11V regulator output. (The three fans had already reduced the regulator voltage to 4.99V).

…

I may have a little trouble with the fans each morning as the power comes up gradually—they may have trouble starting if the panel is not yet putting out enough power. If that turns out to be a problem, I may need to add some circuitry to detect low voltage on the regulator and turn off one or more fans.

The fans I’m using have brushless motors, which means that they include electronics to set the speed of the fan.  Based on the electrical noise I see on the power line, they appear to have a constant frequency independent of whether the fan is spinning or not. This simple design for the electronics is cheap, but not as good as a design that detects the speed of the motor and optimizes the phase and frequency of the rotating magnetic field to maximize torque or efficiency. The whining at low power when the fan is not turning seems to be at the frequency at which the fan is expected to spin with a 5V supply, but I’ve not used a microphone to check.

I tried two different approaches to handling the low-voltage problem:a power-on-reset chip and a simpler FET circuit of my own design.

The first design was to use a power-on-reset chip and an nFET to turn off the motors when the voltage was too low.  This was inspired by the standard Miller engine used in solar motor toys to harvest energy from low-current solar cells, store it in a capacitor, and discharge the capacitor through the motor when there was enough charge. This did not work well, as the voltage from the regulator quickly got up to 5V when there was no load, but dropped almost immediately when the fan was turned on, causing the reset chip to turn the fan off again. The chip I was using, MCP100-450DI/TO has only 50mV of hysteresis, so it would turn off again almost as soon as it turned on, and then wait for 0.3 s before turning on again.  The 50mV hysteresis meant that I’d need a large storage capacitor to get the fans past their initial high-current startup—I estimated around 15F, which would be an expensive supercapacitor.  Without a huge capacitor,  this circuit resulted in pulsed whining when the power supply was not capable of delivering the full 5V, which was even more annoying than the steady whine.

The second design I tried was just using the exponential turn-on of an nFET to provide full current when the voltage was high enough, but limit the current to very low levels when the voltage was insufficient:

The large capacitor on the gate keeps the transistor on for a while even after the voltage starts dropping.

One essential part of the circuit was the large electrolytic capacitor on the gate voltage.  Without it, the voltage on the gate would rise to a point where the motors were struggling to start and stay there (when the power supply was provided with too low an input voltage). With the capacitor in place, the fans would continue to get current even as the voltage dropped a bit, giving them enough time to spin up and reduce their current load.  Eventually the gate voltage would drop enough to start pinching off the current and the fans would hesitate and need to be spun up again. The result was that the fans would start even with a large series resistance between the 9.26V power supply and the input to the regulator (16Ω and sometimes even 32Ω), but not run smoothly unless there was enough power being delivered to the power supply for their continued running.

A smaller capacitor (220µF) worked also, but the fluctuation in speed happened more rapidly.  Much smaller (33µF) did not filter the power-supply fluctuation enough to hold the fans on long enough to start up cleanly.

Adding capacitance across the 5V terminals decreased the electrical noise, but did not seem to change the behavior of the circuit.

The flyback diode (the BAW62) is probably not necessary with these fans, since the brushless-motor controller built into the fans undoubtedly has its own flyback diodes.  I saw no evidence of inductive spikes when turning off the nFET.

So far I’ve only breadboarded the nFET control circuit, but I’ll probably solder it up later this week, when I get some time.