Re: Very low power dynamo (alternator actually).

age

ing

ts

nto

o to

=A0 =A0 =A0 =A0...Jim Thompson

ry

ible

Its very doable. I had no idea those old dynamos put out so much.

NT

age

ing

ts

nto

o to

=A0 =A0 =A0 =A0...Jim Thompson

ry

ible

t of

f my

I'm not seeing "all sorts of entangled loadings"? What kind of response were you looking for?

I want the ADC driver to swing from -0.75 to +0.75 diferential output (into an LTC2242-12 ADC, which alone fills the next sheet of the schematic) as the signal swings 0 to +3. The ADA4950 isn't a high input impedance, and its VCM will be 1.25 volts. And I need to keep the filter properly terminated.

John

essage

dering

as

knots

y into

s so to

was

ttery

eak

ossible

most of

r.

d of my

OIC- well isn't that just constant-k design or something...

message

wondering

into

to

possible

of

my

It would be cool if I could set up performance targets and let Spice do the iterations for me. One of my guys does that sort or brute-force circuit design - lets it run for hours - but he uses Octave and some optimization program around that. He'll design filters by just pushing poles and zeroes around iteratively.

I'm still constrained by wanting to use parts we have in stock, or stock parts plus a minimum of new purchases.

Seems like I spend a lot of time trying to use stock parts, and use parts aleady present on the schematic somewhere else. I did write a program that finds the best in-stock resistors to hit a given ratio and Thevenin impedance.

ftp://jjlarkin.lmi.net/Rugrat.jpg

but that's just a hint of what could be done. I want to tell Spice "here's a list of parts I have in stock, and here's the output I want, wake me up when you're done. Go to Digikey if you have to." I suppose I could accomplish the same with a smart intern.

John

Having noticed quite a few apps on the LT d/l page, I wondered what switchercad does?

Switchercad is - I think - just another name for LT Spice.

John

I remember in my old microwave filter class that we had just started using Touchstone to design microstip filters. A couple of the guys tried using the optimiser, and would set it up late and night, and come in the next morning to see the results. They were doing a band pass filter, but forgot to put in the 'stop' bands, just the pass band.

The next morning, they came in to find it had optimised it to a single transmission line... ;-)

Charlie

On an intellaSys (*) project some village idiot manager hired some clown peddling optimization software, and asked me to give the clown one of my cells to optimize.

So I gave him my internationally famous band-gap that is flat as a pancake and always starts up at very low VDD, any temperature, any process corner.

The "optimizer" completely killed it and couldn't ever get it to work again ;-)

So I was fired ;-)

(*) The rumor mill says they have crashed and burned... just another investment black hole. ...Jim Thompson

Most elegant!

John

message

wondering

into

to

possible

of

my

That's more or less what I did with my electromagnetic simulator--a fast clusterized FDTD engine with a big script wrapped round it to do the multidimensional optimization, among lots of other things.

Cheers

Phil Hobbs

What a weird, foul old hen you are.

John

flattering will get you no where John!

Jamie

He's still obsessed with excrement, too. He mentions it a lot more than he mentions electronics. Yuk.

John

It's unfortunately been over 20 years since I've touched power conversion, so I can't contribute much of a practical basis to this discussion, but here are my theoretical contributions.

Bike dynamos are designed with "excess" inductance, such that as the frequency and internal voltage rises with speed, the unit current limits at approx 0.5 A. A number of people use series capacitors to cancel part of this inductance in order to get more current out of their dynamos. If you wanted to extract the maximum possible DC power out of one of these, I would consider trying the following:

- Since the AC voltage output is low at low speeds, synchronous rectification should be more efficient than using a diode bridge.

- Due to the high series inductance, I think it makes sense to pull as close to sinusoidal current from the dynamo, so something resembling a PFC front end might be useful.

- Tracking the frequency of the incoming AC and using that in the control seems to be an obvious path to follow.

- Frequency dependent series compensation might be useful, although if the PFC front end can have an adjustable leading power factor, that might do it to.

Being insulted by such a creepy weirdo somehow seems rather trivial compared with being isulted by anyone of any conseqence.