FREE: Private Consultations

Makes sense to keep your designs out of public view.

John

Why can't it be public? Years ago I suggested that if you designed a curve tracer everyone would want to build it. How about now?

Come to think about it, I also suggested once that you should write a textbook.

I was thinking the other day about an early-warning monitor for failing PC power supplies. If the caps start to go, the output will have increased ripple. Lots of critical systems could benefit from this, and I personally would build a few to meet my own needs.

A small module could plug into the end of a power cable.

A small board could monitor the voltages on the PCI bus.

Reaching test points for the CPU supply is harder. Not sure if the PCI bus could be used to monitor that.

Indeed. But several others have already done it, here's just two...

But that's no fun, that's work :-(

And I said, "a few hours". I'm just looking for some fun "puzzles'. ...Jim Thompson

I'm reluctant to use a system to monitor itself. I'd be inclined to use a battery-powered monitor... just a receptacle/plug in-line with the power cable to the mother-board, with comparators to measure the P-P ripple.

(Although my immediate pondering suggests that battery-powered wouldn't be necessary.) ...Jim Thompson

"Jim Thompson" wrote in message news: snipped-for-privacy@4ax.com...

Then public would still be acceptable? Just providing the opportunity in case the free client wishes it to remain private rather than, for instance, posting it here?

I've got a neat diversion to calculate, though it's not really a two hour problem: Consider a push-pull Royer oscillator, consisting of two transistors and an ideal centertapped choke, supplied from a constant voltage source V with series inductance Ls. Determine all the voltage ratings necessary (peak voltage on Ls, on the CT choke, transistors). Find the RMS output voltage (end-to-end, so that a load could be connected across the choke). Determine the inductance required for a specified ripple in Ls at operating frequency F.

The ratings are all easy, but the last one is more interesting, as it has (as far as I know) a transcendental solution.

Tim

Hmm. Aren't you and the reindeer sort of busy from now through the 25th?

-- Paul Hovnanian mailto: snipped-for-privacy@Hovnanian.com

------------------------------------------------------------------ Life would be so much easier if we could just look at the source code.

On a sunny day (Sun, 11 Dec 2011 21:43:25 -0800) it happened "Paul Hovnanian P.E." wrote in :

He thinks most kids are leftists weenies. LOL

Perhaps like this...

PC

ed

nally

I bus

=A0 =A0 ...Jim Thompson

=A0 =A0| =A0 =A0mens =A0 =A0 |

=A0 | =A0 =A0 et =A0 =A0 =A0|

=A0|

=A0 =A0 =A0 |

Doing aerospace autopilot designs, we had the requirement that with a single component failure, the system alerted you and continued to MEET spec! With second component failure, the system alerted you and nulled out!

Now THAT has design challenges to it.

I used to do some of that stuff... single event upsets, etc ;-)

Here's what I posted as a possibility for Tom's problem, a start, embellish to your personality needs with indicators, etc...

It detects cheap wine? Only you would design something like that! ;-)

Well, it's about time you've priced yourself according to the value of your contribution.

Good Luck! Rich

It doesn't trigger on Thunderbird ;-) ...Jim Thompson

Then you need more gain in the control loop. ;-)

Thanks Jim. That's simpler than what I was thinking.

No, but it does on Ripple

{;-)

Jim

As it does on inductive spikes.

John

OK, I'm working on a project where I am triggering an SCR pair which = drives=20 a highly inductive load, which consists of a transformer with a high = current=20 secondary which may be connected to a variety of loads. It may be = anything=20 from an open circuit to a bolted short, but usually will be a circuit=20 breaker which may trip at any point in its current waveform. The trigger =

circuit should be capable of handling a variety of SCRs from 55 amp = TO-220=20 types such as the Teccor/Littelfuse SK055R, 55A, 1000V (which will be=20 mounted on the trigger board), to external hockey puck thyristors rated = at=20 typically 500A and 1600V.

From experience I know that the initial phase angle must be something = like=20

We have had several designs and the one we now have uses a PIC18F2420 = which=20 senses the zero crossing of the 120 VAC line and fires the gates after a =

specified delay corresponding to an initial phase angle. And it also = senses=20 the continuity of the gates and disallows initiation if one or both are = open=20 (as sometimes happens, causing DC transformer drive and HUGE currents).

The challenges for the new design are: (1) Using 12 VDC for power (losing the 120 VAC) phase reference (2) Using the voltage across the SCRs for initial phase reference (3) Making an efficient and accurate isolated zero crossing detector for =

If you're interested in doing some parts of this design I can send you = what=20 I have so far, or you could provide whatever portion of the design you = can.=20 If it exceeds your time limit for a freebie, we can still maybe strike a =

deal. This is a small part of a new major design and it would be nice to =

offload it to some extent. But I'm planning to finish the schematics = this=20 week and order prototype PCBs by next week, so time is of essence.

Thanks,

Paul