Transistor booster design

em

on.

How would it fail? I can't see it burning the base via a 470pF capacitor. It would more likely burn the collector, but that's the same in any other oscillator.

To prototype until we get the right circuit into a chip. We will be able to have any number of clamping diodes in it. Transistors are pennies, they are cheaper than fuses.

[snip]

You are such a novice. It looks like you zener the B-E junction of Q1 every cycle. Any idea about how that gradually degrades a transistor into a non-transistor?

Hopefully whoever designs the chip will have more sense ;-)

...Jim Thompson

--
|  James E.Thompson, P.E.                           |    mens     |
|  Analog Innovations, Inc.                         |     et      |
|  Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
|  Phoenix, Arizona            Voice:(480)460-2350  |             |
|  E-mail Address at Website     Fax:(480)460-2142  |  Brass Rat  |
|       http://www.analog-innovations.com           |    1962     |
             
         America: Land of the Free, Because of the Brave

How about this?

It's a variation on the inverting switcher I posted to abse recently.

It starts up at about 50% duty cycle and, once the zener starts to conduct, backs off as needed to keep the output regulated. It should be pretty efficient.

There are other ways to do the feedback, but this isn't bad. A voltage divider into a common-base PNP, with the base on +5, would work too.

The gate resistor is optional. Use a 3-cent 2N7002 for low power levels. Run a uP port pin to the schmitt input for software shutdown.

I like this circuit because it's simple and is very quantitative.

John

Oh, add a bubble.

John

What? Why Jim, you want _another_ blinking LED circuit??! (Somebody need to supply the correct emoticon for that whopper!)

Ok, here's a groaner for you: narrow range of components, violates rules, works only at a small voltage range.

9-12V Vcc ---[1K]---+------+ | | | \\e [330uF] |--- 2N2222 | /c | | | [LED] | | | [1K] | | Gnd ----------+-------+

Base - no connection, Vcc somwehere between 9-12 volts. Damn thing blinks.

Ed

And the sad part is you don't know WHY it works ;-)

...Jim Thompson

--
|  James E.Thompson, P.E.                           |    mens     |
|  Analog Innovations, Inc.                         |     et      |
|  Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
|  Phoenix, Arizona            Voice:(480)460-2350  |             |
|  E-mail Address at Website     Fax:(480)460-2142  |  Brass Rat  |
|       http://www.analog-innovations.com           |    1962     |
             
         America: Land of the Free, Because of the Brave

soon.

OK, I will check out how it degrade. I have seen this circuit running for a long time. They do fail, but much longer than what I need.

I only need to run it for 30 seconds at a time (perhaps a few hundred times life cycles). As long as it does not short circuit the control line, it should be fine. The board can be replaced regularly, just like the batteries. The board is actually cheaper than the batteries. It's not much more difficult to replace than the batteries. Actually, we have also considered soldering in the batteries. The customer can just replace the board when the batteries run out.

This board is primary an extension of the battery, and cost is more important than reliabilty. Alternatively, we can just provide 15 volt batteries, rather than trying to use the button cells.

To me the sad part is that it's no fun for you. :-(

Ed

FUN is understanding the whys and the why nots.

I'm having fun all the time. Why do you think I keep doing chip designs though I'm past retirement age. Because it's FUN solving problems that other people can't... it's an ego trip ;-)

...Jim Thompson

--
|  James E.Thompson, P.E.                           |    mens     |
|  Analog Innovations, Inc.                         |     et      |
|  Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
|  Phoenix, Arizona            Voice:(480)460-2350  |             |
|  E-mail Address at Website     Fax:(480)460-2142  |  Brass Rat  |
|       http://www.analog-innovations.com           |    1962     |
             
         America: Land of the Free, Because of the Brave

Jim Thompson snipped-for-privacy@My-Web-Site.com posted to sci.electronics.design:

Damn straight. It is one of the biggest reasons i like my job. And it is not only just doing things that they cannot; it is doing it better of faster or usually both.

Yes, it's much better than linnix' proposed multivibrator circuit, because the oscillator is well-defined and has nice sharp output transitions. I haven't used the classic multivibrator circuit since my first year or two of working with transistors, in the early 60s.

John's voltage-boosting circuit is a nice replacement.

+Vs in ----#####---+---|>|---+---- HV out | | ,--------/\\/\\----|>|------ | --------' +---/\\/\\----, | | +5 | | | _| | |--' +--|__ >o---+---/\\/\\---|| _|_ | \\ |--, 2n7000 --- gnd schmitt | | tinylogic gnd gnd

Schmitt-trigger ICs do draw supply current while in the linear region, especially if operated from higher supply voltages, like 5V. But the efficiency is still likely quite good. John, do you have any measurements for us?

Indeed!

...Jim Thompson

--
|  James E.Thompson, P.E.                           |    mens     |
|  Analog Innovations, Inc.                         |     et      |
|  Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
|  Phoenix, Arizona            Voice:(480)460-2350  |             |
|  E-mail Address at Website     Fax:(480)460-2142  |  Brass Rat  |
|       http://www.analog-innovations.com           |    1962     |
             
         America: Land of the Free, Because of the Brave

Now that's REAL engineering :-(

...Jim Thompson

-- | James E.Thompson, P.E. | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona Voice:(480)460-2350 | | | E-mail Address at Website Fax:(480)460-2142 | Brass Rat | |

| 1962 | America: Land of the Free, Because of the Brave

Note the added "X"

I assume X is a zener.

I have done the same sort of thing paralleling the sections of a full sized gate. With a few more parts, you can add a crude current limiting. Basically, it is feeding the HV output back to control the off time. As the HV increases, the off time decreases. The on time is nearly fixed. The current in the inductor always gets enough time to stop between on times.

----/\\/\\---HV ! --/\\/\\--+--!o----------+-!>o---------Gate ! --- --- ! GND

Do you remember the 1-transistor unishot circuit? I always thought that was cute. The early model airplane r/c transmitters used a string of unishots to make the transmit pulse sequence.

No, I just made this one up. The inverting version that I posted to abse did +12 to -12, 0.5 amps out, 350 KHz, at 92% if I used a biggish inductor, more like 85% with a surfmount 10 mm^2 one at 450K.

John

This would be fine for discrete, but I am trying to limit the process to bipolar junction only (for the control circuit). My main energy lost is through the clamping diodes, if only I can reroute and recycle it somehow.

The new version is at linnix dot com slash power2

An auto query engine (66.249.67.129) is hitting my spice server every few minutes on linnix dot com slash power

For low output voltages, the forward conduction of the zener could prevent startup. Considering this John's PNP transistor version would be more widely useful.

+Vs in ----#####---+---|>|---+---- HV out | | ,--------------c e------ | --/\\/\\--' +---/\\/\\----, b pnp | | +5 | +5 | | _| | |--' +--|__ >o---+---/\\/\\---|| _|_ | \\ |--, 2n7000 --- gnd schmitt | | tinylogic gnd gnd

Good call, but as long as Vs > +3 or so, the zener won't forward conduct.

For the PNP version, I'd add a resistor from the emitter to ground, to form a voltage divider, which would then use the +5 as the reference for the HV output.

John

Nice.

+Vs in ----#####---+------|>|---+---- HV out | R2 | ,--------------c e------ | --+--/\\/\\--' | R3 b pnp | | R1 +---/\\/\\----, | | '--/\\/\\--, | +5 | +5 | | | _| | |--' gnd +--|__ >o---+---/\\/\\---|| 2n7000 _|_ | \\ |--, --- gnd schmitt | Vo = 5.6 (1 + R2/R1) | tinylogic gnd starting estimation gnd try to use R3 >> R2