Transistor booster design

I don't understand the C2-Q1 thing. The only way this could work is if the transistor base were zenered every cycle, and that's not healthy longterm. Is something missing?

And what is the graph?

Strange.

John

C2-Q1 and C3-Q2 oscillate back and forth. It works, I am using 2 CR2032 to run at 15V (voltage regulation omitted for simplicity). The only problem is drawing 7 to 8mA from the batteries (which are rated

4mA). I am thinking about adding heat sinks to the batteries to cool them down. Is copper the best heat sink?

The output voltage vs. time in second. Approx. 10ms to get stablized.

What is?

Are the batteries getting hot? Seems unlikely with those numbers.

Running DC through a capacitor is strange.

John

[snip]

You've heard of modern math? This is modern electronics ;-)

...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

Whenever I see a capacitor in series with a diode, all my alarms go off. And it's surprising how often I seem to see it.

Interesting component, a capacitor that only conducts in one direction.

John

...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

The spec says 4mA continuous, 20mA pulse for 15 seconds. We need 8mA for several seconds, so just want to be safe.

Don't forget the inductor kicks. The voltage oscillates between +/-

60V, it's not pure DC.

Spice actually shows hundred of volts, but it does not account for breakdown limits. Spice is fakey but seldom lie. The web server is running spice 3f5.

The base of Q1 needs average positive DC current flowing into it to operate. It can't get it through C2. If you used a huge cap for C2, Spice will tell you that it will "work" for a while, but it will quit as C2 eventually charges up.

DC-wise, the base of Q1 is floating. Spice is no match for reality.

How big is C2 in the simulation? How big is R2?

John

[snip] [snip]

Not to mention it probably won't start up in the first place ;-)

...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

But I have tested and measured with real parts.

C2 is 470pF, R2 is 10K.

Yes, it does, but only for narrow range of components for different types of transistors. I have simulated with 2N3904, BC847, 2N2222, BD139, etc. I have actually tested with BC847B (happen to have some in stock).

Even if it does not start by itself, I can kick start it with an micro. The micro start and stop the circuit and monitor the output as well. The emitters are connected to a 2N7002 in that case.

And ac-coupled astables like this tend to have at least one latchup state. This one sure does.

I still think it can't work for long. I never would have Spiced something like this in the first place. The regulated flyback inverter I posted to abse worked, closed-loop, the first time I turned it on, without benefit of simulation. Spice, like PowerPoint, can inhibit thinking, as it has in this case.

John

Yes, if the components are not chosen correctly, it will saturate the transistors. But all I need to do is to jump start it with the micro, which is monitoring the output any way.

Getting it to run is not the problem. Getting high efficiency is the problem. I can of course switch it directly with the micro, but it would kill the micro processing loop. The transistors are also blocking the dangerously high surge from the inductor. Think of them as ringing buffers. They don't need to be working by themselves.

It is currently about 50% efficiency. I need to get it higher without killing the batteries.

Linnix, You came here for a peer review. You asked us all what we thought. We told you. You became argumentative, "Yes it does..."

The circuit is crap. It disobeys the fundamental rule for such oscillators... remove all capacitors, then the circuit should bias up in a condition where there is definite loop gain. Your circuit simply saturates.

You can take your post to "basics", and they'll pat you on the back... a blinking LED is "success" over there... and, unfortunately "success" here as well sometimes :-(

"Narrow range of components" is a hacker's "design" criteria, but not a true engineer's.

...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

You might not believe the simulations, but I have tested it on actual components.

It only saturates if you drive it out of range with unreasonable values. With gains of 800 (BC847C), it doesn't take much to start the oscillation.

I consider 50% tolerance acceptable value. In simulations, it works for many times this tolerance. Unless you believe simulation is total crap.

Go do it. I hope you survive.

...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

Oops, sorry, there is a 10K resistor to the base of Q1 missing in the drawing. It is in the simulation and actual prototype. It works on range of VCC (3.5 to 5.5) and RLoad (12K to 30K). For this transistor (correction: gain of 500 typical, max of 800), it won't work over 6V under certain conditional. The micro (AVR) can't take over 5.5V anyway.

Others (2N2222, BD139) works at higher voltage.

Bingo.

John

Well that can't happen if V+/ESR of L1 is far in excess of Hfe*V+/R1. And both transistor bases have pull-ups to V+ on the base, so no problem with a DC turn-on path. But the collector of Q1 is clamped to Vbe of Q2 at switchover and the base takes a big wallop of turn-on current, so that can't be good, and there is the potential for zenering the base junctions. There is no clear cut mechanism for output regulation either. If he has logic, he should have gone with an inverter relaxation oscillator, or a dedicated boost regulator IC, makes no sense to use discretes here.

It can be limited with a small C3.

I did have clamping diodes on both bases in simulation, but not enough space in layout.

There is, just not shown here.

It will only be on for 30 seconds at a time. We want zero current the rest of the time, by disconnecting the ground (or common emitters) point.

This circuit should cost less than $1 to make and trash in a few months. Hopefully, we will be able to change to the IC version.