Simple dump control circuit

[South Easter]

Hi all. I have assembled and used ghurds dump controller.  (Thanks for all the help ghurd!).  Since the ZM component is hard to find here in Africa, and also ghurd tells me they have stopped making them, I wanted to have a design of a similar circuit using easy to get components.

I don't know much about electronics, but have tried to design something as follows.  I dont really understand why the two FETs are needed as shown, but have copied that from ghurd.

Any comments or suggestions would be much appreciated!  Thanks in anticipation.





--

[Opera House]

Indeed....If parts are hard to come by, why add extra ones.  

When the op-amp output is high the first fet is driven low which dries the next fet high.  Why not just drive the external power fet with the op-amp directly.  There is no speed issue with the 10uf on the sample voltage.

I just don't like these op-amp circuits that are inherently unstable.  Add some positive feedback from 1 to 3 like a 1 meg. Those without instruments just don't see how crazy the transition gets.

Why choose a 5V power supply rail and limit yourself to hard to get fets.  Put a 1K resistor from 2 to 3 on regulator and another 1K to ground.  This will make the supply voltage 10V and expand the number of fets you can use.

Add a resistor to the bottom leg of the control pot so it doesn't adjust down to zero.  figure a resistor ratio that will give you + and - 20% of the desired voltage.

[ghurd]

I still have more than 2,000 of the 'ZM's,

just so people know the kit should be available for quite a while.

The ZM part bypassed words like "inherently unstable" and "crazy transition".

However it operates at logic levels.

I used the 2 inexpensive fets to get around using multiple, expensive, hard to get logic level power fets.  

Plus, I think if there were issues with ugly switching in the first part of the circuit, the double fets should clean it up before the output to the power fets.  The only person who O'scoped it said the output is fast and sharp.

The output of my circuit will drive about any N fet you can find.

It was simply the most cost effective way to make a complete circuit.

And I have a personal dislike for anything with more than 3 pins. ;-)

G-

[South Easter]

Thanks for the feedback!!   Here is a modified version which is intended to drive the fets harder, and give a little positive feedback.

[Opera House]

I'm actually quit fond of three terminal amplified zeners like the LM431 which can also be used in this application.  I keep a stock of a couple hundred for these simple applications.   These can easily drive a LED or relay by themselves.  Of course in a dump application, adding an extra PNP transistor will allow injecting some positive feedback  for a 1/4 volt or more hysteresis.  Still don't like slam bam regulators.

[South Easter]

Here is a slightly altered version. Its skips out the small Nfet by switching inputs to the comparator - thanks again for your comments.  

Also, this way the output of the comparator is floating, and so I assume draws minimal current when the dump is off.  I have chosen the resistors large in order to minimize quiescent power use - I hope not too large?

[commanda]

If this is for a 12 volt system, the 7812 regulator won't work.

If you're really using an LM193, the output pin is open collector. So you can power the circuit with a 5 volt regulator, and connect the P type fet and 100K pull-up straight to 12 volts. (Don't do this with just any op-amp with a totem-pole output stage).

Amanda

[South Easter]

Thanks for your comments Amanda.  Its intended for 24V, but I would like it usable for 12V by changing a resistor or two.

I used the 12V supply to make sure the Pfet turned on hard, but on reflection I think 5V at the gate will do fine.

Here is the latest version:

[Opera House]

I'd drop the value of R6 TO 10K.  1M feedback presents a significant load on the output of the opamp.

[Opera House]

I should explain. The pull down from the feedback is about enough to turn on a 5V fet in a 24V system.  Feedback changes quite a bit from a known 0-5V to 0-28V.  The gate voltage may be high enough to kill the fet at higher supply voltages.

[elt]

On the way to a more-featured dump controller, I came up with this simple

12/24 volt circuit

http://www.fieldlines.com/comments/2007/8/26/173613/622/29#29

With slow switching speeds, the op amp can drive FETs directly.

ScottAI points out that when you go for minimal part count, different solutions tend to looks alike and he gives links to some similar low-part-count circuits in the reply.

- Ed.

[South Easter]

Thanks you Opera.  I have tried... but can't quite follow your explanation.  If you have time, please explain a little more.

TIA

--

[Opera House]

The op-ap is open collector, so the off state voltage on the fet is formed by 100K/1M voltage divider on a 24V system would be about 2.4V on the gate.  It would not be 0V as one would like.  

At turn on that would be full supply voltage on the gate.  Usually it is limited to about 18V.  Boosting the 5V regulator with resistors to 10V would be a solution.  this would also protect the op amp output.

[Opera House]

Notice that you don't use any bypass caps on the regulator. These have a tendency to oscillate without them.  I even found an article a while back on how to make a ham transmitter out of one.  Suggest that it also be buffered with an input resistor to absorb any spikes.  It is pretty easy to over voltage these with an input like a wind generator.

[commanda]

100nF is a typical value to use.

Amanda

[(unknown)]

Interested in a kit...any available?price?shipping to Canada?

Thanks eh

[ghurd]

I can be contacted through ghurd1 (at) yahoo (dot) com.  

Put "Controller" in the subject.

G-