Leaky Op Amp??

[billymc]

I found this dump load controller on the internet a while back and cannot seem to find it again. (Would like to give credit where credit is due) I had to draw this on ms paint as I'm not at the computer where I saved the file. There are actually 3 more 3055s and 1 ohm resistors in the circuit. The problem I'm having is that the output of the op amp only goes down to about 1/2 volt.  As you adjust the 1k pot to a certain value, the voltage suddenly comes up which turns on the 3055s and loads up the battery. That part is working but since the 741 output stays at about 1/2 volt when the 3055s are supposed to be off there is a small constant load and the 3055s are heating up. I am working with a 12 volt system.

My question is, can the output be made to go to zero on this circuit or at least low enough to not turn on the 3055s?

Thanks for any help.

billymc

[ghurd]

Maybe a resistor between them to ground?

I have a feeling the first one is leaking enough to partly turn on the second one.

Just a thought.

G-

[billymc]

Do you mean between the op amp output and the first 3055 base? I thought of that as in a bleed off resistor but didn't try it yet. Kinda thought the transistor itself would act as a load. Will give it a shot though.

Thanks

billymc

[kitno455]

you need a rail-to-rail opamp?

allan

[ghurd]

I mean between the first 3055 and the second 3055.

Not sure if it will help, but it is what I would try.

G-

[Flux]

The 741 is not really suitable for single rail operation. try a 071 or 081 or the cmos amps designed to give rail to rail operation with single supply.

If you want to codge it with the 741 just add a diode between the output of the 741 and the base of the first transistor ( anode ot 741 out) and add a i k resistor from the base of first transistor to ground. The extra diode drop will keep the transistor turned off.

I wouldn't use a 2N3055 as the first transistor of a darlington pair, use a smaller thing like BFX 85 with lower leakage.

Flux

[billymc]

Thanks all.  Several options to try.  I have the diodes and resistors on hand and will probably try flux's fix.

billymc

[Slingshot]

Billy,

Both of the transistors should have a "pulldown" resistor from base to ground.  This ensures that the transistors are off when the op amp's output is low (or as low as it can go).

If you are getting significant load current with the first transitor's base at 0.5 volts, this tells me that you may have a leakage issue with the transistors.  Pulldown resistors would help, but something else may also be awry.

[Slingshot]

BTW, I do not like this design.  Because the op amp's output will never reach the positive rail (how close depends on the current drawn), and there are two diode drops from the op amp output to the dump load, you will probably always have at least 2 volts across the output 3055.  It will never saturate, which means it will run hot.

I'd redesign this circuit to place the load in the collector of the output device, with appropriate consideration of the drive ciruitry.

[Nando]

The basic circuit would be better with a modification

Replace the 2 transistors with a MosFet capable of driving 30 amps in both cases you may need a heat sink if a continuous job has to be done.

Move the load from ground to plus voltage and the Darin of the Mosfet.

Place a 220 ohms in series between the Gate of the Mosfet and the output of the amplifier.

Nando

[Opera House]

If you look at commercial test loads like the ACDC EL-300, most of the heat is dissipated in the transistors.  This may not be what you intended.  Given this design, op amp drive of 10ma X 40 driver gain X 20 output gain you will likely only be able to sink about 8A. I like to use resistors to limit current in sections of the circuit like op amp and driver.  Using a PNP driver and a LM324 will drive the output closer to saturation.  Resistor in the collector would be even better with a small resistor in each emitter.   With infinate gain this circuit is going to sing like crazy when the load turns on.  A little feedback in this ciruit is suggested.  

[dinges]

And shouldn't there be a bit of hysteresis in the opamp? Maybe a resistor from output of opamp to + input of the opamp.

Just an idea I have. may not be necessary for all kinds of dumploads, but some will be happier when not constantly turned on & off at threshold level.

[Opera House]

Wanted to say use the LM339 with an open collector output, the LM324 is just a slightly better conventional op amp.   Using a LM431 amplified adjustable zener would simplify the circuit and eliminate the op amp.

[Nando]

This circuit doe snot need feed back.

It has a "natural" feed back via the battery posts since the amplifier has a very high gain will change states with a few millivolts changes in the battery level.

Nando

[dinges]

Just found the link for you (googled with 'dumpload controller 2n3055 741).

http://groups.yahoo.com/group/HRM_wind_solar_energy/message/11

http://members.rennlist.org/warren/dumpload.jpg

[dinges]

Und 'tis:

http://members.rennlist.org/warren/dumpload.html

[Nando]

Flux fix may not work because the DARLINGTON driver is driving several power transistors in parallel.

The circuits needs to drive at least 2 amps.

The circuit also NEEDS heat sink because the transistors are in linear mode and the output transistors are driving 225 watts = 17 amps and are going to have around 2 volts Vce so no saturation which represents around 34 watts dissipation.

The basic circuit is so basic that is been misused as a Dump ballast controller.

The logical solution would be to change the power devices for Power Mosfets moving the load to the positive side and the Drain of the MosFets.

If you still want to use the 2N3055 transistors, then some changes are due to have transistors saturation.

Get a PNP transistor ( 3 amps or better) Emitter to plus voltage, Collector to bases of 2N3055 then 470 Ohms resistor from ground to bases of 2N3055 .

From plus battery a 1000 ohms resistor to base of PNP, then from base of PNP a 390 Ohms to output of amplifier.

Then interchange the inputs of the amplifier.

This way better saturation of all power 2N3055 .

Still the MOSFET would be a much better solution.

Nando

[kell]

How about using a relay to carry the dump current?

Here's a schematic you can view in a non-proprortional (fixed) font such as courier.  Paste it into windows notepad or something because it doesn't show up correctly on the board here.

    Batt+

     |

     |

     +--------------+------------------------+---,

     |              |                        |   |

     |              |                  relay )   -

     /              /                   coil )   ^ 1N400X

     \              \                        )   |

     /              /    ,--//--,        +---'

     \              \    |          |        |

                   |    |   |\     |        |D

     <----//--------+---|+\    |      |--

                   |        |  >---+-----||  n-channel

     \              +--------|-/           |--   mosfet

     /              |        |/              |S

     |              |                        |

     |               /                      gnd

     gnd           /_  5.6 v zener

                    |

                   gnd

Use the contacts of the relay (not shown) to connect the battery to a dump load.

This way you won't have any  heating of transistors and such.

Note that hysteresis (the feedback resistor to the op-amp pos input) is necessary with this circuit to prevent chattering of the relay.

The dump load will turn on at a certain voltage and hold on until the battery drops to a certain lower voltage, determined by the values of the resistors on the input of the op-amp.  You can look up op-amp hysteresis or similar terms in Google, and you will find the formula to calculate resistor values that will give you the exact voltage window you want.

[kitno455]

lets try an experiment...



  Batt+

    |

    |

    +--------------+------------------------+---,

    |              |                        |   |

    |              |                  relay )   -

    /              /                   coil )   ^ 1N400X

    \              \                        )   |

    /              /    ,--/--,        +---'

    \              \    |          |        |

                   |    |   |\     |        |D

    <----/--------+---|+     \    |      |--

                   |        |  >---+-----||  n-channel

    \              +--------|-/           |--   mosfet

    /              |        |/              |S

    |              |                        |

    |               /                      gnd

    gnd           /_  5.6 v zener

                   |

                  gnd




allan

[kitno455]

arg. scoop, how i hate thee...

allan

[Flux]

On looking at that circuit in detail, as a result of various comments I have to agree that it is pretty terrible.

Although a 2N3055 is nominally rated at 15A it will be virtually impossible to drive it much harder than 6A and even then its gain will be very low. I hadn't realised how hard the circuit was driven and that there were 3 power transistors in parallel.

Nando is right that a BFX85 will not do it. To make it work decently I would have used a BFX85 driving a TIP42, then driving the power transistors. I think you will need 2N3771 to handle the 12A.

It would be far better in the end to use mosfets. The op amp needs changing for a comparitor to switch the fet rapidly and the mosfet needs a driver, as the open collector of most comparitors will not drive the mosfet.

The basic truth is that the simple circuit will do the job, but not for dumping 36A.

Billymc it depends how much load you want to dump, if you need lots of amps then it really isn't up to the job. The relay may be the simplest solution if you don't want to virtually start again with a better circuit.

Flux

[billymc]

I wouldn't mind starting from scratch if a fairly simple solution can be had.  I was shooting for a simple dumpload controller in the 500 watt range that any yard ape with a soldering iron (that would be me!) could build.

Commanda's units are a thing of beauty but out of my league as far as assembly. I want to go to the local electronics supply house and grab what I need.

It would be nice to have a couple of simple diagrams on implementing some of the suggestions above.

Thanks to all, plenty to think about.  Think I'm suffering from brain strain. Been to long since the ac & dc circuit analysis classes of old.

billymc

[kell]

Try this link.  I posted my diagram on a newsgroup where it will show up correctly.

http://groups.google.com/group/sci.electronics.basics/browse_frm/thread/9e35b5aa51ebbfad/3c7e02c77fb65227#3c7e02c77fb65227

I hit preview, then cut and pasted the link into the URL bar in Internet Explorer and it worked.  If it doesn't work for you, you can still get to the newsgroup by going to google, click groups, and find sci.electronics.basics

[billymc]

The link worked.  But I still had to paste the image into notepad to get it to show up correctly. This may very well be just the thing I was after. Will do some checking and find suitable mosfet and resistor value.

Thanks

billymc

[commanda]

Oatley electronics have a nice kit.

http://www.oatleyelectronics.com/kits/k220.html

Amanda

[kell]

Just for information purposes, when you read those newsgroup postings, most reader programs (including Google groups) have a button you can click that says "view in fixed font" or "non-proportional font."

The electronics newsgroups have some very knowledgeable and helpful people on them if you have any questions about the feedback resistor calculations.  

As for the mosfet, the specs you will want to look at are Rds (drain-source resistance) and the power rating.  Calculate the relay coil current first, then to find the amount of heat (power, in watts) the mosfet will have to dissipate, square the current and multiply by Rds.

It's likely that a small bipolar npn would also work.  Look at op-amp output current and transistor gain.