Which opamp would be best for this ?

[Edwin2]

Hello all,

I remember, there was some problem when using opamps with their inputs near
the supplyvoltage.

I now have a cirquit, I hope I can post that too, where the inputs are at +supply - 100mV.

Could I use any OP, or is there something to consider ?
The OPs should not be too expensive, bcs I need a lot.

Kind regards,

Edwin

 

[ericgibbs]

Hello all,

I remember, there was some problem when using opamps with their inputs near
the supplyvoltage.

I now have a cirquit, I hope I can post that too, where the inputs are at +supply - 100mV.

Could I use any OP, or is there something to consider ?
The OPs should not be too expensive, bcs I need a lot.

Kind regards,

Edwin

Hi,
Why dont you adjust the values of the components in order to bring the inputs within the operating range of standard OPA's.?

 

[Edwin2]

Hi Eric, thanks for the very quick reply.

It is used to charge a second battery in a car, and when I go lower, it will not get charged.
Now I should get the 2nd battery up to some 200...300 mV below the main battery.

Edwin

 

[ericgibbs]

Hi Eric, thanks for the very quick reply.

It is used to charge a second battery in a car, and when I go lower, it will not get charged.
Now I should get the 2nd battery up to some 200...300 mV below the main battery.

Edwin

hi,
Most OPA's as may already know dont make 'good' comparators.

A comparator like the LM393 etc, will not work reliably at Vin > Vsup -1V.

 

[Edwin2]

Well, I might have to think of something else then :-(

Thanks anyway Eric,

regards,

Edwin

 

[Willbe]

There are opamps that work "rail-to-rail" rather than within two diode drops of V+ and V-. If you can find an online "selection guide" they will probably show these types.

 

[ericgibbs]

There are opamps that work "rail-to-rail" rather than within two diode drops of V+ and V-. If you can find an online "selection guide" they will probably show these types.

hi willbe,
IIRC most of the rail/rail amps are about 5V supplies.

I would use a power schottky to isolate the two batteries, a 400mV drop should be acceptable.

 

[Oblong]

If you are looking for RR in and out OPA's the TLV2371 is not that pricey.

http://focus.ti.com/docs/prod/folders/print/tlv2371.html

National has a good selection guide here.

http://www.national.com/analog/amplifiers

Possibly for your application you could use a TLV431 low voltage,or 431 2.5V ref as a comparator. Mind you I'm not 100% sure what it is your trying to do.But if your looking for cheap you aren't going to do much better then an LM431SCCMFX 20 cents small quantity.

http://canada.newark.com/82C8197/semiconductors-integrated-circuits/product.us0?sku=fairchild-semiconductor-lm431sccmfx

 

[Edwin2]

Thanks for the replys.

The cirquit should prevent `short-cirquits` when the 2nd battery is completely flat.
Any relay would burn away, with, say, 200 Amps, when you connect a flat battery
to a charged one.
This cirquit is meant to keep the current to a max of 5 or 6 Amp.

Regards,

Edwin

 

[MrAl]

Hi Edwin,

There is a much simpler solution if you can avoid ground referencing the device
you intend to power. If you rearrange the circuit you will be connecting the
resistor in series with the ground leg of the device to be powered rather than
in series with its plus power supply pin. Doing this eliminates the rail to rail
input op amp requirement, and a standard LM358 will work in many cases.

So the million dollar question is...
Can your device to be powered stand to have a small resistor in its ground
lead rather than its supply lead? Measurements can still be made with
a multimeter across the device, it's just that one leg will no longer be
connected to ground.

 

[audioguru]

You have the drain and source of the P-channel Mosfet connected upside-down.
The source is supposed to connect to the resistor and the drain is the output.

Many FET-input opamps like the TL071 work fine when their input is at the positive supply voltage because their FETs are P-channel.

 

[Edwin2]

I know, re D/S, in real the FET will be connected the right way.
It is only the wrong symbol I used in the diagram. Sorry, should have told in the first post.

No, it is not possible to lift the 2nd battery from ground, I already thought about it.
I think I etch 1 PCB, and try the circuit with different opamps, starting with the mentioned TL 071.

There is a much easier circuit with only 2 transistors, but that has a drop of at least 1,5 Volt.
That is why I thought of this one, with the opamp.

Thanks !

Regards,

Edwin

 

[ericgibbs]

I know, re D/S, in real the FET will be connected the right way.
It is only the wrong symbol I used in the diagram. Sorry, should have told in the first post.

No, it is not possible to lift the 2nd battery from ground, I already thought about it.
I think I etch 1 PCB, and try the circuit with different opamps, starting with the mentioned TL 071.

There is a much easier circuit with only 2 transistors, but that has a drop of at least 1,5 Volt.
That is why I thought of this one, with the opamp.

Thanks !

Edwin

hi Edwin,
Dont forget the main car battery could fall to 9V when you pull the starter.!

Anything you have connected in series with the two batteries will be reversed biassed for a while..
Regards,

 

[Edwin2]

Yes, very good to mention, but the chargingcircuit is switched on at least 1 minute after
the motor is running.
The 2nd battery might be even 200 Ah, so, the starterbattery has to get recharged from starting,
before one can start charging the 2nd battery.

With a series of these circuits in parallel, the chargingcurrent is limited to 25 or 30 Amps.
To keep a bit power left for the rest of the car.

Rgds,

Edwin

 

[Hero999]

You have the drain and source of the P-channel Mosfet connected upside-down.
The source is supposed to connect to the resistor and the drain is the output.

Many FET-input opamps like the TL071 work fine when their input is at the positive supply voltage because their FETs are P-channel.

The + and - pins are the wrong way round, there's no compensation so it will oscillate and there needs to be a capacitor on the output to improve the transient response.

 

[Edwin2]

Sorry, but the inputs are correct.

The reference is 100 mV below supply.
With no current through the FET (and the sense-resistor), the voltage on the - input is the
same as the supplyvoltage, so (100mV)higher => the output is LOW => FET is complete open.
(Watch it, it is a P-FET, symbol is upside down, as mentioned in a previous post)
As soon as the voltage over the senseresistor becomes lower than the reference, the output
goes higher, to close the FET.

Kind regards,

Edwin.

 

[Hero999]

EDIT:
You're right, the input pins were the right way round and I messed it up.

You still need a compensation capacitor though as P-channel MOSFET has gain so a compensation capacitor is required or it will oscillate. You could avoid the discrete compensation capacitor by making the gate resistor a high value and use the MOSFET's gate capacitance as a compensation capacitor.

 

[Edwin2]

No problem, re messing up... It is sometimes very confusing.
First, it `hangs` on the positive supply instead of standing on ground, then using an
inverting amplifier.... Sometimes the brain doesn`t co-operate with it

And than... in my poor English.... I wrote `open` FET where I meant conducting and vise versa.
But an OPEN switch is NOT conducting....

I was planning to make the gate-resistor about 1...10 kOhm.
The gatecapacitance is mostly around 2 nF by this kind of FETs.
So, a bit extra is no problem. buckets of 0805 100 nF `in stock`.

And when this is ready, the next problem, that I did not work out yet, is `measuring` the
voltagedifference between both batteries. If < 0,5 Volt, a big relay may be switched on,
to give the 2nd battery the last bit to get charged.

Kind regards,

Edwin

 

[MrAl]

Hi again,


The only problem with the TL071 is that it may or may not meet the input
common mode spec needed to reach to +Vcc.

Here is another approach using a common LM358 which can do high side
sensing...

(In this diagram a 10 ohm load is shown for an example)

Note that even though the supply is 15v, the inputs are working
around a comfortable 7.5v each.

**broken link removed**

 

[audioguru]

OOPs, I had the inputs of the opamp backwards.

A capacitor to ground at the output might cause the circuit to oscillate.