Searching for an Op-Amp

Does anyone have a good suggestion for an op-amp with the following specs:

  1. Single-ended supply between 8V and 24V
  2. Up to 100ma output
  3. Rail-to-rail in and out
  4. Can withstand +14V on the output when VDD is removed
  5. Can withstand +14V on the + input when VDD is removed
  6. No more than 50 microamps leakage with VDD removed and +14V on the output and + input
  7. No more than 10 microamps input bias
  8. Differential gain greater than ~500

Obviously, a number of these parameters are quite easy to meet, but finding them all together in a single chip is another matter. I was thinking of something like a TLV271, but I'm not certain it meets all the requirements.

formatting link

Reply to
Leslie Rhorer
Loading thread data ...

There are parts that can handle #5, but none that can handle #4, unless you let the output diode carry the Vdd rail up to +13.5 volts. Maybe you can power the op-amp through a diode? Or add an output switch?

--
 Thanks, 
    - Win
Reply to
Winfield Hill

What about glomming a fast CMOS inverter onto the output of the op amp + reverse bias protection didoes and resistors to the rails at the output of that?

With propagation delays in the low nS one should be able to wrap a feedback loop around the output of the inverter instead of the output of the op amp without much trouble

Reply to
bitrex

------------------------

** It misses a few.

Spec 1 implies pak dissipation up to 2.4 watts, unless you add a buffer.

Spec 4 might be met by adding a reverse diode from output to supply.

Engineering is all about making what you want from what you have.

.... Phil

Reply to
Phil Allison

Obviously a little more detail is in order...

Reply to
Leslie Rhorer

cs:

Um, yeah, it could if it were driving a resistive load to ground. I should have been more detailed, but I had to head out. In practice the Op Amp is only going to have to dissipate about 300mW or less.

Yes, but that would defeat the purpose of employing the Op Amp in the first place.

Here is the circuit someone else devised:

formatting link

It is designed to have the battery take over whenever Vin drops out. It wo rks, but it's rather brutish and sports a somewhat high parts count, given its purpose. What's worse, in practice it limits the choices for Vin and V batt rather uncomfortably. In order to be stable, Vbatt has to be close to 3V lower at full battery charge than Vin. Since the number of different b attery voltages out there is rather limited, and since one would tend to wa nt to use a commonly available power supply - like +12V - and since the min imum acceptable battery voltage is going to be about 8V, things are a bit t ight.

I would like to replace the circuit with something more like this:

formatting link

This is more elegant, but more importantly, the fully charged battery volta ge could approach to within about 1V of Vin. Meanwhile, the minimum Vbatt can be as low as roughly 6.8V. It allows for a wider selection of power su pplies and battery types.

Reply to
Leslie Rhorer

I was afraid of that.

I considered that. It's doable, but only if the idle current of the Op-amp is less than roughly 100 microamps. That's a pretty tall order for a device able to source up to 100ma.

Or add an output switch?

Reply to
Leslie Rhorer

The first circuit is not really so bad: the 1N4148s 33R 4k and BC858 are making a constant current source to charge the backup battery. The current is approximately 20mA. Those parts have no role in the transition to backup power. If your backup battery is not rechargeable then the circuit you need is just two diodes (1N5404 in that circuit).

The second circuit is indeed elegantly and beautifully drawn but will not work in any way you imagine.

piglet

Reply to
piglet

------------------------

** Op-amp makers normally supply devices that current limit at a safe value - rather than make the chip vulnerable.
** The diode will not affect normal operation, as it is always reverse biased - so your "purpose" is not explained.

** So you needed a simple battery charger.

... Phil

Reply to
Phil Allison

pecs:

.

ld have been more detailed, but I had to head out. In practice the Op Amp is only going to have to dissipate about 300mW or less.

st place.

works, but it's rather brutish and sports a somewhat high parts count, give n its purpose. What's worse, in practice it limits the choices for Vin and Vbatt rather uncomfortably. In order to be stable, Vbatt has to be close to 3V lower at full battery charge than Vin. Since the number of different battery voltages out there is rather limited, and since one would tend to want to use a commonly available power supply - like +12V - and since the m inimum acceptable battery voltage is going to be about 8V, things are a bit tight.

tage could approach to within about 1V of Vin. Meanwhile, the minimum Vbat t can be as low as roughly 6.8V. It allows for a wider selection of power supplies and battery types.

If you were able to relax more re psu voltages, keeping the psu a bit on th e high side, the constant current charger might be replaced with nothing mo re than a resistor plus a diode, and that's your lot. I expect it's not the direction you want to go in, but you never know, it could save money.

NT

Reply to
tabbypurr

The second circuit is no good, for one thing there's nothing to bias the !N4148 diode into conduction except leakage current out of the op amp inverting input, which will be small and could be anything.

Reply to
bitrex

[snip]

It works - the two 1N5404 diodes do most of the work. Swapping the silicon diodes for a lower voltage drop Schottky diode like 1N5819 will help a lot (or an appropriate part for the current it needs to supply).

The rest of it is to charge the rechargable battery at constant current (perhaps not such a good idea anyway). There are three terminal constant current sources readily available or you could roll your own from a FET and a single resistor. They are static sensitive. eg.

formatting link

A cruder constantish current source delete the 1N4148s in the base lead and increase the 4k resistor to something like 100*beta ohms per volt. You would be at the mercy of gain variations but it would be OK for a one off. You could probably safely increase the 4k to 100k anyway.

I am not convinced a constant current charge is the right thing to do. If the rechargeable cell is flat you may as well let it recharge at an exponentially declining charge rate starting from C/5 or similar and tailing off to C/100 top off rather than at a steady 10mA all the time.

If you can narrow the voltage gap between Vsupply and Vbatt it is much less of a shock to the downstream circuit when the supply fails.

--
Regards, 
Martin Brown
Reply to
Martin Brown

pecs:

r.

ould have been more detailed, but I had to head out. In practice the Op Am p is only going to have to dissipate about 300mW or less.

irst place.

t works, but it's rather brutish and sports a somewhat high parts count, gi ven its purpose. What's worse, in practice it limits the choices for Vin a nd Vbatt rather uncomfortably. In order to be stable, Vbatt has to be clos e to 3V lower at full battery charge than Vin. Since the number of differe nt battery voltages out there is rather limited, and since one would tend t o want to use a commonly available power supply - like +12V - and since the minimum acceptable battery voltage is going to be about 8V, things are a b it tight.

oltage could approach to within about 1V of Vin. Meanwhile, the minimum Vb att can be as low as roughly 6.8V. It allows for a wider selection of powe r supplies and battery types.

Yes, I know. There is no doubt I would have to supply a resistor to bias t he diode. It's not really relevant to the ask, however. The diode's only purpose is to supply a reasonably steady potential difference so the op-amp will deliver a steady charging current. The crux here is to find, if poss ible, an op-amp that won't croak when mains power is lost. When VDD is zer o, neither the diode nor any bias resistor does anything at all.

Reply to
Leslie Rhorer

If a design requires a long list of unicorn-like characteristics for a part like the ones given to successfully implement, IMO this doesn't mean that one should go hunting unicorns, it means that it's not a good design.

Reply to
bitrex

That is to say sometimes only a unicorn will do, but you will pay a high cost in money or time to get one. Don't ever kid yourself that you can save cost or design time unicorn-hunting; if there is any way to make what you need out of six ponies with horns glued to their heads then that's the preferable option.

Reply to
bitrex

Sure it works. Not exceedingly well, but it works.

Well, not really. That is to say it will help in terms of narrowing the reference voltage, which is part and parcel of the reason for wanting to e liminate one of the diodes, but narrowing the difference between the bias p otential on the transistor and it's cut-off voltage makes the circuit far l ess stable than it already is. It will be quite susceptible to changes in temperature, and this circuit will need to operate through a very wide rang e of ambient temperatures.

Trickle / float charging is a venerable means of keeping rechargeable batte ries ready for use. Admittedly, it isn't as sexy as monitored fast chargin g, but it's easy, and it works.

That will operate on ~1V between the positive rail and the battery terminal potential? I'm all ears. Do you have some part numbers?

Yes, I know. I considered an FET, but from the very source you yourself ci ted: "Although the JFET can provide a simple solution for the design of a curren t source, it is not the best solution obtainable. Much better designs can b e achieved using bipolar transistors (which you will meet in the second yea r), or by using operational amplifiers"

FETs aren't all that stable when only dealing with around 1V of headroom.

That would not work very well. It would be actively discharging the bat tery whenever VDD disappeared. It would positively suck for cold storing t he device. Ideally, the charging circuit should leak less than C/10,000. This will be a user-supplied battery of indeterminate capacity and voltage, so the design needs to be flexible, andno matter what, it shouldn't discha rge the battery to any significant extent.

Yes, that is in fact preferable. It is a little difficult to manage in thi s situation. The circuit in deployment will have a charging monitor, so I am considering making the load resistor variable so the user can set it to match their chosen battery.

True, but it is not much of an issue in this case.

Reply to
Leslie Rhorer

How about using a small relay, e/m or SSR, to disconnect the input and output? The other specs are tough enough already.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
John Larkin

Perhaps, but in this case, it isn't really a long list of unicornish requir ements. Just one, and I am hoping there is a thoroughbred out there that c an deliver.

Yes, but it never hurts to look around the bushes for a little bit to see i f a unicorn is hiding there. There are a lot of parts out there with attri butes of which I am unaware. Coming here to ask the question is hardly a f ull-scale unicorn hunt. If there isn't, well there is the old circuit whic h works in a more limited scope than desired. Being flexible for one's use rs is a VERY desirable goal, if it can be done at a reasonable cost. So is saving space.

Reply to
Leslie Rhorer

a

od

irements. Just one, and I am hoping there is a thoroughbred out there that can deliver.

h

if a unicorn is hiding there. There are a lot of parts out there with att ributes of which I am unaware. Coming here to ask the question is hardly a full-scale unicorn hunt. If there isn't, well there is the old circuit wh ich works in a more limited scope than desired. Being flexible for one's u sers is a VERY desirable goal, if it can be done at a reasonable cost. So is saving space.

There are lots of current sources (at Digikey.) Here's the first I found.

formatting link

5020T1GOSCT-ND/2705192

I think that and the two beefy diodes is all you need.

George H.

Reply to
George Herold

That wasn't a stated design constraint until now.

You could still permit faster charging when Vbatt is lower.

Pretty much. ISTR LM334 but 10mA will be pushing it. 5mA no problem.

formatting link

I expect there are beefier ones.

You didn't specify temperature stability as a key requirement. You asked for a way to reduce component count.

How do you work that out?

The transistor is still isolated from the battery by the diode in its collector when Vin is < Vbatt. It really depends on how constant you want the charge current to be and over what temperature range.

I reckon you are barking up the wrong tree - you want to guarantee a minimum long term float charge rate that is safe for the battery but it really doesn't matter if it charges faster when the battery is flat or whether it varies a bit with temperature.

It buys you some extra headroom without using unicorn components.

--
Regards, 
Martin Brown
Reply to
Martin Brown

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.