>> I'm making a simple (as possible) op-amp buffer for the PIC
>> ADC and would like to know the effect of unbalanced supply
>>
>> I'm using the amp as a non-inverting unity-gain voltage follower
>>
>> Circuit has available 5.4V (Vcc) and 13.3V (Vbatt). An I/O
>> is being used to generate -5.3V (Vee) for the op-amp
>>
>> I need the input to the ADC to reach at least 5.12V, which is
>> Vref. Common op-amps are quite short of rail-to-rail, so using
>> 5.4V as Vcc isn't working. Op-amps to hand are LM358, LM
>> 1458, LF412, TLS272 and TL072 (all dual - 2 ADC channels)
>
> Adding a pull-up resistor to the output of the LM358 can extend it's useability quite a
> bit, something worth trying? Say 1k ~ 4k7 or so. Without it they are awful - can only
> drive to within 1.5V of +V (from my memory), but on the other hand they drive down
> to 0V ok and input common mode range is good, includes 0V.
>
> --
> Brent Brown, Electronic Design Solutions
> 16 English Street, St Andrews,
> Hamilton 3200, New Zealand
> Ph: +64 7 849 0069
> Fax: +64 7 849 0071
> Cell: +64 27 433 4069
> eMail:  brent.brownKILLspamclear.net.nz
>
>

>>> The datasheets output voltage swing spec is not adequately
>>> defined as they don't say where the load is connected to.
>>> If
>>> the noted 10k is to ground then 5 mV typical is simply a
>>> reflection of the driver off leakage current .
>
>> The LM324 is a bipolar circuit. It's a reflection of the
>> Vce(sat) of the
>> NPN BJT used as a 50uA (nominal) current sink to ground,
>> which is what
>> pulls it down the  last bit, n'est ce pas?  It won't
>> behave a bit like
>> leakage current.
>
> You obviously have a better schematic of the internals than
> I :-). You are no doubt right that that 50 uA source is an
> NPN driven to ground.
>
> I was actually starting with my real world experiences and
> retro-fitting an explanation to them. In retrospect this may
> have given the right answer for a less than fully correct
> reason :-).
>
> My practical experience is that if you want it to drive
> really close to ground you have to help it out fairly
> severely. A quick look at a circuit that came to mind where
> I'd needed to do this (7 years ago as it happens) shows I
> used a 2k7 to ground and was driving a 10k load to an
> intermediate voltage. That 10k matches the 10k load of the
> data sheet and I know that I was not getting close to the 20
> mV max the spec sheet claims.

> I can't remember whether it was Richard Prosser or Brent
> Brown who asked where I'd be at the Mystery Creek Fieldays
>
> Can't guarantee if or when I'll be there personally, the client is
> manning it initially
>
> PE63
>
> Our budding web site
>
> http://www.remstatt.com/
>
> A run-through on Sunday highlighted last-minute changes requiring
> an all-nighter at the keyboard. Haven't done a straight 40-hour stint
> for a while (yaaaaawn). But it paid off
>

> >  Op-amps to hand are LM358, LM1458, LF412, TLC272
> > and TL072 (all dual - 2 ADC channels)
> >
> > LMC6482 will work without twisting your brain...
>
> Thanks Vasile (and particularly the non-brain-twisting bit. I don't
> think I can cope with another just now ;-) ). The whole project
> will be re-visited shortly and there'll be a more leisurely look,
> compared with the hurried flap to get it made in the first place, at
> the PIC's ADC buffering components. Experience with previous
> projects shows that you can use fairly standard components
> without affecting performance too badly for many applications,
> but I'd like to fine-tune this one a little better. Although even the
> current, simple, set-up returns results to 0.2% using 'junk box' parts
>

>> Use the AD reference at 4V or so. Where do you get inexpensive 4V
>> or so reference?
>> TL431 or LM431
>> Specifically, in any PC power suply there is one.
>> How many do you have in your basement ? I think I have hundreds...
>
> I have access to thrown-out 48V Nokia PSUs and each one has a
> control board with a dozen TL431. I haven't tried one as a voltage
> reference source though. The 4.096V references I looked at were
> a little expensive (although I didn't shop around), so I should include
> a TL431 in the next round of experiments then
>
> Currently I'm using a jacked-up 78L05 to make 5.12V for Vref for
> several boards. It's quite easy to set with a couple of fixed resistors
> and a trim-pot. If I continue to use that it'll need testing to make sure
> it's stable from say -10C to 40C. The case is heated a little by the
> PSU so temperature probably won't go too low
>

> Jinx,
>
> The TL431 is likely to be much more temperature stable than the LM7805
> and should be lower noise also.  Also much more consistent between
> samples.
>  I've used the TL431 several times and while it's not perfect, it
> meets my requirements over -10 to +70C.  Just use resistors with good
> temp coefficients!
>
> RP
>
> 2008/6/11 Jinx <RemoveMEjoecolquittEraseMEEraseMEclear.net.nz>:
> >> Use the AD reference at 4V or so. Where do you get inexpensive 4V
> >> or so reference?
> >> TL431 or LM431
> >> Specifically, in any PC power suply there is one.
> >> How many do you have in your basement ? I think I have hundreds...
> >
> > I have access to thrown-out 48V Nokia PSUs and each one has a
> > control board with a dozen TL431. I haven't tried one as a voltage
> > reference source though. The 4.096V references I looked at were
> > a little expensive (although I didn't shop around), so I should include
> > a TL431 in the next round of experiments then
> >
> > Currently I'm using a jacked-up 78L05 to make 5.12V for Vref for
> > several boards. It's quite easy to set with a couple of fixed resistors
> > and a trim-pot. If I continue to use that it'll need testing to make sure
> > it's stable from say -10C to 40C. The case is heated a little by the
> > PSU so temperature probably won't go too low
> >
> > --