which is for the most part taken care of by the feedback loop. I say for the most part because it takes finite time for the opamp output to slew 1.2v and the output devides to respond. Hence the use of reasonably fast parts to minimise the slight glitch.
Compared to biassed output devices, you get less parts count, less cost, and better energy efficiency, and more distortion.
yes
do you have a link to a circuit, I'm not clear what you mean there
NT
386 has had a massive user base, as a rock bottom audio power amp. Why do you think it'll vanish? And even if it did, for repairs there's no lack of scrap with them out there.
NT
Variations on this:
ftp://jjlarkin.lmi.net/Opamp_boost_2.JPG
It can be as simple as adding two transistors and two resistors to the opamp. It resembles the LM317 PNP helper circuit.
John
us
out
ome
yof
eThat one requires some serious loop compensation, especially at large output currents...
Sure. Any closed-loop system has to consider compensation.
John
It might disappear because TI took over and have a (loosely) competing product
A horrid combination of poor distortion performance and instability.
Vladimir Vassilevsky DSP and Mixed Signal Design Consultant
I use a variation of that topology for my NMR and MRI gradient coil drivers. Noise and hum are in the 1 PPM range, and 20 to 100 amp current steps settle to a few PPM in times like 100 microseconds. The trick here, as in any super-precise amplifier, is to get the open-loop transfer function to be extraordinarily linear *before* applying overall negative feedback.
The other trick is to use a really good current shunt.
John
"John Larkin"
** That schem is never gonna work.R2 should go to the output, not R3 or else the mid point voltage is not defined and THD will be very high.
Also R3 needs to be a low value, so the op-amp runs beyond class A range.
... Phil
"Vladimir Vassilevsky"
** Not near as horrid as the POS you posted.... Phil
As I noted, lots of variations are possible. I use this one, without R6, in gradient drivers that have overall NFB from a shunt that senses load current. My sketch is of course a simplification. I was just illustrating the concept of using opamp supply pins to drive the boosters.
R3 defines the base drive into the transistors. Sure, one should pick the right value.
John
"John Larkin" "Phil Allison"
** Should have all been posted before - d*****ad.** That needed posting before, too.
The topology is not used very often, though a few audio power amps have been based on the same idea - notably from Cerwin Vega back in the late 1970s.
Even the simplest two transistor implementation is capable of converting a standard op-amp to rail to rail operation with up to 1 amp peak current available.
To avoid crossover distortion, the resistors feeding each supply pin must be chosen to suit the idle current of the type of op-amp used ( and the actual DC rails ) so the boost transistors are biased slightly on. Emitter ballast resistors may be needed too if operating at 1 amp type currents.
.... Phil
Read what I posted, doofus.
That that resistors should be the right values? Who woulda thunk it.
All you know is audio. There's a lot more to the world than blasting drivel into loudspeakers for the amusement of morons.
It works great with power mosfets.
John
"John Larkin = Liar"
** It was all non specific, hand waving crapology.As usual from a lying, autistic POS like you.
** Still posting non specific, hand waving crapology.As usual from a lying, autistic POS like you.
** Still posting non specific, hand waving crapology.As usual from a lying, autistic POS like you.
... Phil
Isn't repetition a symptom of autism?
John
es
kI'll resist commenting on Phil's issues. The one thing that leaves me a bit uncomfortable about the helper circuit is that opamp quiescent current varies from sample to sample, and as temp changes. I suppose that would be ok if you keep the trannies in class B, but with no nfb around the trs you dont really want to do that.
NT
ts
dEven so, to discontinue a runaway success would seem foolish.
NT
"Tabby"
I'll resist commenting on Phil's issues.
** Then comment on Larkin's.Cos he is the lying POS with sooooo many.
The one thing that leaves me a bit uncomfortable about the helper circuit is that opamp quiescent current varies from sample to sample, and as temp changes.
** The sample variation ion Iq is not large and temp variation even less so.Never heard of " bias servos " ??
I suppose that would be ok if you keep the trannies in class B, but with no nfb around the trs you dont really want to do that.
** FFS - there has to be NFB around the whole thing or it will not work.It should be run as a class AB stage with a few mA bias in the boost transistors.
Did you not read my comments at all ??
.... Phil
Yup, the simple version had better run the bipolars quiescently off, or add some diodes and emitter resistors. But in this respect, it has all the problems of the traditional/ancient bipolar PNP/NPN amp topology. But it can swing almost rail to rail.
If you use mosfets, they can be biased into conduction bit without runaway hazards. I confess that the hairy amps that I design this way use active current mirrors and have an Iq trimpot, strung between the rails of the opamp.
ftp://jjlarkin.lmi.net/Amp.jpg
LT1217 is a great opamp here, if performance matters more than cost. It's fast, can output lots of current, and Iq is 1 mA typ.
Anyway, it's an interesting current splitter topology. Positive signal swing pulls current out of the opamp V+, negative from V-, and the crossover is as close to perfect as you are going to get.
John
What do I need to do to read your files? I get this message; The image ?file:///C:/Users/AppData/Local/Microsoft/Windows/Temporary%20Internet%20Files/Content.IE5/BA3GOP9Q/Amp%5B1%5D.jpg? cannot be displayed because it contains errors. I'm using Mozilla Firefox. Thanks, Mikek
Have something to add? Share your thoughts — no account required.
Ask the community — no account required