DIP IC heat sinks

The k/watt from junction to case is so high that you don't need anything fancy, just glomp something on there. If it covers the whole top & has some extra ali it's good. Trying to get more out of it you get near zero returns for your extra buck.

PCB area can help, but again it's limited. Nevertheless one can triple P_out with the right heatsink - if you don't mind pushing the electrical ratings to the edge.

If Sinclair were around today he might try sanding DILs down to get more power out of them :)

If ambient is high for some reason it should be possible to improve things further with a 2nd heatsink on the underside, eg a strip of ali. But really by that point the electrical specs are getting pushed too hard already, so it's not normally pointful.

NT

Thanks, it ain't pretty but seems to work well enough!

This part seems like it would be useful as the next step up in power op amps, naturally it's been discontinued:

LM675.

Cheers

Phil Hobbs

I use the OPA544 too, (+/-35V supply, not cheap) George H.

The LM675 goes up to 60V, has a 5-MHz typical GBW, and costs $5. Specs are a bit sketchy otherwise, e.g. no noise info at all. OTOH you're using it as an output stage, so how bad can it be?

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

Oh I use the LM675 too. It's not as happy at unity gain.

GH

To save the extra $10, you could just add an RC between the inputs to increase the HF noise gain a bit. (Reduces the input impedance in noninverting mode, of course.)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

Right, I've seen that in the spec sheet,

Never really understood how it worked. I'll have to think about it. (The inverting case looks simple, just throwing away signal and feedback above ~500kHz. ?)

George H.

It's a differential amp, so the inverting situation is exactly the same as noninverting except for minor CMR effects.

Easiest way to look at it is as a 1:N voltage divider followed by a gain-of-N amplifier.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

That's a great opamp. I use the surface-mount package with a big topside copper pour and thermal vias down to an inner power pour.

If you use the DIP package, use the DIP16 and get as much heat as possible out the leads, into lots of PCB copper.

The main virtue of a topside epoxied heat sink is that it will spread out the chip hot spot. You can heat sink the bottom of the dip to the PCB too, with a gap-pad or something.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

Maybe parallel op amps like dis?

The jump in power op amp pricing is a lot when you want to step up to something you can bolt a heat sinkto apparently, it goes from fiddy cent for the TCA to around 5 bucks for an LM-whatever.

You can use them to DC-drive small motors. When the motor is running at speed the back-EMF keeps the dissipation low and the chip stays cool no problem. The problem is a rapid sequence of spin-ups and spin-downs which start heating the die and reduce the amount of power it can push. Motor stalls, then the control loop attempts to "solve" this problem by pushing more current exacerbating the problem.

Maybe parallel two of them?

OK... I've used the rail-splitter circuit too. (similar type thing.) Say, as long as this isn't too mundane, why does one need Rs in the non-inverting case? (figure 12)

George h.

I'm not looking at the DS at the moment, but it's probably to correct for input bias current, which is fairly horrible IIRC. (I'm in a hospital cafeteria waiting for my wife to come out of hip surgery.)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

Hmm, seems like you'd need diodes on the outputs or something to stop them fighting each other.

George H.

Ballast resistors are the usual approach. But it won't help the TCA0372 that much because the dissipation on the die will be about the same, and it's theta_JC that's the problem. Two separate chips would be more the ticket, or else a RRIO plus a couple of transistors (extra points for putting in a current limit).

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

Negative feedback is keeping all the inputs at the same voltage, for ideal op-amps if both inputs of both amps are at the same voltage then the outputs of the two amps couldn't ever be different from each other.

Ideal op amps cost more than LM675s. ;)

You could use a chopamp to make the summing junctions track to within some microvolts, which would reduce the quiescent current to a very low level.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

OK I'm not understanding. Your pic has one opamp getting feedback from motor and 0.1 ohm R, and the other FB from just two resistors. (Stating the obvious.. sorry...) Seems like in general the inverting inputs are going to be different.

George H.