Forums

4quad supply

Started by John Larkin April 6, 2016

I'm designing a gadget that will need a bunch of 4-quadrant (+-
voltage, source/sink) power supplies, DAC controlled, well bypassed on
the output. Range is roughly +-12 volts, 100 mA or so.

So I need a power opamp that doesn't mind a capacitive load. We have
TCA0372 in stock, a dual 40-volt, 1 amp opamp that costs 45 cents.
Great part.

I set one up. It really, really doesn't like capactive loads; it
oscillates like mad and gets finger-toasting hot. So I did the usual
compound feedback trick and that works good. The values are what was
handy, and could be tweaked a bit, but it seems fine.

The actual application will use bipolar power supplies.

https://dl.dropboxusercontent.com/u/53724080/Circuits/Power/4quad/4quad_sch.JPG

https://dl.dropboxusercontent.com/u/53724080/Circuits/Power/4quad/4quad_bb.JPG

Driven from a function generator:

https://dl.dropboxusercontent.com/u/53724080/Circuits/Power/4quad/4quad_sq_wave.JPG

This amp, like a lot of others, goes bezerk if you pull the input just
a bit below the negative rail. I don't plan to do that.


-- 

John Larkin         Highland Technology, Inc
picosecond timing   precision measurement 

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

On 04/06/2016 08:11 PM, John Larkin wrote:
> > > I'm designing a gadget that will need a bunch of 4-quadrant (+- > voltage, source/sink) power supplies, DAC controlled, well bypassed on > the output. Range is roughly +-12 volts, 100 mA or so. > > So I need a power opamp that doesn't mind a capacitive load. We have > TCA0372 in stock, a dual 40-volt, 1 amp opamp that costs 45 cents. > Great part.
It is a great part, except that its theta-JC is ridiculously large. (I expect they crippled it on purpose to restrict it to audio because it's so cheap.) On +-15V supplies, it runs noticeably warm with no load, because its quiescent bias current is 8 mA typ, iirc. If they'd put it in a D2PAK or something, it would be a beast. I've used it for things like rail splitting, but only when the main load goes across the real rails. I'd be nervous about subjecting it to the outside corners of quadrants 2 and 4 for very long at the levels you're talking about--it would be liable to toast itself. Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 hobbs at electrooptical dot net http://electrooptical.net
On Wed, 6 Apr 2016 20:32:48 -0400, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

>On 04/06/2016 08:11 PM, John Larkin wrote: >> >> >> I'm designing a gadget that will need a bunch of 4-quadrant (+- >> voltage, source/sink) power supplies, DAC controlled, well bypassed on >> the output. Range is roughly +-12 volts, 100 mA or so. >> >> So I need a power opamp that doesn't mind a capacitive load. We have >> TCA0372 in stock, a dual 40-volt, 1 amp opamp that costs 45 cents. >> Great part. > >It is a great part, except that its theta-JC is ridiculously large. (I >expect they crippled it on purpose to restrict it to audio because it's >so cheap.) > >On +-15V supplies, it runs noticeably warm with no load, because its >quiescent bias current is 8 mA typ, iirc. If they'd put it in a D2PAK >or something, it would be a beast. I've used it for things like rail >splitting, but only when the main load goes across the real rails.
It really needs a power pad. Too many pins for a dpak.
> >I'd be nervous about subjecting it to the outside corners of quadrants 2 >and 4 for very long at the levels you're talking about--it would be >liable to toast itself.
I think I'll be OK. I'll heat sink the four Vee pins, and put a good ground pour under the body. So if I need to, I can put something thermally conductive under the chip, some Bergquist gap-pad stuff maybe. I already need to blow some air on this part of the board. I could glue a heat sink to the top if things got grim, but I expect it won't come to that. A few opamps admit to being c-load stable. A few more are good without it being advertised. Most aren't. It's hard to predict. Even if people give us an equivalent schematic, it's not always true. -- John Larkin Highland Technology, Inc lunatic fringe electronics
On Wednesday, April 6, 2016 at 8:32:54 PM UTC-4, Phil Hobbs wrote:
> On 04/06/2016 08:11 PM, John Larkin wrote: > > > > > > I'm designing a gadget that will need a bunch of 4-quadrant (+- > > voltage, source/sink) power supplies, DAC controlled, well bypassed on > > the output. Range is roughly +-12 volts, 100 mA or so. > > > > So I need a power opamp that doesn't mind a capacitive load. We have > > TCA0372 in stock, a dual 40-volt, 1 amp opamp that costs 45 cents. > > Great part. > > It is a great part, except that its theta-JC is ridiculously large.
Hmmm. 10*C/W. (DIP-16)
> (I > expect they crippled it on purpose to restrict it to audio because it's > so cheap.)
OnSemi seems to push it as general-purpose-- http://www.onsemi.com/PowerSolutions/product.do?id=TCA0372 "The TCA0372 is a monolithic circuit intended for use as a power op-amp in a wide range of applications, including servo amplifiers and power supplies. No deadband crossover distortion provides better performance for driving coils." One wonders if OnSemi's package 966 version (SO-16) has better thermal performance, but their spec sheet doesn't give thermal data for that package. It *is* a nice-looking part.
> On +-15V supplies, it runs noticeably warm with no load, because its > quiescent bias current is 8 mA typ, iirc. If they'd put it in a D2PAK > or something, it would be a beast. I've used it for things like rail > splitting, but only when the main load goes across the real rails. > > I'd be nervous about subjecting it to the outside corners of quadrants 2 > and 4 for very long at the levels you're talking about--it would be > liable to toast itself.
Cheers, James Arthur
On Wed, 6 Apr 2016 20:31:53 -0700 (PDT), dagmargoodboat@yahoo.com
wrote:

>On Wednesday, April 6, 2016 at 8:32:54 PM UTC-4, Phil Hobbs wrote: >> On 04/06/2016 08:11 PM, John Larkin wrote: >> > >> > >> > I'm designing a gadget that will need a bunch of 4-quadrant (+- >> > voltage, source/sink) power supplies, DAC controlled, well bypassed on >> > the output. Range is roughly +-12 volts, 100 mA or so. >> > >> > So I need a power opamp that doesn't mind a capacitive load. We have >> > TCA0372 in stock, a dual 40-volt, 1 amp opamp that costs 45 cents. >> > Great part. >> >> It is a great part, except that its theta-JC is ridiculously large. > >Hmmm. 10*C/W. (DIP-16) > >> (I >> expect they crippled it on purpose to restrict it to audio because it's >> so cheap.) > >OnSemi seems to push it as general-purpose-- > >http://www.onsemi.com/PowerSolutions/product.do?id=TCA0372 > "The TCA0372 is a monolithic circuit intended for use as a power op-amp > in a wide range of applications, including servo amplifiers and power > supplies. No deadband crossover distortion provides better performance > for driving coils." > >One wonders if OnSemi's package 966 version (SO-16) has better thermal >performance, but their spec sheet doesn't give thermal data for that >package. > >It *is* a nice-looking part.
It seems to be well behaved.
> >> On +-15V supplies, it runs noticeably warm with no load, because its >> quiescent bias current is 8 mA typ, iirc. If they'd put it in a D2PAK >> or something, it would be a beast. I've used it for things like rail >> splitting, but only when the main load goes across the real rails. >> >> I'd be nervous about subjecting it to the outside corners of quadrants 2 >> and 4 for very long at the levels you're talking about--it would be >> liable to toast itself. > >Cheers, >James Arthur
I'm using the wide SO16, which is rated 12 j-c and 80 j-a. I might dissipate a watt worst-case, maybe a little less, so I should make some modest effort to get the heat out. Mount a tiny tea kettle on top? -- John Larkin Highland Technology, Inc lunatic fringe electronics
>I'm using the wide SO16, which is rated 12 j-c and 80 j-a. I might >dissipate a watt worst-case, maybe a little less, so I should make >some modest effort to get the heat out.
>Mount a tiny tea kettle on top?
Hmm. With +-15V supplies, sinking 100 mA from +12 is 2.7 W. Per section. (Tweet! Goal posts in motion. Five yards, second down.) ;) They could make a single that would go in a D2PAK or TO220--it would only need 4 pins plus the pad. The power handling is in the LM317 class, which makes the packaging ridiculous. 30V, 1.4A, SO-16. Weird. Cheers Phil Hobbs
On Wednesday, April 6, 2016 at 11:32:00 PM UTC-4, dagmarg...@yahoo.com wrote:
> On Wednesday, April 6, 2016 at 8:32:54 PM UTC-4, Phil Hobbs wrote: > > On 04/06/2016 08:11 PM, John Larkin wrote: > > > > > > > > > I'm designing a gadget that will need a bunch of 4-quadrant (+- > > > voltage, source/sink) power supplies, DAC controlled, well bypassed on > > > the output. Range is roughly +-12 volts, 100 mA or so. > > > > > > So I need a power opamp that doesn't mind a capacitive load. We have > > > TCA0372 in stock, a dual 40-volt, 1 amp opamp that costs 45 cents. > > > Great part. > > > > It is a great part, except that its theta-JC is ridiculously large. > > Hmmm. 10*C/W. (DIP-16) > > > (I > > expect they crippled it on purpose to restrict it to audio because it's > > so cheap.) > > OnSemi seems to push it as general-purpose-- > > http://www.onsemi.com/PowerSolutions/product.do?id=TCA0372 > "The TCA0372 is a monolithic circuit intended for use as a power op-amp > in a wide range of applications, including servo amplifiers and power > supplies. No deadband crossover distortion provides better performance > for driving coils."
Huh... I've got that part in my drawers* but I've never applied power. (I keep waiting for the right project... more than 30 mA, less than 300 mA.) But I was going to ask if anyone had tried it for driving coils. (which is my typical high current app.) Oh I've got another question too. I've used the classic circuit for slowing down the opamp for driving a big capacitance... (shown in JL first picture above.) Is there something equivalent when driving a coil? (All I've every done is a Zobel thing across the coil.) George H.
> > One wonders if OnSemi's package 966 version (SO-16) has better thermal > performance, but their spec sheet doesn't give thermal data for that > package. > > It *is* a nice-looking part. > > > On +-15V supplies, it runs noticeably warm with no load, because its > > quiescent bias current is 8 mA typ, iirc. If they'd put it in a D2PAK > > or something, it would be a beast. I've used it for things like rail > > splitting, but only when the main load goes across the real rails. > > > > I'd be nervous about subjecting it to the outside corners of quadrants 2 > > and 4 for very long at the levels you're talking about--it would be > > liable to toast itself. > > Cheers, > James Arthur
On 04/07/2016 09:32 AM, George Herold wrote:
> On Wednesday, April 6, 2016 at 11:32:00 PM UTC-4, dagmarg...@yahoo.com wrote: >> On Wednesday, April 6, 2016 at 8:32:54 PM UTC-4, Phil Hobbs wrote: >>> On 04/06/2016 08:11 PM, John Larkin wrote: >>>> >>>> >>>> I'm designing a gadget that will need a bunch of 4-quadrant (+- >>>> voltage, source/sink) power supplies, DAC controlled, well bypassed on >>>> the output. Range is roughly +-12 volts, 100 mA or so. >>>> >>>> So I need a power opamp that doesn't mind a capacitive load. We have >>>> TCA0372 in stock, a dual 40-volt, 1 amp opamp that costs 45 cents. >>>> Great part. >>> >>> It is a great part, except that its theta-JC is ridiculously large. >> >> Hmmm. 10*C/W. (DIP-16) >> >>> (I >>> expect they crippled it on purpose to restrict it to audio because it's >>> so cheap.) >> >> OnSemi seems to push it as general-purpose-- >> >> http://www.onsemi.com/PowerSolutions/product.do?id=TCA0372 >> "The TCA0372 is a monolithic circuit intended for use as a power op-amp >> in a wide range of applications, including servo amplifiers and power >> supplies. No deadband crossover distortion provides better performance >> for driving coils." > > Huh... I've got that part in my drawers*
Hope it's not the DIP. Yeeouch! ;)
> but I've never applied power. > (I keep waiting for the right project... more than 30 mA, less than 300 mA.) > But I was going to ask if anyone had tried it for driving coils. > (which is my typical high current app.)
Audio is one big application, so it ought to be okay driving a speaker, at least.
> > Oh I've got another question too. I've used the classic circuit for > slowing down the opamp for driving a big capacitance... (shown in JL > first picture above.)
> Is there something equivalent when driving a coil? > (All I've every done is a Zobel thing across the coil.)
Since the coil's impedance ought to be largish up near the unity gain cross, I wouldn't think it would be a big issue. Anyway, unless the coil's self-resonant frequency is low, the phase will be leading, which won't hurt stability. (I haven't built a coil driver since my audio projects as a teenager.) Cheers Phil Hobbs
On Thu, 7 Apr 2016 05:11:24 -0700 (PDT), Phil Hobbs
<pcdhobbs@gmail.com> wrote:

>>I'm using the wide SO16, which is rated 12 j-c and 80 j-a. I might >>dissipate a watt worst-case, maybe a little less, so I should make >>some modest effort to get the heat out. > >>Mount a tiny tea kettle on top? > >Hmm. With +-15V supplies, sinking 100 mA from +12 is 2.7 W. Per section. (Tweet! Goal posts in motion. Five yards, second down.) ;)
I don't expect to hit total chip dissipation above one watt. It's complicated.
> >They could make a single that would go in a D2PAK or TO220--it would only need 4 pins plus the pad. > >The power handling is in the LM317 class, which makes the packaging ridiculous. 30V, 1.4A, SO-16. Weird.
It does thermal limit, so mounting/heatsinking experiments can be done. The board that I'm laying out now is a little test tile, and I can futz with that before we attempt the real thing... if the other parts work. I'm making a pulse generator output stage, and the opamp supplies programmable Vhi and Vlo. That's the easy part. -- John Larkin Highland Technology, Inc lunatic fringe electronics
On 2016-04-06 21:40, John Larkin wrote:
> On Wed, 6 Apr 2016 20:31:53 -0700 (PDT), dagmargoodboat@yahoo.com > wrote: > >> On Wednesday, April 6, 2016 at 8:32:54 PM UTC-4, Phil Hobbs wrote: >>> On 04/06/2016 08:11 PM, John Larkin wrote: >>>> >>>> >>>> I'm designing a gadget that will need a bunch of 4-quadrant (+- >>>> voltage, source/sink) power supplies, DAC controlled, well bypassed on >>>> the output. Range is roughly +-12 volts, 100 mA or so. >>>> >>>> So I need a power opamp that doesn't mind a capacitive load. We have >>>> TCA0372 in stock, a dual 40-volt, 1 amp opamp that costs 45 cents. >>>> Great part. >>> >>> It is a great part, except that its theta-JC is ridiculously large. >>
It would almost need a constantly emitting can of cold spray.
>> Hmmm. 10*C/W. (DIP-16) >> >>> (I >>> expect they crippled it on purpose to restrict it to audio because it's >>> so cheap.) >> >> OnSemi seems to push it as general-purpose-- >> >> http://www.onsemi.com/PowerSolutions/product.do?id=TCA0372 >> "The TCA0372 is a monolithic circuit intended for use as a power op-amp >> in a wide range of applications, including servo amplifiers and power >> supplies. No deadband crossover distortion provides better performance >> for driving coils." >> >> One wonders if OnSemi's package 966 version (SO-16) has better thermal >> performance, but their spec sheet doesn't give thermal data for that >> package. >> >> It *is* a nice-looking part. > > It seems to be well behaved. >
Except with capacitive loads. When I have to drive capacitive loads I like class D amps.
>> >>> On +-15V supplies, it runs noticeably warm with no load, because its >>> quiescent bias current is 8 mA typ, iirc. If they'd put it in a D2PAK >>> or something, it would be a beast. I've used it for things like rail >>> splitting, but only when the main load goes across the real rails. >>> >>> I'd be nervous about subjecting it to the outside corners of quadrants 2 >>> and 4 for very long at the levels you're talking about--it would be >>> liable to toast itself. >> >> Cheers, >> James Arthur > > > I'm using the wide SO16, which is rated 12 j-c and 80 j-a. I might > dissipate a watt worst-case, maybe a little less, so I should make > some modest effort to get the heat out. > > Mount a tiny tea kettle on top? >
Or the wort pot, for the next batch of IPA. Can't wait to get started. -- Regards, Joerg http://www.analogconsultants.com/