What's the hi-side current sense chip du jour?

[snip]

I applied Francis Ford Coppola's Law on Monday... chased a root canal with a full bottle of his fine Claret... otherwise I don't know if I'd have slept a wink :-(

...Jim Thompson

-- | James E.Thompson, P.E. | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at

| 1962 | Stormy on the East Coast today... due to Bush's failed policies.

[snip]

Not strange at all. How are you to get common-mode rejection of the flying end of the inductor?

I told you how to do it right, but you have neither space or money ;-)

...Jim Thompson

-- | James E.Thompson, P.E. | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at

| 1962 | Stormy on the East Coast today... due to Bush's failed policies.

Don't worry, Obama will take some money and space from you to give to them!

"

I made the usual rookie tequila miscalculation, once, in 1981. Once, it turns out, is enough...to this day the slightest whiff of the stuff gives me psychedelic flashbacks!

Cheers, James Arthur

Maybe you should sub that to me :-) ==>

I have a really pig circuit that have to be designed in a 7mm dia PCB, plus it have to work -40/+125C.

I think I've just come up with the right solution this morning and the customer's complaining because ah! oh! too many components... ah! oh!

0201! yada yada... Old solution... ah! oh! but you want to stuff both sides?

Yes Mr. customer, 3 more components, but old solution doesn't fulfill the requirements at thrice the estate with COB...

Arghhh...

--
Thanks,
Fred.

AD8605. Please use it in your high volume app, so that it won't die for my low volume app :-)

OK, we'll probably buy a lifetime few wheels...

--
Thanks,
Fred.

le

r

What is your topology?

My circuit dealt with that problem by just lowering the set point current a bit as the input voltage went up. This was just an extra resistor from the Vin over to the comparator. The resistor's primary purpose was to improve the power supply rejection ratio. It shifted the average number to be around zero.

On May 24, 1:51=A0pm, Joerg [...]

I have mounted the D2PAKs on the back of the PCB and put a bit of "gap pad" between them and the chassis. It doesn't work as well as doing it the normal way but it can save square inches. The controlling circuit can be right above it on the Comp side.

[....]

Is this the limit on the peak current or the average? Since the inductor likely has more slop than this I assume it is for reasons other than keeping the inductor out of saturation

That's a pretty common jelly bean part, AFAIK. We use a lot of AD8606s (dual version).

yabut the 5 bumps WLCSP is nice and at 1 by 1mm is pretty unique 'til now. I was joking since I don't see AD ever playing Maxim's game, but we'd be in trouble should it disappear.

--
Thanks,
Fred.

ng

.

One thing to consider about parts showing up in catalogs. There may be a bit of quid pro quo here, i.e. some small vendor paying the distributor to carry the parts. I don't want to name 'em and shame 'em, but it has been done and is probably legal. The real key is how many parts the distributor stocks, not that it shows up in a catalog.

Oh, but haven't you heard? Maxim is now a "good guy". Just ask 'em!

Dunno why, but the Avenet rep was pushing a Maxim part when the AD8655 fit the bill. The AD8566 is $.71 (reel of 3000) and the Max part was closer to $3, though the Maxim would have made us a deal, apparently. Wonder why.

Not only that but also how fast that stock moves. Digikey shows stock, and it's one of my automatic habits during circuit designs to check back later. If nothing moves and stock is three-digit or less, lil' red flags go up. In the same way they go up when price or leadtime says "call", except then it's the really big red flags.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

A German NG poster who has emigrated to Paraguay said his root canal cost about $20, very professionally done. My last one was almost $700. Now that really hurt. Maybe I fly to Asuncion next time.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Both. Basically cut-off but there should also be a warning along the lines of "If you continue hanging more stuff onto this rail things will go dark soon". That would already be two comparators.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Interesting. What's that "gap pad" stuff? It would be nice to be able to sink into the chassis that way. But only if this pad can conduct away a large chunk of the thermal energy. Plus 2-oz copper is frowned upon quite often and with D2PAK it'll almost be a requirement.

Mainly not to fry sensitive loads. So I can't really use inductor peak, considering a large input voltage range.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

:-)

Yeah, I'd really like to avoid 0201. 0603 would be ok with this client and maybe I can convince them to go 0402 but probably not farther down in size. Of course I can check if needed but in this case they'd prefer the simpler choice even if it costs 50c more. But of course not $5 more.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

[ ... but it's a problem with expensive or single-source solutions]

How about a complex R-divider step down? Two transistors with bases at +5, emitter resistors to GND, make a matched pair of current sinks. From first collector, run a 100k resistor to sense resistor (+), from second collector a similar 100k resistor to sense resistor (-). The difference voltage at the collectors is same as on the sense resistor, so you can connect a comparator or difference amplifier there without taking the divider penalty.

Trying to find a single-chip solution is maddening, and there's always the trap of that lovely solution that only one manufacturer can provide...

Not there, of course. Buck = output voltage lower than input. So, what's wrong with providing a lil' sense electronics plus two extra pins that you can hang across a hi-side Isense resistor? Older PWM chips (almost) had that, by including a 2nd internal opamp tied to the error node. I'd use those in a heartbeat but of course they are too low in switcher frequency. Seems the old and now retired league of chip designers thought more about the application.

Oh, for a buck I think I can do it. I just want to know which hi-side sense chip is popular so I won't saddle the wrong horse here. Like what's the VW-Beetle or Toyota Corolla of hi-side sensors?

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.