MC34063 with small inductors
in [Design]
|
From: Tim Wescott on 7 Apr 2010 16:21 Joerg wrote: > Spehro Pefhany wrote: >> On Tue, 06 Apr 2010 18:00:41 -0700, the renowned Joerg >> <invalid(a)invalid.invalid> wrote: >> >>> Hello Folks, >>> >>> Have to correct a design with a 34063 in there. Buck, VIN 16-30V, >>> VOUT slightly under 12V, 400mA max output. And no, I don't like it >>> much :-) >>> >>> One of the issues is a saturating inductor. Can't change the layout >>> and to be able to cram one in there I probably have to drop down to >>> 47uH, maybe 68uH if lucky. >>> >>> SPICE says it's ok but none of the datasheets or app notes recommends >>> going that low. Any reason why not? >> >> Too slow. ... > > > Although it does run in SPICE. > I think that one of the last things I'd trust SPICE for would be as an indication that a chip really could run faster than its rated speed, at least if the chip model came from the vendor. -- Tim Wescott Control system and signal processing consulting www.wescottdesign.com
From: Jim Thompson on 7 Apr 2010 17:07 On Wed, 07 Apr 2010 13:21:06 -0700, Tim Wescott <tim(a)seemywebsite.now> wrote: >Joerg wrote: >> Spehro Pefhany wrote: >>> On Tue, 06 Apr 2010 18:00:41 -0700, the renowned Joerg >>> <invalid(a)invalid.invalid> wrote: >>> >>>> Hello Folks, >>>> >>>> Have to correct a design with a 34063 in there. Buck, VIN 16-30V, >>>> VOUT slightly under 12V, 400mA max output. And no, I don't like it >>>> much :-) >>>> >>>> One of the issues is a saturating inductor. Can't change the layout >>>> and to be able to cram one in there I probably have to drop down to >>>> 47uH, maybe 68uH if lucky. >>>> >>>> SPICE says it's ok but none of the datasheets or app notes recommends >>>> going that low. Any reason why not? >>> >>> Too slow. ... >> >> >> Although it does run in SPICE. >> >I think that one of the last things I'd trust SPICE for would be as an >indication that a chip really could run faster than its rated speed, at >least if the chip model came from the vendor. Particularly since LTspice switcher models have a high behavioral content, rather than being modeled at the device level. That lets them certainly simulate faster. I can't comment on their accuracy... that's _very_ dependent on who wrote them. ...Jim Thompson -- | James E.Thompson, CTO | 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 http://www.analog-innovations.com | 1962 | The only thing bipartisan in this country is hypocrisy
From: Tim Williams on 7 Apr 2010 17:25 "Jim Thompson" <To-Email-Use-The-Envelope-Icon(a)On-My-Web-Site.com> wrote in message news:qadpr5d55p3s1st00ltebi83q20fpie206(a)4ax.com... > I did off-line switchers for GenRad using only LM339's and TL431's for > the control system... plus Opto Couplers for loop isolation. You don't even need LM339's. http://webpages.charter.net/dawill/Images/RegBO.gif Tim -- Deep Friar: a very philosophical monk. Website: http://webpages.charter.net/dawill/tmoranwms
From: Tim Wescott on 7 Apr 2010 17:41 Jim Thompson wrote: > On Wed, 07 Apr 2010 13:21:06 -0700, Tim Wescott <tim(a)seemywebsite.now> > wrote: > >> Joerg wrote: >>> Spehro Pefhany wrote: >>>> On Tue, 06 Apr 2010 18:00:41 -0700, the renowned Joerg >>>> <invalid(a)invalid.invalid> wrote: >>>> >>>>> Hello Folks, >>>>> >>>>> Have to correct a design with a 34063 in there. Buck, VIN 16-30V, >>>>> VOUT slightly under 12V, 400mA max output. And no, I don't like it >>>>> much :-) >>>>> >>>>> One of the issues is a saturating inductor. Can't change the layout >>>>> and to be able to cram one in there I probably have to drop down to >>>>> 47uH, maybe 68uH if lucky. >>>>> >>>>> SPICE says it's ok but none of the datasheets or app notes recommends >>>>> going that low. Any reason why not? >>>> Too slow. ... >>> >>> Although it does run in SPICE. >>> >> I think that one of the last things I'd trust SPICE for would be as an >> indication that a chip really could run faster than its rated speed, at >> least if the chip model came from the vendor. > > Particularly since LTspice switcher models have a high behavioral > content, rather than being modeled at the device level. That lets > them certainly simulate faster. I can't comment on their accuracy... > that's _very_ dependent on who wrote them. I think that, were I tasked with writing a bunch of models for a bunch of chips, that my main goals would be (a) make them never crash, (b) obfuscate or outright hide the inner workings of the chip, and (c) make them accurate _over the specified operating range_. Given (c), I wouldn't hesitate to publish a model that failed to predict correct _or_ incorrect operation if the chip were operated will outside of it's limits -- I'd probably aim for at least 50% out of range at room temperature, but if a customer tried to predict the behavior of a 100kHz part at 200kHz, I'd have some really nice, diplomatic way of saying "you're screwed, get over it". In fact, it'd probably be in the app notes for the models. -- Tim Wescott Control system and signal processing consulting www.wescottdesign.com
From: Jim Thompson on 7 Apr 2010 17:49
On Wed, 07 Apr 2010 14:41:14 -0700, Tim Wescott <tim(a)seemywebsite.now> wrote: >Jim Thompson wrote: >> On Wed, 07 Apr 2010 13:21:06 -0700, Tim Wescott <tim(a)seemywebsite.now> >> wrote: >> >>> Joerg wrote: >>>> Spehro Pefhany wrote: >>>>> On Tue, 06 Apr 2010 18:00:41 -0700, the renowned Joerg >>>>> <invalid(a)invalid.invalid> wrote: >>>>> >>>>>> Hello Folks, >>>>>> >>>>>> Have to correct a design with a 34063 in there. Buck, VIN 16-30V, >>>>>> VOUT slightly under 12V, 400mA max output. And no, I don't like it >>>>>> much :-) >>>>>> >>>>>> One of the issues is a saturating inductor. Can't change the layout >>>>>> and to be able to cram one in there I probably have to drop down to >>>>>> 47uH, maybe 68uH if lucky. >>>>>> >>>>>> SPICE says it's ok but none of the datasheets or app notes recommends >>>>>> going that low. Any reason why not? >>>>> Too slow. ... >>>> >>>> Although it does run in SPICE. >>>> >>> I think that one of the last things I'd trust SPICE for would be as an >>> indication that a chip really could run faster than its rated speed, at >>> least if the chip model came from the vendor. >> >> Particularly since LTspice switcher models have a high behavioral >> content, rather than being modeled at the device level. That lets >> them certainly simulate faster. I can't comment on their accuracy... >> that's _very_ dependent on who wrote them. > >I think that, were I tasked with writing a bunch of models for a bunch >of chips, that my main goals would be (a) make them never crash, (b) >obfuscate or outright hide the inner workings of the chip, and (c) make >them accurate _over the specified operating range_. > >Given (c), I wouldn't hesitate to publish a model that failed to predict >correct _or_ incorrect operation if the chip were operated will outside >of it's limits -- I'd probably aim for at least 50% out of range at room >temperature, but if a customer tried to predict the behavior of a 100kHz >part at 200kHz, I'd have some really nice, diplomatic way of saying >"you're screwed, get over it". In fact, it'd probably be in the app >notes for the models. One thing that would be wise, but is rarely included... power supply ramp-up. Having been burned mightily with a chip with both positive and negative supplies, I make a point of simulating power ramp-up and sequencing... then adding local protections, as needed, into the chip design. Most behavioral models don't do that adequately. I'd like customers to come to me with their device-level circuitry and ask for a behavioral equivalent... but they won't. Somehow they think I'll steal their IP. It generally isn't worth stealing :-) ...Jim Thompson -- | James E.Thompson, CTO | 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 http://www.analog-innovations.com | 1962 | The only thing bipartisan in this country is hypocrisy |