24-42V to 12V regulator - any advice?

You wrote in a later post that the device is used in the cab and apparently not mounted inside anywhere. Then you could be ok, but it will get warm.

American cars, yes, there is unfortunately not too much concern regarding gas mileage. Aircraft, totally different thing these days.

Right, but all those things add up. All the electronic gizmos that have been added to cars in recent time have noticeably pushed up the share that the electrical system consumes. It is the reason why alternators become ever larger. My first car had a DC generator that delivered maybe

150W on a good day and that was sufficient. Unthinkable these days.

That's why I had also suggested separate power devices :-)

1W dissipation in a SOT-223 is too much for my taste, even with DPAK I'd be skeptical. D2PAK, maybe.

Will they start the engine or turn a major compressor on/off during the time this device is plugged in?

In the end it'll also depend on cost and produced quantities. For lowest dissipation and best efficiency a synchronous buck is best. But since that has an internal bootstrap supply it'll have to be able to ride through load dump episodes or you'd need a nifty scheme to disable it for that time. On aerospace designs for 28V power buses I always try to allow it to ride through because clients really like that. But some people provide a cut-out feature.

--
Regards, Joerg

http://www.analogconsultants.com/

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

[...]

Coilcraft is a good suggestion, I'd second that. They do not sell via Digikey, they sell directly. But they do a lot of high-volume product and their prices are very reasonable.

When volumes become really high you can get good deals in China but there MOQs always apply.

--
Regards, Joerg

http://www.analogconsultants.com/

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

For those with no presence.

Best regards, Spehro Pefhany

--
"it's the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

In Siberia, starting a fire under the engine is a reliable way of getting the engine started the next morning.

The Finnish Air Force used a different tactics during the 1939-40 Winter war. After the last sortie for the day, the oil was drained from the engine, During the next morning, the oil was heated above some open fire, and then the sorties were flown.

I remember old people from up north telling similar stories. Dad would come home from work, the oil was let out, into an old kettle. It was then placed on the kitchen stove way in the back. In the morning dad would pour it back into the motor and head to work.

--
Regards, Joerg

http://www.analogconsultants.com/

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

SOP back in Model T days. ...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     |

      Remember: Once you go over the hill, you pick up speed

Thanks!!!

I did have to add a space before the final G in that part number to find it at Digi-Key.

Since the 6.8 nC gate charge if repeated at 250 KHz at 28 volts amounts to .048 watt, I would question these numbers a little. Also, that .038 ohm figure appears to me to be at junction temperature not much over 25 C, and MOSFET resistance varies highly directly with temperature - I want to plan on .08 ohm at a bad-case temperature with that MOSFET. That's .09 watt conduction loss at 50% duty cycle and 1.5 amps, thankfully that's still a nice small number...

Switching loss at 28V: I figure .5 * 250 KHz * (18 ns combined gate drive rise and fall times + 14 ns combined rise and fall times of the MOSFET), times 28V times 1.5A - which amounts to .18 watt. The input capacitance of the MOSFET is maximum ~76% that at which the gate drive output's risetime and falltime are characterized at. However, the risetime and falltime figures that I found are typical, not maximum. Make that .27 watt switching loss when input voltage is 42V.

Then again, .27 watt highish-side switching loss and .09 watt max conduction loss add up to .36 watt in the MOSFET. I can tolerate actuality being twice that should that occur.

And .24 watt for the inductor and 1.14 watts total - I will believe those after I see these with my own eyes. So far, MOSFET loss appears to me highish-side around .2 watt more than reported. However, it does appear to me reasonable for total loss to be around 1.5-1.7 watts. That does mean significantly less heat than I would get from using LM2576.

Thanks again for this, especially for mentioning Webench.

I just went there, and I inputted 20-45 V input, 1.5A load, and 50 C ambient. Result was suggesting primarily LM5085 - I don't know now how I started considering LM5088. LM5085 requires much fewer external components, though the external MOSFET is P-channel - more losses than an N-channel one, but probably very tolerable.

Result in further detail says efficiency 88%, 454 mm^2 footprint size, switching frequency 566 KHz, BOM items 15, BOM cost $3.28. I know that I will need a lot more than 4.54 square cm to get rid of 2.16 watts of heat, but this sounds attractive to me.

Efficiency was increasable to 90% with lower switcghing frequency of 352 KHz and more footprint area and higher cost.

I wonder if this efficiency is minimum for the entire input voltage range that I mentioned? How realistic?

Webench even suggested alternatives. Another attractive one is LM5576, for reported 93% efficiency and 17 item BOM count, with switching frequency of 299 KHz. 1.26 watts of heat is a nice comfortably small number. LM5088 (which I was looking at) is another suggested one, with

92% efficiency and lower BOM cost than LM5085 or the other 8 alternatives, looking to me attractive despite higher BOM count of 22. Switching frequency mentioned was 525 KHz.

One of the other options was LM5116 - rated for 100V. Another was LM2592HV-ADJ, 150 KHz and BOM count of 10 items, but with higher BOM cost and 88% efficiency.

Again, thanks for mentioning Webench!

--
 - Don Klipstein (don@misty.com)

(75 x 44 x 23 mm, willing to upsize in the likely event that's needed)

I would think owners of fleets of vehicles could get concerned about grams per hour per vehicle... When these vehicles are not rented by customers who pay for the fuel...

Now you have me thinking what a 600 watt stereo system could do to fuel economy!

Yes, you did say that before Raveninghorde mentioned a specific part number.

1W does sound to me on the aggressive side for a SOT-223, though half a watt sounds to me comfortable. I am comfy with a watt in a DPAK, though with lots of copper area on both sides of the board and more vias than I can count with 2 hands.

Then again, I like to consider highish side (preferably worst-case) switching losses, plus worst-case conduction losses at some high junction temperature of 100 or 125 C.

At this rate, I am coming up with .27-.36 watt higher-side for a Raveninghorde-suggested MOSFET switched at 250 KHz by an LM5088.

Unfortunately, the answer is yes. A major market of this product is for having it supply ~12V power to electric/electronic automotive service tools from the vehicle's battery.

This gets me in the mood to think of ways to clamp/clip/absorb voltage surges... Any suggestions here?

In my case, incurring slightly larger package size is preferable to the likely-increased cost of going synchronous.

In this application, it does not matter too much if this device shuts down for milliseconds or a fraction of a second due to overvoltage. There is *some* preference for it to "keep on trucking", and I am currently leaning to added external components for overvoltage protection.

My "Plan A" in that area is shunting excessive voltage with a combo of a zener diode (or a couple in series), a MOSFET, a resistor across gate and source, with a 2nd resistor (maybe a .22 or .47 ohm 5W "sandsone" one?) upstream.

Any suggestions there?

--
 - Don Klipstein (don@misty.com)

[...]

Yeah, the airlines always have to pay for their fleet fuel.

Depends on whether you listen to Mozart or the latest chk chk *BOOM* chk chk *BOOM* rap songs :-)

When I picked up my car the dealer had forgotten (!) to mention that it had their top of the line 6-speaker stereo. So I turned it on and for the heck of it onto a hard rock station. On every *BOOM* you could see the RPM needle react and hear the engine controller compensate. That costs serious fuel.

[...]

On a device that is not left powered up in a cab in summer that's probably ok.

The diode is the killer. Hence the suggestion to use sync buck.

This shows in essence how it's done:

Of course one would want more than 100V here. As I said, I prefer to have equipment remain operational while in surge. For that one would need either a PWM chip that is high voltage (pilfered from the LED lighting market et cetera) or use a cheap PWM chip, feed that via a zener/transistor and then have a lighting controls FET driver. Some of those can stomach impressive bootstrap voltages.

If this is used while working in the engine area make sure you have reverse polarity protection. Else it's gonna happen ... phut ... *POP*

Synchronous means one more FET but losing the diode. Usually pretty much a wash from a cost point of view.

Shunting is easier with a TL431 circuit. You'd want snappy action or the FET will walk outside its SOA and blow its lid. A 0.22ohms resistor adds operating losses and it would need to be massive. Or use a thermo-disk to cut power if it gets too hot. In case a trucker has "this great idea" how to get a recalcitrant engine to start up.

--
Regards, Joerg

http://www.analogconsultants.com/

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