@saad Nothing beyond the basic knowledge that it's an alternative to PWM that skips makes the average number of pulses the output rather than the average width. What are you trying to do?

@KevinVermeer Trying to calculate the frequency limits before Pulse Skipping kicks in. The datasheet provides the formulae, but my fmax1 always ends up being equal to my switching frequency! Did the calculation on paper and on excel. Here's the datasheet if you're interested: linear.com/docs/Datasheet/3507fa.pdf. I think I'm in over my head and should probably start with smaller switchers!

@kortuk - I know you're a msp430 guy - you got any words on electronics.stackexchange.com/questions/23381/… or am I just being dumb?

@rfusca Try it with a 100kOhm, I doubt it will have problems, if it does, solder a second 100kOhm on top, double stack it.

Stacking two on top is an easy way to get a lower resistance.

@rfusca that 2.2nF, as you probably read, is your maximum capacitance, I would use 1nF if you have it instead.

@rfusca or skip populating it for a bit.

@Kortuk i wasn't going to add the cap there

@Kortuk hmm I guess I'll try that

@rfusca I am pretty sure 100k will work, but really, just put another on top if there is an issue.

@Kortuk any particular reason you think 100k will work? I just don't know much about how it programs

@rfusca it is only twice as high. I believe it is just a pull up

using a weaker pullup will not stress the programmer, and without the capacitance there you should be able to recover to the higher voltage without any real delay, but without the capacitance there you might get more bounce.

wants the advantage to a cap on the reset line?

well i guess just the same reason as normal

@rfusca it is a filter, I doubt you need it.

@rfusca I would put a 1nF and be safe

@Kortuk hmm ok. I'll have to give it a go without the cap to start, the circuit is already etched and completed sans that reset pullup

if I get funky stuff, I can always etch another and get a 47k resistor and a 1nF cap to make sure those weren't the issue. but its a stupid simple circuit and completely non critical even if it did randomly reset - in fact it could be better lol

@Kortuk did you see this week's tideals.com?

Does anyone here know a good DC-DC regulator with 3 outputs? I only require buck converters as my Vin is higher than Vout. I'm hoping that's a bit easy to design.

@saad Unless you want a linear device or a flyback regulator, you may just be better off with 3 regulators.

@W5VO Oh that in case, I think I would be done quicker! I really like National's Simple Switcher and already designed a 5V one today. I'll just design 3.3V and 1.8V now. But, don't you think they would take up more space? Though space isn't that high of a priority.

I think a flyback might be smaller, but it would be a lot harder to regulate all 3 voltages. I would probably go with 3 simple switchers or something along those lines. You can get the parts they need pretty small.

you might check, but those simple switchers look pretty good

@saad Do you need any sequencing?

@KevinVermeer Nope. No device on the board does, fortunately.

@saad That makes it a lot easier!

Many FPGAs and large microcontrollers have precise requirements for sequencing.

@KevinVermeer I know! I was pretty amazed to see all the features modern DC-DC converters have.

You may be able to use one of their switchers, though.

What kind of current do you need?

@KevinVermeer Yes, I was relieved to note the Max V doesn't need sequencing. Regarding current, about 700mA on 5V, 1A on 3.3V and I've measured 200mA on the 1.8V but also read that this rail can need a surge during powerup, so I think 500mA sounds safe.

I actually may get away with linear regulators but I'm gonna test how they get first before I put them into the final design. I could calculate it, but the above currents are really just estimates for the 3.3V and 1.8V rail. If they do get too hot, time to put in some switchers.

@saad That's a decent idea. There are a lot of modules nowadays that are designed to plug into a TO-220 footprint.

However, let's assume your estimates are accurate. You're using .7*5+1*3.3+.2*1.8=7W of power. The power dissipation in those linear regulators would be 1*(5-3.3)= 1.7W in the 3.3V regulator, which is a little much. Without a heatsink, a TO-220 has a theta-JA of about 75 deg. C/W.

So, at 1.7W, you'd be cooking at 125 degrees above ambient. Not good.

@KevinVermeer Yeah, that's why I'm already investigating switchers. :) I also don't have problems redesigning the power-section of the board - we're only getting 4 made initially, for a medium-term test.

However, with a simple heat sink like this one (<- First result for "TO-220 heasink") you can drop that to 15 deg.C/W, so it would be <30 degree rise.

That would be quite warm, but probably permissible. A bigger sink could dissipate even more power.

Yeah, I quite like that. Definitely the easiest solution!

Hmmm, one thing I forgot to add, the area where the machines are intended for use can get very hot during the summer. 35C ambient, at least. Probably more in a box where the electronics would be housed. Would a rise of 30C cause issues in a not so ventilated box? We're trying to avoid ventilation because of dust (incredibly dusty where I live).

@saad Depends on the box. Are we talking about a cabinet, like this: upload.wikimedia.org/wikipedia/commons/f/fa/… or a small plastic box?

Small metal box.

@saad Oh, you've got a metal box? Heatsink it to that.

Oh yeah, could do that! Solves all issues!

As long as you can keep the tabs isolated / at the same voltage

Yes. What if it's a wooden box (it isn't, but suppose it is)?

then you wouldn't see much benefit from putting the heatsinks on it

Because the heat would be trapped? The only recourse would be switchers then?

high efficient switchers that is.

because wood isn't nearly as thermally conductive as metal, and there is a fire risk as well. Metal would help reduce the thermal resistance of the junction to ambient, but wood would probably not.

@saad By your earlier specs, you're producing 7W of power in that box even if you use 100% efficient switchers.

Your box will have a certain level of thermal resistance. The metal one would probably be on the order of a few degrees C/W (depending on size and other factors)

but the wooden one would be larger, again depending on the dimensions, thickness and species of wood, amount of air gap, etc.

Ah yes. Glad we're not using wood! Just makes things more difficult.

@KevinVermeer Would one put the footprint of the regulator near the edge of the PCB so that they can be bolted onto the metal case? That's the only way I can visualize this. To electrically isolate them, I think we can use thermal paste.

@saad Right, they'd need to be near the edge and have some extra room for the legs. You'd slip the parts into their holes, mount the PCB on the standoffs in the case, lift them a cm or so, attach your electrical insulating/thermally conducting substance to the case, screw the heatsink to the holes in the case (which you've already drilled for this purpose), and then solder them in place. If you're doing a lot of boards, find the height they need to be at and build a jig.

@saad Thermal paste isn't very dependable. It should usually be thin enough that it only fills the voids, not that it provides a layer between the part and the wall. Plus, it's often conductive. I'd recommend a pad produced for this purpose, there are plenty: search.digikey.com/us/en/cat/fans-thermal-management/…

Any of the Kapton or silicone ones will be insulative.

Thanks a lot, Kevin! Appreciate all the help!

@saad Welcome! It's nice, sometimes, to be able to chat about something without actually having to do it. Today marked the completion of two semester projects, and I'm pretty burnt out. I've got a little more to do on one of them, but it seems I'm not getting anything done tonight, so I'm heading out.

See you around...