Forrest, you rock! Per your suggestion, I reached out to eevblog to find someone that could help. One guy that responded recognized the thermal problem you described. He was the only one that brought it up. Upon looking at my design, he explained why a few other problems were happening.
I've had several others look at it in the past and not one noticed the problem with that chip. Thanks again for your insight and suggestions. Chuck and Ken, you both helped me realize the design was flawed even when some other "experts" just recommended tweaking it. I just wish I had reached out to the list sooner. I keep forgetting the depth and breadth of knowledge here. On Wed, Mar 6, 2019 at 1:55 PM Forrest Christian (List Account) < li...@packetflux.com> wrote: > Simplified mental picture statement: a linear regulator like the > AZ2117 works by burning off excess voltage as heat. > > Assume you have a circuit which takes 500mA @ 3V. If you want to > feed it with 12V, and use a linear regulator, the circuit is still > going to take 500mA but at 12V, with 500mA@9V being turned into heat > by the linear regulator and 500mA@3V being turned into heat by your > circuit. > > 500mA x 9V = 4.5W. > > Each type of case can dissipate (or get rid of) a certain amount of > power. The AZ2117 couldn't get rid of power quickly enough so it > overheated. It happened to have over temperature protection so it > shut down. Less sophisticated regulators will destroy themselves. > > The VXO8705-500 is a switchmode regulator, which effectively works > like the following: > > Assume you have a circuit which takes 500mA@3V. If you want to feed > it with 12V, and use a switchmode regulator, the switchmode regulator > very quickly turns the power on for 25% of the time, and off for 75% > of the time, making an average of 12V*25%=3V. The average current > used at 12V will also be cut in a quarter, so you'd have 125mA@12V. > Because it's switching it on and off there is very little heat > generated. > > On Wed, Mar 6, 2019 at 11:11 AM canopy--- via AF <af@af.afmug.com> wrote: > > > > Forrest, I added the VXO7805-500 between the battery and the board and > now the board appears to > > be working reliably from battery. It's always worked fine via USB. Can > you explain why going from 12V > > down to 5V makes the circuit work reliably? When the AZ2117 was reducing > voltage was it also reducing > > the amps going through as well? > > > > I like the TPPM0301 but yes, I could use a few more mA. Also, I'd like a > wider temperature > > range. I found another 500mA version but it has the same 0-70C operating > range. > > > > > > > > On Sat, Mar 2, 2019 at 1:15 PM Chuck McCown <ch...@wbmfg.com> wrote: > >> > >> I never looked at the data sheet until Forrest noticed the thermal > problem. > >> Surface mount. I use similar packages in some products. My pcb is > double > >> sided with large lands with a bazillion vias perforating it trying to > use > >> the PCB as a heatsink. It sucks for any real amount of heat. I am > able to > >> get rid of about 5 watts without turning things brown and having the > parts > >> unsolder themselves. > >> > >> In your case, the reg has a thermal shutdown circuit so you never got > to the > >> obvious smoke and self disassembly clues. > >> > >> The data sheet says you can flow -20C air over it at 8 watts... > >> Just spray it with freon and you are good to go! > >> > >> > >> -----Original Message----- > >> From: Forrest Christian (List Account) > >> Sent: Saturday, March 02, 2019 10:23 AM > >> To: AnimalFarm Microwave Users Group > >> Subject: Re: [AFMUG] OT: EE Consultant > >> > >> Like Chuck said, pretty much any switchmode regulator aka buck > >> converter aka dc-dc converter. This is where you end up with > >> cost/benefits - there are fully integrated ones out there which are > >> expensive but easy. A quick mouser search turns up a VXO7805-500 as > >> an example. You can also do all of the design yourself, which is > >> where datasheets come in handy. > >> > >> Note that this isn't going to be a drop in replacement since the USB > >> voltage isn't going to be enough to power this with the switch method > >> you are using. I'd reccomend setting this up so the bat goes > >> directly into the regulator (it might even be reverse polarity > >> protected, if not add a diode on the input to protect it). This > >> would then have a 5V out which you would then feed into the rest of > >> your circuit. This would require that you rearrange a bit of > >> everything else. > >> > >> If I was designing it I would probably try to use something like a > >> TPPM0301 which does all of the switching and regulating for you on the > >> 3.3V side - You'd feed the USB into the 5Vcc pin, feed the output of > >> the regulator described above into the 5Vaux pin, then take your > >> output out of the 3.3Vout pin, ignoring the GND. However, if you > >> really need 2.2W this particular part won't be enough, and you might > >> actually have to add a 3.3V switcher as well since 2.2W is around > >> 750mA and that's pretty beefy for a linear regulator of any sort. On > >> the other hand, if you meant 2.2W currently at 12V, then this might be > >> be fine. > >> > >> This pretty much neuters your source identification leds though. And > >> the monitoring stuff to go with it. There's ways around this as > >> well. > >> > >> Depending how much you want to share, I will say that the eevblog > >> forum is a really good resource for this type of stuff. They also > >> have a jobs board if you're looking for someone who might just take > >> this off your hands. > >> > >> On Sat, Mar 2, 2019 at 7:22 AM can...@believewireless.net > >> <p...@believewireless.net> wrote: > >> > > >> > USB is preferred input so the battery isn't drawn down during normal > >> > operation. I'm drawing > >> > about 2.2W under normal operation. Is there another chip you'd > recommend > >> > other than > >> > the AZ2117? > >> > > >> > On Sat, Mar 2, 2019 at 3:51 AM Forrest Christian (List Account) > >> > <li...@packetflux.com> wrote: > >> >> > >> >> Oh this is an easy answer. > >> >> > >> >> At 12V, the AC2117 is dropping about 8 of those volts. > >> >> 8V*400mA=3.2W. Thermal resistance to the heatsink on this part is > >> >> 100C/W. So with a room temperatureish heatsink at 30*C, you're > going > >> >> to end up with a junction temperature of 100C/W * 3.6W = 320C + 30*C > >> >> = 350C. This is above the absolute maximum operating temperature > >> >> range of 150*C. > >> >> > >> >> The most you can draw through this with an 8V drop and an appropriate > >> >> heatsink not exceeding 150*C (I'm assuming it gets a bit warm) is > 150C > >> >> - 50C = 100C. 100C/W means this part can dissipate at most 1W with > >> >> an adequate heatsink. This equates to 1W/8V = 125mA max with a 12V > >> >> input voltage. > >> >> > >> >> If you don't have an adequate heatsink or the temperature gets higher > >> >> this is even lower, since the rating without a heatsink is 125C/W. > >> >> > >> >> Note that this isn't a problem off of USB because the AZ2117 doesn't > >> >> drop much voltage, so even at 1V of drop, you'd only have > >> >> 1V*400mA=400mW of power dissipation, and the unit should be able to > >> >> handle this even without a heatsink. > >> >> > >> >> ALSO: Could you clarify here if the USB is the preferred input? If > >> >> not, one could rip out IC2, R26, R3 and Q1 and replace it with > another > >> >> diode. > >> >> ALSO: Not sure why the indicator leds are done the way they are - > they > >> >> probably work, but seems like overkill. > >> >> > >> >> On Fri, Mar 1, 2019 at 12:14 PM can...@believewireless.net > >> >> <p...@believewireless.net> wrote: > >> >> > > >> >> > Schematics attached. The main issue I'm having is when powered via > USB, > >> >> > everything works fine. When powered via a 12V car battery, the > >> >> > chips seem to not have enough power. There are two comparator > circuits, > >> >> > one to monitor USB/Batt and report how the unit is being powered. > >> >> > The 2nd actually provides the power. One chip will occasionally > draw > >> >> > ~150-400ma of power causing the board to lock up. Chips are powered > >> >> > via Vin. > >> >> > > >> >> > What I don't understand is why everything works perfectly on USB > but > >> >> > has issues when powered via 12V. > >> >> > > >> >> > On Tue, Feb 26, 2019 at 11:53 AM <ch...@wbmfg.com> wrote: > >> >> >> > >> >> >> Circle of stability, miller capacitance all good stuff. > >> >> >> > >> >> >> From: Mark Radabaugh > >> >> >> Sent: Tuesday, February 26, 2019 9:48 AM > >> >> >> To: AnimalFarm Microwave Users Group > >> >> >> Subject: Re: [AFMUG] OT: EE Consultant > >> >> >> > >> >> >> Oh sure, now you probably want me to believe transistors have > states > >> >> >> other than off and on. > >> >> >> > >> >> >> Mark > >> >> >> > >> >> >> On Feb 26, 2019, at 11:20 AM, Ken Hohhof <af...@kwisp.com> wrote: > >> >> >> > >> >> >> > And good luck finding an analog circuit guy anymore ;-) > >> >> >> > >> >> >> That’s what I was. Not much demand, that’s why I’ve been doing > other > >> >> >> things for over 20 years. Actually made the switch to engineering > >> >> >> management in the 80’s but still did some design work for another > 10 > >> >> >> years because all the engineers wanted to do digital and processor > >> >> >> designs. I still have the books, the brain is rusty though. > >> >> >> > >> >> >> I tried to convince the new engineers that high speed digital > design > >> >> >> meant knowing analog, RF and microstrip techniques, but they > weren’t > >> >> >> buying it. Getting their products through regulatory emissions > >> >> >> testing did give them a bit of religion. > >> >> >> > >> >> >> I still remember laying out PCBs using red and blue tape. > >> >> >> > >> >> >> From: AF <af-boun...@af.afmug.com> On Behalf Of Mark Radabaugh > >> >> >> Sent: Tuesday, February 26, 2019 9:09 AM > >> >> >> To: AnimalFarm Microwave Users Group <af@af.afmug.com> > >> >> >> Subject: Re: [AFMUG] OT: EE Consultant > >> >> >> > >> >> >> Yep - I’m qualified to answer some things and would be totally > lost on > >> >> >> others. EE has become a huge field with many many disciplines. > >> >> >> > >> >> >> And good luck finding an analog circuit guy anymore ;-) > >> >> >> > >> >> >> Mark > >> >> >> > >> >> >> > >> >> >> On Feb 26, 2019, at 9:56 AM, Ken Hohhof <af...@kwisp.com> wrote: > >> >> >> > >> >> >> Analog? Digital? Even with those there tend to be specialties > like > >> >> >> emissions, safety, PLD/ASIC design, > microprocessor/microcontroller, > >> >> >> RF, etc. Or do you need a generalist? > >> >> >> > >> >> >> From: AF <af-boun...@af.afmug.com> On Behalf Of > >> >> >> can...@believewireless.net > >> >> >> Sent: Tuesday, February 26, 2019 8:33 AM > >> >> >> To: AnimalFarm Microwave Users Group <af@af.afmug.com> > >> >> >> Subject: [AFMUG] OT: EE Consultant > >> >> >> > >> >> >> Having an issue with a circuit design, anyone know a good > electrical > >> >> >> engineer they can > >> >> >> recommend? > >> >> >> -- > >> >> >> AF mailing list > >> >> >> AF@af.afmug.com > >> >> >> http://af.afmug.com/mailman/listinfo/af_af.afmug.com > >> >> >> > >> >> >> > >> >> >> -- > >> >> >> AF mailing list > >> >> >> AF@af.afmug.com > >> >> >> http://af.afmug.com/mailman/listinfo/af_af.afmug.com > >> >> >> > >> >> >> > >> >> >> > >> >> >> ________________________________ > >> >> >> -- > >> >> >> AF mailing list > >> >> >> AF@af.afmug.com > >> >> >> http://af.afmug.com/mailman/listinfo/af_af.afmug.com > >> >> >> > >> >> >> -- > >> >> >> AF mailing list > >> >> >> AF@af.afmug.com > >> >> >> http://af.afmug.com/mailman/listinfo/af_af.afmug.com > >> >> > > >> >> > -- > >> >> > AF mailing list > >> >> > AF@af.afmug.com > >> >> > http://af.afmug.com/mailman/listinfo/af_af.afmug.com > >> >> > >> >> > >> >> > >> >> -- > >> >> - Forrest > >> >> > >> >> -- > >> >> AF mailing list > >> >> AF@af.afmug.com > >> >> http://af.afmug.com/mailman/listinfo/af_af.afmug.com > >> > > >> > -- > >> > AF mailing list > >> > AF@af.afmug.com > >> > http://af.afmug.com/mailman/listinfo/af_af.afmug.com > >> > >> > >> > >> -- > >> - Forrest > >> > >> -- > >> AF mailing list > >> AF@af.afmug.com > >> http://af.afmug.com/mailman/listinfo/af_af.afmug.com > >> > >> > >> -- > >> AF mailing list > >> AF@af.afmug.com > >> http://af.afmug.com/mailman/listinfo/af_af.afmug.com > > > > -- > > AF mailing list > > AF@af.afmug.com > > http://af.afmug.com/mailman/listinfo/af_af.afmug.com > > > > -- > - Forrest >
-- AF mailing list AF@af.afmug.com http://af.afmug.com/mailman/listinfo/af_af.afmug.com
