My first basic question is regarding the best locations for storage capacitors on a board. I am not concerned with de-coupling caps (i.e. the common 0.1uF approach) for various IC's on the board, rather more about bulk storage. From the reading I have done, it would appear that storing these capacitors near the voltage regulator (the shorter the distance the better) is preferable to having them distributed across the board. On the other hand, I would think that having localized storage (near potential spike loads) would help voltage stability by minimizing the distance that the power has to travel to the point of use. Or should I not be concerned with this and just put the largest storage cap that may be needed near the LDO and call it a day? FWIW, the boards in question are pretty small (i.e. 10x10cm) and the power lines are generally 32 mils wide and usually form a grid across one side of the 2-sided board. 

My second question regards the LDO tunnel of death, i.e. the areas of stability and instability that are a function of LDO design, ESR, and input/output capacitance. Depending on the LDO and the manufacturer, capacitances beyond the minimum indicated in the data sheet are encouraged as promoting stability. I also gather that some LDO designs (presumably older ones) need some ESR and hence need to rely on electrolytic or tantalum output capacitors for output smoothing, while other (newer?) designs have eliminated the need for ESR and rely on ceramic capacitors on the output stage instead. However, few manufacturers publish a graph that shows the allowable ESR vs. capacitance curves (aka the tunnel of death) and even when they do, these charts are usually limited to the recommended range of output caps, not for output caps that are larger than the minimum recommended by the data sheet. Usually, the best one can hope for is for LDO manufacturers to publish ESR and capacitance recommendations and little else.

FWIW, I like placing a relatively large capacitor on the output (i.e. 470uF) of my LDOs to minimize the potential for spikes and so on. However, a 470uF cap is well beyond the minimum output caps recommended (and in some cases discouraged) by LDO manufacturers. So what is the downside of a large tantalum or aluminum electrolytic capacitor on the output of an LDO? And how worried should I be if I place the recommended output cap in addition to this storage cap about ringing, etc.? Many thanks for any insights!

Mihun,

Nice job running OSX on a Dell. Must have been an interesting project also!

One of my funnier experiences with the (generally awesome) German postal system was sending a gift package to my god-daughter over there... except it didn't arrive... instead, the mother got a notice to pay up... long story short, the Droid at the customs hall took a look at the contents listed on the customs form, noted "gift", read it in German (in which case you're shipping 'poison') even though the package came from a English-speaking country and items shipped internationally are always declared in English, and decided that a duty applied.

A terse letter from the mother regarding the English-reading skills of the droid in question unlodged the package without the need for duty payments. Naturally, without comment.

Hi all,

Just wondering if any of you have used the MCP3901 before or not? It's an analog front end for energy measurements that looks like a neat intermediary chip between dedicated energy IC's like the AD7753 from Analog Devices and general-purpose dual-channel ADC's like the ADC122 series from National Semi/TI, for example. The device is interesting in multiple ways that I find quite intriguing:

1) It doesn't feature a bazillion internal registers and impossible (for a hobbyist, IMO) calibration routines like the AD7753. Just a few internal control registers that mere mortals can handle.
2) It does feature a PGA for each channel (up to 32x gain), a Voltage Ref In and Out, a phase compensation register, up to 256x oversampling, an external clock, dithering, and a data ready pin (for interrupt-driven data sampling).
3) The analog end runs on 4.5-5.5V, the digital end can handle everything from 2.7-5.5V. This makes interfacing with the chip a lot simpler, regardless of MCU used.
4) It uses SPI, which I dislike, but there is a library available to make configuring / listening to the chip easier.

The data sheet suggests that if you put a 8MHz crystal on this thing, set up the PRE/OSR registers appropriately, and you'll allegedly get 7812 samples per second for both channels with 14 ENOB... and if you can live with 3.9ksps per channel, up to 14.9 ENOB... all in a chip that costs less than $2.5 ea in single quantities. Seems like a pretty awesome solution, wonder why not more people are talking about / using this chip?

Along similar lines, the MCP3911 offers 3.3V-only operation but a much wider latitude re: oversampling, gain calibration, offset calibration, etc.

32 TQFP, but in a 5mm x 5mm format. Leaves 0.5mm pitch per pin and even with a stencil that is a PITA!!!

Reason being that unless you have just the right type and amount of solder paste, there is so little solder mask between pins to prevent shorts. And when you try to clean up accidental solder bridges with anything but a robotic helper and a giant magnifying glass, you have the joy of observing pins wandering from pad to pad as you brush them with your soldering iron when you try to suck up / drag solder the excess away. Thanks but no thanks, I am never touching those types of chips again, if I can help it.

I note that MinuteMan used to make a UPS with ~12V unregulated battery output. However, that model appears to be discontinued and much of the electronica out there that is UPS-protected really wants regulated 12VDC and 5VDC inputs (i.e. like a ATX power plug into a motherboard). In an ideal world, the manufacturers on both sides (i.e. CPU as well as UPS) would come to an agreement for a standardized plug that offers several voltage outputs and a means to communicate between the devices to establish just what needs to be sent, how much can be sent, etc.

I really like the Pico line of products, as they also offer the option of efficiently running all sorts of standard PC motherboards off of unstable power sources like boat, car, and like systems.

Have a look at the data sheets I linked to above re: the Pico UPS and the Dimension Engineering gear, I think they should answer your questions.

The Dimension Engineering DC-DC converters are pin-compatible (but larger) than traditional LM780x voltage regulators and they do not need external capacitors (though I still gave them one on the input side). What I like about these DC-DC converters is that they are easy to adjust with a high-quality rheostat. They appear to be very stable re: output, i.e. little to no difference in voltage whether an iPhone is drawing an ampere of power or there is just the trickle of power running through the LEDs.

The datasheet for the Dimension Engineering unit indicated a range from 5-33V. The alleged conversion efficiency for my application appears to be about 85% and 90%+, respectively.

Very simple circuit, I suspect that looking at the board file is easier though than the schematic (sorry!)

FWIW, the empty rectangles on the board picture are the PTCs. The resistors are show their resistance (not name). All components are through-hole by design so I wouldn't have to bother with a stencil or fiddly soldering.

HTH

For a simple solution, consider line-voltage thermostats for sale in the USA that are designed for electric baseboards. Fairly cheap, can handle high loads (though the inductive kickback of a compressor may present a problem for any device interrupting its line service). Once the temperature setpoint is exceeded, the thermostat will 'open' the line, no microprocessor needed. This way you can directly interface the compressor with the thermostat. Single-pole versions are good for low-hp, 120V compressors, while the dual-pole versions can be used for 230V service. A quick google scan indicates unit prices ranging from about $15-26 for manual versions.

A downside of a manual thermostat is the (usually limited) temperature range. 80F may be below your cutout temperature. On the other hand, I'm not sure it's a good idea to run a compressor within a space with a very hot air supply (i.e. much above 90F).

Now, if you want to insist on micro-processor solution, one additional option is a watchdog timer that automatically resets the microprocessor unless the microprocessor program is running normally (and updating the values that the watchdog timer is accumulating). Several GPRS shield solutions use such watchdog timers since sometimes the gears get gummed up and the microprocessor hangs.

True, Digikey tends to charge for minimal orders. I usually easily exceed the $25 minimum and I really appreciate the fast delivery (usually a day or two) and the very reasonable prices. IIRC, their minimum shipping charge is less than $3 for USPS delivery in a nice sturdy box, etc. I guess it comes down to time value of money... it's not worth $5 to deal with Newark on top of Digikey to me. E-Bay can be even less money but the delivery times can stretch into a month or two, depending on how much the DHS folk want to sit on the customs clearance forms instead of processing them... and thanks to the sequester, that aspect is only going to get worse. I sincerely wonder what the USPS and other shippers will do with the building backlog that a slowdown in processing would create at airports, etc.

Using the royal mail is like challenging Dirty Harry to a shootout. Do you feel lucky, punk, do ya?

Seriously... like many aspects of life in Britain, expect to pay top first-world coin for service that would make most third-world service providers blush with shame. Even in the deepest darkest post offices of India (with their legendary red tape and all that) I always had faith that the stuff would arrive - and it did. Ship packages from England or France and you're taking chances. 7 out of 9 packages shipped after B-school were meticulously opened with a razor knife, looted, and then re-closed using mailing tape made exclusively for the postal service (i.e. not available to the general public).

Given where morale is headed in the USPS system, I can only hope that the postal police will keep that kind of behavior in check in the US.

FWIW, I thought the presentation was very interesting. Not that I understood more than parts since I have yet to program in assembly, but the issues with reverse-engineering are near and dear to my heart - I do it for appliances and industrial equipment on a daily basis. Their efforts are commendable just as archeology is a commendable science - not necessarily because it's super relevant but because it helps explain some obscure aspect of life past - such as early microprocessor design that can't be replicated easily because the documentation is largely gone.

Newark is a vendor of last resort. They may stock a fair amount of kit not found elsewhere but the prices are very high, based on the (admittedly anecdotal) comparisons I have made between them and Digikey. The sort of experience described above would make me a former customer - life is too short to pay a premium for crummy service.

Hi All,

Just wanted to report on my latest iTead/iMall acquisition, a stainless stencil to accompany two boards I was ordering also. The stencil came mounted on a (welded!) sturdy Al frame and nicely sealed against the edges (white rubber cement and Al-Tape on the other side). I simultaneously ordered two 10x10cm boards and asked them to merge the two top pad layer Gerbers into one stencil, as the allowable area is 20x10cm, i.e. perfect for this application.

I am very grateful to iTead, which did the merge and also sent me a confirmation PDF. The stencil came with the boards, all are beautiful. The stencil is the 'chemical etch' kind, i.e. the one advertised for products that do not require a fine pitch, while the laser stencil is advertised for high-precision applications. FWIW, one of the parts on the stencil is a 28-pin TSSOP, and the holes are perfect (0.65mm pitch).

While this service is not nearly as quick as Pololu can be with its laser-cut Mylar stencils, the price (if you buy a 20x30cm stencil for two 10x10cm parts at a time) is virtually the same. Don't get me wrong, I am very happy with Pololu and its stencil service - I recommend it to everyone - but iTead has just given me a very good reason to send them my stencil as well as my PCB business.

FWIW, packaging was a sturdy cardboard box and shipping to NA using DHL was $28, IIRC.

Hi all,

About a year ago, a neighbor's tree interrupted our power service for the better part of a day. Since the APC power supply that was supposed to keep the phone lines and cell phones charged had several annoying features, I decided to cast out for my own UPS for the house phone as well as the cell phones. The APC UPS is built for a totally different purpose - Lots of circuits, inductors, etc. to produce the semblance of a sine wave from a DC power source. The crazy thing is of course that for most applications all that AC power is turned back into DC, with the added bonus of an annoying recurring beep, vast inefficiency at lower power operation, and hence very short runtime despite the use of two 7Ah 12V SLA batteries inside (cabled in series)

So I decided to take a different route, electing to use a PICO-UPS charger board (green) for the battery and a small 5x5cm board (blue) of my own design to distribute power to the phones. The black phone jack on the left of the board is for the house phone (6.3V) while the dual USB jack provides 5VDC and features the resistor divider that adafruit was so kind to document to enable Apple users to recharge their iPhones. All outputs are protected with individual PTCs, feature individual LEDs (to alert me to a PTC shutting down a circuit) and two Dimension Engineering DC-DC power supplies provide the respective DC outputs. An external wall wart provides 15VDC @ up to 1.6A to charge the battery and provide power to the phone board.

I really like the Pico-UPS board so far thanks to its small size and very cool operation. My only improvement suggestion would be to change the power inlet from a .187" blade design to something compatible standard wall warts (i.e. 5.5mmx2.1mm center positive jack). But that is a totally minor issue. I am also super happy with the Dimension Engineering DC-DC converters (here and here), among others. They are well-designed, stable, and allow excellent adjustment of the output voltage. Yet to be implemented is a safety feature (i.e. 2A inline fuse on the battery line) and an enclosure.

Yes, there cheaper ways to do all this. Our friends at Dimension engineering have since released more power supplies, etc. But the 3A charger has ample capacity to charge two iPhones at 1A each while also allowing the house phone base (DECT) to continue to operate. Unlike the APC approach, there is no beeping in UPS operation (just some silent LEDs), the conversion efficiency is on the order of 75%, and the battery will be super easy to replace as needed in the future. FWIW, Pico UPS also offers units with programmable outputs, allowing enterprising users to run devices like routers off a similar setup with a stable voltage.

Lest it come across that way, I am not trying to bash APC and their lines of of products. Rather, I believe that there is a market for UPS's that provide efficient power at a regulated 12VDC and 5VDC output to complement what is already out there. Unfortunately (and this will likely kill the idea) industry in turn would have to agree on (a) standardized connector(s) to allow direct hookups and hence for devices to bypass the need for AC-DC-AC conversion. However, imagine the benefits for all sorts of electronic devices, from lower potential EMI to longer run time for the same size battery.