The LiPo Project

Sep 26, 2005 at 8:58 AM Post #31 of 46
Ok, the low-voltage cutoff circuit works very well. I'm doing a board layout for a MINT-sized standalone version of it now. This will use conventional quarter-watt-sized resistors (as well as SMD semiconductors) for ease of sourcing and construction, and to keep the price quite low. When we move to a charger and a cutoff circuit on the same board I'll switch the non-precision resistors to SMD to save space.

I tested it this weekend using a TL081 and a variable supply, and even using 1 per cent resistors selected with a 4-digit meter it was easy to get it to trigger right on the 3.0V point. I couldn't see any noise on the scope (some very low-level noise was visible at the voltage reference cathode pin, but not anywhere on the supply, at least above the inevitable noise level occurring on an open-air protoboard circuit).
 
Sep 26, 2005 at 9:09 AM Post #32 of 46
The only charger chip that seems to work for our purposes is the bq24105 from TI (usable at 11.1V, few external components, available). But it only comes in a QFN-20 package -- that is, a 20-pin flatpack with no leads sticking out. A couple questions:

- Has anyone worked with QFN in DIY work? Is it possible to solder without resorting to exotic toaster-oven methods?

- Has anyone worked with this package in EAGLE? The standard parts library doesn't come with anything in QFN, and I don't really feel up to designing a new package... Is there anything out there that will make a PCB design possible?

EDIT: I'm also open to suggestions as to different charger chips... I'd been considering the MAX846A, in a much friendlier SOIC-14 package, but it doesn't seem workable with 3-cell, 11.1-V applications -- it is switchable between one and 2 cells, but not 3. Anyone know of any others?
 
Sep 26, 2005 at 5:13 PM Post #33 of 46
a little late to this discussion... but this place seems to have a selection of protection boards for various size li-ion packs, as well as reasonably priced smart chargers and li-ion, as well as poly li-ion cells and packs. if you were really lazy (and could build a low voltage amp with a single cell), the 1st gen ipod battery has the IC on the battery, and a single cell charger is cheap. otherwise, the PCBs and poly cells seem like a reasonable option for a reasonable price. can;t vouch for the quality though
confused.gif


you could throw together a 170 maH four cell (14.8V) poly li-ion pack with the protection circuit and a smart charger for less than $60. and the size would be about half that of a credit card and 4mm thick. nice.
icon10.gif


http://www.batteryspace.com/index.as...S&Category=711

http://www.batteryspace.com/index.as...S&Category=489
 
Sep 26, 2005 at 9:49 PM Post #34 of 46
I've never tried soldering QFN, QFP is alright. I think QFN does have leads sticking out, but not very far and on very tight pitch

The relevant Eagle library is smd-ipc
 
Sep 26, 2005 at 11:10 PM Post #36 of 46
Dougi: Yes, there are a number of ok pre-made options... I'm right now using a LiPo charger on a 2" square board (though I don't want to use it heavily until I have discharge protection). None of them are exactly perfect for built-in headphone amp use, which is our goal. But they're good for experimentation, and if you're happy with

It is true that some cells have built-in discharge protection. You have to watch it though, becaue a) Most of he ones in desirable amplifier voltages (over 10V) don't, and b) if you take three 3.7V cells with discharge protection and operate them in series, there's a good chance that the discharge shutdown circuits in each cell won't work -- I think it's really only safe if you have something in the amp box itself to keep a dangerous discharge from occurring.

Guzzler -- It sounds like that package could be a real headache. I'm tempted to switch to the MAX745, which does require a few more external components but is pretty easy to implement and a really good versatile charger chip for headphone-amp purposes: easily switchable from 1 to 4 cells, up to 24V input voltage, doesn't heat. It's an SSOP-20 chip and is available at RS in UK or Newark in US.... Here's datasheet:

http://www.maxim-ic.com/quick_view2.cfm/qv_pk/1661



http://pdfserv.maxim-ic.com/en/ds/MAX1772.pdf
 
Oct 13, 2005 at 10:42 AM Post #37 of 46
Ok, after a bit of a delay (to teach myself EAGLE), here's a working schematic and PCB fragment for a low-battery cutoff. A charger, to go on the same MINT-sized board, is in the works.

Any critiques would be most welcome. I'll happily send you the EAGLE files if you're interested.

Again, this is meant for use with 11.1-V lithium polymer batteries of the sort that are readily available from radio-control hobby shops, typically in 2200 MAH sizes. (These actually charge to around 12.3V using standard LiPo chargers; 11.1 is a nominal number).

You MUST use a circuit like this if you want to have lithium batteries in a headphone amp. It cuts the battery out of the circuit, using an ultra-low-resistance MOSFET switch, if it drops below 9V.

This circuit is ultra-low current drain, in the single microamps. It uses more or less widely available, inexpensive components. I've used non-SMD resistors because the first 7 resistors need to be 0.1 per cent tolerance and they're in megaohm values (in order to keep current draw to a bare minimum), which is a combination that isn't commonly available in SMD. The IC, the voltage reference and the MOSFET are all 8-pin SOIC devices (even though the voltage reference is really a simple 2-pin device). I've mounted the latter two on the back of the board, behind the op amp, so the whole thing occupies very little space.

By PCB fragment, I mean this uses up the left third of a MINT-sized PCB. There seems to be enough room left over for a MAX745-based charger (the 33mh inductor is the largest item involved, aside from a biggish capacitor that may or may not be strictly necessary).

On the other hand, I could use the other half of the board for a) a battery-level meter using the venerable LM3914 and b) a circuit that turns on the amp when there's music present on the input, and off when there hasn't been music for 10 or 20 minutes -- see my thread on this circuit.

That's why I'm posting this fragment -- so if there's interest, people can combine it with other things they've got on the go.

Please let me know if this board design seems at all reasonable.
 
Oct 13, 2005 at 4:18 PM Post #38 of 46
It definitly looks like a start, but I think that 1206 smd resistors would be much more suitable for this project. I dont know what your sources are like, but at Mouser, there are 0.1% resistors into the Mohm range, for about $0.70 each. Also, it wouldnt be that hard to order several 1% resistors and match them, a bit of a pain, yes, but would definitly aid in shrinking the board, especially when the charger is installed. What kind of current draw increase would we be looking at if one used only 1% unmatched resistors? The MAX745 appears to be a Li-ion only charger, is this a problem for us? The large cap you are talking about, I assume this is C1 in the MAX745 datasheet? It doesnt appear to specify the capacitor type required, or voltage, but I would tend to assume that a relatively linear cap with a voltage rating above that what the charger outputs is required. Panasonic SP looks like it would work, except that it only goes up to 8v in 68uf sizes, and they are hard to source. I dont suppose tantalum would be appropriate either. Also, you talked about using a 2200mAH 11.1v LiPoly, I havnt seen any batteries of this capacity that would fit in any case that I would consider portable. 1200mAH 11.1v is about the largest size I would want to carry around, unless you have a source for some that are much more compact. You also have a via right in the middle of the pad for the ref, if at all possible, could you move it, I would suggest taking it in so it is fully under the body of the IC. Glad to see this project progressing, any work on the schematic of the charging section yet?
 
Oct 14, 2005 at 9:27 PM Post #39 of 46
Excellent notes, Cetoole, thanks. I'll have the charger half of the PCB done soon -- it's just that I have to design the MAX745 for EAGLE, which means learning how to assign pins to a pre-existing case (SSOP-20). A pain, that.

Quote:

I think that 1206 smd resistors would be much more suitable for this project. I dont know what your sources are like, but at Mouser, there are 0.1% resistors into the Mohm range, for about $0.70 each.


I'm still considering that possiblity. Mouser does indeed sell .1 per cent 1206 resistors in a 1M value (though not in higher megohm values) in single quanities... but I can't find anyone else in North America or Europe who does. I'd still have to take the same approach, stacking 1M resistors, though SMD would save a small amount of space.

But for some of the non .1 resistors, SMD would actually take up more space. You can easily get a 5M resistor with leads, but an SMD one is a bit of a hunt. So, in brief: If the charger half of the circuit proves tight, I'll switch more parts to SMD.

Quote:

What kind of current draw increase would we be looking at if one used only 1% unmatched resistors?


The issue isn't current draw, it's low-voltage cutoff point. As the application note says, a difference of 1% could be the difference between 9V and 9.3V, or several hours' listening... on the other hand, this isn't such a big deal for a headphone amp. I breadboarded this circuit using 1% resistors (or worse), and it shut off VERY close to 9V without hand-picking resistors. So I'm not so sure if precision is so important here... maybe leave it as an option.

Quote:

The MAX745 appears to be a Li-ion only charger, is this a problem for us?


It works fine for li-polymer -- all Lithium-type chargers use the same scheme, current charging switching over to voltage charging. The voltage thresholds are exactly the smae. Li-poly batteries are actually a lot more resilient than li-ion, which are notoriously fiddly and tend to have "memory," which li-poly don't.

Quote:

The large cap you are talking about, I assume this is C1 in the MAX745 datasheet? It doesnt appear to specify the capacitor type required, or voltage, but I would tend to assume that a relatively linear cap with a voltage rating above that what the charger outputs is required.


Yeah, the 68uf cap across the battery. I'm not so sure how necessary it really is. Or how much smaller I can go. Since the cap will still be connected across the battery when it's powering the audio circuit, there's some reason not to use tantalum. Of course, a good electrolytic cap across the battery supply is a good thing to have in an audio circuit -- batteries are noisier than you'd think -- but I'd rather not have it here, for space reasons.

Quote:

Also, you talked about using a 2200mAH 11.1v LiPoly, I havnt seen any batteries of this capacity that would fit in any case that I would consider portable. 1200mAH 11.1v is about the largest size I would want to carry around, unless you have a source for some that are much more compact.


I'm prototyping with a 2200mAH battery -- it's just less than 20mm high, not exactly pocket-sized. Fits nicely into the sort of case used by some people for PIMETA amps. I'm going to use it for a nice amp with a big Alps pot and a Firewire jack so I can boost-charge my IPod.

But once I've got this small board nailed down I want to do one with three of the tiny 900mAH mobile-phone cells -- these are available from Asian merchants on EBay for about $2 each. (search for Sony Eriksson K750I -- it's the battery for that phone, about the thickness of an After 8 and smaller).

Quote:

You also have a via right in the middle of the pad for the ref, if at all possible, could you move it, I would suggest taking it in so it is fully under the body of the IC.


Yeah, good point. I'll find a better place for the via.
 
Apr 2, 2006 at 10:58 PM Post #40 of 46
Any progress on this?
RC
 
Apr 2, 2010 at 7:48 PM Post #45 of 46
AFAIK it never reached prototiping or production, but I could be wrong.

Given that a few years have passed and that lipo technology has improved much since then, it would be a neat idea to resurrect this topic and go http://www.head-fi.org/forums/newrep...p=6527637ahead! There are some nice R/C lipo batteries in very small sizes.

I'm envisioning that we could put around 400-600mA at around 9v into the space of a standard 9v battery, including charging and protection/monitoring circuitry.

That would be ~double the current capacity of todays 9v batteries.

Making this available as a project (including pcb layouts, schematics, BOM's and so on) so people could add it to their amp or dac pcb's would be awesome.
 
You must log in or register to reply here.

Users who are viewing this thread

Total: 4 (members: 0, guests: 4)
Back
Top