Raising voltage from a power source (eg battery)?

Nazo

Nazo

2[H]4U
Joined
Apr 2, 2002
Messages
3,672
So I have a headphone amp I use with my computer -- a CMoy using an AD8620 OPAMP (this particular combination sounds incredible as if it were far better than a CMoy IMO and seems to suit my headphones very well.) I've been using 8xAAAs to get the voltage high enough, but I just recently bought a super-cheap battery + charger for my phone (Samsung Galaxy S3, so the battery is fairly beefy for a phone and while this one is a cheap made in China battery that probably delivers far less than in advertises -- especially considering how little it weighs) and it turns out that I got a better deal on Amazon for a two pack of a higher quality battery with a better charger...

I couldn't help but to think that it might well be a LOT more convenient to just use a single quality rechargeable battery from a phone that's this popular (it probably goes without saying that they'll be making batteries for this thing for a very long time indeed) and maybe possibly even grab another of the Amazon deal with the two higher quality batteries later on if this works out. It would probably be pretty easy to adapt one of the chargers to use a switch to provide power or charge depending on where it's set and just essentially use it as the battery holder as well, so it might theoretically work.

So I've heard that it's possible to raise voltage. I do of course realize this isn't 100% efficient (and that more current is required to compensate as well) and even if the use time is worse than the AAAs (which was I think as many as three or four months or so probably -- I lost track -- so even if I assume 1/3rd the time it's still very good) but this is actually a LOT more convenient even at the worst if it works. Does anyone know the specifics of how to actually do this? Could I raise a battery that's "3.7V" (though it seems they charge as high as 4.3 these days apparently) all the way up to 10+V? And, if I could, given the nature of the project (a headphone amp) would it produce any noise that could be picked up by such sensitive circuits? It probably doesn't take much at all really...
 
Yeah a charge pump can do that. You just have to see if the limited current that comes out of it is still enough. And yeah its not 100% efficient, but 90% is pretty good :D

Charge pumps use a clocks like 200MHz, and since thats nowhere near listening range, the amp should be filtering those frequencies out.

heres a data sheet of one: (as you see its with 100mA output) http://www.ti.com/lit/ds/symlink/lt1054.pdf
 
That does sound like the right sort of thing. Is the minimum supply voltage an issue? That thing recommends 3.5V as the bare minimum. I don't know if voltage continues to drop once the battery gets down to nominal ranges, but if it does the range between 3.5V and 4.25V or so does seem like the lifetime might be a tad more limited than I expected (though with these batteries that might still not be too bad.)

Is there anything of this nature that's pre-built in any way? I've never really had success building anything with chips myself, lol. (I don't count this CMoy since I just used a kit and even had a socket to put the chip in rather than soldering it directly -- the AD8620 is in a Brown-Dog adapter bought on eBay pre-assembled for me, lol.) Basics with capacitors and resistors don't usually present a problem, but so far everything with a chip that I have to build from scratch -- even simple voltage regulator designs -- just seems to end in something going wrong.

And yeah, output current isn't a huge issue. The AD8620 seems to be kind of inefficient (it seems to drain batteries a decent amount faster compared to, say, the all too popular OPA2132) but any inefficiencies in it plus the amount used by the headphones themselves probably still is only around 15mA or something really tiny like that (I don't really have the means to accurately measure it, but we're talking very little here! I know the headphones are extremely sensitive -- the highest number I could find in any review was a bit over 8mA for the 120 ohm model at the test voltage level which is supposedly considerably less efficient than this 50 ohm model, so I'm assuming some pretty massive inefficiency in making up that theoretical number, lol -- and the whole amp can still be driven for around a month or so with a single 9V battery.) Actually, it might be a problem the other way around. Would it count as a load for such a circuit given that we're talking about such a tiny load? Or would it go overboard and maybe even overshoot the intended voltage? Don't get me wrong, the design can handle some very high voltages and a lot of people use two 9V batteries in series instead of one, but that has to reduce efficiency, and I would assume maybe be bad in other ways (the gain is already pretty high as it is and I have to set the volume extremely low, so I definitely don't need more output voltage and, in fact, would like to get it pretty close to the ideal 10V -- eg that minimum 5V to the chip with the CMoy's virtual ground design dividing it in half -- if it can hold that in a stable manner since any more than that is unnecessary.)
 
Yes, you can do it. It's called a boost or step-up converter. Look up a datasheet for a MC34063 (like this datasheet), and it'll take you step-by-step through the process of designing the circuit. The circuit you want is on page 5 of the datasheet, and calculations to determine the values of the various components can be found on page 10.

I'm not sure why you'd need 9Vp-p for headphones, though--it seems to me you'd never run headphones off a signal that powerful...
 
I guess you didn't catch the posts above?

As for the voltage, you do NOT drive headphones directly from a DC power source! The voltage is to power the operational amplifier which in turn amplifies the line-level signal and drives the headphones. Also, the CMoy uses a virtual ground which effectively divides the power supply's voltage in half, so the 9V of one 9V battery at its prime is actually only 4.5V power supply to the OPAMP and insufficient (which is part of why I have 8xAAAs. I get more lifetime since they have more capacity but also the voltage is 12V at the start and takes a while to drop below the minimum.) Actually, I seem to be able to get away with lower than the supposed 10V (eg 5V to the chip) minimum before the sound starts clipping and generally sounding terrible, but 9Vs just weren't cutting it since they run out so quickly. Perhaps I got lucky and got an especially good chip twice in a row (but more likely they just have a bit of tolerance beyond what the official specs say, especially since these headphones are low impedance and fairly efficient, so pretty easy to drive overall.) So while I might have some leeway, I need to aim for that voltage range to power that chip sufficiently.

Actually, I haven't ever tried to hook up a voltmeter or anything, but from what I understand, the actual output to these headphones is probably actually less than one volt... (The specs say 112dB for 1KHz @ 1V. Ok, that's not the full range, but if I'm not entirely misunderstanding things, that means that since I'm obviously not crazy enough to listen to my music at 112dB volume levels, I'm using less than 1V...) Higher impedance headphones (some actually go up to as much as 600, and 300 ohms is not uncommon at all with quality headphones) need more voltage though which is the main reason so many designs actually go so high as to use two 9V batteries. Mine are 50 ohms however and really don't need that much voltage at all (the catch being that lower impedance headphones need more current which can sometimes be tougher for lower quality OPAMPs to handle -- though usually it's more an issue with analog amplifiers than digital ones.) The real voltage here is clearly being used by the chip for its work rather than for the actual output.
 
Sorry, I should have been more clear about my post. I guess I've been working with rail-to-rail op-amps lately. Yeah, I can see how a 5V supply would be insufficient, if the op-amp has a minimum output of GND + 0.5V and a maximum output of Vcc-1.5V.

If you're looking for a kit, you might start with a MintyBoost, and replace a couple components to get the output voltage you need. (if you need help, feel free to ask)
 
What would you suggest to double the output and to work with a higher input voltage? I need 10V out or higher for it to be ideal. Wouldn't such a big change basically mean a completely different set of components though?
 
Last edited:
Take a look at this page and scroll down to where it shows the schematic. To get 10V out, here's what you do with the kit:
1) cut the trace going from pin 6 of the controller
2) solder a 10kOhm resistor between pin 6 and pin 7 (or any other contact that's connected to pin 7) of the controller
3) solder a 10kOhm resistor between pin 6 and any of the pins it used to be connected to (before you cut the trace

Note that this is with a LT1302-5, which uses a 5V feedback voltage (which is great when you want 5V out, since it lets you eliminate two resistors). If you use the LT1302, you'll use a 4.7kOhm resistor in step 2 and a 33kOhm resistor in step 3.
 
Hmm... Looks like the kit sells for $20. Well, the PCB image there is for a single sided PCB (I was never good at schematics, but I can adapt a PCB design to my own different design, lol) so would it work out cheaper to adapt that design buying the parts separately and using a very small breadboard (I have several) then just replace his MAX756 with a LT1302 or LT1302-5 (is one inherently better than the other for this?) to get 10V for cheaper than that? I'd hate to spend $20 on a kit then replace and add parts to raise the cost even further...
 
Resistors are a few cents apiece, so it wouldn't cost much to adapt the kit. As far as I can tell, the only difference between the LT1302 and LT1302-5 is the feedback voltage.

If you were to put your own kit together, it'd definitely be cheaper than $20. I'd be surprised if the total cost of the parts were more than $10. Mouser or digikey can get provide you with all the parts you need.
 
Go on ebay and look for a DC-DC boost converter. Probably all of $2.50 built if you can wait two weeks for shipping from Hong Kong. LM2577 is the normal chip used and there's a pot on them to adjust the voltage to what you need. I've used them and they're just fine for this kind of application.
 
Well, the idea was to use the extra separate battery charger and one of the extra batteries I got for my Samsung Galaxy S3. All I'd have to do is make a switch for charging or something and I could use the charger as the mount for the battery as well. When charging is switched off it would just connect the battery directly to the power and when on it would connect to the charging circuit instead (the idea being to never share anything -- not even ground -- since USB can be so messy.) I'm not sure if it's a "protected cell" or not though.

Anyway, I have to admit I just haven't really gotten around to doing all the work on this. The AAAs have held up better than I thought they would and I'm not sure if it's worth all of the trouble or not at this point. I hadn't been listening to music as much for a while there though, so I guess I'll have to see how it goes in the near future.
 
Ok, so I've looked into it in more detail since that last post. First, I don't need to buy a separate battery. Like I said, part of the point of this was that I had extra batteries for my phone including one cheap one from China (with probably 1/2 the density of standard batteries) that I wouldn't mind tearing up (and if it really worked out it's surprisingly cheap to get third party batteries that are every bit as good as the stock one.) The DC-to-DC converter was surprisingly cheap though, so I did go ahead and grab one. I didn't realize how small it is either. I won't have much trouble fitting it together with the amp and charger (the charger will be the battery connector as well -- I'll have a toggle switch to switch it between charging mode and power to keep things really simple and cheap.)

I do have one question though. You say the CMoy should filter out any ripple. How sure are you that its own simple filtering will be sufficient to not add noticeable noise? Is there any really simple step I could take to filter the output without having to build any sort of real circuit or whatever?
 
The ripple from the DC-DC converter will likely be in the 50kHz (or higher) frequency, so it wouldn't affect your audio signal anyway. Bigger capacitance across the power supply for the op amp will reduce voltage ripple.
 
Ok, thanks for the help. That DC-DC converter looks to be too convenient and small to not at least try. I'll admit I'm a bit concerned about trying to always keep it a completely sufficient voltage. Well, these batteries might be a bit weak since they are nominally rated at 3.7V (though they get overcharged to 4.2+ it seems -- this is intentional on Samsung's part, so all I can say is I'm glad these batteries are very cheap and will be so common that I should still be able to buy them on places like Amazon for years after the SGS3 goes out of production. I've noticed that batteries stick around for a long time for phones that are popular enough. I can actually still buy batteries for the LG Chocolate3 I used to have, and that thing has been out of production for probably more than five years at the minimum...) Hopefully they'll hold up well, but there's only one way to really find out, lol.
 
Lithium-ion batteries have a nominal voltage of 3.6 to 4.2V, so the 4.2 you're seeing is nothing unusual. I've also been pleased to find readily-available batteries for older devices.
 
Well, that was kind of what I was saying about the voltage, however, to be more accurate, these batteries are set to overcharge all the way up to as much as even 4.35V (which is pretty much critically overcharged. Not quite enough to cause them to explode, but enough to significantly impact the battery lifetime.) I never found any way at all to make them charge to a lower value (I suspect it's actually in the batteries' management chips themselves given that when I use an external charger the result is much the same.)
 
You must log in or register to reply here.
Back
Top