I'm building a "transportable" battery powered dac+amp combo. The energy source will be one or two 14.8v rechargeable batteries (the ones jumbled together in the form of a square)
As far as I'm aware, power supply is one of the most important components of the analog sections. I'd like to know the list of battery types and their load driving capabilities so I can start designing the power supply network.
Online searches lead me to multiple competing opinions. One says batteries are better suited to constant load. Another says batteries are better suited to pulsed load so that the space where it doesn't operate gives buffer for the depletion region to regenerate. This is apart from the linear models where we have internal resistance models etc.
I'm not super well versed in batteries, but I kind of have two approaches I can think of. In case the battery can sustain a load very well along with transient powers, I could be using a LDO + some filtering and that would be it.
If not, I shall be using two batteries and a switching logic + active components to generate somewhat kind of an AC like waveform (one battery doing positive, one doing negative) and then an active power filter+ rectifier to generate stable dc. Since a single battery is not operating continuously, it might be able to get more stable if that aberration exists, and make the non linear issues of the batteries disappear when seen at the power input end. I could also possibly recharge the battery when in use without disturbing fidelity. The rectification circuit is present in my burson fun where it does a switching operation at about 170khz and then rectifies it, hence pushing any ripple noise etc well beyond the 170khz band.
As far as I'm aware, power supply is one of the most important components of the analog sections. I'd like to know the list of battery types and their load driving capabilities so I can start designing the power supply network.
Online searches lead me to multiple competing opinions. One says batteries are better suited to constant load. Another says batteries are better suited to pulsed load so that the space where it doesn't operate gives buffer for the depletion region to regenerate. This is apart from the linear models where we have internal resistance models etc.
I'm not super well versed in batteries, but I kind of have two approaches I can think of. In case the battery can sustain a load very well along with transient powers, I could be using a LDO + some filtering and that would be it.
If not, I shall be using two batteries and a switching logic + active components to generate somewhat kind of an AC like waveform (one battery doing positive, one doing negative) and then an active power filter+ rectifier to generate stable dc. Since a single battery is not operating continuously, it might be able to get more stable if that aberration exists, and make the non linear issues of the batteries disappear when seen at the power input end. I could also possibly recharge the battery when in use without disturbing fidelity. The rectification circuit is present in my burson fun where it does a switching operation at about 170khz and then rectifies it, hence pushing any ripple noise etc well beyond the 170khz band.
What battery chemistry are we talking? LiIon, LiFePO4, lead acid / gel cell?
What sort of output power do you need? Class D or AB amplifier?
Like everything involving chemistry, batteries are more or less slow. You want no shortage of parallel capacitance for load peaks. Look at what the car audio guys are doing - they use massive supercapacitors. Since you probably just want a few dozen watts, you'll probably get along with more modest sizes, but I'd still expect >10000 µF.
Burson Fun appears to be a headphone amplifier, Class A no less? Looks like it can be supplied with a PC Molex, too. Apparently the first thing it does with its input voltage is run it into 2x 2 DC/DC converters, probably a step-up job plus an inverter each to obtain a split power supply of +/-18 V or so.
If that's the case, it shouldn't be too picky. I'd say get a 1000-3300 µF / 25+ V capacitor, a DC/DC step-down board (maybe a common 2.5 A type), plus some sort of battery management / charger, and you should be set.
DC input is rated 5 A @ 12 V and the thing apparently ships with a 6 A power supply - does it really need that? Maximum power output is 2x2.1 W into 32 ohms, so that's at least 16.8 W of power input (Class A push-pull efficiency = 25% max) and probably closer to 26 W. (The idle current of at least 180 mA would mean 13 W of idle power dissipation in output transistors alone.) That would be about 30 W in terms of linear power supply and maybe twice that in a SMPS to account for a nervous over-current protection.
That said, if it really were that power-hungry it would get quite nice and toasty. The enclosure is not that big.
Do you have any way of of determining what its actual idle power consumption is? We are potentially approaching the magnitude of a QRP rig or portable boombox here. That's going to be quite a chunky battery if you want some decent play time. It would still be "portable" in the literal sense (I mean, the definition of "portable" in the late '70s was "one man can carry it and it'll fit underneath an airplane seat"), but have you seen portable audio devices from the last 2 or 3 decades? Do you really want to be pressing a device that is in no way optimized for power efficiency into a portable application?
What sort of output power do you need? Class D or AB amplifier?
Like everything involving chemistry, batteries are more or less slow. You want no shortage of parallel capacitance for load peaks. Look at what the car audio guys are doing - they use massive supercapacitors. Since you probably just want a few dozen watts, you'll probably get along with more modest sizes, but I'd still expect >10000 µF.
Burson Fun appears to be a headphone amplifier, Class A no less? Looks like it can be supplied with a PC Molex, too. Apparently the first thing it does with its input voltage is run it into 2x 2 DC/DC converters, probably a step-up job plus an inverter each to obtain a split power supply of +/-18 V or so.
If that's the case, it shouldn't be too picky. I'd say get a 1000-3300 µF / 25+ V capacitor, a DC/DC step-down board (maybe a common 2.5 A type), plus some sort of battery management / charger, and you should be set.
DC input is rated 5 A @ 12 V and the thing apparently ships with a 6 A power supply - does it really need that? Maximum power output is 2x2.1 W into 32 ohms, so that's at least 16.8 W of power input (Class A push-pull efficiency = 25% max) and probably closer to 26 W. (The idle current of at least 180 mA would mean 13 W of idle power dissipation in output transistors alone.) That would be about 30 W in terms of linear power supply and maybe twice that in a SMPS to account for a nervous over-current protection.
That said, if it really were that power-hungry it would get quite nice and toasty. The enclosure is not that big.
Do you have any way of of determining what its actual idle power consumption is? We are potentially approaching the magnitude of a QRP rig or portable boombox here. That's going to be quite a chunky battery if you want some decent play time. It would still be "portable" in the literal sense (I mean, the definition of "portable" in the late '70s was "one man can carry it and it'll fit underneath an airplane seat"), but have you seen portable audio devices from the last 2 or 3 decades? Do you really want to be pressing a device that is in no way optimized for power efficiency into a portable application?
Thank you very much for the detailed reply. Yes I'm looking at maybe 2-3 watts into 32 ohms or so load for headphone amplification. My headphone is shure srh1540 and hd800 and with high dynamic range music, I'm near maxing out the pot on my burson fun so I do think I'll need to supply 1 watt peaks. I plant to purchase srh1840 which is a lot more demanding and hence the need for 2-3 watts.
I can just repurpose the burson fun using its molex for a beginner trial but my bigger desire is to build an amp of my own, preferrably current feedback, like the ones from my apogee groove but in a more powerful unit. Regarding the burson fun, i haven't been able to measure it's input power requirements, but it's far from getting hot. I doubt it consumes that much power.
What battery chemistry are we talking? It was actually part of my question. I'd love to know more about different battery chemistry and see how it works.
Regarding battery life, I'd be more than happy if it could sustain 2-3 hours on a charge. I rarely listen for more than an hour or more a day and I almost never have sustained listening sessions above 3-4 hours. I want it to be just movable so I could listen in my garden or balcony while working on my laptop. I hate being strapped to a wall outlet.
I can just repurpose the burson fun using its molex for a beginner trial but my bigger desire is to build an amp of my own, preferrably current feedback, like the ones from my apogee groove but in a more powerful unit. Regarding the burson fun, i haven't been able to measure it's input power requirements, but it's far from getting hot. I doubt it consumes that much power.
What battery chemistry are we talking? It was actually part of my question. I'd love to know more about different battery chemistry and see how it works.
Regarding battery life, I'd be more than happy if it could sustain 2-3 hours on a charge. I rarely listen for more than an hour or more a day and I almost never have sustained listening sessions above 3-4 hours. I want it to be just movable so I could listen in my garden or balcony while working on my laptop. I hate being strapped to a wall outlet.
You haven't identified the equipment being powered, or its power supply requirements.
KISS - keep it super simple - a linear regulator to provide a constant supply voltage to your equipment, with an indicator or auto-cutoff for a suitably low battery voltage. It doesn't matter what the battery is then, except for having enough voltage and energy, and you having a suitable charger to maintain battery health.
KISS - keep it super simple - a linear regulator to provide a constant supply voltage to your equipment, with an indicator or auto-cutoff for a suitably low battery voltage. It doesn't matter what the battery is then, except for having enough voltage and energy, and you having a suitable charger to maintain battery health.
Yes, the Chinese amp is a very different beast from the SI unit we know and love. Watts too in many cases. My advice is never buy anything important that doesn't have a datasheet, this applies to cells and batteries too. Of course I refer to the fraudsters not the legit. Chinese brands and suppliers, but the counterfeits are everywhere too so its very hard to know unless you buy via specialist importers who source the real stuff and protect you from the cowboys.
You still haven't said what the load is for the batteries, or the run-time between charges you are wanting...
Its likely an easy job to match suitable batteries given this information. For instance some lithium chemistries can discharge at 20C or mote, current handling won't be an issue.
The one thing you absolutely need is over-discharge protection, without this you'll risk wrecking your batteries permanently.
Its likely an easy job to match suitable batteries given this information. For instance some lithium chemistries can discharge at 20C or mote, current handling won't be an issue.
The one thing you absolutely need is over-discharge protection, without this you'll risk wrecking your batteries permanently.
No, this is pointless and complex and unnecessary. The key properties of batteries are the internal resistance and polarization, the former limits the peak current transients, the latter the continuous current rating (not all chemistries polarize). Datasheets should tell you what you need to know.
Thank you very much. I went around articles, and also saw that NiCad have measurably lower noise than the other cells (apart from Hg cells). However I am a bit afraid to use NiCad due to risk, is there any other second alternative you would recommend, that are fairly safer to dispose and handle while still having low noise? I want it to deliver +/-7V (ie 14.4V battery or 2x 7.2V battery) and current output needs to satisfy 150mA peaks to my headphone load, so overall peak requirement including rest of the resistors etc might be about 250mA, and static constant one might be maybe 40-50mA I guess (too hard to make a guess). (or +/- 9V batteries regulated to about 7 or 8V, if the opamp can handle +/-8V source, the ad8397 I am planning to has been described to get stability issues at high voltages)
Last edited:
Thank You. I have been reading more over the last few months and understood that switching is unnecessary. My load is my headphone which is 45 ohms (almost a flat line, a minor bump to 50 ohm with a fairly wide Q) and consumes near 1 watt peaks during dynamic passages (provided my amp is high dynamic range). So I would ideally want a near +/-7V rail which would mean a 14.4V battery, or 2x 7.2V batteries in +/- configuration. Regulation aside, I might get an output of 6V which I think should be fine for a portable, only minor hit to the overall power. (or +/- 9V batteries regulated to about 7 or 8V if the opamp can handle it)
The amp I am thinking of building using AD8637 opamp since it seems to be able to do 310ma current, quite a bit of headroom over the around 130-150ma peaks needed in my case. I would likely try to optimize it for loads of 40 to 80 ohms. A moderately high output impedance wouldn't be a problem with my headphone since it presents a near flat impedance profile. Other suggestions would also be welcome.
Last edited:
Thanks. I never considered this aspect. My dac already puts out 5v line level out. My gain would not be that high, probably 2x or so. I'll get a NiMH or Li Ion from the store nearby, something with low internal resistance(high current), probably repurpose some camera batteries.
I am not building a dac anytime soon, but in case I build one, would I need to consider battery noise or that is also tolerant?
I am not building a dac anytime soon, but in case I build one, would I need to consider battery noise or that is also tolerant?
Maybe my dusty-old-fartness is showing, but I don't see any possible way that battery noise could pose a significant problem, no matter what the audio signal level. There are so many other much worse (higher amplitude / harder to suppress / filter / shield for) sources of noise that plague electronics designers .. pretty sure that your design effort would be more usefully spent on other issues.
rayma, thanks - first hearing of those. We have a clutch of NiMH's in various duties and their self-discharge rate is an ongoing aggravation.
Cheers
rayma, thanks - first hearing of those. We have a clutch of NiMH's in various duties and their self-discharge rate is an ongoing aggravation.
Cheers
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.