Hi All
I've recently undertook a project of building my first amp.
I plan on using the AA-AB32191 2 x 300w Class D amp.
Looking over the specs this will need 52v @ 16amp.
This is quite a lot of oomph and finding a PSU that supplies this kind of power is proving expensive (more than the amp itself).
So this has forced me to consider other options - one being to build a 14s2p lipo battery using 18650 lipo cells.
This would supply more than enough oomph.
The problem is now how much do I need to worry about keeping those cells charged?
I plan on using a cheap lipo charging circuit to keep them topped up.
Due to the way an amp plays music, ie, it's really only the occasional full volume bass that will push the amp and PSU to the max, I'm thinking I don't need the full 16amp current to keep the batteries topped up?
I'm curious if anyone has created a PSU out of lipos, and if so, any gotchas?
I've recently undertook a project of building my first amp.
I plan on using the AA-AB32191 2 x 300w Class D amp.
Looking over the specs this will need 52v @ 16amp.
This is quite a lot of oomph and finding a PSU that supplies this kind of power is proving expensive (more than the amp itself).
So this has forced me to consider other options - one being to build a 14s2p lipo battery using 18650 lipo cells.
This would supply more than enough oomph.
The problem is now how much do I need to worry about keeping those cells charged?
I plan on using a cheap lipo charging circuit to keep them topped up.
Due to the way an amp plays music, ie, it's really only the occasional full volume bass that will push the amp and PSU to the max, I'm thinking I don't need the full 16amp current to keep the batteries topped up?
I'm curious if anyone has created a PSU out of lipos, and if so, any gotchas?
I would not try to operate this from batteries. The charge time would be longer than the use time and the charger would likely cost more than a proper power supply.
But lets ask a couple of questions here...
What kind of speakers are you planning to run with the amp?
Is it safe to run your speakers on 300 watts per channel?
Do you really need 600 watts in your listening environment?
In my experience most people overbuy on power by orders of magnitude and then discover, to their surprise, their speakers just can't handle it. For speakers rated something like 30 to 100 watts, they are suggesting greater than 30 watts to avoid clipping and less than 100 watts to avoid burnouts.
Most living room listening happens at about 1 or 2 watts. Home theatres often use less than 50 watts total for a 5.1 system. Yes it's nice to have some "overhead" but not at the expense of having to develop monster power supplies and convoluted system setups.
So, my suggestion is to re-examine your needs... you might find you will be perfectly happy with 20 or 50 watts per channel... and you can get those and a proper smps brick power supply for less than the cost of that amplifier.
Why not?
I'm thinking about to do the same thing as high power power supplies are a lot more expensive then lipo cells and a charger 1/10 the maximum power to keep up with the average power usage even at full power due to the crest factor in music.
Lipo's don't like to be fully charged at 4.2V continues/long time. Charging them to 4.1V would double their lifespan. Advised long term storage voltage is around 3.85V
My idea is to mod a BMS to cutoff voltage at (or a bit below) 4.0V per cell (~80% charge) and to make the on/off switch cut off both power from battery to amp and power supply to battery.
13S x 4.0V=52.0V
13S x 3.85V=50.05V
would work well with my tpa3255's.
Some balanced charging leads and connectors and a simple and cheap voltage checker would make it possible to see if cells are failing and to discharge and balance charge the pack every now and then. But I'm not sure if that would be needed.
It would run for hours at low power from these batteries only and give the cleanest power and thus SQ possible.
Any (mobile) power source and a simple current limiting/constant current step up convertor could extend the run time; i.e. a 12V/24V/37V (ebike) battery or even a solar panel which make it more flexible and cheaper.
For mobile usage, this should be better then running directly from a big 12S pack as the voltage drop will decrease output quite fast: 12x3.7v= 44.4V. I can't use a normal 13S pack as it charges to high @13x4.2V= 54.6V and that is to high for tpa3255 so I would have to go 12S, where 13S is more standard and thus cheaper.
Has anybody done this already or why wouldn't this work?
This is exactly what I've been thinking and I was missing the pertinent word in my initial question..."crest" ie. you only need the highest power to cover the large crests (when the bass kicks in) in music when at full volume, but 95% of the time the music wouldn't be at this level, thus leaving overhead in the charging circuit to keep the batteries at required voltage/charge.
The beauty of this approach is the batteries can supply huge amounts of current - a 13s lipo battery could potentially supply 650 amps of current.
This is way more than a 48v PSU could supply.
Net result is the amp is never starved of power.
And all I need to do is buy a cheapo 50 volt step up converter to take my cheapo 12 volt PSU up to required charge voltage for my batteries.
I reckon the whole setup would cost around £50 - roughly half the cost of a PSU.
Anyways - I'm surprised this hasn't been considered before, and as I like to experiment, I'm gonna give it a go
)I have a lot of experience with such a li-ion powerpack since its widely used in electric skateboards and I think the battery solution is going to work very well. But be aware of the voltage sag on high loads might be higher than you think.. So I suggest using batteries with the lowest impedance you can get and keep the wires short and thick. In our skateboards we draw about 3000W peak and the voltage sags so much that it impacts the speed noticably. And thats using li-ion 10-12s battery pack in 3p-5p and costing hundreds of dollars. If I would to make a battery pack I would look into buying two 6s lipo racing drone batteries and connect those in series. Lipo 12s bms is cheap on banggood and other online skateboard stores.
The step up charger you mentioned is extremely noisy! Even more noisy than the server psu mentioned earlier. If going this route rcr filter should be used between step up and battery to dampen that noise. It would work without but then your depending on the amps noise rejection capability.
Another alternative is to make a smallish unregulated linear psu feeding the bms. That would provide a more high end psu worth the extra hassle that comes with batteries. A fuse, transformer, diode bridge and a capacitor is really all you need. A 50-100w linear psu would probly be enough and not cost much money.
The step up charger you mentioned is extremely noisy! Even more noisy than the server psu mentioned earlier. If going this route rcr filter should be used between step up and battery to dampen that noise. It would work without but then your depending on the amps noise rejection capability.
Another alternative is to make a smallish unregulated linear psu feeding the bms. That would provide a more high end psu worth the extra hassle that comes with batteries. A fuse, transformer, diode bridge and a capacitor is really all you need. A 50-100w linear psu would probly be enough and not cost much money.
Which actually means Class D amplifiers are incredibly cheap, not that power supplies are expensive.
oomph? ... maybe.
The Million Dollar questionis: for how long?
A cheap charging circuit will top up NOTHING.
Not the battery pack you intend to use anyway.
And it will be a power supply in its own right.
How do you think that using an inadequate (big time) power supply + expensive battery pack is better than using the proper one is beyond me
Probably not, but same applies to a straight power supply.
Just thinking aloud, you might use a much smaller, say, 5A capable supply to cater to average music and add some big caps so it handles short peaks.
Much better option than the battery kludge.
Fine for portable amps, if anything because there´s no other option.
But if there is a wall outlet available, forget it.
I agree, the "how long" question is the one thing that I'm struggling with quantifying. There are so many variables that I think the only way of answering this is to try it and see how I get on.
As for keeping the batteries charged, again, this goes hand in hand with volume & music type.
I've considered the supercapacitor option but from an energy density point of view, they're nowhere near Lipos. My lipo pack will be able to supply 650amps, which is way more than is actually needed.
some basic cost comparison...
A 52v 16amp PSU would cost around £100
whereas...
13 x 18650 batteries cost around £30
12volt 10a PSU around £13
DC-DC Converter 15A 400W (to take the 12v to 52v to charge the batteries) around £6
Total = circa £50
As for keeping the batteries charged, again, this goes hand in hand with volume & music type.
I've considered the supercapacitor option but from an energy density point of view, they're nowhere near Lipos. My lipo pack will be able to supply 650amps, which is way more than is actually needed.
some basic cost comparison...
A 52v 16amp PSU would cost around £100
whereas...
13 x 18650 batteries cost around £30
12volt 10a PSU around £13
DC-DC Converter 15A 400W (to take the 12v to 52v to charge the batteries) around £6
Total = circa £50
A couple thoughts/questions from a(nother) battery guy. I've dealt with 18650s for years, including a whole lot of safety testing and failure analysis so I'm always cautious when people are thinking of using them, especially in largish sized configurations like this. 18650s can be very nasty when something goes wrong due to the ability to build up massive pressures.
So enough scare mongering
How are you planning to connect your cells together? Do they come with tabs welded on?
What cells are you planning on using?
Do you know that they are authentic?
If you have a 14S2P battery, or 13S or whatever you aren't getting 650A out of the pack. If you are looking at power tool type cells they generally top out at 30A in 18650 size. Since you have 2 parallel strings you could get 60A. (Still more than you need for amp operation)
So let's say you have a 14S2P pack. You have a working voltage range of 58.8V on the top end, and 42V on the bottom end assuming this is a 4.2V charge cell. Average voltage is right around 50.4. Based on the quick look at the amp spec it says 25-50V input, what happens above that? You could limit the charge voltage (which is a great thing from both a longevity and safety standpoint) but your runtime would drop significantly.
Speaking of charging, how are you controlling the charge? Does the DC-DC converter do a CCCV charge control, or are you just leaving it at 52V? How does the DC-DC converter behave if the batteries are empty? These are things to think about.
For runtime calculations lets say you are using the Sony (sorry, Murata) VTC6. It is a 3000mAh cell and you have 2 in parallel. So you have 6Ah in total. The cells in series give you voltage, but they do not give you more Amp-hours. So if you fully charge the battery pack and your average load is 6A, it will last for an hour. If you must charge them lower, or choose to for longevity, you will get less than an hour. So if you really need and use that 300W x 2 your runtime will be measured in minutes, not hours. If you don't need that 300W x 2 that amp will run fine on a lower voltage, lower current supply
And just to make a quick comment on the overall cost. If 13 batteries are $30 and you need 26 of them, plus the PSU, the DC-DC converter, the battery management unit and you're somewhere right around that cost of the PSU.
So enough scare mongering
How are you planning to connect your cells together? Do they come with tabs welded on?
What cells are you planning on using?
Do you know that they are authentic?
If you have a 14S2P battery, or 13S or whatever you aren't getting 650A out of the pack. If you are looking at power tool type cells they generally top out at 30A in 18650 size. Since you have 2 parallel strings you could get 60A. (Still more than you need for amp operation)
So let's say you have a 14S2P pack. You have a working voltage range of 58.8V on the top end, and 42V on the bottom end assuming this is a 4.2V charge cell. Average voltage is right around 50.4. Based on the quick look at the amp spec it says 25-50V input, what happens above that? You could limit the charge voltage (which is a great thing from both a longevity and safety standpoint) but your runtime would drop significantly.
Speaking of charging, how are you controlling the charge? Does the DC-DC converter do a CCCV charge control, or are you just leaving it at 52V? How does the DC-DC converter behave if the batteries are empty? These are things to think about.
For runtime calculations lets say you are using the Sony (sorry, Murata) VTC6. It is a 3000mAh cell and you have 2 in parallel. So you have 6Ah in total. The cells in series give you voltage, but they do not give you more Amp-hours. So if you fully charge the battery pack and your average load is 6A, it will last for an hour. If you must charge them lower, or choose to for longevity, you will get less than an hour. So if you really need and use that 300W x 2 your runtime will be measured in minutes, not hours. If you don't need that 300W x 2 that amp will run fine on a lower voltage, lower current supply
And just to make a quick comment on the overall cost. If 13 batteries are $30 and you need 26 of them, plus the PSU, the DC-DC converter, the battery management unit and you're somewhere right around that cost of the PSU.
I've gotta admit that your concerns are the same as mine. Even though I'm sure the battery idea will work, I am concerned about the safety side of this kind of setup.
The batteries will be soldered by me using heavy gauge wire.
The batteries will be come from some old laptop batteries I have lying around. Each cell will be tested to ensure they haven't dropped below 2.5volts, in which case I'd discard them.
I also plan on using 14 in series but only charging to 52volts - this will ensure each cell is only 3.7volts - well within the safe limits.
I may have overestimated the 650 amps quote but the point I was trying to make is that by using batteries the amp would never be starved of power (provided the batteries are charged).
In terms of charging the batteries, I've decided to not use a BMS. I don't see this as much of a problem as I don't plan on taking the cells to max (4.2volts). Instead I'm just gonna simple charge (to positive and negative) with the step up converter I have. I've gotten my hands on a 19v 9amp PSU that'll power the step up converter, which itself can supply 15amp. Yes - I have a shortfall of around 7amps, but the idea is the amp doesn't use 16amps 100% of the time. Obviously, if during a particularly hard session of heavy rock music at full volume, it may struggle, in which case I may have to rethink the whole thing.
I plan on installing an analogue volt and amp meter so I can monitor things, and also cos I think they'll look retro and cool!!
I'll be honest, I have no idea if this will work, but it'll be an interesting build and I'll learn a lot.
If you are concerned about safety, you should be using a BMS of some kind that monitors each individual cell in the string. Even if you aren't charging to max voltage, cells of different age and health will behave differently. Over time this could lead to individual cells becoming overcharged or over discharged when the whole string voltage looks fine.
+10 jc2; cells can fail and start to discharge themselves and overvolting/overcharging the others; use a bms or a balance charger and balance wires, to prevent fire.
Some things I learned since my last post in this topic:
Instead of lowering the cut off voltage, an active balancing board can be added to a bms to keep the cells all at the same voltage while not being fully charged.
Many bms I looked at with 'active' or 'balancing' in the name don't do this, but I see no reason this would not be possible.
Li ion and lipo is essentially the same and can contain many chemistries.
PO stands for POlymer bag.
Both can be found as high capacity OR as high discharge variants.
Lifepo4 cells are safer and their discharge curve is flatter.
Defective (ebike) batteries with 18650 cells are a cheap source.
The Opus bt-3100/3300 charger seems to be a cheap and good candidate if you have to test many cells. (4 cells at the same time; 4x as fast as the well known Imax b6 and does an automatic charge after discharge) If you put in dummy batteries with wires you can connect any kind of lithium or lifepo4 cells to test up to 20Ah/20000mAh. (I ordered one @ali for €26)
Some things I learned since my last post in this topic:
Instead of lowering the cut off voltage, an active balancing board can be added to a bms to keep the cells all at the same voltage while not being fully charged.
Many bms I looked at with 'active' or 'balancing' in the name don't do this, but I see no reason this would not be possible.
Li ion and lipo is essentially the same and can contain many chemistries.
PO stands for POlymer bag.
Both can be found as high capacity OR as high discharge variants.
Lifepo4 cells are safer and their discharge curve is flatter.
Defective (ebike) batteries with 18650 cells are a cheap source.
The Opus bt-3100/3300 charger seems to be a cheap and good candidate if you have to test many cells. (4 cells at the same time; 4x as fast as the well known Imax b6 and does an automatic charge after discharge) If you put in dummy batteries with wires you can connect any kind of lithium or lifepo4 cells to test up to 20Ah/20000mAh. (I ordered one @ali for €26)
A good 18650 cell easily pumps out 20 amps, some 30.
I have found the best battery solution is to use Ryobi one plus drill batteries. They are very well rated and designed with all sorts of protections built on and charge super quick. A standard 2.1 Ah battery easily runs all night on a 3116 or 3118 at high volume and can really pump out the bass.
I don't understand everyone here who says batteries will be no good, and additionally all problems of noise and earthing go away.
I have found the best battery solution is to use Ryobi one plus drill batteries. They are very well rated and designed with all sorts of protections built on and charge super quick. A standard 2.1 Ah battery easily runs all night on a 3116 or 3118 at high volume and can really pump out the bass.
I don't understand everyone here who says batteries will be no good, and additionally all problems of noise and earthing go away.