MIT Research Tweaks Smartphone Amplifier Voltage To Gain Battery Life 47
hypnosec writes "Two MIT electrical engineering professors, Joel Dawson and David Perreault, have claimed that they have cracked the age old efficiency problem related to the power amplifier in smartphones by designing a new amplifier that consumes just half the power as compared to their current counterparts. Current transistor-based power amplifiers consume power in two modes – standby and output signal mode. The only way to reduce power consumption and increase battery life is to use the least possible power when in standby mode. The problem here is that if the power is kept very low when in standby mode, because of sudden jumps from low-power standby mode to high-power output mode, signals get distorted. This is why current technologies waste a lot of electricity as standby power levels are kept at a relatively higher level to avoid distortion. The new technology, dubbed asymmetric multilevel outphasing, is basically a blazingly fast electronic gearbox that would select the best possible voltage to send across to the transistors that would minimize power consumption."
Wow! (Score:1)
Twice [slashdot.org] in three days? How soon before phones are actually generating net power?
Re: (Score:3)
I hope they've done the needful... (Score:4, Interesting)
...and filed a patent for the implementation or call it execution....otherwise some clever company, that is better known for suing others in the industry,; that's had one of its famous patents invalidated recently, will file...and sue.
Re: (Score:1)
...and filed a patent for the implementation or call it execution....otherwise some clever company, that is better known for suing others in the industry,; that's had one of its famous patents invalidated recently, will file...and sue.
Because little things such as properly filed IP documents have stopped apple so well in the past, right? http://economictimes.indiatimes.com/news/international-business/apples-iphone-loses-name-battle-with-mexicos-ifone/articleshow/17092765.cms [indiatimes.com]
Re: (Score:1)
Holy smokes! It's not like Apple owns a patent on the proper use of punctuation. Did you type that with your eyes closed?
Re: (Score:2)
I know you're probably joking, but Apple buys chipsets and analog hardware for their cellphones from companies like Qualcomm, Triquint, and so on, so I think that these researchers are safe.
Duplicate story from Oct 31 (Score:5, Informative)
This is a duplicate story posted on Slashdot on Oct 31: http://hardware.slashdot.org/story/12/11/01/0021213/breakthrough-promises-smartphones-that-use-half-the-power [slashdot.org]
Luckily this time the summary includes a link to an actual technical paper. The summary and the news article make it sound like this is an Envelope Tracking [wikipedia.org] amplifier, but if you read the paper this is actually something different, it is more complicated and more interesting.
They are starting from an outphasing amplifier which divides a variable envelope signal into two constant amplitude but variable phase signals which can be amplified more efficiently since the amplifier doesn't need to output both large and small signals. But combining the signals is inefficient because the combiner must absorb some of the power when the two halves of the signal are very out of phase with each other. What the MIT researchers are doing is extending the outphasing technique to allow multiple discrete amplitudes on each amplifier to minimize the combiner inefficiency. It's more efficient than plain outphasing, I'm not sure how it compares to envelope tracking since the authors did not compare it to this in their paper.
Re: (Score:2)
Re: (Score:2)
Re:Duplicate story from Oct 31 (Score:5, Informative)
Zero crossings and low output amplitudes with outphasing techniques are a bitch. If you use a hybrid combiner, nearly all of the energy goes into a dummy load on the difference port. If you use a "lossless" combiner such as a matching network, the VSWR seen by the amp elements skyrockets and - best case, efficiency is crap. Worst case, you fry the output stages of the amp.
Around a decade ago I worked on an outphasing system - I fried a LOT of hardware by accidentally feeding the system a low or zero amplitude, which effectively caused two amps to fight each other 180 degrees out of phase. http://www.google.com/patents/US6930547 [google.com]
What they seem to be doing here is reducing the voltage provided to the amp elements, reducing the power output, which allows them to reduce total output power efficiently while keeping the two amp elements relatively in-phase, and also allowing them to "fight" each other without frying themselves like they would if running at full power.
E.g. it's probably something like outphasing for 70-100% envelope levels, and power supply modulation below that. Likely not a "hard" threshold - they probably transition from one approach to the other somewhere in the 30-70% operating region.
http://slashdot.org/comments.pl?sid=3222705&cid=41842229 [slashdot.org] for some more details.
hybrid A/theta modulator (Score:5, Interesting)
One way to do high efficiency is to do separate phase and amplitude modulation, the so called A-theta modulator. you get your amplitude control by adjusting the PA voltage _continuosly_ and do your phase modulation with an IQ modulator.
For wide bandwidth modulation formats this is a bit of a pain since you need a very wideband, high current, power supply. so they are doing an A/theta modulator but trying to simplify the bias control on the PA to avoid that.
What I'm not clear on is why they are doing this when they have a predistortion loop anyway. a pure predistortion loop should be able to achieve very similar results without any need for the PA bias adjust. you can also do it with 1 PA instead of two.
maybe it simplifies the PD loop substantially, I can't tell yet from the paper, but it trades this off against needing two PAs and PA bias control. And that PA bias control is a bit of a headache two, because you can't just stick a switch in there. The selectable bias levels have to come from a switching power supply which switches between different output voltages. doing this is also non-trivial.
I don't know, looks like somebody's thesis to me. Doesn't look like it's particular practical.
Also, first rule of looking at schemes like this. How much of that power they saved is being used in the more complicated digital circuitry. That's the reason you don't see PD loops in cell phones. It's a wash, you spend so much power analyzing the signal to do PD that you burn up the savings . Now if you have a 10W transmitter, PD makes lots of sense.
Re: (Score:1)
One way to do high efficiency is to do separate phase and amplitude modulation, the so called A-theta modulator. you get your amplitude control by adjusting the PA voltage _continuosly_ and do your phase modulation with an IQ modulator.
For wide bandwidth modulation formats this is a bit of a pain since you need a very wideband, high current, power supply. so they are doing an A/theta modulator but trying to simplify the bias control on the PA to avoid that.
What I'm not clear on is why they are doing this when they have a predistortion loop anyway. a pure predistortion loop should be able to achieve very similar results without any need for the PA bias adjust. you can also do it with 1 PA instead of two.
maybe it simplifies the PD loop substantially, I can't tell yet from the paper, but it trades this off against needing two PAs and PA bias control. And that PA bias control is a bit of a headache two, because you can't just stick a switch in there. The selectable bias levels have to come from a switching power supply which switches between different output voltages. doing this is also non-trivial.
I don't know, looks like somebody's thesis to me. Doesn't look like it's particular practical.
Also, first rule of looking at schemes like this. How much of that power they saved is being used in the more complicated digital circuitry. That's the reason you don't see PD loops in cell phones. It's a wash, you spend so much power analyzing the signal to do PD that you burn up the savings . Now if you have a 10W transmitter, PD makes lots of sense.
Good point, but pre-distortion loops like Cartesian feedback loop can't handle much in terms of channel bandwidth (i.e. narrow band 12.5khz is fine). Maybe this is the reason that most higher data-rate modems use simple AGC to keep the PA within the linear operating range. Of course, linear operation usually means inefficient. I'm guessing they have figured out a way to get linear output and higher data rate than what is normally possible with pre-distortion loops. Guess I should RTFA first.
Re: (Score:2)
they can handle as much as you want - they just get a lot harder to implement because the feedback loop typically has to be 5x the bandwidth of the operating bandwidth to make sure it can compensate for 3rd and 5th order products.
AGC has nothing to do with linear operation of the amplifier (well it does but only to the extent that operating at really low power levels is more linear) . The problem with these modulation schemes with high Peak to average is that they require the amplifier to be "backed-off" i
Re: (Score:1)
They're using two PAs with outphasing so that they can use a switched-mode PA design. Switched-mode PAs operate with extraordinarily low loss, but only at full output voltage. By phase modulating two switched-mode PAs, you get the best of both worlds: high PA efficiency and the ability to modulate your output power.
Re: (Score:1)
aha - I missed that note. Now it makes more sense. Those sorts of PAs would be close to unusable even with a PD loop.
still requiring a PD loop is still a big problem.
and there's still the issue that you can get these efficiencies with a pure PD loop. This scheme still has the problem that wide bandwidths are going to require a wide bandwidth feedback loop.
Re:hybrid A/theta modulator (Score:5, Insightful)
No, they're doing something more complicated than A/theta modulation (aka envelope tracking aka envelope elimination and restorataion aka ...). They are doing an outphasing amplifier which splits the original signal into two constant amplitude variable phase signals which will reproduce the original signal when they are recombined, but they are adding some envelope tracking elements to further improve efficiency since wideband combiners will absorb the differential section of the input signals as loss.
A predistorter doesn't have much of anything to do with PA efficiency (the point of this MIT research), PD's are for linearity. High efficiency PA topologies typically take a hit to linearity. Fancy techniques like outphasing and envelope tracking are better than some but they still reduce linearity. Poor linearity means increased distortion. Distortion increases error rates in digital transmissions, and it also leads to signal leaking into adjacent signal bands, which isn't allowed in the tight cellular spectral environment. So they put a PD in there to linearize things a bit further. That's probably why they are using discrete bias levels instead of continuously variable - they can optimize their PD loop to work best in these four well-characterized states which gets most of the efficiency gain while making the linearization easier.
These are not fundamental limitations. It's making more and more sense as digital processing and DSP chips get faster, cheaper, and lower power. The point of research is to try to push the state of the art and it may not be "practical" when the research paper is hot off the press. If it's good enough then it will be practical some time down the road, and if it isn't good enough then it will be left aside as a lesson learned. Apparently the researchers think its good enough to found a startup company, so they're either foolish or they understand it better than you. Time will tell.
Sounds like ... (Score:3)
Lots of novel stuff with state of the art DSP techniques. But every once in a while I spot the ghosts of tubes lurking around.
Re: (Score:1)
You can buy these right now at Amazon [amazon.com] and others.
I have a couple of these and they work great for running remote PA systems on gel-cells.
Re: (Score:3)
The "Class T" amplifier is a name-branded class D amp. Class D amps rock for both efficiency and energy density. As I sit here right now, i am listening to CAKE playing on a pair of Peavey speakers with built-in 400W Class D amps. Not only do they sound magnificent, but they also pack a whallop if I should just frob the Volume slide on my preamp. I know that they don't draw anywhere near the specplate figure when I have them loafing like they're doing now.
The problem with using PWM in any of its variati
Age old? (Score:2)
How can there even be an "age old efficiency problem related to the power amplifier in smartphones"? Sounds like an oxymoron to be. Or maybe it's just me getting age old.
Re:Age old? (Score:4, Informative)
It's clumsily worded. It's an age-old problem in power amplifiers that are used in situations where power varies, such as radio receivers and transmitters that need to work at multiple power levels. But I don't understand why they identify smartphones as a place where this has a great potential to improve battery life. It's really more applicable to non-smart cellphones where the audio amp and transmitter PA are bigger proportions of the total battery draw.
Cellphone power is often dominated by processor and display power.
Re: (Score:2)
This is what you think.
But the fact that my smartphone gets very warm and the battery charge falls off precipitously when conducting a voice call (screen off, CPU more-or-less idle) says differently.
Further anecdotal evidence: I have an OG Droid which no longer has cell phone service. Its battery lasts for about a week in light usage (web browsing over Wifi, music player, that sort of stuff).
The exact same phone, with cell service, used to
Re: (Score:3)
Cellphone power is often dominated by processor and display power.
Compare the standby time to the talk time of your phone.
For example: The iPhone 5
# Talk time: Up to 8 hours on 3G
# Standby time: Up to 225 hours
I have no doubt you will kill the battery quickest by running the processor and display,
but the transmitter is no slouch when it comes to drawing down your reserves.
Re: (Score:2)
So it won't help much in increasing battery life when you are using the device a lot. But it should help if you're leaving your phone in standby and in "airplane mode". As for standby but not in "airplane mode" it depends on what percentage of standby consumption is due to polling the station(s).
Re: (Score:2)
Re: (Score:1)
I have a few of these I got from China. Its amazing how much sound I can get driving these with a single 18650 lithium cell. Sure not as loud as my bigger class T amps, but about as loud as that vacuum-tube radio I had as a kid ( 50L6 output... remember 'em?).
I have been tracking low-power stuff as I intend one day to go off-grid, but I do not want to give up my creature comforts to do so.
Re: (Score:2)
If you scroll up a little, you'll see some comments i made along these lines. The big problem is that Class D amps, while truly magnificent at audio frequencies, really aren't feasible at UHF frequencies.
Direct link to PDF in summary (Score:1)
Editors: Please mark all links to PDFs (and any other content that is not easily digested by browsers)
Re: (Score:1)
The ".pdf" on the end of the link didn't mark it well enough for you?
And what browsers in common use don't easily show PDFs?
Re: (Score:2)
The ".pdf" on the end of the link didn't mark it well enough for you?
I don't usually hover over each link in a /. summary to inspect it before clicking it. Emails caught in my spam folder? Yeah, probably a good idea. Vetted (well, theoretically) links on /. I usually trust enough to just click-away.
And what browsers in common use don't easily show PDFs?
Firefox is working on it right now [mozilla.org], but it looks like they're targeting FF18, which is currently scheduled to move to beta on Nov 19th [mozilla.org].
I believe that Chrom{e|ium} has some support, but I haven't seen it (might not be configured on default installs).
Re: (Score:1)
Wow! (Score:2)
50% is a lot of power. Even if they can only achieve half of that in a real world implementation, it's still a significant leap forward. The latest smartphones, such as the Nexus 4, are boasting 10 hours talk time right now. 12-14 hours would be phenomenal.
And I'll bet the geniuses aren'te hams (Score:2)
Re: (Score:2)
Actually, CDMA-based technologies like UMTS require strict power control to handle the near/far handset problem.
Unfortunately, many amplifier topologies are inefficient when operating at reduced power unless you rebias them for the new operating point.
Get the physics right (Score:2)
One does not send voltage, one uses voltage to send amperage.
Congratulations, Baldrick (Score:2)
You've invented the triac preregulator, as seen in Carver amplifiers from the 1980s and probably earlier!