Ricci

Yep yep. Agreed on your points Mark. That's why I posted the impedance curve. I would disagree a bit. Not with anything in general only that while the DCR is indeed a bit high-ish, I'd contend this is normal use driving speakers. Impedance fluctuates greatly with frequency and current is only high near the minimums. Unfiltered sealed systems are about the worst case. They only have one impedance peak and the amplifier is asked to produce high current into the impedance minimums in the lowest frequencies. Typical use for higher order systems means generally higher impedances, more impedance peaks and a low end that is filtered away in order to protect the drivers. Sure I could get some drivers with the DCR a bit lower down near 2.8-3 ohms for a 4ohm load but they will also typically have much higher motor force than the boat anchors used so the impedance peak will be much more severe and cover a broader plateau raising the average impedance over a large area of the bandwidth. I'd expect more current only required in the very deepest bass and perhaps at a narrow band above the resonance. I know you probably have at some point, but I doubt many others will have looked at actual average power applied to the speakers using even what is considered "intensive" content, over a significant amount of time like 30 minutes or an hour. Even 400w average applied power to a subwoofer over more than a few minutes is a lot for the reasons discussed above. There are huge peaks in the content but the duration and overall average power is actually much lower than most people would assume. I almost went with the SP amps. I didn't because of the whole GP testing thing and also because when the Euro went into the trash for a while I got a smoking deal on the K's. I needed the most amp I could get on one set of terminals and the SP can only use one side so that kind of clenched it. Kind of wish I did for the HT just because either amp has all the power needed but the SP fans are silent. Anyway this wasn't really meant to be a tediously detailed comparison but something different and fun. I still need to do some testing on the Ipal 21's so I may just go ahead and take both amps outside and let them both have a shot into 1 ohm and 2ohm minimums. I definitely want to use the 12K on the single 21 because that truly dips to 1ohm in spots and I don't think bridging the K20 into that will result in its best performance.

Sort of. There are a few innocent bystanders that came of walls or vibrated off and fell. Lost a few light bulbs. Some rattles I don't think can ever be fully taken care of. It's just too much energy for that.

Oscilloscopes and current/ voltage captures are cool but how do these behave when strapped to a bunch of speakers? That was the point of this to begin with. Let's get away from the test bench captures while driving resistors in a water tank and use a real world load with all of the complexity and possible nastiness of multiple, heavy coil, high inductance drivers driven really hard. FR traces at a volt are useful but no guarantee of maintaining their shape when the amp is asked for full output or when loaded with complex reactive loads. Now a FR at the onset of clipping with a very slow and taxing measurement would be nice, but no one provides those. I wanted to dispense with paper ratings of things like watts and see the bottom line of how each amp handled driving a set of subs when the throttle was mashed to the floor. I don't even care about how clean the amp is or at what point it gets dirty with this type of setup. With a finished system design you either have enough gear to stay out of clipping and keep it clean or not IMO. That kind of thing was not within the scope of this type of setup. No doubt about it the sub 20Hz range is way more demanding on the amps and once things get down to 10Hz and below it is even worse. The duty cycle is much higher simply because the waveforms are so long in duration and the impedance is the lowest so current is both high and demanded for longer amounts of time. I don't know how many people are really familiar with REW's signals, but a maximum length sweep set to a bandwidth of 2-120Hz or similar is vicious when dumped unfiltered from soundcard to amp to speaker. The 2-20Hz band takes what is probably the first 7 seconds or so and when the amp is being driven with an input signal that is a couple of dB beyond what would light the clip lights it is a rough test. Nothing in any movie soundtrack I've ever seen yet has the duration and sustained draw from the amp to compare. It's not as bad as just leaving a 5Hz sine wave on at clipping indefinitely, but it is worse than anything I've seen in music or movie content. The fact that both the amps produce as much juice as they did and are protected against blowing themselves up even when pushed to their limit with this type of <10Hz stuff was exactly what I want to see. I truly would worry about blowing some amplifiers if driven in this way. Try it at your own risk. Also the CEA-2010 style bursts...I'd like to note for those reading along that they are 6.5 cycles so at 50Hz it is equivalent to 0.13 seconds. Not long at all, but comparable to a kick drum thump overall. However down at 20Hz it is 0.33 seconds duration and at 10Hz it is now 0.65 seconds. Not real long but it might as well be an eternity compared to some of the 1 cycle 1kHz signals used to produce the power ratings for many amplifiers so it is more demanding than the signals used by many manufacturers for their ratings. I would like for amplifier manufacturers to adopt a standard to provide burst ratings using these standardized signals at something like 10Hz, 31.5Hz, 100Hz, 1kHz and 16kHz.

Thanks Mike. I would love to do this with some smaller, more affordable amps but the setup and all that for this was a real PITA. Not something I want to take on again any time soon.

The K20 roll off is the same as the K10 FWTW.

One quick update. I am putting my system back together this morning and was reconnecting everything and double checking all of my settings between the 2 K20's and the K20 used for the comparison's current setting was actually set to 32A Fast instead of 32A Slow. That may have affected the 4 ohm bridged long term test by having it limit current draw a bit quicker. I don't know for sure and I'm not setting everything back up to find out. Maybe down the road some time. Dave by stereo I meant that the amplifier was running 2 separate channels. It is 2 bridged amps of course but it is run with 2 channels. The k20 was run into a mono load. The SP could not be so it was 2 loads with a mono signal.

I did post the FR of the PS amps and also the SP amp in comparison in the electronics FR thread I believe. You can use the comparison with the PS to derive the SP FR. It is pretty good below 10Hz.

You are correct Spacebug. Thanks for catching that. I always assume SMPS is synonymous with Class-D but it is not.

No worries man. Almost done!

Crap...Maybe should've reserved one more.

The wrap up. Looking at the data as a whole the K20 does have a higher short term output capability and also long term into higher impedance loads. The SP2-12000 is certainly no slouch either and closes the gap at low impedance loads and with high duty cycle signals into extreme low frequencies. Based on the power ratings the results are pretty much as expected. Both of these amplifiers are monsters and offer among the highest potential output on the market with the capability to drive multiple of the heaviest duty drivers or bass cabinets to their limits. The K20-DSP-Aesop certainly offers more bells and whistles with a huge variety of settings, on board DSP, remote control and monitoring, networking, AES digital inputs, cable compensation, universal power supply, etc. Most of which would not be used in most cases unless you are running them as part of a huge networked system in a stadium or something. The SP2-12000 is more straight forward and offers comparable amounts of power. The SP2-12000 feels a bit lighter and is physically a little smaller. The K20 is a single rack space but it is a very deep amplifier requiring a deep rack. The SP2-12000 is 2 rack spaces but is a much shorter case so the overall effect is that it feels smaller despite being 2 rack spaces. The SP2-12000 is also much quieter. The K20's fans always run and while they are not as loud as some other amplifiers they are audible in a quiet home setting. The SP2-12000's fans don't even seem to run most of the time and didn't even turn on until asked to produce very large amounts of power for an extended time. That being the case it is virtually silent which is awesome for a quieter home or studio setting. It also ran cooler. It produces barely any heat until really loaded down for an extended period at which point it got barely warm. The K20 idles with a bit of warmth after being on for a while and will warm up a bit after being used heavily. It has never gotten what I'd call hot by any means but it does run a little warmer than the SP2-12000. The SP2-12000 also has a little less roll off below 10Hz for the ultra low frequency fanatics. The Powersoft is designed and manufactured in the EU while the SP2-12000 is a North American design and manufacture. The SP2-12000 is also much cheaper than the K20. Even if you get an extremely good deal on the K20 or K10 series like I did and you eschew all of the extra DSP or AESOP options they are still notably more expensive than the SP2-12000 on average. Both of these amps tolerated all of the abuse with ease and are obviously well protected and designed such that you will really have to do something stupid to blow one up. I mean if running them way into clipping below 10Hz multiple times with sine waves won't do it it's going to take something really extreme. Both are absolute mules and I would proudly own either. A final word for owners of Class-D amplifiers. I always tried to run my amps bridged due to the advantages I've read about but after the clear data and first hand behavioral differences I noted with the K20 when running full bridged versus not, I am truly a believer. If you own a Class D based amplifier that has the ability to be bridged that pretty much means that it is half bridged when not running in bridged mode and you could be severely limiting the amplifiers performance for bass duties. If you want the best performance possible from the amplifier for driving subwoofers you need to bridge your amplifier. This is especially true if you are driving a sealed or IB bass system and running it unfiltered on the bottom for the most extension possible. The extreme low frequencies combined with high current demand seem to be especially problematic. The K20 seemed to be alright at 20Hz and above with the loadings used here even when unbridged, but who knows how other amplifiers may behave, so I'd advise against taking the chance with diminished head room and instability in the low bass range and just go straight to bridged mode.

After talking with Brian about the results for the single channel tests and the issues with the K20 and low frequencies he mentioned that it was likely due to either DC protection or bus pumping. My thoughts were that it was one of these or perhaps current limiting also. I'd heard of bus pumping in Class-D amplifiers before and was familiar with the idea but had never really experienced it happening with an amplifier before in such a conclusive way. Brian recommended doing a dual channel test with both channels run and the Powersoft bridged. The SP2-12000 is a dual mono design with two separate amplifier modules each with their own power supply which are already in a full bridged configuration. The K20 and other amps in the Digam series have each channel sharing a power supply and are half bridged when run in stereo but can be placed in a full bridged configuration. In a nutshell what happens when a Class-D amplifier is in a half bridged configuration is that there is an induced current back from the load to the amplifier which can cause fluctuation and instability in the amplifier. The issue is worst at very low frequencies. There are a number of technical papers on Class-D amplifier design that explain the issue in depth with a quick internet search. By placing the amplifier in a full bridged configuration the bus pumping issue is resolved. This sounded like the issue being experienced with the K20 during the single channel testing and indeed after thinking about it I virtually always run the K series amps I've owned in a bridged config. The few times that I hadn't were during testing of a single speaker where it could not begin to handle the full power from even a single channel of the amplifier so the limits were never approached. That would explain why the issue didn't appear in those couple of tests using just a single channel. Armed with this bit of information the dual channel tests would be with the SP2-12000 run in stereo since it is already a pair of bridged amplifiers. The K20 would be run in a mono bridged configuration. In bridged mode the amplifier will see half of the impedance on each channel so the loading works out the same for say an amplifier run into two 4 ohm loads in stereo and the same 4 ohm loads presented as 8 ohms mono to a bridged amp. 8 Ohm Stereo SP2-12000 / 16 Ohm Mono K20 / Both Channels Driven For this test the SP2-12000 was run first. The burst testing was without issue. The long term sine wave test was also without issue. The K20 also ran through both tests without issue. Good so far. The burst results are presented below. As with the single channel 8ohm results the K20 produces a bit higher output with an average about 1.9dB higher. The long term sine wave results show much the same thing with the K20 maintaining a similar lead over the entire range of the test. By the way the response shape shown is the raw response in my room so it isn't pretty but it works for the purpose of illustrating the comparison. 4 Ohm Stereo SP2-12000 / 8 Ohm Mono K20 / Both Channels Driven For this test the K20 was run first. The burst testing was without issue. The long term sine wave test was also without issue. WOOHOO! I was relieved that bridging the amp fixed the issue to say the least. The SP2-12000 did its thing as usual without complaint. The burst results are similar to the other tests so far with the K20 able to push the speakers about 1.7dB louder on average. The two channel driven long term testing follows a similar pattern as at 8 ohms and the K20 is able to push the speakers to slightly higher output levels for the duration of the measurement signal. 2 Ohm Stereo SP2-12000 / 4 Ohm Mono K20 / Both Channels Driven For this test the SP2-12000 was run first. The burst testing was completed first. I did manage to get the SP2-12000 to briefly mute or protect with the 10Hz burst while trying to get everything possible out of it. As before this was pushing the amp deep into clipping and reducing the level just a hair took care of it. The long term testing was then done and the SP2-12000 could be caused to mute or protect briefly below <10Hz if the level was too high but again this was with a signal that was a good 4 or 5dB past the point that the clip lights would be lit up. Reducing the input signal level a dB or two allowed the SP2-12000 to complete the sweep at full power without any problems. The K20 completed the burst testing without issue even at 10Hz and I could not get it to protect or mute even running the input signal up 6dB or more past what would cause clipping. The 2ohm long term test also was completed without any muting or protection triggering. Looks like bridging the amp did indeed entirely remove the issue seen in the single channel unbridged testing. This test series was the only one that got the SP2-12000's fans to engage at all and produced any heat in the amp. It got barely warm. The K20 did spool the fans up a bit on this one as well and also became a bit warm. Warmer than the SP2-12000. The K20 seems to run with the fans set to cool the amp much more aggressively and the case does run warm to the touch even at idle where the SP2-12000 won't warm up significantly unless you are really beating on it for awhile. The 2 ohm, both channels driven, burst tests show much the same story as we've seen so far. The K20 is able to coerce the XXX's into producing about 1.6dB higher output on average. The long term sine wave testing at this lower impedance load is a little more interesting. The K20 exhibited some limiting during this and was held about 3-4dB below clipping for large sections of the sweep. The general trend is that the K20 would produce near full power for a section of the sweep then reduce output for a while and then ramp back up and then limit a bit for awhile. We can see that the K20 starts off with a bit higher output down in the resonance of the earth frequencies of 2-5Hz and then the SP2-12000 takes over from 5-15Hz. The K20 rises back up and takes back over from 16-34Hz. Then the SP2-12000 takes over or runs even with the K20 over the 35-70Hz octave. The K20 comes back on from 70-120Hz and produces a bit higher output again. To me this looks like long term current limiting going on in the K20. The Digam series of amplifiers has a number of settings for limiting the amplifiers output or matching it to the circuit that it is on. The maximum current setting is 33A which is a long term average which is there to prevent tripping breakers. With typical content I doubt that this is ever an issue but the sine wave measurement used here is 100% duty cycle for 23 seconds or so and with the relatively low impedance minimums and the amplifier being run at full output I suspect this may actually be coming into effect and the software may be limiting the current a bit. The most notable area of the K20's measurement is between 30-40Hz where clearly the output is limited abruptly for a bit. If this set of sweeps is compared with the impedance curve of the driver provided earlier we can see that the output squishing happens in the areas of high current demand. I will say that there is no way that I could hear anything happening or tell anything different during this measurement by ear. The entire house was being shook violently and a reduction of a couple dB in output simply doesn't audibly register to me when my home and body is being punished like that. Perhaps outside...I don't know. Anyway I'd call the 2 ohm long term sine wave test a draw or a slight win for the SP2-12000.

Let's start with the single channel measurements as they were done first. 8 OHM Single Channel Results The SP2-12000 SpeakerPower was up first at 8ohms. The burst tests were done first followed by the long term sine wave test. Nothing notable to report. The SP2-12000 had no issues even when driven several dB into clipping. The fans didn't even turn on. The K20 followed also at 8ohms again nothing to report here. The amp chugged along. The single channel 8 ohm burst data is represented in the chart below. The K20 is set as the baseline measurement in the blue column and it's results are set at "0". The SP2-12000 result is shown in dBW relative to that recorded by the K20. A positive number indicates higher output a negative number equals less. Again this is the recorded acoustic output of a group of heavy duty speakers when driven with both amplifiers. We can see that during this test the K20 recorded about 1.6dB higher output on average. No long term comparison data. The reason will be explained below. 4 OHM Single Channel Results The K20 went first at 4 ohms. The burst tests were done first followed by the long term sine wave test. The burst tests went as expected. No issues. However when running the sine wave test which starts way down near 2Hz, the K20 muted the outputs. The output level had to be dropped way down by 7-9dB to get it to complete the sweep without muting the output during the beginning of the sweep in the extreme low frequencies. I had never encountered this behavior before so I tried the other channel with the same result. I even tried the other K20 I have and it repeated the same. I started moving the start frequency of the sweep up to see if this would help and once the sweep started closer to 20Hz the K20 would complete the sweep at full output without protecting. The SP2-12000 however, after running through its burst testing, had no issues with the full sine wave sweep even when driven several dB into clipping as before. Since the K20 wouldn't complete the full sweep in this setup unless the bandwidth was restricted or the level was reduced some 6dB or more below the SP2-12000 capability I saw no point in presenting the long term data comparison for this test. The SP2-12000 clearly won that portion of the test. The burst testing is presented below in the same manner as previously outlined. The K20 is the baseline "0" result and the SP2-12000 output is shown referenced to that baseline. Similar to the 8ohm single channel results the K20 has about 1.9dB greater output on average with the 4ohm burst tests. 2 OHM Single Channel Results The SP2-12000 went first at 2 ohms. The burst tests were done first followed by the long term sine wave test. The burst tests went as expected. The only thing that cropped up was that I was able to get the SP2-12000 to mute briefly with the 10Hz burst with it driven WAY into clipping while trying to get every last 0.1dB. Otherwise it was uneventful and backing off a little bit removed that issue. The long term test was also not an issue for the SP2-12000 after backing off the signal a bit from the burst testing. The clip lights were still illuminated for nearly the entire sweep. I knew that the K20 would likely repeat the issue from the 4 ohm long term sweep test at 2 ohms, so the SP2-12000 would win the long term single channel portion with ease. This was the case. The K20 did not like being asked for full power in the extreme low frequencies with one channel driven. It muted the outputs unless the level was greatly reduced or unless the bandwidth below 20Hz was skipped. The issue even started to show up in the 2 ohm burst testing. The K20 would mute at the 10 through 16Hz bands if pushed too hard. At 10 and 12.5Hz the output prior to muting was down some 8 to 10dB from the SP2-12000's recorded results. At 16Hz the K20 kind of got its act together and was around 3dB low from what it would have been. The SP2-12000 produced notably more output at that band as well. At 20Hz and above the K20 finally started doing its thing and averaged about 1.3dB higher output. Ok so at this point a couple of things are clear. On a single channel versus single channel basis the K20 can burst higher numbers short term but the SP2-12000 kicks its ass with longer term or higher duty cycle signals at very low frequencies and lower impedance loads. The limited K20 output and aversion to extreme low frequency information is why the long term data is not compared for the single channel tests. It's a clear cut result. At this point I sent some of this information to Brian and asked his opinion on some things since he is a certified amplifier guru and I'm more like the guy who stayed at the Holiday Inn Express last night when it comes to amplifiers. I know a little about thery and design but I don't kid myself when comparing my limited understanding to someone who does it for a living. This will be covered in the next post.

Attached here is an example of the wiring plate used and also an example impedance of one of the drivers in a sealed cab of about the same airspace with the coils in series. Obviously the impedance of the cabs shifted somewhat with heat and high displacement of the drivers but this is the baseline. The use of so many drivers should have kept the amount of heat buildup in the coils from being excessive and causing notable compression or parameter shifting. Regardless both amplifiers were driven into the same loading so should have experienced the same things or close enough for the rough general nature of this type of testing. By looking at the impedance we can get a rough guideline of the areas that required the highest current from the amps and which areas were near the peak and required more voltage.

Last year at about this time Brian Oppegaard at Speakerpower sent me one of his 240V, dual mono, SP2-12000 rack amplifiers so I could try it out and do a comparison against one of my Powersoft K10's. This was after discussing back and forth for a while. Of course a few weeks after that I became very busy and made a move to sell my K10's for K20's with all the bells and whistles which put the project off for a while. So here we are a year later and I've finally gotten around to what I intended to do from the get go. (My apologies for the extreme delay Brian.) I'm more of an objective data driven person rather than subjective so to that end the plan was to actually gather some form of rough comparison data on the amplifiers. Also I do not have an amplifier test bench and need another test regimen and equipment setup like I need a hole in the head, so the plan was to test the amplifiers while powering subwoofers. This should be more representative of a complex real world load on the amplifiers but also has a number of challenges. First is that these amplifiers are among the most powerful on the market and it will take a number of drivers to be able to handle their full output without being the limiting factor. #2 is that it is going to be loud. I already had 8 sealed, dual voice coil, RE Audio XXX 18's in my HT space so I figured that should take care of the speaker load. I never planned to do any testing or listening to either amplifier above the bass range. Really who buys 10kW rated amplifiers and uses them for anything other than bass? I wanted to use a sealed system in order not bandwidth limit the amplifiers and to require them to supply power down deep. I originally planned to do the measurements outdoors to remove room effects and to keep from subjecting my house to the absurd level of bass that would be produced, but after looking at the logistics of transporting the four 260lb dual opposed cabinets, the amplifiers, the wiring change overs, setup time and worrying about a good window in the weather I gave up on that. Indoors was so much easier I decided to take the lazy route. The test regime was to wire up as many XXX drivers in groups of two that would meet nominal loads of 8, 4 and 2 ohms. The amplifiers would be connected to each nominal load and then be run up to maximum output with 6.5 cycle shaped CEA-2010 type bursts in order to look at dynamic output and also with the same 23 second long sine sweeps that I use for testing speakers outdoors in order to look at something with an extremely high duty cycle, much worse than any music or movie content I'm aware of. this would give a good glimpse of long term output and short term. The amplifiers were both run from the same 240V AC line, the output was recorded by an Earthwork's M30 microphone placed at the headrest of the main listening position. The signals used were all software based and the microphone preamp providing the signal to the amplifiers was a Presonus Firebox. The gain for both amplifiers was set to the highest available setting. No DSP or other filtering was used for either amplifier. Originally I only planned to do a set of single channel driven tests since the SP2-12000 is a dual mono design with each internal channel already bridged. However after discussions with Brian and encountering issues with bus pumping with the K20 (Which is not a full bridge design) when running a single channel a set of tests running both channels was also done. For the dual channel tests the SP2-12000 was run in stereo since each channel is already a full bridge design and the K20 was operated in bridged mode. An effort was made to use wiring that was as close to each other as possible for each amplifier. It could not be exact but in general was very, very close. I used my existing 4 runs of 4 pole Speakon cables in the room and put together a wood panel with a set of Speakon jacks that would allow me to use a variety of jumper schemes to switch the load presented to each amp. In general I would always use the same drivers for both amplifiers if all 8 were not used and maintained parallel / series wiring of voice coils and drivers for both amplifiers with any changes needed in the wiring scheme accomplished at the amplifier or at the Speakon switch plate so that the loads were as close as possible. Obviously by using speakers as the load and amplifiers with this much juice thermal and dynamic compression in the drivers is a concern. However the tests using the full compliment of 8 drivers seemed to suffer from this very little and those using 4 drivers much less than I expected. Still it will be noted which amplifier was tested first at each nominal resistance as the second one to be tested may have been dealing with a bit of heat in the drivers. More to come...

I've got some fellow DIYers who have offered to help build out the prototype pair of cabs for close to cost. They do good work and have a substantial amount of bass horn chi themselves, so that helps out versus trying to explain every detail to typical cabinet shops who have no idea what they are looking at in the plans. It's probably going to be a while, but I hope to have something by the end of the year. I'm finalizing the print on something else I've been working on for a couple of years currently. Something to replace my GH's with. Hopefully it works as modeled.

I haven't been to a GTG or visited another enthusiasts home in years. Haven't made it to any of the big Expo's either. Slacking!

I have got to make it down to Dave's one of these days.

Won't work with the 21SW152 at least not ideally. I'm not going to distribute these plans anyway. Sorry. This one is mine. Of course there's enough in this thread already that if you really wanted to have a go at it there's enough to figure it out or close enough. A couple of nice gents have offered to help me out with the cabs so I hope to have some movement on this by the end of the year.

Nice looking space Bmoney. I wish mine was that clean.

Should be getting back with Jeff sometime before the end of the year and I'd expect that the new version of the OS-LFU should be first on the list. The old version was pretty beastly itself so I'd expect it to be impressive.

No doubt. I'm used to big subs but I swear each manufacturer is sending me their biggest heaviest sub now. That sucker is 170lbs!

Welcome Steve. I recognize your handle from a few forums. A Funk sub huh? That's good stuff there. Nathan does amazing work.

Fi SP4 18D1 Sealed testing done. Orion HCCA-152 sealed testing done. XTZ Sound 3X12 review will be coming to Audioholics in the next couple of weeks.

I'm really not that impressed with the M-force drivers personally. Of course I have not used or heard them so I may be proven badly wrong later. They are huge and their motors are ridiculous and the tech is very impressive but they are designed to get loud and not go deep. Their huge 30" has less overall displacement than the 24" driver from SI so it will likely have less very low bass output despite it's ridiculous efficiency. I do expect them to be VERY loud in the pro audio wheelhouse of 35-80Hz but we've had no trouble getting extremely loud over that bandwidth for years. I think they missed an opportunity to really push things a bit further with that motor design. I also wonder about the cone flexing on the 30 and 40" drivers. I still want a chance to play with one of these for sure though. Like I said a horn for these would be extremely difficult to design due to the driver design, size and mounting scheme.