Quiet PA amps

It just occurred to me that "quiet" could also mean, "free from excessive hiss that's prevalent with many venue sound systems." Although, I suppose the source of that hiss is almost always the multichannel mixer and not the power amplifiers. I have a lot of experience combining consumer, professional, and audiophile gear, so it's interesting to piece together the best of all the options in a single system.


Meanwhile, on the subject of quiet fans, I searched Mouser for fans at 28 dBA (none were lower) and dropped $50 on a sampling that would fit my QSC RMX 850a.

Sunon MF80252V3-1000U-A99 Vapo Bearing, MagLev Motor

Sunon EE80252B3-000U-999 Ball Bearing

Orion Fans OD8025-24MB Dual Ball Bearing

Delta Electronics AFB0824M Ball Bearing


I'll report back about the results. The QSC gets too hot if it stays on for hours and hours, so I can't run it without the fan. The challenge will be testing the four fans before installing one. I imagine that the one that seems quietest outside the case might not necessarily be the quietest fan when mounted in the specific construction of the RMX 850a. Yeah, I'm definitely suffering First World Problems here.
 

kipman725

Member
Paid Member
2007-06-10 12:41 pm
Warrington
You have to get very high end before you get no noticeable hiss from a directly connected compression driver even ignoring noise from before the amps. On my speakers I have a passive crossover taking them down to ~100dB/1W and get some hiss with a CTS4200, its not noticeable from 2m in a quiet room but within a meter of the horn I can hear it. My desktop speakers have directly connected comps to my battery system and the hiss is too much, I haven't investigated if its the amps or DSP yet though (3e audio stuff). Most PA systems just have the amp gain controls on maximum to stop tampering but this is very sub-optimal for noise.

I have a full rack of Yamaha P series now (3*P7000s, 1*P5000s) so I will post about how that works out when I have them fitted. I'm thinking of buying a quant asylum analyzer as I'm getting to the point in performance where spec sheets are not going to provide a clear indication of performance and I want to measure things like my DSP and check its actually good (I think its good but is it?).

If your hooking up comps directly and want it to be silent at close range I think you have to look at stuff like Purifi or Benchmark AHB2 to get below the noise floor of a quiet room. I have a kind of 'dream' system in mind for a future club install with 60*60 horns with AXI2050 directly driven from >300Hz. Formed into 120x60 arrays, 4 point system (so 8 comps total), all driven by Purifi amps. Underneath that a folded version of my midbass horn design, perhaps with longer path length (100 - 300Hz). But if it happens its still a long way off.

Ball bearing fans are noisier than fluid dynamic bearings. You will have better selection of fans if you use a 12V fan and a small 24v->12v buck converter (ebay). I did this for my DSP fan as I needed a silent 40mm fan and the installed one was 24v, I just heatshrinked the voltage converter onto the fan lead.
 
Last edited:
You have to get very high end before you get no noticeable hiss from a directly connected compression driver even ignoring noise from before the amps.
I never thought about the fact that venues often use compression drivers, and I have none of those at home. I didn't even realize that the hiss could be due to the driver type and not the other unique electronics in a venue.

I've certainly heard many venue systems, but I've only ever built out one. Even in that case, I used pro amps and consumer audiophile speakers (and the subs were underpowered). Not quite typical.

Ball bearing fans are noisier than fluid dynamic bearings. You will have better selection of fans if you use a 12V fan and a small 24v->12v buck converter (ebay). I did this for my DSP fan as I needed a silent 40mm fan and the installed one was 24v, I just heatshrinked the voltage converter onto the fan lead.
I ended up with a MagLev fan that runs at 24V. I am rather reluctant to install a buck converter (switching noise source) inside my amp enclosure.

I'm assuming that 12V fan manufacturers might offer 24V versions of the same mechanical designs, but I'm just getting started sourcing fans. Actually, I'm going to make a bold guess that the only difference between 12V and 24V fans from the same family of a given manufacturer are the values of the passives on the internal circuit boards. Thus, I'm extremely reluctant to buy a 12V fan and attach it to a 24V system. It would seem preferable to let the manufacturer alter the internal PCB stuffing to adapt a quiet 12V fan to 24V input. Then again, I'm sure I have a lot to learn on this topic.
 
Last edited:
I searched Mouser for fans at 28 dBA (none were lower) and dropped $50 on a sampling that would fit my QSC RMX 850a.
Here are the results of my very poor tests. I do not even have a decibel meter. I do have an app for my iPhone, but I didn't even think about using it. That would require very careful distance between fan and mic, plus I'd have to find out how to get everyone in downtown Seattle to be quiet.

First tests were of each fan just sitting on my desk, powered by a very old linear +/-12V supply (24V total from + to - outputs). I noticed some noise that was probably due to the fans vibrating on the desk.

Second tests were with the four fans grouped together in line, and weighted down to reduce vibration. I didn't bother mounting each fan in the QSC RMX 850a because that would have taken far too much time.

RPM varies slightly between these models, and that could account for some seeming louder than other even though they're all rated 28 dBA.

Sunon MF80252V3-1000U-A99 Vapo Bearing, MagLev Motor
This is the one I eventually selected to install. It was the quietest in the group test, although it seemed louder than at least one other fan before I weighted down the fan bodies. Perhaps it was vibrating on the desk more.

Sunon EE80252B3-000U-999 Ball Bearing
Perhaps second loudest?

Orion Fans OD8025-24MB Dual Ball Bearing
Perhaps second quietest?

Delta Electronics AFB0824M Ball Bearing
This one seems to have strong magnets, and almost clicks into place as the blades are manually rotated. Just spinning the blades without power is noisier - mechanically - than the others. Not surprisingly, this model was the loudest of the four. There might be an advantage to this mechanical/magnetic design, but in this application it lost the competition.

I suppose my results are not terribly useful, given how poorly calibrated (i.e. not at all) they were. The MagLev does seem quieter, but I think I can still hear it. I'll have to live with it for a few days before I'll know whether there's enough noise to distract (during quiet moments, of course).
 
I have a QSC RMX 1450 and a QSC ISA 450 that I use in my home theater, but I changed the fans in both. Both have Noctua fans. One has a 12 Volt Noctua and an exterior DC to DC buck converter control module that I put into a small box to vary fan speed, and the other has a 24 Volt Noctua (this was an industrial use model and a bit pricey) with the leads between the amp's fan output on the circuit board and the new fan's leads running to the exterior of the amp, where I have attached a resistor to slow the fan down. The 12 Volt fan with the controller can be slowed to a complete stop. For what it's worth, even for movie viewing with four passive subs running at high volume, the air coming out of the amps is never even warm at the lowest speed. I have actually accidentally unplugged the fan cable on the amp with the external control, and run the amp for a week or so with the fan completely off without it shutting down due to overheating. When I noticed the fan was off, I had just finished watching a movie and the air coming out of it wasn't noticeably warm.

When I was doing the fan mods, I was worried about the reduced airflow that the new fans had, even at full strength, but now I think that we tend to underestimate how hard these pro units are required to work in their intended applications. In my 24' X 16' X 7' home theater, the volume levels the amps are required to produce are not going to cause overheating. My recommendation is to buy older used pro amps, QSC, Crown, etc., swap out the fans for quieter models, and use a resistor inline or some sort of fan controller to get them as quiet as you need. If you are using the amps for two channel music listening rather than loud home theater subwoofer use, you can probably turn the fans nearly off and still not overheat the amps.
 
Last edited:

kipman725

Member
Paid Member
2007-06-10 12:41 pm
Warrington
If you use the amp at so little of its full potential all the time though wouldn't you be better off with a lower powered amp that is fanless? E.G I started this journey with the APART PA2120 which is 2*120W and fanless. Its only because I use my PA as my home system that I have this requirement for silent PA amps, at home I'm barley lighting the signal present lights on my sub amp. Its easy enough to hook up a voltmeter in peak hold mode to the output of your sub amp and see what your actual power requirements are.

My testing of lower power Crown models shows that from cold they manage around a minute of full output (sine wave) on a single channel before thermal shutdown. Running them at gigs on tops they run cool but the fans do turn on while at home they don't turn on (although other people with the same models have fan turn on at idle so there are firmware differences). I'm of the opinion that most of the time PA amps have more cooling than needed but under some circumstances they don't have sufficient cooling even for music content (but these circumstances wouldn't arise at home).
 

H713

Member
2017-11-24 11:21 pm
Madison WI
A lot of amps will run into issues with their circuit breakers before they run into thermal issues, at least in my testing for a lab project. That's the case with RMx series amps or the Macro-Tech. The RMX2450 will do 2400W for short pulses (few seconds), but more than that and the breaker on the back trips. Same goes for the Macro-Tech, though to a lesser degree since it has a 15A breaker for each channel.


The Macro-Tech 3600 also has pretty aggressive SOA protection on the output devices. I've never actually seen a 3600 with blown outputs, probably for this reason. The SOA protection is more aggressive when the amp gets hot. It is also more aggressive with increasing frequency.
 
I use my QSC amps for DIY passive subs. They are all 4 Ohm drivers, and the drivers have very low sensitivity. The amps can put out 450 watts for each channel into 4 Ohms. For home theatre use, I prefer the 450 watt per channel amp over the 120 watt per channel home amp I had previously. At one point, I had the 120 watts per channel amp, and two of the subs. For music, the home amp was adequate with the two subs, but for movies, it wasn’t sufficient. Changing the home amp for the QSC fixed this. When I was looking for a more powerful amp for my two subs, the used home amps I was able to find locally were more expensive than the pro amps. It is all a balancing act and all very system and setup dependant. My sub drivers are Alpine Type R, and have terribly low sensitivity. If the subs were more sensitive, a lower powered home amp would have been fne.
 

H713

Member
2017-11-24 11:21 pm
Madison WI
There's nothing inherently wrong with using a really big amplifier for your home system. I use an MC2 MC650 for most of my listening, and it's 650W into 4 ohms. I don't use it because it's a 650W amplifier, but rather because it's a really, really great amplifier that happens to be very powerful.

The other amplifiers I discussed (the Crown and the Crest) were in for other projects, and I thought it would be interesting to compare them. The Crown is being used for a physics demo and the Crest was an amp I rescued from E-waste and rebuilt the driver boards.
 
There's nothing inherently wrong with using a really big amplifier for your home system. I use an MC2 MC650 for most of my listening, and it's 650W into 4 ohms. I don't use it because it's a 650W amplifier, but rather because it's a really, really great amplifier that happens to be very powerful.
Same here. I selected the smallest amplifier that has both balanced XLR inputs and high-current SpeakON outputs. I would have purchased a mono block if there were one available within reason, but settled on something stereo that could be bridged mono. I could have gone with the 1U half rack SPA2-60 bridged for 200W of convection cooled power, but I like sticking with the same input and output connectors that my other amps use in case I want to swap back.

Considering how much headroom the Pass Labs XA30.8 provides to my tiny LINN TUKAN speakers, it seemed only fair to give the sub a lot of headroom as well.

p.s. QSC support mentioned that it's actually possible to get more power out of a single channel without bridging, because the single power supply in the stereo amplifier doesn't have to feed both channels when the second channel is unused. I don't quite follow the math, since it seems like the supply should be capable of the same amount of total power either way (half+half versus all-on-one+nothing-on-two), but I'll take their word for it.

They also assured me that it would be fine to run the RMX or SPA with the second channel powered but disconnected (something I would normally be reluctant to do). I'm slightly tempted to try the RMX 850a in stereo mode with only one channel connected, but then I would have to rewire the SpeakON.
 
I may have mentioned that I ran the QSC RMX 850a without its fan for a day or two before purchasing assorted fans. At first, it seemed cool or not too hot to touch. My rack has 1U of space above the RMX, which makes it easy to check on the heat. The top panel of the RMX case screws to the stereo heat sinks, so the top plate probably contributes to heat dissipation.

I turned the amp off at night, but on the second day I had left it running longer than the day before. There wasn't even any music playing, but the amp was way too hot. It didn't shut down or anything, but it was certainly too hot to comfortably leave my hand there. The Pass Labs XA30.8 is toasty, but not so hot that you feel a need to remove your hand.

With the 28 dBA Sunon MagLev, the RMX 850a doesn't even reach "normal" temperatures. My 1U DAC actually runs hotter than the RMX 850a (probably because of the 8 microphone preamps in the ADC section of my DAC).

I'll eventually do some extended listening at power levels that might exercise the subwoofer amp, and then I'll check to see how warm it's getting. At least I don't notice the fan noise any more.
 
I'm assuming that 12V fan manufacturers might offer 24V versions of the same mechanical designs, but I'm just getting started sourcing fans. Actually, I'm going to make a bold guess that the only difference between 12V and 24V fans from the same family of a given manufacturer are the values of the passives on the internal circuit boards. Thus, I'm extremely reluctant to buy a 12V fan and attach it to a 24V system. It would seem preferable to let the manufacturer alter the internal PCB stuffing to adapt a quiet 12V fan to 24V input. Then again, I'm sure I have a lot to learn on this topic.
Here's an interesting note: The schematic for the QSC RMX 850 (not 'a') includes the following part = "8900-9050-1 DC FAN 24V 80X80 (+5V)"

So, now I'm rather confused. What does it mean for a 24V fan to have (+5V) appended to the description? Does that mean the fan runs on 5V internally, but takes 24V on the control lines? Does that mean the fan runs internally on 24V but take +5V control and boosts the voltage internally to 24V? I suppose it could just be a typo in the schematic... or it could mean that the PWM circuit is 5V, controlling a 24V output signal (but it seems like that detail would be documented on a different line in the parts list than the fan itself).

I know that the fans I bought accept +24V DC without letting out their magic smoke, so the (+5V) doesn't mean control voltage.
 

H713

Member
2017-11-24 11:21 pm
Madison WI
Welp, I just bought a Crest 3301 on ebay. Sold "as-is", but the listing showed it powered up with both "active" LEDs illuminated, so that's a good sign. Once it arrives and I get it running I'll report on the fan noise. It should be a good indicator of how noisy any of the Crest xx01 series 2U amplifiers with a single fan (so that should apply to the 3301, 4801 and 6001) are.

The 3301 is one of the two amplifiers from the xx01 series to use a Class AB (rather than Class H) output stage, the other being the 4601. These two models are supposedly targeted more towards the studio sector, rather than the touring sector.

My plans for this one are to install a quiet fan (if needed) and do some other modifications, then use it for driving passive midfields / nearfields. If I deem the performance of the input board to be inadequate, then I'll redesign the input board. I need to look more carefully at how they've implemented the NE5517 peak limiter circuit, but that could definitely have some not-so-nice sonic characteristics.

First I need to sit and cross my fingers that the shipping company doesn't figure out a way to destroy this one before it gets to me.
 

kipman725

Member
Paid Member
2007-06-10 12:41 pm
Warrington
attachment.php


Installed new amps now running 3*P7000s +1*P5000s. Top amp rack has 2*80mm fans blowing across the top (holes in sides) to stop accumulation of hot air. I used slim ball bearing fans but even at 6V I can hear them so I might switch to thicker fans even if I then run the risk of damaging them in transit (as they would stick out about as far as the side runners on the rack box). The noise level is also low enough I can sometimes hear very high pitched whining from the DSPs power supply. I already re-capped any suspect caps in that supply so I suspect its just intrinsic to the design, its not loud but its a dent in perfection. I would like to upgrade to a symetrix radius NX, I did have the opportunity to buy one for £400 but it was untested and sold as seen and I didn't want to risk that amount of money. At the moment I run optical SPDIF into the DIGIO which goes over the 8x8 over symnet which outputs to the amps.

When setting up the new amps I found a flaw in the design of the P-series. With my Crown CTS4200 I had set the gains on the amp channels such that 18dBu on the DSP output clipped the outputs of the amp. I did this for optimal noise performance. When looking at the P5000/P7000s output on a scope with an 18Dbu signal and the attenuators turned down to prevent output stage clipping however I saw a clipped waveform on the scope display (also no clip light). So I had to turn the DSP down to 8dBu full scale output and turn the amps up to max, at which point the waveform only clips when the output of the amp clips. This has the side effect of reducing system SNR though so at this point I get way too much hiss from my tops when sat on the sofa (even though the P series has a lower noise than the crown these changes have amplified the noise from the DSP).

Anyway I have managed to solve the situation as I don't need anywhere near full output at home so I just turn the amp attenuators down at home and if and when I get to take everything outside again I will turn the amps back up.
 

Attachments

  • amp_rack_new.jpg
    amp_rack_new.jpg
    310.6 KB · Views: 327
I had run a pair of two ways for a while using a Peavey FH1 with an Eminence Kappa 15c for LF, and an Altec 511B horn with a Renkus-Heinz SD-1800 compression drivers for HF. Using a passive crossover, and two stereo low wattage SET amps, there was no hiss from the speakers. When I installed a pro electronic crossover, the hiss started. After a lot of fiddling with gain in my system, I now have a tolerable level of hiss out of the HF horns. My listening room is a purpose built room in my basement, and is very quiet. This means that the hiss needs to be very low in order for it not to be audible. I have it at a tolerable level, but it’s still not as completely quiet as it is through the passive crossover. It’s mostly about the gain structure, but it’s also about every additional piece of equipment between source and speaker. Also, none of this matters nearly as much for your regular run of the mill 88 DB/W/M speaker. A high sensitivity speaker is a harsh mistress...
 
Last edited:

kipman725

Member
Paid Member
2007-06-10 12:41 pm
Warrington
yeah I'm at 100dB/1W (8ohms) due to the passive mid/high crossover in my speakers. I have several projects though which will use directly connected compression drivers and I think the only solution for this is going to be:
1) High end amps with low self noise (although directly connecting TPA3255 to a comp is almost there in terms of noise)
2) Relay based multi channel volume control
I suppose a 120dB SNR DSP would also work but they seem to be lacking from the market*

*Unless you use a PC but then lots of other issues that make this not work for me

**Also love the esotericism of valve amps and DSP.
 
Last edited:

conanski

Member
2013-03-31 3:53 am
When looking at the P5000/P7000s output on a scope with an 18Dbu signal and the attenuators turned down to prevent output stage clipping however I saw a clipped waveform on the scope display (also no clip light).
Means you are clipping the input of the amp right.


I'm at 100dB/1W (8ohms) due to the passive mid/high crossover in my speakers. I have several projects though which will use directly connected compression drivers and I think the only solution for this is going to be:
You don't need the full power of those amplifiers in that case though do you. So instead of turning down the DSP output level for the CD's leave it at the same level as the mids and turn the HF amp gains down that amount. Yamaha has a cover that goes over the gain controls for this exact purpose I'm sure.

I have the same issue with my setup.. P3500 or P5000 driving CDs directly but so far I have just been ignoring it because once the DJ starts playing it's a non issue.
 
yeah I'm at 100dB/1W (8ohms) due to the passive mid/high crossover in my speakers. I have several projects though which will use directly connected compression drivers and I think the only solution for this is going to be:
1) High end amps with low self noise (although directly connecting TPA3255 to a comp is almost there in terms of noise)
2) Relay based multi channel volume control
I suppose a 120dB SNR DSP would also work but they seem to be lacking from the market*

*Unless you use a PC but then lots of other issues that make this not work for me
Every time I read your posts on this it cracks me up how parallel our setups/thinking/challenges are... acoustic and electrical noise, gain staging issues, etc.

On DSP acoustic noise (since you are a fellow Symmetrix used-equipment fan)... I have a Jupiter, a couple of Radius AECs, and an Edge. The Jupiter has no fan. The Radius and Edge do have fans. They are variable speed, but in-rack, the Radius AECs' fans tend to run, and are audible. Not sure about the Edge, I haven't racked it yet.

I'm guessing a Radius NX *might* run quieter than the Radius AEC, as I suspect the AEC processors generate more heat than the 'plain old' analog IO cards. Ventilation is cross-flow. See attached photo of an Edge frame, populated with 16 ch of analog outputs.

As to the DSP and/or amp electrical noise, and which is the bigger problem... I haven't bought my QuantAsylum analyzer yet either (waiting for the QA402 to come out), so I can't be as quantitative about that as I'd like without trying harder than I want to. But here is what I can say anecdotally...

My PA tops use SA8535 HF drivers, directly connected in a 4-2ay setup. With their waveguides installed they are ca. 107db/W. The mids are 8" JBL CMCDs, about the same sensitivity. MB is lower sensitivity at about 100dB/W, and the 'domestic use' subs are fairly low sensitivity... about 92.

With my "A-rig" rack, (2x Powersoft 1204 300W/Channel, about 300W/ch for MF and HF) 2x Powersoft 5202 1250+/ch for MB and Sub), you can hear hiss at my listening position, 5m away.

Same issue with my powered Danley SM60's, which IIRC, use a 1KW PowerSoft module and a passive crossover. Both systems sound great but aren't silent.

In the above cases, I can't tell you whether the amp(s) or the DSP is the dominant source of noise without some work. But...

I also have a 'C' rig rack using smaller, lighter amplification. The intent is to use this rack for tiny venues. It will mostly get to stay put in the studio, running my nearfield monitors.

The amps in this one are little LabGruppen E-series 1U units of various outputs... nominally (L-G's numbers are pretty optimistic for anything but brief peaks) 100W/ch for the tweeters, 200 for the mids, 400 for the MB, and 600 for the LF. Same DSP (Radius AEC) as the bigger rack. The un-trimmed gain of these amps are somewhat better matched to the different driver's sensitivities.

At the moment I have this rack connected to the tops I described above, with its subwoofer amp's input gain dialed in to just barely clip on most music when the Radius is getting full-scale digital input. The other amps are dialed down to match that level with the 4-way crossover program I developed for the Radius, with mostly neutral digital gain.

The streaming source, mixing, and transport to the DSP are all digital (Dante).

With this rig, electrical noise is not a problem, and it is still frighteningly loud. Could admittedly use more power on the subs, I'll probably get another L-G E12:2 if I can find one cheap enough. Which I hadn't sold all of my Crown XLS2502s. Oops.

I can't hear hiss unless I have an ear directly facing the tweeter and am less than two feet away. Granted I suspect I'm noticeably older than you ;-(

To put a number on it, I just set up my OmniMic about 1 foot away from the tweeter, digitally muted the DSP's input, and measured an A-weighted noise level of 45dBA. That's a little above the noise floor in that end of the room, but not by too much.

I then made a 200 ohm shorting plug for the tweeter amp's phoenix input connector, and put it in place to bypass the DSP. The noise measured... 45dBA.

So I don't *think* the Radius's electrical noise is the limiting factor, at least with these amps and these particular gain settings.

All that said, I still want that QA402... I want some definitive measurements of distortion, rather than bulk noise, so maybe I can stop fretting about whether or not I really 'need' a post-DSP VCA or switched-attenuator gain control in this system.
 

Attachments

  • EdgeGuts.jpg
    EdgeGuts.jpg
    1,014.5 KB · Views: 117
  • EdgeDAConverters.jpg
    EdgeDAConverters.jpg
    873.2 KB · Views: 118

kipman725

Member
Paid Member
2007-06-10 12:41 pm
Warrington
Yep input stage clipping, I just turned my amp gains down which is fine for home use. However when I play outside I have to turn them up and there will be some hiss (much less than most PA systems) and I want perfection. It makes sense regardless there would be more noise as the amps have a higher potential output voltage and so need more gain to reach full output.

tonyman108 very interesting our setups are so close. I knew the radius/edge had fans in them but had presumed they would be like the 8x8 (40mm unit). On my 8x8 I have fitted a quiet fan and replaced most of the cover with mesh and then I have the rack cross air flow running over it. It runs cooler than stock this way, stock its very noisy. From memory though the power consumption of an edge/radius is less than an 8x8 so the cooling is probably only needed for the psu. The 8x8 does the DSP processing using loads of xilinx spartan 6 fpgas which are not very power efficient. I think my plan of attack with a radius would be to sacrifice a rack space and pull air out of the top.

JBL CMCD seems super cool from the JBL papers. I haven't heard them but I have heard a modified JBL HLA rig that runs at a local club and found it quite impressive in the right spot (space is a bit crap acoustically and too long with a bad delay system).

I'm 31 my profile picture is from when I joined Diyaudio and didn't know anything, its funny to not be taken seriously sometimes.

I'm going to get the QA analyzer soon as I'm also developing a semi commercial project (as in free designs but you can also buy complete boards off me) for an 8ch DSP amp and will need to measure it.
 
Last edited: