My quest to eliminate the noise/rumble of my SVS SB-1000

[trl]

It still seems the problem is more likely on your end....

I have the same issue, I got a replacement amp plate that still has the noise, but a vit lower. Several folks are "suffering" by the same issue.

Repositioning the sub in the room, away from the ears, could hell a bit.

 

[MCH]

Hello guys, i have made some progress with my subs but i could do with some advice.

I went ahead and bought an Aiyima A07 max -the new version of the Aiyima A07 that has mono mode- and bypassed the plate amp of the SVS soldering the driver cables directly to the high level inputs of the sub. Then i can use the sub as if it was passive leaving the plate in place but unplugged:

Et voilà, the rumble is completely gone and, to my ears, it sounds fantastic! Easiest and most satisfying mod ever!

As i have two identical SVS SB1000, i took the chance to measure the modded one vs the stock one:

Regarding SPL, my UMIK reports that the stock sub can reach 105 dB at max vol (with REW pink noise) and with the Aiyima at max vol in exact same conditions it reaches 110 dB, so no issues there. Man, the little aiyima literally dances on top of the sub when maxed!

Now, of course, i am bypassing as well whatever DSP the SVS has built in. Reading the datasheet of the DSP chip i had the impression that there was nothing complex going on, probably only the volume control and low and high pass filter. But to check it out I measured some sweeps with REW and i can see some differences. I don't know how relevant (good or bad) these differences can be so i am posting the measurements seeking for advice.

Here is the full sweeps. There is no EQ at all, I just adjusted the volume of both the aiyima and the stock plate amp. I took two measurements of each just in case. Both subs were measured exactly in the same position with the mic also in the same position:

the SVS is set with the xover frequency at its max value, obviously the aiyima plays full range. Other than that, the only differences i see is the boost between 20-40 Hz in the SVS and a lower response between 45 and 90 Hz. Both filters have a strong influence in the distorsion:

SVS: the boost at 20-40 comes with a boost on distortion:

Aiyima: much less distortion at 20-40 but significantly more at 45-90:

finally phase. I dont know how to read this so i just leave it there. They look quite different. Comments are welcome.
SVS:

aiyima:

what do you think? there is no way i revert the mod and go back to the stupid rumble, but will i be missing something?
(of course, i will be applying proper low pass filter at 100 Hz - mains are kef LS50 meta - and room eq)

 

[ppataki]

Hello guys, i have made some progress with my subs but i could do with some advice.

I went ahead and bought an Aiyima A07 max, the new version of the Aiyima A07 that has mono mode and bypassed the plate amp of the SVS soldering the driver cables directly to the high level inputs of the sub. Then i can use the sub as if it was passive leaving the plate in place but unplugged:

View attachment 328299 View attachment 328300 View attachment 328303

Et voilà, the rumble is completely gone and, to my ears, it sounds fantastic! Easiest and most satisfying mod ever!

As i have two identical SVS SB1000, i took the chance to measure the modded one vs the stock one:

Regarding SPL, my UMIK reports that the stock sub can reach 105 dB at max vol (with REW pink noise) and with the Aiyima at max vol in exact same conditions it reaches 110 dB, so no issues there. Man, the little aiyima literally dances on top of the sub when maxed!

Now, of course, i am bypassing as well whatever DSP the SVS has built in. Reading the datasheet of the DSP chip i had the impression that there was nothing complex going on, probably only the volume control and low and high pass filter. But to check it out I measured some sweeps with REW and i can see some differences. I don't know how relevant (good or bad) these differences can be so i am posting the measurements seeking for advice.

Here is the full sweeps. There is no EQ at all, I just adjusted the volume of both the aiyima and the stock plate amp. I took two measurements of each just in case. Both subs were measured exactly in the same position with the mic also in the same position:

View attachment 328312

the SVS is set with the xover frequency at its max value, obviously the aiyima plays full range. Other than that, the only differences i see is the boost between 20-40 Hz in the SVS and a lower response between 45 and 90 Hz. Both filters have a strong influence in the distorsion:

SVS: the boost at 20-40 comes with a boost on distortion:

View attachment 328314

Aiyima: much less distortion at 20-40 but significantly more at 45-90:

View attachment 328316

finally phase. I dont know how to read this so i just leave it there. They look quite different. Comments are welcome.
SVS:
View attachment 328319
aiyima:
View attachment 328318

what do you think? there is no way i revert the mod and go back to the stupid rumble, but will i be missing something?
(of course, i will be applying proper low pass filter at 100 Hz - mains are kef LS50 meta - and room eq)

Excellent results!!
I would definitely leave it like that
Distortion will be tamed once you tame the frequency response

 

[MCH]

Excellent results!!
I would definitely leave it like that
Distortion will be tamed once you tame the frequency response

Thanks, I will definitely leave the aiyima there, actually I plan to buy a second one for the other sub. Now I feel I have killer small subs!
I now wonder, as the minidsp & co. get popular, why don't they sell dumb subs with just one of these small amps inside?

 

[trl]

I see a high pass filter done by the DSP at around 10 Hz, probably to protect speaker's diaphragm. Of course, it's easy to be seen the low-pass filter as well on the upper frequencies. If you can figure out a way to setup the crossover somehow it seems a great amp replacement.

BTW, did this amp swap solved any noise issues in the end?

 

[MCH]

I see a high pass filter done by the DSP at around 10 Hz, probably to protect speaker's diaphragm. Of course, it's easy to be seen the low-pass filter as well on the upper frequencies. If you can figure out a way to setup the crossover somehow it seems a great amp replacement.

Yes, normally I have my mains and subs with low and high pass filters and all is room EQed. I hear your comment about the high pass at 10Hz, sure, I can replicate that with EQ as well.

BTW, did this amp swap solved any noise issues in the end?

Yes! The noise is completely gone, and no pops when turning on/off the aiyima. The stock svs is good at that too, and this is important for me as I plan to have them switch on and off when a signal is detected like the stock svs do

 

[mdsimon2]

With the audio interface you already own, it is pretty easy to make frequency response measurements of the SVS amplifier output using this technique -> https://www.diyaudio.com/community/threads/how-to-distortion-measurements-with-rew.338511/. Should be pretty easy to match once you have that measurement. You could also try at different levels to see if they were doing something weird from a dynamic EQ perspective but my guess is they just have some sort of limiter / compressor if anything.

Glad to hear the noise was resolved! Back when I was using a miniDSP OpenDRC-DA8 (unbalanced outputs), I had audible hum on my DIY subwoofers, so I understand your pain.

Michael

 

[MCH]

With the audio interface you already own, it is pretty easy to make frequency response measurements of the SVS amplifier output using this technique -> https://www.diyaudio.com/community/threads/how-to-distortion-measurements-with-rew.338511/. Should be pretty easy to match once you have that measurement. You could also try at different levels to see if they were doing something weird from a dynamic EQ perspective but my guess is they just have some sort of limiter / compressor if anything.

Glad to hear the noise was resolved! Back when I was using a miniDSP OpenDRC-DA8 (unbalanced outputs), I had audible hum on my DIY subwoofers, so I understand your pain.

Michael

Those are great ideas, never thought there could be dynamic EQ (I have my doubts though) or about measuring the amp response directly.
Not that I want to replicate the exact original response, I will be applying room EQ on top anyways, but now the amp is there doing nothing, why not?
What I don't have is a dummy load, but was thinking, if this amp is made to be used with this driver only, can I just use the driver as load? Or would it change the response the interface sees? (excuse my ignorance)
Second question, that I didn't see in the link (other than 1% or better). Any requirements for the resistors in the voltage divider? I understand there is close to no current flowing through them, but in case I am missing something.
Thanks for the suggestions, will post my results when I get them.

 

[MCH]

With the audio interface you already own, it is pretty easy to make frequency response measurements of the SVS amplifier output using this technique -> https://www.diyaudio.com/community/threads/how-to-distortion-measurements-with-rew.338511/. Should be pretty easy to match once you have that measurement. You could also try at different levels to see if they were doing something weird from a dynamic EQ perspective but my guess is they just have some sort of limiter / compressor if anything.

Glad to hear the noise was resolved! Back when I was using a miniDSP OpenDRC-DA8 (unbalanced outputs), I had audible hum on my DIY subwoofers, so I understand your pain.

Michael

@mdsimon2 all,
I am, i think, ready to measure the frequency response of the stock plate amp of the SB-1000 but i could do with some advice before adding a 300W amp to the rig. I read the linked thread and i got myself a 4 ohm 100W power resistor and a cheap used interface.

this is the intended measurement rig:

the small resistors are 100K - 10K - 10K - 100K. The middle point goes to Sleve and the two side points to Ring and Tip. I understand a regular voltage divider but i struggle to understand this one. Will measure the output with the oscilloscope before plugging it to anything.

1. I set the interface as both output and input in REW
2. I generate a tone in REW and capture the RCA output with the oscilloscope. I used the manual level knob of the interface so that the peak to peak in the oscilloscope is roughly 100 mV (why 100 mV, i don't know, as i don't know the gain of the amp i am going to measure either, suggestions are welcome. The objective is not to destroy anything, but i will be measuring the output of the amp with the oscilloscope before plugging to the interface in any case. The amp has a volume knob, but i don't know what is the maximum attenuation, maybe it is just a few dB, no idea)

3. with these settings, i measure the frequency response of the interface plugging the L channel of the line out to the tip and sleeve of the TRS line in, just for this test. Previously i check with the oscilloscope that the sweep REW generates to measure is also 100 mVpeak to peak.

this is the result, very much in line to what Julian Krause measured for this same interface (right):

All good until here. My question at this point is if i am doing something fundamentally wrong, remember i have not used any amp nor the diy load yet.
If all ok, the plan is the following:

with the card output signal of 100 mV (as in the picture):
1. Use an Aiyima a07 max (has a proper volume knob) to do the first test with the load. Start from minimum volume and increase until i measure, say, 2 V (1W) at the load. Should measure flat
2. If nothing expledes, do the same with the SVS amp, starting also from min volumeand if the output at the load is still too hot, lower the input signal with the soudcard knob.

Does it sound reasonable? suggestions are welcome.

 

[MCH]

ok guys, today i said, bah! what can happen? and plugged the plate amp of the SB-1000 to the measurement rig:

first thing to check, the rumble, the only reason i did all this:
I plugged my scope to the resistor terminals, no input signal connected, my scope is handheld powered by battery, no chance of ground loops, this is the reading:

not precisely a flat line. Plug the soundcard and capture the RTA in REW:

This is with no signal input, volume at minimum, computer is battery powered. There it is, the very real SVS SB1000 rumble.

Now the frequency response. Adjusted the output of the soundcard to 100 mV and captured measurements starting from the lower volume that didn't give the "low signal error", ca. 5-6 notches of the volume knob until the soundcard clipping led lighted up. I took one measurement for each notch of the volume knob. These are all the sweeps superimposed:

And this basically confirms what we saw measuring the in room response: a boost at 20-40 Hz and a dip at 45-90 Hz + that super dip at 10 Hz.
Another observation is that it does not seem to be any dynamic change, but the truth is that i didn't test strong signals at high volumes.

I wonder how and why they do that dip at 10Hz. @trl suggested it is there to protect the driver. Is such dip something that i could implement with CamillaDSP when using the sub as passive? like with a very narrow q and a very negative gain? do i need to bother or can i just forget?

PS: the power resistor, without heat sink, didn't even get warm. Super easy setup to measure frequency response of amps, and you only need 5 resistors.

 

[trl]

With DSP you can, of course, setup your own high pass filter aka subsonic filter.

Ideally, all passive subwoofers (drivers) should be inhibited to reproduce frequencies below their abilities to play properly, this way will protect the driver. However, music is not happening below 20 Hz for sure, so this subsonic filter (if installed) is there just to protect the driver from possible damaged amplifiers (which might output DC voltage if damaged and no protections installed) or unwanted subsonic frequencies coming from a pick-up.

 

[MCH]

With DSP you can, of course, setup your own high pass filter aka subsonic filter.

Ideally, all passive subwoofers (drivers) should be inhibited to reproduce frequencies below their abilities to play properly, this way will protect the driver. However, music is not happening below 20 Hz for sure, so this subsonic filter (if installed) is there just to protect the driver from possible damaged amplifiers (which might output DC voltage if damaged and no protections installed) or unwanted subsonic frequencies coming from a pick-up.

Very helpful, thanks, I didn't know all this.
I guess it has been a good idea to measure the frequency response after all, now I know where exactly to set the subsonic filter. What is calling my attention is that it looks more like a dip filter Vs a high pass filter.

 

[DWPress]

What is calling my attention is that it looks more like a dip filter Vs a high pass filter.

Yes, curious but if you're implementing a new filter I would just use a HP at 25Hz with a 24-48dB rolloff and call it good. No musical information down there and you definitely don't want your subs to attempt to play it.

I have a SB-1000 as well and have never had a problem with it. Glad you were able to sort your problem and kudos to you for going so far!

 

[MCH]

Yes, curious but if you're implementing a new filter I would just use a HP at 25Hz with a 24-48dB rolloff and call it good. No musical information down there and you definitely don't want your subs to attempt to play it.

thanks for the suggestion

I have a SB-1000 as well and have never had a problem with it. Glad you were able to sort your problem and kudos to you for going so far!

I have two, and 3 plate amps, all show the same rumble. The third plate amp sent by SVS to replace one with the rumble issue. I guess i had bad luck. At some point i thought that maybe it was an issue with the European voltage (there is a jumper you need to change to switch from 110 to 240V), as i believe these subs are more popular in US, but i know at least @TheBatsEar has one without issues in Germany.

 

[trl]

@MCH, did you happen to have an audio recording with the rumble / rumbling noise? It would be interesting to hear it and also to analyse on what exact frequency is the rumble heard.

 

[MCH]

@MCH, did you happen to have an audio recording with the rumble / rumbling noise? It would be interesting to hear it and also to analyse on what exact frequency is the rumble heard.

I made a video back in the day to send to SVS. Let me see if I can find it.
If not, I could assemble back the amp+driver, no problem, but as it involves soldering and I am currently quite busy it will take a couple of weeks. Does this work?

 

[trl]

If you can find the video you've sent to SVS then great, but if not then please don't stress out much...or at least do take your time and don't rush it. Thank you!

 

[MCH]

Found it. It is like this because I wanted to show there was nothing plugged to the sub (=no ground loop possible) and needed to be short enough to be sent via email. But there are a couple of seconds of rumble in there:

 

[MCH]

Hi,
I was thinking on going a bit further with this project and i would like to ask for comments or advice on my plan.
As i mentioned, i have now running my two SVS subs as passive powered by Aiyima A07 max. However the original plate amp remains in place, unused, but intact.
I was thinking on ways to reduce a bit the clutter what basically means to try to get rid of the power bricks that feed the amps and i thought, why not repurposing the PSU of the original SVS plate amp that is there doing nothing? my reasoning is the following:
- The plate amp is rated 300 watts RMS, 700+ watts peak power, that means that its PSU must be more than enough to power one aiyima each
- The PSU is nicely placed inside the sub, with a nice means socket, a power button (that i will not use as i use a master/slave power strip but nevertheless is nice to have), and heatsinked to the thick metallic plate itself.
- I still have two unused terminals for the high level in/outs that i could use to bring the DC power out to the aiyima without having to dig a hole.

What is the problem then? according to @trl the PSU supplies 53VDC, and this is a bit too much for the aiyima, i would need to bring down the voltage to 48VDC.
I have no experience with high power step down converters, so i was wondering if what i want to do is even possible or recommended. I understand that the step down converter will dissipate some power, but i only need to decrease the voltage by 5% (5 VDC), i am guessing it should be feasible (?). There is space enough to heat sink it to the back plate, unless the converter is very big. Any recommendations for 300W (at least) converters that can do such conversion? if the output is less than 48V is also ok, currently i am running the aiyimas with 36V. Does all this make sense at all? am i missing something? Thanks for any comments.

PSU and output connector (the one to the left is the one delivering 53VDC.

 

[Wolf]

My initial question having read through this, is whether or not this is a mains hum issue being at 50-60Hz range. I am less tech savvy with powered things, so I could be all wet, but I don't see that in the THD sweep. I'm guessing that is not an issue.

A friend of mine has the older ported non-DSP version, and to my knowledge, no issues like this.