Have any of you intentionally rolled off the bass in your amplifiers?

[Saurav]

My main speakers go down to about 50Hz, and I use a subwoofer. I've tried taking the subwoofer signal from the amp's speaker outputs and from the linestage, and right now it's hooked up to the linestage. My speakers use the Eminence Beta 12CX, and I read a comment last night which said that their sound cleans up a lot if they aren't fed anything below 60Hz, because that significantly reduces IMD.

So, I was wondering about building a low frequency rolloff into my amps. In my case it would help my speakers, and I'm wondering if it would help the output stage too. I currently use the ASL Wave 8 (PP 6BM8), and plan to build a 2A3 SET eventually. In either case, it seems like the best way to do this would be to lower the value of the coupling capacitor so I get a -3dB point of around 60Hz with the grid leak resistor. That way, I won't introduce any additional capacitors into the signal path. For now, a quick-and-dirty solution would be to add an extra cap between the input RCA and the input resistor, so the required value would be different.

Have any of you done anything like this? What would be the negative consequences of doing this (besides the fact that the amp becomes tailored to my system, but that's easily changed if I give it to someone else)? What about phase shift, would that start to show up significantly higher than the -3dB point? I wonder if that'll be a big issue, since the phase will be messed up around that region anyway since it's transitioning to a subwoofer, and mine doesn't have a phase control.

Also, maybe this would slightly improve the performance of a cheap OPT like the 125ESE, since it won't have to handle any bass?

Thanks,
Saurav

[SY]

Yes, putting in the rolloff will make life easier for your PS and OPT. But it will also audibly weaken the bass unless you set the pole at least a couple of octaves lower than the f3 for your speakers. I'd shoot for a pole in the 10-15 Hz area. If your amp has feedback, that will extend the bandwidth a bit (assuming you put the pole within the loop, rather than at the input), so you can set the pole proportionately higher until you measure an f3 for the amp at the target frequency.

Sonoma, eh?

[Sch3mat1c]

In Revision 3 (6L6 SE) I used a .022uF cap for the first coupling cap, that's in series with the source and the 100k volume pot. Should roll around 50Hz, probably needs to be higher but anyway, I figure it reduces distortion since a big bass hit will send the thing spinning into saturation (OPT is only a P-T31 from AES, not really made for 6L6s).

Tim

[Saurav]

OK, I'll give it a shot then. And yes, I'm in Sonoma. I noticed your location too... maybe we should try and meet up sometime.

[SY]

Sure, anytime I'm not travelling. Drop me an email.

[fdegrove]

Hi,

I fully agree with SY here, best do it at thee first cap coupled stage.
Calculate the cap value according to the roll-off required.

You'll definetily end up with a happy amp and surprise, surprise...the whole system will sound much more like one piece.

Cheers,

P.S. next step up will be bi-amping...that's another chapter though.

[Saurav]

OK, I tried it, and I does sound cleaner. I ended up with a frequency of around 80Hz based on the caps I had at hand. I put the cap at the input, I hadn't thought about feedback interacting with the cap value if I put it after the splitter.

I still have a bit of a hole from around 60Hz to 40Hz, because if I turn the sub up to fill in this region, the hump at 35Hz gets just too loud. So, it's set to a compromise of 35Hz being a little loud and 55Hz being a little soft, which is working fairly well.

P.S. next step up will be bi-amping...that's another chapter though.

That, and active crossovers. Don't want to go there yet though. I've got many other things to try before I fall off that deep end

SY - why would you recommend a 15Hz pole? Wouldn't that be like a subsonic filter? My speakers have fairly high distortion below 60Hz (the AX review of the Adire HE10.1 mentioned this, I'm assuming the 12.1 would be similar), so that's what I'm trying to remove. So, weakening the bass is exactly what I'm trying to do, the subwoofer will cover that.

Or were you trying to say something else and I misunderstood you?

Thanks,
Saurav

[fdegrove]

Hi,

Assuming you tried a 0.022uF cap, you may want to to try a 0.033uF or a 0.047uF cap to plaster that hole.

What is the gridleak of the next stage anyway?

Cheers,

EDIT: Make sure the polarity of the entire system is respected, i.e. make sure your subs, main speaks etc. all are the same polarity... otherwise you'll be running around in circles 'till kingdom come.

[SY]

Plain old sloppy reading on my part. I wasn't thinking about the sub at all, just the main speakers.

Yeah, you've hit one of the limitations of mono subwoofing; if you cross it over low enough so that the mono nature of the bass is inaudible, you haven't relieved the main speakers and amps of too much of their burden. Here's a suggestion that might help that bass hole: Measure the f3 and the Q of the main speakers' rolloff. Set the pole so that you get an acoustic third order (if the main speaker is a sealed box) or an acoustic 5th order if you've got a ported box. Then modify the sub xover to give you the corresponding low pass. Don't worry about what theory says the sub polarity should be with respect to the main, just find the polarity that gives you the smoothest transition in your room.

[Saurav]

Yes, the main speakers are sealed, and I think I have something close to a 3rd order rolloff now, because the pole of the RC filter is almost the same as the f3 of my speakers. As far as adjusting polarity of my sub goes, I usually find that one way gives me lots of peaks and dips, while the other way keeps it smoother, and that's the setting I use.

Assuming you tried a 0.022uF cap, you may want to to try a 0.033uF or a 0.047uF cap to plaster that hole.

That region's been a problem with every speaker I've had, so I think it's partly a room thing. I could bring the sub up to fill it if I EQ'd the 35Hz peak out, that's something I've been thinking about.

What is the gridleak of the next stage anyway?

You can see the schematic if you follow this link: http://hifiaudiolabs.tripod.com/waveav8info/ - right click on the schematic, copy the URL and paste it into the address bar. For some reason, linking straight to the image doesn't work, Tripod blocks it.

I put the cap between the input RCA and the input resistor, which is 100K. The driver is direct coupled to the phase splitter, which is RC coupled to the PP output stage, the grid leak here is 470K.

I have .017uF in there now (.01 with .0068), I didn't have any .022s. You're right, increasing the capacitance there would help, the pole is probably too high right now.