Running bridged amplifiers hard into 4 ohms

[Kieran Hall]

Hi, us again with more subwoofer questions...

We were previously asking about getting more output out of our existing three spaced-apart subwoofer cabinets. After much, much planning and negotiating (and a little bit of late-night "what the Estates department doesn't know won't hurt them" carpentry... shh!) we've now doubled the number of drivers, so we have two stacks of two JBL 4645B cabinets (left and right) and a 5749 in the centre. We still weren't able to move them all together hence the scary-looking towers of subs.

We're a bit cash-strapped now to say the least(!), so we were trying to run each pair of drivers as a 4 ohm load into ISA 750 amps. This sounds really great and the sub upgrade was well worth doing (we doubt there are too many cinemas in the area that can play this low at reference level when called for) , but we now have the problem that with extremely loud, deep and persistent LFE (rocket launches in Interstellar at 7.0 are the culprit here!), the breakers on the back of the amps trip after a while due to the massive current draw.

Luckily we took the opportunity to rewire when installing the new subs and ran the full four cores, so it looks as though the solution now is to add more amps. This is fine for left and right, but because of yet more quirks of the room it looks as though we'll continue to need to run the entire 5749 as a 4 ohm load. I was thinking either a DCA 3022 or an ISA 1350 might be able to hold its own more? Does anyone have any experience running these into subs at 4 ohm bridged? Obviously it's far from the nicest way to treat an amp, but if needs must?

[Marcel Birgelen]

I guess you have your answer right there, if the breakers on the amps pop, then you're overloading them. A DCA 3022 or ISA 1350 in bridged mode should have no trouble driving a single 5749 cabinet. Keep in mind that with such loads your wiring will also cause quite a loss. Having about 30% headroom on your specification is no luxury.

[Steve Guttag]

The ISA line shows a 2Ω stereo/4Ω bridged spec with a 1% distortion. So, it will do it...just not too cleanly. Power is power. It is showing a 2400W output into 4Ω bridged. For a 120V/20A circuit, that would be max. Your local power may differ from that and there would be a power factor in there to convert between VA and Watts. Also note, for those specs, they are referencing a 1KHz tone, not a typical subwoofer signal...so, you might get a bit less than 2000W really available.

So, how big is this room that you need, essentially, three double-18s? You don't get mutual coupling by spreading them and you likely get some odd lobing in the response and coverage, which may work against you (would need to model it with EASE.

[Kieran Hall]

Just to clarify, it's not external breakers tripping, it's the ones on the back of the amps themselves (although we tried the three amps on three different supplies and they did seem marginally more keen to stay powered on, so I guess there is some supply voltage sag at play -- still staying at a nominal 230V though, nothing drastic). We can get all the power we need to the amps, we're just pushing the ISA 750s too hard (in the edge case that is Interstellar's LFE). Good to hear that either of those two amps should hold up a bit better.

The room seats around 340 people and is 21.5m from screen to porthole (we did measure width but I can't remember off the top of my head and don't have it to hand). We'd have loved nothing more that to have the subs mutually coupled but alas, the room is what it is (we've essentially just stuffed a sub anywhere where one could plausibly go!) -- I've no doubt there's a weird radiation pattern too; the room is pretty atrocious acoustically (weird sloping roof) and we're quite pleased that we surprise pretty much all of the techs who come with how good it sounds despite all the 'on paper' atrocities going on. We now get down to 22Hz and can do it at reference, but the amps are working hard so I think Marcel's on the money about not having luxury even though 3 twin-18s in a not-all-that-huge room sounds like more than plenty.

We have 70mm Tenet coming up in a couple of weeks so we're crossing EVERYTHING that we might get something sorted by then!

[Marcel Birgelen]

Yeah, it was clear it's the breakers on the amp which are tripping, which is a clear indication that you're overpowering the amp. A sagging voltage will increase the number of Ampere going through and thus cause breakers to trip early. If you can't upgrade your power supply or it's generally unstable in your neighborhood, a solution may be to install UPSes, but bigger amps, not driven to the extremes of nearly tripping, will also help leveling out the grid side of things.

Keep in mind that distortion will increase once you start operating amps at such extremes and your speaker drivers may also suffer as a consequence of this.

Like Steve indicated, if possible, you should try to get the subs as close together as possible to avoid combing effects. It's hard to predict interference patterns, especially in rooms with complex shapes and a lot of reflective building materials, but spreading subwoofers arround should only be done if a good coverage cannot otherwise be established from a single location. (Or if you really have different subwoofer signals, but I guess it's a normal x.1 setup.) Looking at the specs you provided, it's not really a small room. but 6 x 18" drivers seems like more than sufficient as long as you drive them with sufficiently spec-ed amps.

[Steve Guttag]

The room is only 21.5m deep and, effectively, three double-18 systems with a nominal efficiency 100dB 1w/m per double. They are not clustered so you don't get full mutual coupling but with three systems you should be up to about 102dB 1w/1m...so you'd need just over 2100W total. Each amp should see about ⅓ of that (when running full tilt). So...about 700W/amp. And again, that is at 0dBFS, fader at 7.0. Worst case. Something isn't adding up.

[Kieran Hall]

When we did these calculations we took the 6 dB off of efficiency to allow for the B6(-alike – the Dirac is doing the boost) filtering. Previously with no EQ on the subs at all we were not too bad with just the three single-18s. Is this sensible thinking? (I should probably also mention that the room is much wider than it is deep)

[Leo Enticknap]

For a 120V/20A circuit, that would be max.

This is in the UK, so likely a 230V/13A circuit ... officially. The last time I lived there (though in York, not Coventry), in practice, I could get anything from 215ish to 245ish coming out of the wall. It's not like California, where I've never measured a consumer circuit more than 2-3 volts either side of what it's supposed to be, and most of the time it's 120 on the nail.

[Steve Guttag]

Width has no bearing on power needed. The 6dB occurs at one frequency with a Q of 2. You are also going to get a bit of room gain unless the room resembles an anechoic chamber. Odds are, you need worry little about the B6. Personally, given that equipement, I'd probably go to more of a DCA1222 bridged. It can handle gobs of current though QSC has been lazy about publishing its specs properly. It can drive down to 1.6Ω/side stereo and it will do an honest 700W bridged into 4Ω. Being a 230VAC country can hurt you some. QSC uses a voltage doubler for 120V so its rails can be generated from that higher voltage...for 230V, if you are lower than that, it will likely come off over the power rails, which would limit maximum voltage swing (which is lower on the DCA1222 than the higher versions...in fact, they all have higher rails than the models below them. Being able to cluster the speakers better will also benefit your cause. Floor mounting and baffle walls make things even better. The more boundaries the better without going into the corners (that gets you new problems).

[Tony Bandiera Jr]

The room is only 21.5m deep and, effectively, three double-18 systems with a nominal efficiency 100dB 1w/m per double. They are not clustered so you don't get full mutual coupling but with three systems you should be up to about 102dB 1w/1m...so you'd need just over 2100W total. Each amp should see about ⅓ of that (when running full tilt). So...about 700W/amp. And again, that is at 0dBFS, fader at 7.0. Worst case. Something isn't adding up.

I am willing to bet that one of more of the drivers (or set of drivers) is out of phase. I ran into a similar problem many years ago when I was starting as a tech, and even though I could get the proper SPL reading, it seemed that the amps were having to push way too hard...then it was suggested to test each sub individually, noting SPL readings, then turn them BOTH on. Low and behold, with them running individually, I was pushing a high but equal SPL. With both on the level dropped. After figuring out which one was out of phase (I mistakenly only compared one sub with the center and ASSumed that both were correctly phased) a quick swap of wires got everything booming away.

Right after that I bought a phase checker set and never had a recurrence of that issue.

Kieran, check your phasing on all six drivers, I am willing to bet you'll find one or more out of phase.

[Marcel Birgelen]

An out-of-phase driver may easily trip an amp, though you'd expect that you'd find those during room calibration.

But don't underestimate the potential "damaging" results of combing/interference at those low frequencies.

Also, I didn't know that Dirac did B6 filtering. It does that out of the box? You said you compensated for the B6 filtering with another 6dB boost? I hope that's not across the entire spectrum.

[Kieran Hall]

I will check the phasing again just in case... they were put in at stupid o'clock in the morning so the voice in the back of my head is starting to nag.

Marcel, I misspoke -- I mean, I took 6dB off of the expected sensitivity to compensate for the 6dB(ish) boost applied by the EQ, don't worry we've not gone mad and added more! The Dirac's probably not a proper B6 as such, but you can see that it's doing the boost at 25Hz when it's EQing the sub channel to be flat at the low end.

Thanks everyone!

[Marco Giustini]

What matters is the final frequency response. The system needs a B6 to be flat in an anechoic chamber but your auditorium is not as such. So let the Dirac do what it thinks is required to have a flat response - keeping an eye on the boosts it applies at low frequencies to avoid damages, ie, don't let the Dirac apply stupid boosts at very low frequencies.

Regarding the breaker tripping, Interstellar at 7.0 is probably a VERY demanding segment. Did you also add a high pass filter on your system? I think you have an option in the AP20/25 but the Dirac can also do that for you: make sure your curve goes down sharply around 20Hz - in fact, better to observe the response curve and match the target curve to the response curve below 25Hz: you don't want the signal to be pushed at those frequencies.

Can you share the Dirac screen here? I'd like to see the measured curve and the target applied.
Thanks!