Powersoft Digam 5000 on T48-3015LF

[thijs666]

Hi all,

I recently purchased a Powersoft Digam 5000 second hand for a nice price. It's rated at 750W/8 Ohms RMS/Channel and stable to 2 Ohms (2050W/channel). So it should be able to power up to 10 T48's .

Anyhow, one of its 'features' is a "clip limiter and permanent signal compressor". What this does, is protect the speakers from excessive clipping. I can say that it works. If you drive the amp hard (red LED lights up), it gradually decreases it's gain until you're in the 'safe zone'. But the detection of a too large signal is a little odd: sine(like) waveforms are compressed/decreased the most. If driven with a pure sine, the amp turns down its gain until it puts out no more than 30~35Vrms . Effective for 'normal' music, counterproductive for dubstep ...

The facts:
750W/8 Ohms ~ 77 volts
450W/8 Ohms ~ 60 volts (= max voltage for Eminence 3015LF)
Peak 'music' power 3015LF: 900W ~ 85V
35 volts ~ 150W/8 Ohms (= quite overdone 'protection' if you ask me, but safe for a 3015LF, that's for sure)

The thing is, this Digam 5000 can output 750W shortly during bassdrum peaks and doesn't compress the sound (this is what Powersoft intended for these amps, amplify music, not sinewaves). Now would it be safe to rely on the clip limiter/compressor of the amp when powering a T48-3015LF loaded? Or posed in a different way: can the 3015LF take 750 watts of music power (not sine waves) when mounted in a T48? Or: is 60V the limit for linear excursion of the cone or for Xmech?

I have a DCX2496 and set the limiter for 60V output (quite difficult when the amp starts decreasing its gain due to the sine wave signal after only a few seconds). If I switch the limiter off, I hear quite some more power in the kick, while any continuous sinelike wave gets compressed/turned down. It seems to sound cleaner without the DCX2496-limiter...

I find it quite difficult to write down exactly what I am tryin to say, so please bear with me. If something needs more explanation, please ask!

[Grant Bunter]

I think there's more than a few amps out there now that peak limit as your's does, to prevent overheating and consequent thermal shutdown of the amp. Not just to prevent damage to your drivers, it also protects the amp. It means the show goes on, albeit at reduced power for varying times.

I'd be wanting to use the DCX to limit the driver according to the plans...
It would be interesting to know how your amp sees transients, and if it limits them also.

[Gregory East]

Yes, transient peaks would scare me since I don't know enough about the practicalities of limiting. If you were routinely slamming over 60v I think that would be bad news.

[thijs666]

It would be interesting to know how your amp sees transients, and if it limits them also.

What exactly do you mean? You mean transients that would result in clipping/an output voltage that the driver stage wouldn't be able to deliver? Or just peaks above the 30~35 V limited continuous RMS signal?

I'm pretty sure it lets most of the transients through when in 'limiting mode'.. That's also why it sounds 'cleaner' at high volumes'; there's no 60V brick wall limit. I'll put a scope on the output again to verify and report back.

[thijs666]

If you were routinely slamming over 60v I think that would be bad news.

Why? Because of the thermal limit of the driver or because of over excursion or too much strain on the cone? That's what I'd like to know. The specs of the 3015LF allow for 900W peak which is probably its thermal limit (EIA 426A/measured, explained on (mostly page 6) of http://www.diy-audio.narod.ru/litr/Keel ... -426-B.PDF). So I´m very interested in learning a little more detail about what that 60V limit is based on .

[Grant Bunter]

It would be interesting to know how your amp sees transients, and if it limits them also.

What exactly do you mean? You mean transients that would result in clipping/an output voltage that the driver stage wouldn't be able to deliver? Or just peaks above the 30~35 V limited continuous RMS signal?

I'm pretty sure it lets most of the transients through when in 'limiting mode'.. That's also why it sounds 'cleaner' at high volumes'; there's no 60V brick wall limit. I'll put a scope on the output again to verify and report back.

The transients I'm talking about are those that Bill describes in the "protecting your drivers" section of the plans for the Titans. According to that section, if your amp has 10dB headroom, then your transients could be reaching 7,500W
It seems to me that these transients may be of such short duration that clip detection circuitry and inbuilt limiters don't have sufficient time to engage before the transients have come and gone, and done their damage...

It would be interesting to see what you find from scoping it...

[Grant Bunter]

If you were routinely slamming over 60v I think that would be bad news.

Why? Because of the thermal limit of the driver or because of over excursion or too much strain on the cone? That's what I'd like to know. The specs of the 3015LF allow for 900W peak which is probably its thermal limit (EIA 426A/measured, explained on (mostly page 6) of http://www.diy-audio.narod.ru/litr/Keel ... -426-B.PDF). So I´m very interested in learning a little more detail about what that 60V limit is based on .

Over excursion in horn loaded cabs.
The 60V limit imposed by Bill is arrived at by cab modelling software, and a product of many factors. The most significant 2 of those factors are Xmax and impedance.

I think there's a lot of drivers not getting to thermal limit in Bill's designs. Due to the efficiency of the horns, they don't need to!

[Gregory East]

If you were routinely slamming over 60v I think that would be bad news.

Why? Because of the thermal limit of the driver or because of over excursion or too much strain on the cone? That's what I'd like to know. The specs of the 3015LF allow for 900W peak which is probably its thermal limit (EIA 426A/measured, explained on (mostly page 6) of http://www.diy-audio.narod.ru/litr/Keel ... -426-B.PDF). So I´m very interested in learning a little more detail about what that 60V limit is based on .

You and me both.

I don't know that the excursion limit voltage is set exactly at xmax.

[Grant Bunter]

You and me both.

I don't know that the excursion limit voltage is set exactly at xmax.

Xmax and impedance being the 2 most critical factors in determining voltage setting came up in another post, and that answer came from Bill.
The original question was something like "why are the excursion limit voltages the same for the lab 1x and 301x when Xmax and VD for each driver are different...
Bill replied: Impedance. I asked if Vd came into play and Bill said yes, but the 2 most important factors are Xmax and impedance...

I've just done did a bit more reading.
While you may already know this, I didn't.
Thermal limit is purely a figure at which the VC will fail with prolonged exposure to that level of power. Peak RMS may or may not be based on thermal limit.

And more reading suggests that your amp may clip without the clip light showing. I assume that comes back to either the response time of clip detection circuitry, or the duration of the clip. Doesn't mean it's not happening however...

While it's handy to know how voltage limits are set, banging away above a set/recommended voltage limit is a bad thing for no other reason than it will most likely lead to driver failure.

To me, all that is simply stronger arguments for limiting properly at the specified voltage.

[SirNickity]

Or posed in a different way: can the 3015LF take 750 watts of music power (not sine waves) when mounted in a T48? Or: is 60V the limit for linear excursion of the cone or for Xmech?

Think of it this way:

Let's say you have a piece of notebook paper glued into a picture frame. If you put your fist up to the paper until you meet resistance, it will bulge a little. This is Xmax. If you keep pushing, it will stretch a tiny bit, then tear. That's Xmech.

The difference between Xmax and Xmech varies by driver. I don't remember off-hand what it is on the 3012/3015, but on one of the drivers I looked at, it was like 1mm difference. (That was more likely the Alpha 6s... or something else even.) If there's 1mm difference between "non-linear" and "unable to move any further", that's not much margin for error.

Now, imagine that paper suspended in a frame again. Think about how well it would tolerate not just being pressed into by your first, but being punched 60 times a second.

[thijs666]

Think of it this way:

Let's say you have a piece of notebook paper glued into a picture frame. If you put your fist up to the paper until you meet resistance, it will bulge a little. This is Xmax. If you keep pushing, it will stretch a tiny bit, then tear. That's Xmech.

The difference between Xmax and Xmech varies by driver. I don't remember off-hand what it is on the 3012/3015, but on one of the drivers I looked at, it was like 1mm difference. (That was more likely the Alpha 6s... or something else even.) If there's 1mm difference between "non-linear" and "unable to move any further", that's not much margin for error.

Now, imagine that paper suspended in a frame again. Think about how well it would tolerate not just being pressed into by your first, but being punched 60 times a second.

Xlim of the 3015LF is 9.6 mm; Xmech is 17 mm... little more margin, that's why I thought my question was legit.

And I believe it's not particularly the paper that's torn; it's mostly the voice coil (bobbin) hitting the back plate rather than cone or surround tear, although there are exception of course.

[SirNickity]

Well it's always a legitimate question.

You could be right about the voice coil for the 3015's -- I know generally speaking that's a common mode of failure, but I haven't researched what gives first on any particular driver. I assume the design dictates which component is the limiting factor, rather than it being a constant? It could be the limit of the spider, the surround, or the length of the pole piece. It's gotta be something, and I don't suppose it really matters which. The end result is the same.

[thijs666]

Finally done some measurements.

Here's an overview:


All screendumps are seperately viewable/downloadable from a Photobucket album: http://s446.photobucket.com/albums/qq18 ... Digam5000/.

Explanation:
Files 1xx/first row are from Kyza - Go (Bar 9 Remix), a dubstep song, http://youtu.be/bZumuZZEBrE. Amp driven into hard clipping (red LED continuously on, at least until the compressor kicks in.

Files 2xx/second and third row are with the same song and settings as the 1xx one, only a little more zoomed out and showing a longer period. 'Breathing' of the compressing circuit can easily be seen, as wel as the absence of hard clipping, except for clipped elements already present in the song.

Files 3xx/fourth row are from the song You Don't Know Me by Milow, http://youtu.be/jCErmS3Q-YQ. Amp driven into clipping (red LED lighting up occasionally with the kick). No clipping to be seen in the peaks.

Oh, never mind the scale. I had to trick my scope because neither of the speaker outputs is connected to ground; they are DC offset by about 63 volts, so I had to watch my fingers .

Measurement suggestions welcome .

[thijs666]

Well it's always a legitimate question.

You could be right about the voice coil for the 3015's -- I know generally speaking that's a common mode of failure, but I haven't researched what gives first on any particular driver. I assume the design dictates which component is the limiting factor, rather than it being a constant? It could be the limit of the spider, the surround, or the length of the pole piece. It's gotta be something, and I don't suppose it really matters which. The end result is the same.

I agree. But in this thread I was hoping to find out which limit it was the sets the max voltage to 60V and whether it would be a problem if it would be exceeded for a short time/short peaks that are present in a musical signal.

[Grant Bunter]

Did Xmax and impedance give you your answer, or were you looking for something else?