LM3886 at 40 volts ??

[ashok]

Before the flak starts let me make a small clarification.
I want to drive a 16 ohm load !
So this is close to its max supply ratings but within the dissipation ratings and well within its current rating.
Will it withstand it or break down immediately or after some hours of use ??

 

[traw]

I've run them up to 41 volts into 8 ohm loads. Let it run for days. I like how they sound on high voltages, really open up. Just have securely heatsinked.

 

[tlmadsen]

I dito with traw.

Using a 30V AC transform is perfect. Just remember massive cooling.

Unloaded they easely take more than the 42 V DC statet as max.

Thomas

 

[jackinnj]

you're close to the edge for 8 ohms, OK with 16 as Pdmax is about 20 watts -- make sure that your impedance is actually 16 ohms, however.

the amp does perform better with higher rail voltages -- this is easy to demonstrate if you run the power supply from a variac and measure THD vs rail voltage.

 

[AndrewT]

Hi,
I have this odd notion that most amps perform better into higher loads.
I suspect this applies even more so with a chipamp.

I think your 16ohm idea could be heavenly.
No, I don't mean dead, buried and gone to the next life!

 

[Nordic]

double the speaker impendance = half the distortion.

 

[DJ Exprice]

I'd say you could even go to 48 VDC if you wanted to! Just make sure that you remember to use the AC to DC formula:

VDC = 1.414 * VAC

OR

VDC / 1.414 = VAC

OR

VDC * 0.707 = VAC

OR

VDC = VAC / 0.707


(I love that Formula! )

So:

48VDC ? 33.9VAC

 

[AndrewT]

Hi,
if you intend pushing the supply voltage to the limit, then do keep in mind that the transformer has inherent regulation (important if one channel gets disconnected, even if only temporarily) and that the mains input voltage varies (normally within the tolerance allowed by the authorities).
The maximum PSU voltage assuming highest voltage conditions can be very much higher than the nomimal voltage predicted using that formula.
Take care to do all the sums and not just the ones you like (the answers for).

 

[ashok]

Thanks for all the suggestions.
So I guess if I use a +/- 40 volt regulated supply and use a 16 ohm load I should be OK and not expect any kind of failure of the chip.
The regulated supply should take care of supply sag from the transformer and capacitor on load and also mains voltage surge if the mains does go up by say 15 % (?) from the nominal value .

 

[AndrewT]

Hi,
that sounds sensible.

However, building setting-up and bebugging a regulated supply that in the end sounds good is not easy.

A cheaper and much quicker alternative would be line fuses and a crowbar to ground to protect your chips from severe supply overload.

 

[jackinnj]

Thanks for all the suggestions.
So I guess if I use a +/- 40 volt regulated supply and use a 16 ohm load I should be OK and not expect any kind of failure of the chip.
The regulated supply should take care of supply sag from the transformer and capacitor on load and also mains voltage surge if the mains does go up by say 15 % (?) from the nominal value .

Why not use a bigger transformer ? -- the regulator consumes energy which could go into listening.

 

[ashok]

Yes the chip does sound very nice at higher voltage. Not sure if the rest of the circuit contributes.
6922 input tube running at +100V dc and not 35Vdc.
20,000uF per rail.
Certainly gets much hotter even with music signals and needs a much bigger heatsink . A very 'small' draft of air make a huge diff to the sink temperature.
So maybe a temperature triggered computer fan running at low voltage ?