help on lm4780

[daniel jimenez]

Hello everybody,

I'm currently designing a guitar amplifier around national semiconductor's LM4780 and I've I got a couple of questions from studying the datasheet (that thing is pretty long!).

I have built a couple of effects units and I'm currently in the process of designing a preamp for the unit that will handle all of the tone shaping so I'd like the output stage (LM4780) to be as clean as possible.

1. Is the bridge amplifier application circuit (page 5 of the datasheet) recommended? Is it reliable? I've at the application notes that heat can be a problem, could a fan-heatsink combo realistically overcome that limitation?

2. I seem to be a little confused regarding output power. I would have thought that starting from rail voltage Vcc and Vee and load impedance R_L I could calculate peak output power as
Ppeak=Vcc^2/R_load
and then divide by the square root of two to get rms power.
Some of the graphs at pages 10 and 11 just don't seem to add up to me...

3. At page 11 there are some graphs plotting Total Power Dissipation Vs Output Power/Channel. What exactly is Total Power Dissipation? Why does it go down a point? How can it be less than the Output Power at a high enough value (for R_L=8 Ohm, Output Power=60W, Total Power Dissipation=~55W)?

Hope someone can give me some answers!!!

thanks in advance

daniel

p.s. any info on how much power has been achieved by a single chip in bridge configuration?

 

[Sangram]

1. It could. In general with the LM4780 transferring the heat from the chip to the sink is a bigger issue due to limited package size and transfer area. In practice, I've run a BPA200 built using this chip as a base and with normal running it is not a problem, though the chips do run hottish (~65 degrees).

2. Peak output voltage is usually lower than supply voltage. The difference is dissipated as heat. The available audio power will always be less than the power drawn from the supply, which is roughly what your equation would result in, give or take. There are also issues with reactive loads which have been covered in the past, and also usable power output at reasonable distortion is usually around 44W into a 4 ohm load (panson covered this in a nice post).

In short, it depends

3. Power dissipation = power drawn from supply (-) power supplied to load. For some reason this is the maximum at around 30% power output for a Class AB amplifier but I'm not sure, and not sure why.

For a bridged 4780, given the package size, I don't think you should expect much more than 100W power output into an 8 ohm load in bridge mode, theory notwithstanding. The graph for bridge mode seems to bear me out, though even 1% THD would be too much. Clipping these chips is seriously ill-advised.

 

[daniel jimenez]

Allright, that seems perfectly reasonable.

Since I'm trying to get some nice, slightly distorted guitar sound, then it would be much better to have all the limiting/soft-clipping/distortion at the preamp and just make sure to adjust the preamp output a peak voltage to some level just below what would have the LM4780 clipping. Is that correct?

I'd expect 100W output power should be pretty decent for a guitar amplifier, I've heard some 50W solid state before and even some of those were mighty loud. Besides, going any further could be problematic with respect to power transformer and speaker ratings.

By the way, I'm planning on using a center tapped power transformer rated at some 150VA. A bridge rectifier and some large capacitors should (in principle) reduce ripple to a pretty small amount. How much ripple is too much?

The datasheet says it's unnecessary to have a regulated supply. Can it work like I just described above? Is there a counter-indication to my plan?

 

[Sangram]

Yes, the maximum output of the pickup + preamp + fuzz should be below the level that would cause clipping the amplifier. In practice with a 95dB+ guitar speaker you will find 100W way too much unless you plan to gig (and then there may be better choices there). I also think you should estimate approximately 80W into 8 ohm load with a 25V supply and great heatsinks, so there is headroom before clipping.

My MGR30CD is 30W using some TDA chip and I can never push it beyond about 20% of the dial settings, but it's better to crank it up on the amp a bit there so I can use the controls on the guitar for a little extra hash when it's needed.

I've used as little as 1500uF capacitors on a LM4780 and though there's nothing catastrophic, I wouldn't use anything less than 5000uF or 10000uF per rail for your application.

The crest factor of a heavy crunch sound is pretty low even though there's not much LF energy in a guitar, so you need to estimate for the fact that the amp will need to have lots of juice. And good heatsinks.

 

[daniel jimenez]

Thank you so much sangram. You've been very helpful, just as i had expected from this forum.

Now that I'm armed with knowledge and my doubts are solved I can get to touching up my design.

When I'm through with that I'll post it back here and see what you guys think.

Best regards

Daniel