Looking to get a 22v transformer for my Peter Daniel parallel LM4780, and am wondering what the issues would be if going with 400VA over 300VA which was original suggested? Do I have to change anything in the design? My search hasn't really turned up anything concrete that I feel willing to act upon. Thanks
300va means the transformer is capable of delivering 300Watts. 400va means 400Watts.
So, as long as the voltage of the windings is the same, the wattage is not an issue.
So, as long as the voltage of the windings is the same, the wattage is not an issue.
I used 2 antek an3222 300va 22v transformers for my audiosector lm4780 boards. I didn't use a softstart and switch bounce is minimal but present. I don't know about a 400va though, try it without. Add one later if you need it. I have been very happy with the amp listening nearly every day for a year.
If you have the same voltage transformer, and VA = volts times amps, then the amperage is the only issue. So say for example that the circuit draws a maximum of 5 amps. It doesn't matter whether the power supply can supply 5 amps, 10 amps, or 100 amps. The circuit draws a maximum of 5 amps.
Compare it with a mains wall socket that is fed through a 15 amp circuit breaker. Everything that is plugged in doesn't throw the breaker. A cell phone charger will draw a few hundred milliamps; a toaster will draw a few amps. Even though 15 amps is available (and that limit is only due to the breaker).
Compare it with a mains wall socket that is fed through a 15 amp circuit breaker. Everything that is plugged in doesn't throw the breaker. A cell phone charger will draw a few hundred milliamps; a toaster will draw a few amps. Even though 15 amps is available (and that limit is only due to the breaker).
This is fine under normal conditions but if the amp blows then unless it is fused you will fry the speakers sooner with the extra current.
I always use a transformer of just slightly larger than the required current including losses in the heatsink.
In class AB power goes 60% to speaker and 40% into heatsink at worst case.
Last edited:
i had problems with 2 250VA toroids, they poped t10 with no problem.
So i took large 3,3R noninductive resistor to add some smoothness
And thats all starts fine and no drops.
BTW LM gainclones are impressive, they are effecient around 70-75%!
So i took large 3,3R noninductive resistor to add some smoothness
And thats all starts fine and no drops.
BTW LM gainclones are impressive, they are effecient around 70-75%!
Actually, this isn't true. There is more current available, but the current is determined by the Voltage and the impedance of the speaker, both of which are the same.
Is there a reason for going for 22V transformers? If you are making life so easy for the chip with effectively two 3886s in parallel I see no reason not to go to 25V or even 27V and get a good deal of extra headroom. (and that is taking into account the 4780 not being able to dissipate quite as much as two 3886s, else I would have said 30V.)
I don't know if anyone has a reason to disagree, but it seems to me a crying shame to decline the extra power at no cost at all.
I'm building it in parallel configuration.
22v So I can drive 4ohm loads safely as they require less voltage... 25v is good for 8ohm, but the LM4780 allows 4ohm operation, as opposed to the 3886 which is best suited to 8ohm, no?
Indeed they do require less voltage. However what is generally done is that the voltage is dropped, to something like 22V, in order to accomodate loads of 3 or 4 ohms and not strain the chip. The chip itself can take it in terms of nominal voltages and the current demanded, but what tends to happen when the dissipation demanded is marginal is that the chip gets almost locked into running too hot, and even quieter passages don't allow the chip to get out of a regime where the current limits come down and protection cuts in lower. Also, by itself, while warmth is a good thing for the sound, over-temp seems not to be.
But you are paralleling the chips so the load each half sees is effectively double what it would see alone. Half the current is demanded and the dissipation is halved. This means you can raise the voltage above 22V, back towards what you would have if you were driving an 8 ohm load. Personally I would take it to 27V. You are dissipating across the whole double chip exactly what you would in a single 3886 driving a 4 ohm load on each channel, but you have more area, though not quite double the 3886.
The 3886 has no problem driving 4 ohm loads and, as I understand it, the LM4780 is simply two 3886s on a single die. Also, if you go for 27V you can always wipe a volt and a bit off by using two bridges instead of one, and that could constitute an improvement too.
I can't see any reason to disagree. I think it's always best to use the highest-voltage transformer that can safely be used. Then you can have the most headroom possible, to avoid clipping, which is the most-likely way to damage your speakers. And it won't take as much reservoir capacitance to get the headroom, although you might still want a lot just to have lower ripple and sufficient transient current capability.
Hi there,
With 400va you may want to add a softstart. The problem is the initial current inrush. In my first LM3886 the breaker occasionally came in. In my latest gainclone I use a softstart that I bought off the bay for $20 (ready to use) and it works fine with 2x400 va mono setup.
Best
Michael
- Status
- Not open for further replies.