advertisement


PSU voltage reduction

PaulMB

PaulMB

pfm Member
I'm making yet another LM3886TF amp and I've ordered a 300VA 21+21 volt transformer, the closest I could find ready-made to the ideal voltage I had decided on of 18-20 volts.
21 volts, when rectified, should give 29-30 volts, right? (multiplying by 1.41)
This is a bit higher than the 27-28 I was aiming at, so I'm wondering if by connecting a display LED to the output of the reservoir caps this might drop the voltage that goes to the LM3886 by a volt or two. If not, could I drop the voltage with a diode?
(Hope all this makes sense, don't really know what I'm writing about.)
Thanks for any ideas.
 
A display LED is only going to draw a few milliamps and will make no difference to your rectified and smoothed supply.

My mains voltage is on the high side and I find that a 24Vac secondary gives 37.5Vdc, on those figures your transformer would give 33Vdc. The 1.41 factor refers to the output of the rectifier which has a certain amount of ac ripple superimposed on it; the reservoir/smoothing capacitors smooth that out which gives a slight increase in the output voltage.

malcolm
 
Voltage output depends also on the voltage regulation of the transformer. The voltage rating is usually given at full load, so 'idling' - as it will be 99%+ of the time driving an amplifier - the offload voltage rises a bit. From 5-10% is quite normal which agrees with malcolm's findings; and as much as 15-20% for very small and nasty transformers. You do lose about 1.5v from the two forward diode drops in a bridge rectifier though.

So a '24vac' transformer might well appear to provide about 35vDC with the load idling - looks OK from the LM3886 datasheet.
 
Thank you all very much. The voltage regulator idea is interesting, only downside, as Henry Ford said, "If it ain't there it can't break". Also the LM3886s I've built before seemed in no need of one.
 
I may be mistaken but I thought the rcommendtion of rails for the Lm3886 was 24-0v-24 to 27-0-27v AC before rectifying. I built one -which is currently in (ab)use by my daughter- and tried the powerreg at +25-25 (DC) but didn't like it. This being my experience. Though it sounds great, I'm open for improving it by lowering the trafo v-out.
 
The NS data sheet quotes 28 VDC max for 4 Ohm speakers, and 35 VDC for 8 Ohm speakers. So although my speakers are theoretically 8 Ohms (7 Ohms measured across the binding posts) I wanted to be cautious in case the impedance falls drastically at whatever frequency. But from what I've read here it would seem that a transformer with 21-0-21 secondaries should give around 30 VDC which should be pretty safe. (I hope!)
I don't know if lowering the voltage should improve the sound. What I did find improved the sound on a previous LM3886 amp when listening to CDs was a 1200 Ohm resistance in series and a a 2.2nF cap in parallel with the CD input. (With thanks to Martin Clark for the idea and the values). This made the sound much more mellow and warm, since it rolls off some of the very high frequencies slightly.
 
That supply limitation is really about the chipamp's own, fairly aggressive, self-protection (especially thermal). Running it hard into a 4ohm load would create more dissipation in the rather small IC package than it can stand. Hence it is rated to 50w into 8ohms and only 60-odd, not a full 100W into 4ohms nor at the same supply voltage.

So it is really a question of how loud you listen. If your speakers are fairly sensitive, and you don't have a vast room to fill, and prefer music with real dynamic range (eg classical) to heavily compressed commercial pap, it will probably all work just fine on the higher voltage rails.

(for an IC amplifier where gain is defined by feedback then the DC voltage of the supply really, really should not affect sound quality at all - if it appears to, it is probably a side-effect of this onboard protection at work)
 
While we are on the subject, do you think that a piece of aluminium, 10cm high, by 30cm long, by 5mm thick, in fact the whole side of the otherwise wooden box the amp will be built into, is enough for cooling if the chip is screwed to the inside of the metal plate? (with thermal paste)
 
PaulMB said:
Thank you all very much. The voltage regulator idea is interesting, only downside, as Henry Ford said, "If it ain't there it can't break". Also the LM3886s I've built before seemed in no need of one.
Click to expand...

Correct, but if you find the thread involving Teddy Pardo where he discussed the PowerReg he and a few others stated that it made the LM3886 change from sounding very good to one of the best amps he had heard....

I will be building a set for my intergrated amp (Gainclone+B4+AlienDac).

**edit

Here: http://www.pinkfishmedia.net/forum/showthread.php?t=39306&highlight=powerreg
teddy_pardo said:
Results are jaw dropping, it's comparable if not superior to many high end power amps discussed in this forum including my own one.
Click to expand...

Sam
 
You must log in or register to reply here.


advertisement


Back
Top Bottom