Meanwell SMPS and LM3886?

THUMP LUMP

Member
2007-08-10 3:25 am
Not being that good with electronics, I'm having a bit of a time getting my head wrapped around this. I was thinking that I could use two of the Meanwell 24v 350w supplies to get the V- and V+ to power a pair of LM3886 based gainclones. One, would this be enough amperage for both LM3886 amps and two, can I actually get the V- and V+ from the two Meanwells SMPS? I am thinking yes on both counts but, my mind is mush at the moment and I'm not that good at this type of thing anyway. Thanks for the help.
 
ask what short term peak or transient current the 350W units can provide at 24Vdc?

Each 3886 is specified as having a minimum peak current output of 7Apk.
A pair could demand 14Apk on fast transient signals.
The 3886 is also specified as having a typical peak output current of 11.5Apk. That comes to 23Apk for a pair.

I would expect a pair each with a short term output of ~350W to 550W to satisfy a two channel chipamp.
 

THUMP LUMP

Member
2007-08-10 3:25 am
No idea if there is a internal connection between V- and PE. I haven't ordered anything yet and no schematics to be found on the web. If there is a internal connection between V- and PE, would it be an easy mod? Possibly, someone else can take a look in theirs and see. Like I said in my OP, electronics are not my strong suit.
 
Possibly, someone else can take a look in theirs and see. Like I said in my OP, electronics are not my strong suit.

I have a Meanwell SMPSU - its not identical to yours being only 200W, 24V. I just measured it so it looks like pacificblue is blowing smoke - no connection between PE and V-. I have SMPSUs from various el-cheapo sources and none has this alleged connection to my knowledge. I'd have shorted half of them if it were true as my scope ground is often connected to 0V and I run two as +/- supplies.
 

THUMP LUMP

Member
2007-08-10 3:25 am
That would be nice if the one I am looking at are like that but if they are like what Pacific is describing, it doesn't sound like the mod is much to do. Either way, it's do-able. Now, how to do it is another question. V+ from one to V- on the other with that being the 0v and the V+ and V- that are left being just that?
 

THUMP LUMP

Member
2007-08-10 3:25 am
Just wanting to be sure I am covered running 4 channels of LM3886, based on this from the "Decibel Dungeon"

"Calculating the current requirement is a bit trickier. There is a formula but you need to know some other technical information such as how much power you want your chip amp to provide. As this is a 'dummies' guide, I am going to pass on that one and give you some general guidelines instead. For each channel of a chip amp, I suggest a minimum of 80 VA. So for a stereo chip amp with one power supply I would suggest the transformer is rated at 160 VA. You can go much higher of course,with extra cost and physical size being the limiting factors. But I would not recommend going above 300VA per channel because you would be unlikely to get any further sonic advantage, and you would be into the area where you would need to consider a soft-start system that adds a whole lot more complexity to the supply. Even if you use a 300VA for each channel, I would suggest switching them on and off with separate switches."
 
Those guys appear to be assuming a transformer based supply, not SMPSU so its not entirely relevant to your project. Myself, I'd run the SMPSUs with extra output caps to deliver the short term current peaks music requires.

<edit> For my development of chipamps, I use a 30+30V 3A bench supply for a stereo pair and do very occasionally run out of current (have 2 * 10,000uF on the output). But then I don't listen very loud and I'm no bass freak. So if your listening levels and speaker efficiency (mine are active) are similar to mine I'd say 350W supplies are possibly overkill for 4 channels. Factoring in that you have passive XOs, probably about right so long as you don't forget the extra caps.
 
Last edited:

infinia

Member
2005-05-15 9:51 am
SoCal
Just some things to keep in mind about SMPS.
Be careful about adding arbitrarily larger caps on the outputs of SMPS, the control loop/s can go unstable. Usually for well designed products there are some margins, so you might get away with an additional 50% of the total C of the SMPS secondary and be fine for light loads (worse case).
Also keep in mind you may get some kind of interaction of the loops being connected together at the inputs or even with the outputs in series. I'm sure these could be made to work, but you might need some mods and adjustments in the end. Some mods could involve adding home brew common mode chokes on either/both the AC inputs and DC outputs. These chokes can work wonders for reducing hash noise of single supplies and even solve some weird loop interaction issues especially on multiple connected SMPS.
 

THUMP LUMP

Member
2007-08-10 3:25 am
The intended usage for all this is a bi-amped system. The sub already has a Hafler 200w plate amp so that is covered. The rest (what the SMPS and LM3886 amps will be used for) is for the 5" woofers and 1" dome tweets. 4 ohm on the woofer w/ 91db 1w/1m and the tweet at 8ohm at 90db 1w/1m. They are already in use and bi-amped but I just want to relieve the amps that I am currently using. Wayyyyy overkill. Kenwood Basic M2A on the woofers and a NAD 2100 on the tweets (350w @ 4ohm for the woofers and 60w @8ohms for the tweets) The output meters on the Kenwood peak at about 40w when I am playing it loud, for me and my close listening position.

I believe I will be fine on the SMPS output.
 

infinia

Member
2005-05-15 9:51 am
SoCal
post4 explains what you must check if you plan to connect two PSUs in series.

You have no option. If you want two in series, you must check. It's that simple.

This is rarely an issue for these type of frame switchers , esp. if you read the data sheet the OP linked. Besides it has advantages to keep FG floating at the output (ripple specs), and would readily be apparent to the casual observer after preliminary check out in the common chassis FG+ SE( one shorted output for series connected ).