Is it possible to parallel transformers

[Mike]

can I run two cetner tapped tranformers in parallel to get double the amperage? I have a subwoofer built with an LM3886, and it distorts at low volumes. I know it isn't the speaker and think that the amplfier isn't getting enough power. Will this work?<p>Thanks, Mike

[cato]

Yes, you can parallel transformers. <p>No, I don't think it will solve your problem.<p>Which transformer are you talking about? <p>It sounds like you are trying to draw more current than the amplifier can deliver. You either need to use an amplifier with a lower output impedance or you need to increase the power supply voltage to the amplifier you have (if the amplifier can deal with the voltage increase).

[dyarker]

The thing to watch when paralleling transformers is getting the phase right. Two identicle transformers probably have the same relative phases on the leads.<p>To check, connect the primaries in parallel, then one lead of the secondaries. Apply signal to the primary. The two open secondary leads should have zero volts between them.

[k7elp60]

I have successfully paralled 4 transformers of the the same manufacture and part number. It worked fine, for a high current (10 amp) adjustable power supply.

Ned

[Mike]

Isn't it true, though that the LM3886 needs about 2A input for the full 68W output? My speaker is a 4-ohm Lightning Audio P1.10.4 10-inch car sub. I also know that the 3886 works best with 4ohm loads. I think that my transformer at top puts out 1amp. I am not really sure as I bought it at a surplus store. All I know is it is 48VCT. I guessed on amperage by comparing it to a different transformer that puts out half the voltage at one amp. Mine is about twice the size. <p>How much amperage do I need for the full 68W? And, distortion is caused by not enough power, right? Finally, I noticed that when I crank the volume, even when it isn't hot at all, and the speaker is moving out a good 3/4 inch, it does make a popping noise in the speaker. I have heard that that is caused by trying to put out more power than the transformer can put out. <p>Will in my case, if I purchase another of the exact transsformers, and connect the outputs and inputs together, work? And what happens if the phase isn't right? <p>-Mike

[Externet]

Hi Mike.
You have some misconceptions on what is happening:<p>>>>>"How much amperage do I need for the full 68W?"
Divide 68W by the supply voltage.
That is the minimum current capability needed by the power supply without any derating for inefficiencies. Twice that would be convenient.<p>>>>>>...speaker is moving out a good 3/4 inch, it does make a popping noise in the speaker. I have heard that that is caused by trying to put out more power"
The popping noise may easily be the coil hitting its travel end to the limit of its capability by either too much signal driving it or too little dampening by using the speaker without an enclosure at resonance. <p>>>>>"distortion is caused by not enough power, right?
Distortion is ANY deviation from the original source material and can originate from too many reasons to list here.<p>Put it in its proper, tight enclosure and look at the signal with an oscilloscope. If the signal clips at the rated supply voltage, you reached maximum power. If it clips at a lower voltage, the power supply needs to be capable of more Amperes. <p>Reverse paralelling the transformers phase will emit flashy smoke.<p>Miguel

[dyarker]

Ah, more info...<p>The current equals the square root of the power divided by the resistance.<p>I = (68Watts / 4Ohms)^0.5 = 4.123Amps<p>The voltage eguals the square root of the power times the resistance.<p>E = (68Watts * 4Ohms)^0.5 = 16.492Volts<p>If the 68W speaker rating is PEAK WATTS a 5 to 6 Amp, +-20 to +-24 Volt will do. If the LM3886 is a single supply chip, then 48VDC is okay
[ 16.5Volts peak * 2 = 33Volts peak to peak ].<p>On the other hand, if the 68Watt rating is CONTINUOUS Watts, that is RMS (Root Mean Square) and the PEAK is 1.414 times greater (the square root of 2 is 1.414),
68WattsRMS * 1.414 = 96.152Watts Peak.<p>Now:
I = (96.152Watts / 4Ohms)^0.5 = 4.903Amps peak
and:
E = (96.152Watts * 4Ohms)^0.5 = 19.611Volts peak.<p>A dual supply voltage of +-22V to +-30V, or 45V to 50V single; at 6A to 8A minimum.<p>-----------------------------------------
The 48V rating of your transformer is also RMS, so when it is full wave rectified and filtered the DC Voltage is around 64V,
E = (46V loaded * 1.414) - 1.4V diode drop = 63.6V. Can the LM3886 handle that?<p>That leaves room for a second RC filter, or for regulation with a 2N5886 or power FET.<p>I suggest a 1 Ohm, 10W resistor in series with the 110VAC line before the transformer(s) primary(ies). A power supply like this needs big capacitors; the resistor restricts in-rush currents that blow fuses, melts tranformers, etc.<p>Not as much fun, but more practicle, check Jameco for a 48V, 5A open frame switching power supply. Put that in the box with your speaker and LM3886, and sit back and listen.<p>Cheers,<p>Dale

[rshayes]

If two transformers are connected in parallel, their secondary voltages need to be tightly matched or some means must be added to control circulating currents.<p>If the secondary voltages are slightly different, a current will flow out of the higher voltage transformer into the lower voltage transformer. This current will be proportional to the voltage difference between the two secondaries divided by the sum of their secondary impedances. This current adds to the load current in the higher voltage transformer.<p>"Identical" transformers may be matched well enough to be directly paralleled with only a small circulating current. Different transformers probably are not matched well enough.<p>For 68 watts (continuous) into 4 ohms, you will need to supply an RMS current of 4.123 amperes. The peak current will be 5.831 amperes, and the average current will be 1.856 amperes from each power supply.<p>This is the current that will be drawn from the filter capacitors in the power supply. An equal average current must be supplied to the filter capacitors from the power line through the power transformer. With large filter capacitors, this current does not flow continuously, but instead flows in short pulses of much higher peak current. The RMS value of this current may be several times the average current. The transformer rating is in terms of RMS current, so the secondary current rating should probably be over 4 or 5 amperes to allow delivering 68 watts to a 4 ohm load.

[Mike]

If I give it the full 5 or 6A needed for the full 68W, it will have more power to drive the speaker and also less distortion because of that, right? I dobut that the input is too high. About the popping, I believe that it is the amplifier trying to draw more power than it has. There is no way a 10 inch subwoofer can only move about 1/2 inch to 3/4 inch out max.<p>How much wattage am I getting out of the system at 1A?<p> <blockquote><font size="1" face="Verdana, Helvetica, sans-serif">quote:</font><hr>The 48V rating of your transformer is also RMS, so when it is full wave rectified and filtered the DC Voltage is around 64V,
E = (46V loaded * 1.414) - 1.4V diode drop = 63.6V. Can the LM3886 handle that?
<hr></blockquote><p>Yes, the LM3886 can handle up to 74V. <p>Could the fact that I am using only 1000 uf power filter caps affect anything?<p>Thanks, Mike

[Mike]

Oh, one more thing. How can I tell if the signal clips on the scope? Will it drop or raise or something?<p>Thanks, Mike

[cato]

Draw a sine wave on a piece of paper. Now use a pair of scissors (no running) and snip off the top and bottom arches of the sine wave. Connect the disconnected wave pieces of the sine wave by drawing straight lines along the cut edges.<p>That's what clipping looks like on a scope. There can be further distortions, but thats the main event.<p>[ October 26, 2003: Message edited by: cato ]</p>

[dyarker]

<blockquote><font size="1" face="Verdana, Helvetica, sans-serif">quote:</font><hr>How much wattage am I getting out of the system at 1A?<hr></blockquote>
P = (I^2) * R
P = (1 Amp ^2) * 4 Ohms = 4 Watts<p>If you double the current through a fixed load (like 4 Ohms) it means double the voltage too, so it is four times the power
P = I * E becomes P = (I * 2) * (E * 2).<p>To avoid calculating the voltage, use formula at top.
P = (2 Amps ^2) * 4 Ohms = 16 Watts
P = (3 Amps ^2) * 4 Ohms = 36 Watts
P = (4.25 Amps ^2) * 4 Ohms = 72.25 Watts, etc.<p>---------------------------
Who makes the LM3886 so I can get the spec sheet?<p>Anyway, I suspect that the LM3886 has a self-protection circuit built in. When it goes hard into clipping (running out of power), or clipping too often, it thinks something serious is wrong, and immediately returns the output to zero. Going from max to zero quickly is a pop.<p>1000uF may barely be enough if you turn down the volume for 10 Watts max output. There is a "rule-of-thumb" formula that I can't remember just now. It tells how much capacitance is needed to keep ripple below a certain amplitute for a current load. But I have a gut estimate that a 5Amp supply will need 10000uF or 20000uF for decent sound to the speaker.

[Mike]

First of all, www.national.com supplies the LM3886. Second of all, are you kidding? I am only getting 4W?!?!?!?! And it is as loud as it is? <p>Will increasing the amperage to 5A like somebody suggested also reduce the distortion greatly, or does wattage have nothing to do with distortion? <p>Finally, I tore the amp circuitry out of it, rewired it with newer better wires, and new solder connections. As this was on of my first projects, the soldering was not good. I also noticed that when I changed to a larger heatsink, I forgot to apply thermal grease. So the popping was overheating, since it doesn't do it now with the grease. <p>Since the application is a subwoofer, and great sound and clarity is not neccecary, do I really need the large caps? Won't the 1000uf models work? <p>Finally, how much is a 5A transformer?<p>Thanks, Mike

[dyarker]

Thanx for the link, I down loaded the spec sheet.<p>Have you looked at page 16, right column? It discusses a 40 Watt, 4 Ohm, amplifier.<p>Note 12 on page 4 says the tab is internally connected to the negative supply pin. That means the heatsink cannot be connected to ground, or the LM3886 must be attached to the heatsink on a thermal washer (mica or thermofilm), and a fiber shoulder washer to keep the screw and nut electrically isolated from the heatsink.<p>What size and type heatsink are you using now? Of course if adding grease fixed it, it fixed it.<p>Amazing what 4 Watts can do ain't it? More power prevents the clipping that cato mentioned. The only kind of distortion worse than sever clipping is the popping you had. The cure is more power as you suspected in your first post. Don't believe me about the 4W peak at one Amp?
E (in Volts) = I (in Amps) times R (in Ohms)
P (in Watts) = E times I
substitute I * R for the E in the power equation and you get:
P = I * I * R = I^2 * R
run the numbers thru a calculator. <p>Mouser stock number 546-266N48 is a 48V CT 4Amp transformer for $50. Didn't see a bigger one, but that new catalog is THICK. Add a 6-10Amp bridge rectifier and larger caps, and it should make bass drums loud enough for you.<p>Still sure you don't want a pair of prebuilt switching power supplies? (Or one dual output?) Add two hand wound inductors (not hard) and the 1000uF caps (you already have) to filter switching noise.

[Mike]

No, I believed you, you showed my the equation. I just can't believe how much it shakes the room now with only 4W!!! I can't even begin to imagine 68W!! Thats about 20 times the power!! I don't know the exact type of heatsink, but it does it's job well. I had the volume set to as close as I can to max before clipping and ran it for about an hour and and a half. No problems at all. I opened the back, and felt the chip, and it was cool to the touch. The popping doesn't happen at all anymore. It was caused originally by the input wires being loose plus the volume conrol wire being loose, and when I rewired it, I re-soldered everything. Now it doesn't ever happen anymore. <p>Are there any transformers that will get me a decent power rating without spending that much money? $50 is a little too much so spend on a transformer for me. <p>I don't know if I mentioned this, but I am not really sure what a switching power supply is. If it would work better, how do I go about building one?<p>Thanks for the help, Mike