will this transformer work for bridge into 4 ohms?

[aletheian]

I have a dual prim / dual second toroidial rated at 300VA at 16+16 at 115&115v/5oHz. I am running on 60Hz, 120+v from the wall. I want to bridge two lm3886 chips into a 4 ohm speaker. If I run the primaries in series, I should get 8+8 on the output windings, which will sit the supply somewhere in the ballpark of 12v per rail after rectification and considering my wall voltage.

It looks like I should be able to run two chips at that, but I wanted a "yea or nay" form someone else before I go ahead and etch my boards.

Also, what do you think the power will look like at that voltage/impedance? I am thinking around 50W.

Thanx

[duelbox]

Well using the overture design spreadsheet it looks like what you would want to do is use your 16VAC transformer for +-22.56VDC and parallel the two LM3886's for 89.01 Watts into a 4-ohm load.

I am no expert here, can someone double-check what I said?

[sidehack]

With 16V the rectified output would be about 21.2V after filtering and diode loss [(16*1.414)-1.4], wouldn't it?

The LM3886 chart has the output for a ±21V into 4ohm at about 40W.


Wouldn't paralleling double the available current? Bridging the chips doubles the available voltage (which increases the current load on the chips for a given load impedance), correct?

[aletheian]

Originally posted by duelbox
Well using the overture design spreadsheet it looks like what you would want to do is use your 16VAC transformer for +-22.56VDC and parallel the two LM3886's for 89.01 Watts into a 4-ohm load.

I am no expert here, can someone double-check what I said?

Well, I wanted to do that, but in parellel, the 4 ohm load would look like 8 ohms, so I would get under 50W that way anyhow... otherwise I would do that.

[Cro maniac]

With 16V the rectified output would be about 21.2V after filtering and diode loss [(16*1.414)-1.4], wouldn't it?

I'd say it would be 20,6 V. Voltage drop on diodes happens first, so
U=[(16-1,4)*1,414)= 20,64V. Not a big difference

With +/- 20V on LM3886 power output would be around 39 W on 4ohms with single chip, or 4 times greater in bridge mode with careful thermal design (by that I mean BIG heatsink).

[Arx]

Originally posted by Cro maniac


I'd say it would be 20,6 V. Voltage drop on diodes happens first, so
U=[(16-1,4)*1,414)= 20,64V. Not a big difference

With +/- 20V on LM3886 power output would be around 39 W on 4ohms with single chip, or 4 times greater in bridge mode with careful thermal design (by that I mean BIG heatsink).

Aren't those chips current protected? I think you would hit that LONG before you got to 156W

-Nick

[sidehack]

Each chip would only be seeing half of that though, right? Would current protection kick in before 78 Watts? That's about 4A at the 20V supply. Am I wrong in reading in the datasheet that the current limit is 11A?

[Cro maniac]

Each chip would only be seeing half of that though, right? Would current protection kick in before 78 Watts? That's about 4A at the 20V supply. Am I wrong in reading in the datasheet that the current limit is 11A?

Yes, the current limit is 11A. That equals to almost 250W of continuous power into 4 ohms. The chip would probably melt in few seconds if pushed so hard.
I made an amp with TDA7293 (same package) and raised supply above recommended in the datasheet so I was getting more than 100W from one chip to drive a subwoofer. The amp is working perfectly even today, but the chip is sandwithed between two big heatsinks to remove the heat.
68W rating of LM3886 is based on continuous sinusoidal signal. Because of dynamic nature of music, the average power is much lower, so I think you can easily push this chips a little harder.
I tried one of my LM3886 based amps with +/- 29V supply on 2 ohm speaker, and could not detect any problems. The amps weren't getting too hot, and the protection never kicked in.

[Eva]

I've been using several TDA7294 (the older version of TDA7293) for a long time. Each one has been driving a 4 ohm speaker in an active system (no crossover that would make things harder) with +-41V regulated rails and suitable heatsinking. Power output exceeds 150W per IC.

It works fine, if we consider "fine" the fact of having four ICs working heavy duty for 7 years and blowing a total of three ICs in that time. It seems that the "weaker" units blow prematurely while the "stronger" ones work forever.

[Bazukaz]

Bridging will be good in this case , because the supply voltage is low.
Expect Spike protection response at the highest output levels , but this is not a big issue.
If you would make a very good active cooling , you could probably even prevent protection circuitry from activating.Adding a fan helps a lot , but produces noise.