Trouble Designing -/+12 & 24V power supply

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hello Everyone,

I'm working on a rather intensive active monitor project from the ground up and I'd like to run by a power supply issue I've run into.

I was planning on just further biasing the voltages I created with a -/+24V power supply so that there were also -/+12V rails as well but I'm noticing that this will require so pretty hefty resistors and I don't want a lot of power lost there. Also I don't want to find some crazy transformer that has this many taps. Right now I will be using a toroidial transformer that outputs -/+ 24V and I make that DC. so there is a limited number of taps and such. Any advice on a better way of going about to get the -/+12v rails would be greatly appreciated. The 24V will be going to power amplifiers and the 12 to the op amps for the filters.

I attached a picture of the design so far, just threw it together last night.

An externally hosted image should be here but it was not working when we last tested it.


Thanks in advance!

-Zack
 
How much current do you require at 12V

The input 24V is well within the range of 7812/7912 or LM317/337 regulator.

You dont need a 34V RMS transformer, more like a 18-0-18V.

You also need to take a 0V from the centre tap of the transformer.
 
Last edited:
Well The current will be around 1 -1.5 amps, I plan on just leaving it unregulated for price and simplicity.

The transformer is actually center tapped with 24 V in series or parallel, I had to do rms for the source in my schematic because that was the only option for the Voltage of the source but 34V rms ~~ 48V p-p.
I plan on grounding the center tap and everything, but my only problem is getting the -/+12v in a possibly more efficient way? right now it'd be at least 12Watts and I feel like this is inefficient.
 
A bit more detail about the amp might help. You wrote that the 12V (+/-) is for opamps but 1-1.5A is quite a lot of current for opamps, many of which consume only 10mA or less each... 1A/10mA = 100 opamps??

Without more windings on the transformer, any method used is going to either be a loss of (voltage drop * current), or need a substantially more complex, large, and costly switching regulation subcircuit.

If at all possible you do want to regulate the supply to the opamps. With many regulators this can remain pretty simple and low cost. $1 per regulator IC and a few dozen cents for capacitors before and after the regulator IC is a reasonably small cost and part count relative to that of an entire audio amp + power supply.

However even with this, you're dropping 12V @ 1A to 1.5A... 12W to 18W is a non-trivial amount of heat, pretty significant space and heatsinks would be needed.

IMO, if you must use this transformer, AND if you do really need 1.0A or more for the opamp filters, you are better off using a smaller 2nd transformer to power it than using the resistors or regulators on huge heatsinks... but I would still regulate the output from the 2nd transformer.
 
Well Originally I thought this would just be for filtering opamps (I will be using LME49720 for the filters). However, I just noticed that my hf power amp has a supply voltage limit of -/+ 22V so I don't want to risk nearing its limits with a -/+24V supply.

Since this hf chip is only for a tweeter I don't need a lot of power sent to it. So I think it would be fine to use the -/+12V that the filters will be getting.

The -/+ 24V is for the lf power amp which is a LM3886TF.

so the breakdown of the project is:

-/+12V for ~~ 4 LME49720 chips
and the HF power amp which will be sending about 10 Watts to a tweeter

-/+ 24V for 1 LM3886TF chip
This will be sending ~ 30 Watts RMS to an 8 ohm sub

I'm building each active monitor separate just like they are normal built with separate everythings.

I have looked into regulating the supply with a chip, however it is stated that it provides a constant 1 - 1.5 amps, wouldn't this be excessive? Doesn't it only need to be given that amount of current when pulled by the chip to drive a strong signal?

I'm also having trouble calculating the current, say for the tweeter.

4 ohms 10 Watts

that would be a voltage of about 6.32 (10*4)^(1/2) = V or is this wrong? and yeah this is 10 Watts peak not RMS

I guess I need to be pointed in the right direction for regulating it with a chip

-Thanks too you have been helpful so far!
 
Hi,

Seems to me you know very little about power supplies and your numbers
are all over the place. A 10K/10K voltage divider to halve the rails won't
provide any current at all without the voltage catastrophically falling.

In an active speaker for class AB op-amps you must use regulators,
though with fully discrete fully class A op-amps CRC voltage dropping
and ripple reduction is possible, choosing the series R to drop 12V.

This would depend on the total class A circuitry current
draw, if its 1A, you'd need a 12R 20W resistor in the CRC.

The first C can be be the main caps, the second about 1/10 of the
main value, and then local decoupling caps to each class A op-amp.

e.g. 4,700uF main, 470uF after the R, 47uF for every class a op-amp.

rgds, sreten.
 
So we are clear, what is the actual rating for the transformer (from the manufacturer)? Is it rated as 2 x 24VAC windings?

Your opamps can accept up to +-17VDC rails. If you were regulating down from 24VDC to 17VDC, only 7V drop, as you mentioned your tweeters won't be consuming 10W RMS so heatsinking may not as much of an issue except that if these are built into sealed rather than ported cabinets, it may heat up inside a bit.

Using a different transformer with more secondary windings is what I'd prefer. Using a second transformer and regulated power to the opamps would be second preference. Using an LM317 regulator on a bit larger heatsinks and regulating down from +-24VDC from the one existing transformer would be third... "IF" that 24DC is the actual DC output under load from the rectified transformer you have, or it at least stays over 3V above the desired, regulated output voltage, yet not much above 24VDC.

However, if you're going to use the same/similar voltage range for the opamp filtering and the tweeter gain/power amp stage, I wonder if you need the power amp IC at all... you could just throw a basic push/pull topology buffer into the feedback loop for the (last) opamp, though if you already have the power amp IC but not the parts for the buffer then it could be a reason not to do it that way.
 
Yes the transformer is rated as 2 x 24VAC windings. I guess to clear some things up:

Would it be fine to use the current -/+ 24V rails for the larger LF opamp?

and as for the HF IC and the op amps for the filter I should regulate the voltage down using something like a LM317?

These are AB class amp chips and they will both be heatsinked
 
another question, these regulators provide a constant current and a specified voltage (via resistors) ?

So the current to the amp is always say 1.5 amps is this the same case for un regulated?

I thought the amp pulled current when needed.. how would regulating change this or is it preferred?
 
another question, these regulators provide a constant current and a specified voltage
A voltage regulator regulates voltage. That means if the load resistance goes down, it outputs more current; if the load resistance goes up, it outputs less current. If the voltage into the regulator goes up or down, the regulator adjusts to maintain its set output voltage.
You're building these active monitors "from the ground up." Design and build them to use separate transformers for the op amps and power amps.
 
Yes the transformer is rated as 2 x 24VAC windings. I guess to clear some things up:

Would it be fine to use the current -/+ 24V rails for the larger LF opamp?

and as for the HF IC and the op amps for the filter I should regulate the voltage down using something like a LM317?

These are AB class amp chips and they will both be heatsinked

24VAC on a (corrected) implementation your schematic shows would be about 32VDC. That is too high for (easy, reasonable) linear regulation or a resistor voltage divider to reach 17VDC, let alone 12VDC, whether it be an LM317 or some other linear regulator. There are ways you could make it work, but it's far from ideal and would waste a lot of power, create a lot of heat.

I'm not sure what you mean by "for the larger LF opamp". If you mean "-/+ 24V for 1 LM3886TF chip", yes LM3886TF can run from 32VDC, though this voltage is a bit higher than ideal if your (speaker) driver is less than 8 ohms. If your heatsink is on the back of the speaker cabinet (highly recommended if not in a separate amp enclosure) then cooling will be easier than if it's entirely enclosed in the cabinet. Same goes for linear regulator heatsinking or power resistor heatsinking but IMO power resistor voltage divider isn't reasonable at all except in some stranded-on-a-desert-island-and-that's-all-the-parts-I-have, type scenario.

The above is under the assumption that the transformer current rating is appropriate for the load. I'm not implying it's not, rather I don't recall seeing any spec for that mentioned.

As a couple people mentioned already, there are enough issues with the project/parts that it now seems a matter of what your primary interests are, whether they be learning to design & build, or copying an existing design to get done faster and with least issues to resolve.
 
Last edited:
So lets say I go with Sofa Spud's advice and use two separate transformers with center taps for two separate supplies, one for the power amps and on for the filter board.

I could, instead of using a hf and lf power chip, use a LM4780. One half of the chip for HF one half for LF. This would solve the problem of the two separate amplifier chips needing separate voltages.

Now when the power is rectified and smoothed with caps the voltage is higher than the transformer (forgot about that), which isn't necessarily bad, not what I was planning on, but isn't this more headroom? (The speaker can handle "35RMS/70 max" watts and is 8ohm) and the amp chip can accept quite a bit of rail voltage.

So I guess the separate transformers would be my best option, which isn't necessarily bad since each one will need to supply less current than one supplying everything.

I plan on just working out a design for a few months before I go into implementing so I'm just learning as much as I can. Thank you all for the advice so far and anything else would be greatly appreciated, I will post another (better) schematic in a few days once I'm done with finals and things are less crazy.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.