(Another) Transformer Question

[LyleCaldwell]

I plan on putting a Neovibe and a rehoused EH Deluxe Memory Man in the same 1RU enclosure.  I rehoused a DMM two weeks ago with no problem, using the 24 vac transformer that comes with it.

My thought was to split the 24 vac after the transformer (maybe getting a new 24 vac transformer with more current capacity), send the AC to the DMM, and take the AC to a bridge rectifier and then to a regulator to feed 18-24 vdc to the Neovibe.

Some questions: 

First, if the 24 vac becomes about 33 vdc after regulation, what's the best way to get this down to 18 vdc?  Standard regulator (seems a lot to ask of it)?

Second, which might invalidate my whole current plan, the DMM reverses the polarity of the AC *before* the transformer.  Does that pose a problem with the rectifier and DC side of things?  I'm not sure why the DMM does that, since AC is supposed to work either way.  When rehousing, I just kept everything stock and it worked fine.

Third, if the 33 vdc is too much to regulate down to 18 vdc, should I find a transformer that will give both 24 vac and 12 vac?

Fourth, am I making this too complicated and there's a much better solution?  Yes, I am prepared to use two separate transformers if needed.

[nelson]

Lm317 and a  heatsink.

[LyleCaldwell]

And the DMM reversing the inputs to the transformer won't affect it?

[R.G.]

DMM reverses the polarity of the AC *before* the transformer

I think you mean that the DMM makes a negative DC voltage to run from, while the Neovibe wants a positive DC voltage.

I don't have the full schematic of the DMM. Does it do the rectification with a single diode? If so, you can make the DC for the Neovibe with another single diode with a trick I can show you.

The conceptually simplest thing is to put another transformer in.

You could also get a 42 to 48Vct transformer and make +/-30Vdc with a single rectifier bridge.

So - which way?

[LyleCaldwell]

OK, here's a schematic based on the datasheet at www.st.com.  I have two regulators, one taking 33 vdc down to 24vdc and then another taking the 24vdc down to 18vdc.  The first regulator may not be needed, but I thought it would stress the second less (though I'm not sure about the total current draw using two regulators - heat sinking is not a problem).

[LyleCaldwell]

Sorry, RG, you posted right when I was putting up that schematic.  Here's the DMM schematic:  http://www.freeinfosociety.com/electronics/schemview.php?id=381

It uses -15 vdc internally, and I don't understand exactly how it gets there from the AC.  Here's the portion of the schematic in question:



By reversing the polarity of the AC before the transformer, I mean the DMM has white to black and black to white where the transformer is wired to the mains voltage.  Then it's followed by the circuit above. 

I also have a functional voltage doubler circuit that I might use instead of all this, that takes 9 vac and produces +/- 15 vdc.  From looking at the DMM schematic, the only thing affected by bypassing this stage of the DMM is the power LED, which is simple enough to rerout.  That would let me use my Pedal Power AC instead of using a transformer - no more wall warts for me, please.

DMM reverses the polarity of the AC *before* the transformer

I think you mean that the DMM makes a negative DC voltage to run from, while the Neovibe wants a positive DC voltage.

I don't have the full schematic of the DMM. Does it do the rectification with a single diode? If so, you can make the DC for the Neovibe with another single diode with a trick I can show you.

The conceptually simplest thing is to put another transformer in.

You could also get a 42 to 48Vct transformer and make +/-30Vdc with a single rectifier bridge.

So - which way?

[LyleCaldwell]

OK, had a better idea.  This should do it:

[LyleCaldwell]

Deleted

[R.G.]

Your last shot should be fine if you don't run into too much sag in the raw DC supply.

[LyleCaldwell]

Thanks RG.  How would I measure "too much" sag, and is there anything I could do to prevent it (short of using two separate transformers)?

BTW, I know the +/- 24 vac labels at the left are strictly speaking incorrect.  Those are the connections to the secondaries of a 24v transformer, and "+" is just hot.

[R.G.]

Too much has two flavors  - too much to let the regulators regulate, and so much that your transformer overheats.

Let's take the transformer first. A transfomer has a power limit that it can provide. Get beyond that and it overheats. Generally, if you can't keep the fingertip of your index finger on the transformer core, it's too hot. By using a second load on the transformer, you cause it to heat more. The only good way to tell if it will support the load is to try it and watch for overheating.

On the regulators: what you have there is two half-wave rectifiers, so each polarity of supply gets a pulse of power from the transformer every 1/60th second, or every 16 milliseconds. Between pulses, the 1000uf capacitors have to supply the full load current. That makes the actual waveform on the 1000uF caps be a sawtooth, up to a 33V or so peak and then ramping down at a rate of deltaV = Iload*16ms/1000uF. So bigger loads make the sawtooth deeper.

Three terminal regulators need about 2V across them to work right, so if your positive regulator ever gets less than 26V at its input terminal, it falls out of regulation and the blip gets passed through to its output. For this setup, that happens at I = 1000uF* (33V-26V)/16mS = 0.437A. No, it won't fall out of regulation. The negative 15V regulator is even safer. However, that analysis doesn't take into account that the transformer winding resistance is going to lower the peak voltage from 33V to something lower.

I'd say you're OK if you don't overheat the transformer.

[LyleCaldwell]

Thanks very much.  You must hear that a lot around here.   

How does one predict the effect of the transformer winding resistance? I can go up to a higher voltage transformer to be safer with the +24v regulator, but will that make the -15v regulator work too hard?  I plan on using a TO-220 with a fairly large heatsink, which I could connect to the aluminum chassis if needed.

[R.G.]

The difference between the regulator output voltage and the input voltage times the load current is dissipated in the regulator. If you have 33V into the 15V regulator and 100ma out, then the 15V regulator is burning off 18V*100ma = 1.8W of heat.  I don't know if the DMM uses 100ma or not; that seems kind of large, but maybe. In any case, a TO-220 can dissipate a bit over 1W in free air. With a heat sink on it, it will not die. If the regulators get too hot, put a resistor between them and the raw DC supply to burn off some of the heat.

[LyleCaldwell]

The supply for the DMM is rated 100ma, but I doubt it uses that much.   Thanks for the advice.