Two regulators from one transformer

[MikeMl]

..I've did this but with 2w resistors cause that I had available, voltage was 11v on both sides.

Using 1KΩ resistors? If so, then the 18Vrms you quoted for the transformer secondary voltage is measured between the ends of the center-tapped secondary winding, making the voltage only 9Vrms between the center-tap and either end. See this:

To have enough head-room to keep the 78xx/79xx regulators from dropping-out, you would have to use a transformer that has a minimum of ~28Vrms across the ends of the center-tapped secondary winding.

 

[spec]

Hi Spec I've did this but with 2w resistors cause that I had available, voltage was 11v on both sides.

Thanks Andrew,

Afraid there is a problem using that transformer. The minimum input voltage that would be needed for a 7812 would be 15V, and that would be to the lower excursion of the ripple on the supply. Similarly the minimum voltage required by a 7809 would be 11V. The reason for this is that the 78/79 regulators have a drop out voltage of 3V.

Another complication is that when you take more current from the bridge the voltage will drop further.

As a general rule the DC rectified voltage from a bridge rectifier will typically equal the RMS value of the secondary voltage. But this relationship does not always hold up,

spec

 

[Andrew Borg]

Thanks Andrew,

Afraid there is a problem using that transformer. The minimum input voltage that would be needed for a 7812 would be 15V, and that would be to the lower excursion of the ripple on the supply. Similarly the minimum voltage required by a 7809 would be 11V. The reason for this is that the 78/79 regulators have a drop out voltage of 3V.

Another complication is that when you take more current from the bridge the voltage will drop further.

As a general rule the DC rectified voltage from a bridge rectifier will typically equal the RMS value of the secondary voltage. But this relationship does not always hold up,

spec

Ok thanks Mike and Spec. The thing is that 12v amp worked ok last time I used the separate version. I did not do last combined circuit yet. I will try to find another trafo but still I try your combined circuit to see.

 

[MikeMl]

Ok thanks Mike and Spec. The thing is that 12v amp worked ok last time I used the separate version.

That is because the full-wave bridge was connected across the entire 18Vrms end-to-end transformer secondary with the center-tap not connected. Doing so, makes the bottom end of the transformer GND.


This entire thread has been about the fact that you cannot use a single CT transformer winding to simultaneously feed the single-ended 12V supply, and the split +-9V supplies, because one requires the bottom end of the transformer to be GND, while the other requires that the CT be GND.

 

[spec]

Hi Andrew,

Here is a circuit to overcome the low voltage of the transformer.

The two lower diodes in the bridge rectifier are not used and can be removed if you have a discrete bridge.

What bridge rectifier are you using?

Make sure that the reservoir capacitors are at least 25V working voltage.

You can make C5 1000uF to try the circuit, and it could be that a 1000uF reservoir capacitor will do when your amp is connected to the 12V supply (you still haven't told us what the amplifier is, or is it a secret.)

spec

​

 

[audioguru]

EEK! Half-wave rectifiers? With the high value 1000μF filter caps and plenty of voltage headroom then the regulators should reduce the ripple to almost nothing.
This is the first post where I have used the "μ" Greek letter instead of the "u" English one.

 

[MikeMl]

Here is another way you can still use your existing 18V transformer. This uses full-wave rectification for the high-current 12V output that drives the PA. Half-wave rectification (charge pump) is used to get the unregulated negative voltage. As long as the -9V is lightly loaded, this should work... Watch the power dissipation in the 7812. It will have to mounted on a big heatsink...

Plenty of head-room for all three regulators...

 

[spec]

EEK! Half-wave rectifiers? With the high value 1000μF filter caps and plenty of voltage headroom then the regulators should reduce the ripple to almost nothing.

What are you saying AG. I have come across this fear of half wave rectification before and it does not make sense. There is nothing wrong with half wave rectification in the right circumstances. It just means that the ripple voltage is double, but subject to knowing what the 12V amplifier is, the current drain on the power supply is very low, hence the ripple voltage is correspondingly low.

This is the first post where I have used the "μ" Greek letter instead of the "u" English one.

very avant garde

spec

 

[spec]

Here is another way you can still use your existing 18V transformer. This uses full-wave rectification for the high-current 12V output that drives the PA. Half-wave rectification (charge pump) is used to get the unregulated negative voltage. As long as the -9V is lightly loaded, this should work... Watch the power dissipation in the 7812. It will have to mounted on a big heatsink...

View attachment 100696

Plenty of head-room for all three regulators...

Mike,

Rs (0.1R) and R1 (22R) are pure speculation: you do not know what the equivalent source resistance of the transformer is, nor current load on the 12V supply line. .

By the way, I like your circuit.

spec

 

[MikeMl]

Mike,

Rs (0.1R) and R1 (22R) are pure speculation: you do not know what the equivalent source resistance of the transformer is, nor current load on the 12V supply line. ...

Not pure speculation. I have been around the block before, so I can make a reasonable guess for the effective source resistance of a "small" transformer...
Besides, the specific value is not that critical.

As to the expected load current, I am assuming that the +-9V supplies are lightly loaded (<50mA), but the +12V supply runs an audio amp connected to a speaker, so requires at least a few hundred mA...

Here is a sim of the half-wave approach. Note how I connected the FWB. At the test currents shown and with C1=1000uF and C3=100uF, there is just enough head room to keep the regulators from dropping out. If Andrew can tell us what the load currents are, then we can come up with the required minimum values...

Making them bigger than required just increases the heat dissipation in the regulators with no benefit...

 

[MikeMl]

Not pure speculation. I have been around the block before, so I can make a reasonable guess for the effective source resistance of a "small" transformer...Besides, the specific value is not that critical...

Here is the sim repeated for different values of Rs = 10mΩ, 20mΩ, 40mΩ, ..., 320mΩ. Doesn't have much effect on V(plus), eh?

 

[Andrew Borg]

The amplifier is a 500w kinter bought it from ebay. Although true power is a 30w one when I tried it.
The bridge is a one chip with heatsink.
Regarding transformer there is a sort of inductor before it ; don 't know if this is lowering voltage.
Could not post images to this forum .
I will try the circuits of spec and mike and see.
Thanks

 

[atferrari]

This is the first post where I have used the "μ" Greek letter instead of the "u" English one.

Nothing like using a new pair of tennis shoes!!

 

[JimB]

Could not post images to this forum .

You should be able to, unless they are excessively large.
Images etc are attached to a post using the Upload a File button, located at the bottom right hand side of the reply box.

JimB

 

[kinarfi]

The thhing I have against LDOs (7812 & 7809) is the power loss (heat). That's where the use of DC to DC modules work well, IMHO. If you can keep the input volt just high enouph for the head room, the heat's not to bad, but in this case, your power losss is nearly as high as the power out. This is why I like those DC to DC modules, you can build your own also using a LM2576. My 20 amp power supply uses them.
Jeff

 

[MikeMl]

The amplifier is a 500w kinter bought it from ebay. Although true power is a 30w one when I tried it...

Is it**broken link removed**
It says it needs only 2A at 12V, which means that the 2X26Wrms and 500W??? is total BS!


If y0u are producing 30W of audio power from a 12V supply, allowing for a bit of heating in the amplifier, the 12V power supply will need to deliver at least I=P/E = 36/12 = 3A. No way is your existing transformer going to do that (for very long). A LM7812 is not suitable at this power level, either.

 

[audioguru]

Ebay spelled the 500 Watts wrong. It should be spelled 500 Whats. The details say R. M. S. 2x26W because it has two channels. Its case looks like a pretty good heatsink so it must waste a lot of power making heat even though it is "digital" (class-D?).
A single-ended fairly efficient amplifier powered from 12V produces 1.9 real Watts into 8 ohms or about 3.6W into 4 ohms. If the amplifier is bridged then the powers are about 3.5 times higher at 6.7W into 8 ohms or 12.6W into 4 ohms.
Then the 26W per channel listed is probably the "peak power" which is simply double the average RMS power and at a horrible 10% of clipping distortion.

 

[spec]

The thhing I have against LDOs (7812 & 7809) is the power loss (heat). That's where the use of DC to DC modules work well, IMHO. If you can keep the input volt just high enouph for the head room, the heat's not to bad, but in this case, your power losss is nearly as high as the power out. This is why I like those DC to DC modules, you can build your own also using a LM2576. My 20 amp power supply uses them.
JeffView attachment 100701

Hy kinarfi,

Your posts about the use of switch mode regulators are correct, cost effective, and innovative. But there is one overriding problem: switch mode regulators generate a lot of noise, typically 100mV peak to peak, while linear regulators only generate about 1mV. Because the switch mode noise is from a very low impedance it is a pig to filter out/screen and can effect sensitive circuits like audio amplifiers and RF receivers.

spec

 

[kinarfi]

I haven't dealt too much with RF or audio, so I can not say just how much noise is created in audio circuit, but it would seem that the frequency of the switch mode step down would be well above the audio frequency. As for RF, I know nothing!!
If any one would like to enlighten me, I'm definitely listening. One example of where switch mode is very common is the 12 to USB.
Jeff

 

[audioguru]

The ICL7660 negative generator IC shown in post #28 operates at 10kHz with many higher harmonics. 10kHz is plainly audible to almost everyone.