Advantages/disadvantages of using LM350/LM317 linear regulator over RC filter to remove PSU ripples

[rwind]

Hi,
I am trying to decide which is better - using LM350 or RC filter to remove ripples. I need clean DC of 18V. Using LM350 I loose about 1.5V but I think it will be cleaner dc. With the RC filter I can use 1Ohm, 8300uF cap and reduce the cutoff to about 20Hz which is good enough. Can someone highlight other pros and cons.
Thanks

 

[crutschow]

The regulator provides a regulated output that does not vary significantly with input voltage change or output load.
It also requires a smaller rectifier filter to get a low output ripple.

If you need a regulated output voltage closer to the input voltage, you can use a low-dropout type regulator.

The RC filter is simpler but the output voltage is not regulated.

 

[rwind]

thanks.

 

[Tony Stewart]

If you bypass the 3 terminal regulator with a fixed R so the regulator current is small, you will get lower dropout and better regulation.

This wont work on pulsed loads..

 

[rwind]

If you bypass the 3 terminal regulator with a fixed R so the regulator current is small, you will get lower dropout and better regulation.

This wont work on pulsed loads..

Can you please explain that further. What do you mean by bypass with a fixed resistor value? Thanks

 

[Tony Stewart]

It's just a poor man's boost when the 3 terminal regulator cannot supply the current needed.

When there is adequate minimum load near the maximum a power resistor calculated to pull-up most of the load current unregulated. But the 3 terminal regulator picks up the slack and pulls the rest regulated.. Hence bypassed load to result in a lower drop.

Active bypass is better with a series pass MOSFET or ultralow Rce (Sat) transistor and appropriate driver. Otherwise get a more expensive ultralow LDO regulator.

 

[moty22]

Can you please explain that further. What do you mean by bypass with a fixed resistor value? Thanks

Tony means put a resistor between the input and output terminals of the regulator.

It's just a poor man's boost

Our generation used to do that, now regulators cost less than power resistors. Regulators are easy to use and leave engineers with more hairs on their head.

rwind, if you need the 18V for delicate electronics use a regulator. If you need the 18V for an audio application then use RC. If you are already using the 18V for audio and you are getting low frequency noise then it's likely to be earthing problem rather than PSU.

 

[Tony Stewart]

Ultralow LDO's use MOSFETs instead of bipolar and naturally the cost goes up according to the MOSFET RdsOn or dropout at rated current.

I found one rated for 0.15V dropout @3A ( 75mΩ) and costs $5.70 (1) in a SMT package D2PAK5 but then it only does 5V.

If moty22 can find a cheap adjustable 3A LDO with similar low dropout for $any that works up to 18Vout, I'll send him an adult beverage.

I don't think they exist.

But then I must have been reading another thread for a 3A Audio regulator and assumed incorrectly that is what you needed, when you didn't even specify. tisk. tisk.

Consider a speaker of 8 Ohms and a power Amp that requires an output impedance at least 30 to 100x lower for good dampening, it also stipulates you have a very inefficient linear power supply with lots of headroom for 100/120 Hz ripple filtering from no load to full load.... like 30% more than Vout or 24Vdc average with < 3.5Vpp ripple before RC filter to get 18V at rated current and ripple filtered by >15dB.

I'm not sure how you plan to manage your losses and noise, but it always intuitive until you've done it before.

Define all your resistance variables first before completing any design
Transformer secondary.
Bridge
Cap ESR
R filter
Rated current
Speaker
Desired low frequency response and dampening factor >30 (poor) (good >>100)

If you desire 10% voltage ripple then bridge current will be 10x Imax average thru the bridge diodes or have significant DC drop.

In the end you may wish you had bought a PS either DC-DC https://www.banggood.com/5Pcs-Mini-...ly-Buck-Module-Step-Down-Module-p-952402.html
or use a PC PSU +/-12V

 

[moty22]

If moty22 can find a cheap adjustable 3A LDO with similar low dropout for $any that works up to 18Vout, I'll send him an adult beverage.

I don't think they exist.

I'm ready for the challenge even for a reward of "poor man's" drink. The only problem is that I have no chance of winning.
Before you cause yourself a injury, it wasn't a big mistake to assume that the current is 3A because if he needed 1A then he would have used LM317.
LDO are useful for battery operated 5V logic, when the battery is 6V you can't use a switcher.
If you are desperate for a LDO for 18V you can use the trick that I'm sure you know about, put the common terminal of the regulator on top of 13V zener.

The assumption that the supply is for an audio amp is based on my imagination. Probably like me you can also guess that it isn't for NASA.
I also assumed that rwind is using one of those one chip power amp, they have good power supply rejection. If he hears 60Hz in the speakers it is likely to be the layout.
(It is silly for a bald engineer to teach another one with hairs on his head about layout and interference. )

 

[KeepItSimpleStupid]

Do a search for a "capacitance multiplier" e.g. **broken link removed**

One of my colleagues built one to reduce the ripple on a DC power supply

 

[Tony Stewart]

I like the concept of the Cap multiplier. The ideal choice for the NPN is a Diodes Inc (nee Zetex) ultralow Rce(sat) device, which has inherently high hFE so that when the device saturates instead of a gain of 10 like normal transistors, you get a gain 4x to 5x more and 25% of the Vce drop or Rce (ESR)

**broken link removed**

This tiny NPN device ZXTN2010GTA for $0.90 at Digikey

has Vce (sat.) = 260mV @ 300mA, 6A BVCEO = 60V : RSAT = 35m ; IC = 6A ; 3W max (good for a 15W supply)
and hFE (min) = 100 @ 2A, 1V
Rce(sat) = 35 mΩ @6A for a Low Equivalent On-Resistance

Consider a bigger part for more power.

 

[rwind]

I'm ready for the challenge even for a reward of "poor man's" drink. The only problem is that I have no chance of winning.
Before you cause yourself a injury, it wasn't a big mistake to assume that the current is 3A because if he needed 1A then he would have used LM317.
LDO are useful for battery operated 5V logic, when the battery is 6V you can't use a switcher.
If you are desperate for a LDO for 18V you can use the trick that I'm sure you know about, put the common terminal of the regulator on top of 13V zener.

The assumption that the supply is for an audio amp is based on my imagination. Probably like me you can also guess that it isn't for NASA.
I also assumed that rwind is using one of those one chip power amp, they have good power supply rejection. If he hears 60Hz in the speakers it is likely to be the layout.
(It is silly for a bald engineer to teach another one with hairs on his head about layout and interference. )

Thank guys for you inputs.
Yes I am just an hobbyist messing around with electronics, no NASA engineer
My PS|U is for an audio amp based on the LM1875 chip, I have a couple of 12.6V 3A transformer that I am trying to use have very little voltage to loose. So was wondering if the LM350 was worth the voltage drop. The LM1875 itself has a good PSRR. And even using a RC of 1 ohm with 8300uF, I still loose volts on the R. So my question is for a chip amp like LM350 will RC filter be better or worse than LM350?

Thanks

 

[rwind]

I'm ready for the challenge even for a reward of "poor man's" drink. The only problem is that I have no chance of winning.
Before you cause yourself a injury, it wasn't a big mistake to assume that the current is 3A because if he needed 1A then he would have used LM317.
LDO are useful for battery operated 5V logic, when the battery is 6V you can't use a switcher.
If you are desperate for a LDO for 18V you can use the trick that I'm sure you know about, put the common terminal of the regulator on top of 13V zener.

The assumption that the supply is for an audio amp is based on my imagination. Probably like me you can also guess that it isn't for NASA.
I also assumed that rwind is using one of those one chip power amp, they have good power supply rejection. If he hears 60Hz in the speakers it is likely to be the layout.
(It is silly for a bald engineer to teach another one with hairs on his head about layout and interference. )

I have a few zeners with me, can you please explain why using the 13V zener is better than using resistors for the adjust terminal of the regulator. Thanks

 

[Tony Stewart]

The LM1875 delivers 20 watts into a 4Ω or 8Ω load on ±25V supplies.
Using an 8Ω load and ±30V over 30 watts of power may be delivered.

Seems like 12.6V 3A transformer or 37.8VA at 60% efficiency might be able to get 22W if you do everything right.

YOu would be better off putting them in series. to get 24Vac, or get a proper power supply. Even bringing out +/-12V from a PC PSU is better and almost free.

 

[Tony Stewart]

nuf said..

 

[rwind]

I like the concept of the Cap multiplier. The ideal choice for the NPN is a Diodes Inc (nee Zetex) ultralow Rce(sat) device, which has inherently high hFE so that when the device saturates instead of a gain of 10 like normal transistors, you get a gain 4x to 5x more and 25% of the Vce drop or Rce (ESR)

**broken link removed**

This tiny NPN device ZXTN2010GTA for $0.90 at Digikey

has Vce (sat.) = 260mV @ 300mA, 6A BVCEO = 60V : RSAT = 35m ; IC = 6A ; 3W max (good for a 15W supply)
and hFE (min) = 100 @ 2A, 1V
Rce(sat) = 35 mΩ @6A for a Low Equivalent On-Resistance

Consider a bigger part for more power.

Thanks.
I have couple of these at hand, hope that will do ?

TIP101: 8.0 A, 100 V NPN Darlington Bipolar Power Transistor

• High DC Current Gain - hFE = 2500 (typ) @ IC = 4.0 mAdc
• Collector-Emitter Sustaining Voltage--@ 30 mAdc
  VCEO(sus) = 80 Vdc (Min)

• Low Collector-Emitter Saturation Voltage
  VCE(sat) = 2.0 Vdc (Max) @ IC= 3.0 Adc
  VCE(sat)= 2.5 Vdc (Max) @ IC= 8.0 Adc

 

[Tony Stewart]

Do you want a 260 mV drop or a 2V drop @3A?

Big difference.

 

[moty22]

I have a few zeners with me, can you please explain why using the 13V zener is better than using resistors for the adjust terminal of the regulator. Thanks

This comment was aimed at Tony for regulating 18V using 5V regulator. I was only trying to win a beer from Tony. I don't think it's what you want.

The LM1875 wants a supply of 16 to 60V or +/- 8 to +/- 30V. If you use single rail supply, like Tony suggested you will get 32V no load and about 24V loaded. this is ok. The disadvantage is that you have to look for a circuit diagram for the amplifier that works on single supply. You will also need large cap for the speaker.
I recommend to make a dual rail (+/-) supply and use the circuit diagram that is on the datasheet.
You can make +/- 18V supply using your 2 transformers 1 bridge of 100V or more and 2 caps of 10000uF or more. You don't need any regulator, they are wast of money and time. LM1875 was designed to work with unregulated supply.
If you want circuit diagram for the supply I will draw one for you.
After you build your amp and it works ok you can try to make improvements, in other words, use it for experiments like all good electronics lovers do.

 

[moty22]

I just downloaded the datasheet. It has a circuit drawing for single unregulated supply. If your preamp is single supply then you are better of with it.

There is something else that you have to bear in mind: Tony got all his "like"s for helping people, I got most of mine for telling jokes.
Most people over 16 normally listen to music at volume of 50 mW.

 

[audioguru]

I have made many audio amplifiers using ICs and never used a voltage regulator. None of them produced hum in the speaker.
Why are people talking about 50V supplies and 20W or 22W into an 8 ohm speaker?

Rwind said he has a 12.6V transformer. Its peak is 17.8V and the full wave bridge rectifier drops it down to 15.8V which is too low for the LM1875. With a 16V supply a graph shows that its power into an 8 ohm speaker is only 2.5W when it is already clipping a little.