Author Topic: Negative version of LP2951 LDO  (Read 4178 times)

0 Members and 1 Guest are viewing this topic.

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Negative version of LP2951 LDO
« on: September 09, 2019, 12:55:48 pm »
I'm looking for a negative output LDO, 100ma.
I found the positive version LP2951, however I'm unable to locate its negative version.
Like the LM317 has a LM337 version, I was expecting the same for the LP2951.
Am I missing something here or the negative version just does not exist for the LP2951?

 

Offline Siwastaja

  • Super Contributor
  • ***
  • Posts: 9336
  • Country: fi
Re: Negative version of LP2951 LDO
« Reply #1 on: September 10, 2019, 06:28:03 am »
Most regulators on the market don't have a "closely related" "negative version" with a similar part number. Positive regulators are much more widely needed, thus a larger selection exists.

Just do a parametric search on your supplier to find the negative voltage regulator you need, doesn't need to be closely related to any positive regulator you are using in the same circuit.
 
The following users thanked this post: ZeroResistance

Offline Andreas

  • Super Contributor
  • ***
  • Posts: 3350
  • Country: de
Re: Negative version of LP2951 LDO
« Reply #2 on: September 10, 2019, 06:50:44 am »
LT1964 if you can live with -20V maximum input voltage

with best regards

Andreas
 
The following users thanked this post: ZeroResistance

Offline RES

  • Regular Contributor
  • *
  • Posts: 109
  • Country: 00
Re: Negative version of LP2951 LDO
« Reply #3 on: September 10, 2019, 08:10:51 am »
LT1175 ?

-20V (-25V abs. max.), adj. Vout, 200-800mA, sense, current limit pins, shdn, SOIC8, 45uA op. curr., 10uA shdn curr.

https://www.analog.com/media/en/technical-documentation/data-sheets/1175ff.pdf
« Last Edit: September 10, 2019, 08:17:22 am by RES »
 
The following users thanked this post: ZeroResistance

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Negative version of LP2951 LDO
« Reply #4 on: September 10, 2019, 08:23:57 am »
Thanks to all the suggestions.
Price also is an issue.

The
LP2951 => $0.5
LT1964 => $3.6
LT1175 => $5.8

I was looking for something in sub-dollar range!
 

Offline Siwastaja

  • Super Contributor
  • ***
  • Posts: 9336
  • Country: fi
Re: Negative version of LP2951 LDO
« Reply #5 on: September 10, 2019, 08:30:05 am »
Sub-dollar things are for mass market devices, which almost always run on single-ended supplies. A small positive regulator IC may be manufactured in tens of millions; a negative regulator in hundreds of thousands, maybe. Hence the price difference.

A regulated negative rail is something a mass market device designer really wants to get rid of, this is one of the first steps of product optimization.

Hence we have some modern opamps, comparators, ADCs etc. which can read down to ground, or even slightly below (say -0.2V), with a single-ended supply.

That's why negative regulators are a bit of niche, and as such, more expensive, with smaller selection.

For price reduction, try to think how you'd get rid of the negative rail. If you really need it, does it really need to be regulated?
« Last Edit: September 10, 2019, 08:32:40 am by Siwastaja »
 
The following users thanked this post: ZeroResistance

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Negative version of LP2951 LDO
« Reply #6 on: September 10, 2019, 08:31:47 am »
OK, so I did a recheck on Mouser,

for negative voltage LDO's looks like there is nothing in the sub-dollar range.
The best it shows is TPS72301DBVR which slots in at $2.7 but its max input is -10V.
The next best in terms of voltage is LT1964 at -20V max input.
 

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Negative version of LP2951 LDO
« Reply #7 on: September 10, 2019, 08:50:55 am »
For price reduction, try to think how you'd get rid of the negative rail. If you really need it, does it really need to be regulated?

I am trying to build a audio mixer in the virtual earth configuration.
So my thinking was virtual earth needs the non -inverting input to be at ground potential, so that pulls the inverting input to a virtual ground.
Is my thinking flawed and if so is there some means to eliminate the negative power supply in a virtual earth configuration.

 

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1408
  • Country: us
Re: Negative version of LP2951 LDO
« Reply #8 on: September 10, 2019, 02:54:22 pm »
...
I am trying to build a audio mixer in the virtual earth configuration.
...

What Siwastaja has said is correct, but audio - at least at better than low-end consumer grade - is one application where having bipolar supplies really helps simplify circuits and improve performance. That said, it is also an application where the regulation of the supplies doesn't need to be particularly good; just try to keep the ripple to a minimum.

 
The following users thanked this post: ZeroResistance

Offline edavid

  • Super Contributor
  • ***
  • Posts: 3464
  • Country: us
Re: Negative version of LP2951 LDO
« Reply #9 on: September 10, 2019, 03:28:24 pm »
1. For a simple design like this, a rail splitter would be as good as split supplies

2. Unless you know exactly what will be connected to the inputs, it's better to AC couple them

3. Don't you need a buffer after your volume control?

4. For the negative supply, why were you thinking you needed an LDO?  Would an extra volt drop really be significant?
 

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Negative version of LP2951 LDO
« Reply #10 on: September 10, 2019, 05:31:51 pm »
1. For a simple design like this, a rail splitter would be as good as split supplies
How does one do that?

Quote
2. Unless you know exactly what will be connected to the inputs, it's better to AC couple them
The inputs would be headphone outputs from other devices. What is the advantage of AC coupling, isn't it better to DC couple and get rid of the DC offset.

Quote
3. Don't you need a buffer after your volume control?
Since the outputs are phone outputs, I didn't consider it necessary.

Quote
4. For the negative supply, why were you thinking you needed an LDO?  Would an extra volt drop really be significant?
I'm currently considering the good old 317 / 337 combo but was interested in the LP2951 due to its low dropout and thought it would be easy to find the negative version of it...
Also i'm currently generating the negative supply with a charge pump, because I wanted to keep noise low and didn't go in for the switched inductor power supplies.
 

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1408
  • Country: us
Re: Negative version of LP2951 LDO
« Reply #11 on: September 10, 2019, 07:43:17 pm »
1. For a simple design like this, a rail splitter would be as good as split supplies

Functionally, sure - it just turns the mid-point of a positive supply into the 0V reference - but this also elevates the "supposed to be grounded" shell of shielded interconnects to the mid-point voltage which can cause problems when connecting to other devices, especially if they share a chassis ground and/or are earthed.

2. Unless you know exactly what will be connected to the inputs, it's better to AC couple them

Agreed, though with proper bipolar supplies it is perfectly acceptable - and even preferred - to just use a coupling capacitor on the input, where even a small capacitance will still allow for a low frequency 3dB point in the single Hz range. The output coupling capacitor can - and should - be dispensed with.

Always keep in mind that wiring 3 or 4 RC networks in series with overall gain present is a great way to make an oscillator...


1. For a simple design like this, a rail splitter would be as good as split supplies
How does one do that?

With just a voltage divider, technically, but it's better to buffer the divider output with an op-amp follower so you don't have to use inconveniently small value resistors. Note that this circuit works best when the current drawn from each rail with respect to 0V is balanced, but that is pretty much always the case with audio.

 
The following users thanked this post: ZeroResistance

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Negative version of LP2951 LDO
« Reply #12 on: September 10, 2019, 09:24:32 pm »
Agreed, though with proper bipolar supplies it is perfectly acceptable - and even preferred - to just use a coupling capacitor on the input, where even a small capacitance will still allow for a low frequency 3dB point in the single Hz range. The output coupling capacitor can - and should - be dispensed with.

Always keep in mind that wiring 3 or 4 RC networks in series with overall gain present is a great way to make an oscillator...

So If I have 2 stages,
1 Summing amplifier and
1 Power amplifiers

DC blocking caps at the Summing amplifier is sufficient?
Is there No need to add it further down the line? Won't the summing amp stage add a dc offset at its output?
 

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1408
  • Country: us
Re: Negative version of LP2951 LDO
« Reply #13 on: September 10, 2019, 10:04:09 pm »
...
So If I have 2 stages,
1 Summing amplifier and
1 Power amplifiers

DC blocking caps at the Summing amplifier is sufficient?
Is there No need to add it further down the line? Won't the summing amp stage add a dc offset at its output?

No, I was saying that as long as bipolar supplies are used you don't need a coupling cap at the output of each stage. In other words, the output of one stage doesn't need a coupling cap if the input of the following stage has a coupling cap (which it should, for reasons already pointed out).

For interstage coupling in an amplifier string with modest overall voltage gain* you can usually dispense with the DC blocking capacitor, but since you will only need a few uF to maintain a low-end frequency response down to a few Hz why not use one there, too? It's getting rid of the DC blocking capacitor between the power amplifier and speaker that is the real benefit** (for distortion, reducing "thump" at turn on/off, and cost).



* - for example, you need 20Vrms to drive 8R speakers at 50W, and the standard line level signal is 1Vrms (0dBV), so that requires a voltage gain of 20; even relatively high offset op-amps like the NE5534 (5mV worst case) will only subject the speaker to a DC bias of 0.1V here. That's about 1.25mW of DC dissipation, which is basically negligible.


** - as well as the DC biasing resistors needed for every op-amp if a single supply is used.
« Last Edit: September 10, 2019, 10:06:11 pm by MagicSmoker »
 
The following users thanked this post: ZeroResistance

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Negative version of LP2951 LDO
« Reply #14 on: September 11, 2019, 07:44:38 am »
So I know impedance of at least one of the outputs that will connect to my Audio Mixer input.
Its output impendance is 200ohms with max 4V rms.

So would I need to match the impedance for this source. The circuit which I had posted earlier had just a 10K audio taper to attenuate the phone output followed by a 1uF dc blocking cap?
 

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1408
  • Country: us
Re: Negative version of LP2951 LDO
« Reply #15 on: September 11, 2019, 10:25:20 am »
You don't worry about matching impedance between outputs and inputs in audio, nor do you even want to. In fact, you want to present any signal source with a relatively constant (and relatively high) impedance so peaks aren't distorted from increased loading.

You also want the DC blocking cap to work into a relatively high and constant impedance to minimize its size and maintain a consistent frequency response, so the better arrangement would be to put the cap first - that is, in between the source and the top of the pot - with the wiper from the pot going to the mixer input.

In your example the volume control pot is 10k and the DC blocking capacitor is 1uF so the -3dB point (that is, when the reactance of the capacitor equals the resistance of the pot) will be 16Hz*. For high-fidelity applications you'll want the cutoff under 10Hz and this is best done by increasing the capacitance, rather than the resistance of the volume control pot, because the latter makes the circuit more noisy. I'd use at least 3.3uF for the DC blocking capacitor here, though there is no harm in going up to 10uF so long as you don't use ultra high-k ceramic types (like Y5V and Z5U) as they are piezoelectric and suffer a number of other pathologies that make them terrible for use in audio or other sensitive signal amplification applications.


* - the equation to determine the reactance of a capacitor at a given frequency is Xc = 1 / (2 * Pi * f * C)
 
The following users thanked this post: ZeroResistance

Offline exe

  • Supporter
  • ****
  • Posts: 2647
  • Country: nl
  • self-educated hobbyist
Re: Negative version of LP2951 LDO
« Reply #16 on: September 11, 2019, 11:29:41 am »
1. For a simple design like this, a rail splitter would be as good as split supplies

Functionally, sure - it just turns the mid-point of a positive supply into the 0V reference - but this also elevates the "supposed to be grounded" shell of shielded interconnects to the mid-point voltage which can cause problems when connecting to other devices, especially if they share a chassis ground and/or are earthed.

Can shield be connected to "real" ground? In my understanding, any fixed potential is good for shielding.
 

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1408
  • Country: us
Re: Negative version of LP2951 LDO
« Reply #17 on: September 11, 2019, 12:04:13 pm »
Can shield be connected to "real" ground? In my understanding, any fixed potential is good for shielding.

Shield is usually connected to earth ground in XLR (balanced) interconnects, but shouldn't be - yet still often is - in RCA and similar type interconnects. In the latter cases this results in the dreaded ground loop.
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20363
  • Country: gb
  • 0999
Re: Negative version of LP2951 LDO
« Reply #18 on: September 11, 2019, 12:56:51 pm »
For price reduction, try to think how you'd get rid of the negative rail. If you really need it, does it really need to be regulated?

I am trying to build a audio mixer in the virtual earth configuration.
So my thinking was virtual earth needs the non -inverting input to be at ground potential, so that pulls the inverting input to a virtual ground.
Is my thinking flawed and if so is there some means to eliminate the negative power supply in a virtual earth configuration.
Just AC couple all of the signal paths and bias the amplifiers to work at half the supply voltage.

The volume control should be between the two amplifier stages. I've converted the second one to a unity gain buffer and put the volume control before it.
 

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Negative version of LP2951 LDO
« Reply #19 on: September 11, 2019, 01:12:41 pm »

Just AC couple all of the signal paths and bias the amplifiers to work at half the supply voltage.

The volume control should be between the two amplifier stages. I've converted the second one to a unity gain buffer and put the volume control before it.

This is still great! but would it still count as an virtual earth mixer?
The earth point has now shifted to 1/2 VCC.
From what I understood about the concept of a virtual earth mixer, is that if all the input signals see a true virtual earth. There won't be any crosstalk between them.
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20363
  • Country: gb
  • 0999
Re: Negative version of LP2951 LDO
« Reply #20 on: September 11, 2019, 03:16:26 pm »

Just AC couple all of the signal paths and bias the amplifiers to work at half the supply voltage.

The volume control should be between the two amplifier stages. I've converted the second one to a unity gain buffer and put the volume control before it.

This is still great! but would it still count as an virtual earth mixer?
The earth point has now shifted to 1/2 VCC.
From what I understood about the concept of a virtual earth mixer, is that if all the input signals see a true virtual earth. There won't be any crosstalk between them.
I don't know what you mean by virtual earth mixer? Are you talking about using an op-amp as a rail splitter?

Note that ideally the potentiometer should be AC coupled anyway, to cut down on noise when it's adjusted, but in this case the DC current should be no more than the op-amp bias current and the gain is unity, so it shouldn't be too bad.
« Last Edit: September 12, 2019, 07:47:04 am by Zero999 »
 

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Negative version of LP2951 LDO
« Reply #21 on: September 12, 2019, 12:18:41 pm »

I don't know what you mean by virtual earth mixer? Are you talking about using an op-amp as a rail splitter?


I read about it here https://sound-au.com/articles/audio-mixing.htm#s3
 
The following users thanked this post: exe

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20363
  • Country: gb
  • 0999
Re: Negative version of LP2951 LDO
« Reply #22 on: September 12, 2019, 02:40:21 pm »

I don't know what you mean by virtual earth mixer? Are you talking about using an op-amp as a rail splitter?


I read about it here https://sound-au.com/articles/audio-mixing.htm#s3
That's got nothing to do with whether the circuit is operating from a single supply rail, a bipolar supply or a virtual ground. Both of the circuits I posted use the same virtual earth principle, as described in that article. The difference is that first one I posted was AC coupled and the second one DC coupled. In the AC coupled design, the DC is blocked, so only the signal matters, the fact that the negative input is fixed at half the supply voltage vs 0V doesn't matter.

The circuit works by fixing the inverting op-amp input at a constant DC voltage. In most cases it's the same as 0V, because it's what the non-inverting input is connected to, hence virtual earth, but it doesn't really matter. Take the input resistors away and you'll see that the impedance into the inverting input is very low, due to negative feedback. If a current is injected into the input, the op-amp's output will move in the opposite direction, pulling current through the feedback resistor to counteract the change in voltage due to the input current.

The circuit is a transimpedance amplifier. It converts a constant current into a constant voltage. In this instance, with a feedback resistor of 10k, 1mA of current in, will give -10V out, because the input will stay at 0V and 1mA of curenet will give a drop of 10V across the resistor.
https://en.wikipedia.org/wiki/Transimpedance_amplifier

Kirchoff's current law states that the currents through all the resistors add together, at the node where they're all connected. Suppose you have 5V at two inputs, with 10k input resistors. The current through each resistor will be 0.5mA, giving 1mA in total, which also flows through the feedback resistor, giving an output of -10V.
 
The following users thanked this post: ZeroResistance

Offline Bassman59

  • Super Contributor
  • ***
  • Posts: 2501
  • Country: us
  • Yes, I do this for a living
Re: Negative version of LP2951 LDO
« Reply #23 on: September 13, 2019, 01:19:00 am »
I'm looking for a negative output LDO, 100ma.
I found the positive version LP2951, however I'm unable to locate its negative version.
Like the LM317 has a LM337 version, I was expecting the same for the LP2951.
Am I missing something here or the negative version just does not exist for the LP2951?

Everyone else is pretty much on the money about negative LDOs.

But I should ask: does your circuit need an LDO? I mean, I understand why you might want one, but unless you're really trying to minimize heat dissipation in your regulators, the 317/337 will do ya fine.

That said, I just did a power supply that needed ±15 V, +5 V and +3.3 V, and I used LT1963A for the positive rails and LT1175 for the negative rail. And yes, the 1175 is frightfully expensive.
 

Offline ZeroResistanceTopic starter

  • Frequent Contributor
  • **
  • Posts: 585
  • Country: gb
Re: Negative version of LP2951 LDO
« Reply #24 on: September 14, 2019, 05:48:05 am »

Everyone else is pretty much on the money about negative LDOs.

But I should ask: does your circuit need an LDO? I mean, I understand why you might want one, but unless you're really trying to minimize heat dissipation in your regulators, the 317/337 will do ya fine.

That said, I just did a power supply that needed ±15 V, +5 V and +3.3 V, and I used LT1963A for the positive rails and LT1175 for the negative rail. And yes, the 1175 is frightfully expensive.

I'm limited by my input voltage that is going to be 9V dc.
Then the LM317L / LM337L would drop around 2.5V.
So that brings its down to max 6.5V at their outputs. Lets keep it 6 for giving it a little safety margin.
So now I would be powering the op-amps with +6/-6V.
The opamp is a non rail to rail type so its output would swing to +/- 2V  of the supply rails.
So 6 -2 gives me 4V as the max output swing.
And I have to handle some line level inputs, which I though would be 1.2Vrms but I realised that one of the inputs would be around 4Vrms max.
Having said that I'm still game for the 317 / 337 but I say that the LP2951 was at a similar price range so wondered it would be better to go in for that, if I get a suitable negative voltage alternative to the 337.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf