+15V/-15V power supply design question

[Tim88]

Am building a ps for an opamp based preamp that uses +15V/-15V. Everything is good except the LM7815 regulates at +14.99V and the LM7915 regulates at -15.31V. Both have 18V going into the regulator so input voltage is good. Tested each one with direct power supply voltage and upping voltage made no change. So, put an adjusting circuit in and bump up the + side to +15.31V, no problem. It now draws an extra 15mA but that shouldn't be an issue.

Question is, is it common for these regulators to be off like this? I know it is within specs but everything else I recall building was pretty spot on. Second,  I assume op amps would function the best with equal +/- supply rails? Maybe I have my undies in a twist over nothing...

I learned a lot building this, particularly the capacitance multiplier addition which reduced the 120Hz hum to the inaudible range. If anyone has any comments or suggestions on my design, please feel free to opine, looking to learn. 

[Alex Eisenhut]

They're not off, they're spot on for their specs. Why do you want such accurate supplies?

[Tim88]

Like I said, maybe I am just being too picky.  Was under the assumption with opamps circuits, you want the supplies to match. This is the preamp schematic, tiz for a microphone. Thanks!

[TimFox]

You need to read the data sheets.
With respect to tolerance on the +/- 15 V supplies for op amps, the relevant spec is the DC PSRR and CMRR, which allow you to calculate the shift in DC input offset voltage for a change in the DC voltages.
PSRR is relevant to the magnitude of the total voltage between the supplies, and CMRR is relevant to the mismatch in absolute values.
High precision in DC values is only relevant to normal op amp applications if a precise DC output is required when the DC signal input is derived directly from the supplies.
More important for normal applications is the hum, noise, and ripple (AC components) of the supply voltages, since PSRR and CMRR fall with increasing frequency.
Still, this is a quantitative question.

[MikeK]

What's the purpose of the pass transistors in front of the regulators?

[themadhippy]

change out the fixed regulators and use lm317 and 337's,supposedly there quieter  for audio work,and you can tweak them to match precisely,or leave  alone as its  good enuff for rock n roll

[Tim88]

You need to read the data sheets.
With respect to tolerance on the +/- 15 V supplies for op amps, the relevant spec is the DC PSRR and CMRR, which allow you to calculate the shift in DC input offset voltage for a change in the DC voltages.
PSRR is relevant to the magnitude of the total voltage between the supplies, and CMRR is relevant to the mismatch in absolute values.
High precision in DC values is only relevant to normal op amp applications if a precise DC output is required when the DC signal input is derived directly from the supplies.
More important for normal applications is the hum, noise, and ripple (AC components) of the supply voltages, since PSRR and CMRR fall with increasing frequency.
Still, this is a quantitative question.

Here is a screen shot of the data sheets. Not exactly sure how to interpret it?

[Tim88]

What's the purpose of the pass transistors in front of the regulators?

Capacitance multiplier, not in the correct spot?

[Tim88]

change out the fixed regulators and use lm317 and 337's,supposedly there quieter  for audio work,and you can tweak them to match precisely,or leave  alone as its  good enuff for rock n roll

Was wondering about those. I have some 317's but no 337's, will have to add them to the next order. It's pretty quite now but can always improve, thanks!

[Zero999]

Like I said, maybe I am just being too picky.  Was under the assumption with opamps circuits, you want the supplies to match. This is the preamp schematic, tiz for a microphone. Thanks!

(Attachment Link)

You don't need the supplies to match. It won't make any difference. The circuit attached to the original post can be trimmed down considerably and will still work perfectly.

[iMo]

With opamps you do not need exact matching as it was said above.
Anyhow, your above schematics is not good with the 7815 level adjusting - it should be wired as below..
PS: the 7815 output voltage will be always higher (with trimmer resistance set >0 ohm) than without the trimmer, 15V plus 0 .. 1.7V with your values.
You want to add up only 0.32V, therefore I would use 3k3 instead of your 1k and you'll get those 0.32V in aprox. the middle of the 100ohm trimmer setting..

[Benta]

Rule #1 for beginners:
Do not try trimming 78xx/79xx regulators like that. The ground pin current varies wildly with load current and input voltage.
The LM317/337 are the right parts for doing that.
Live with the 78xx/79xx voltages as they are. They are 1970s devices.

[floobydust]

For some reason people think exact +/-15V is required and trimming to exact values = better fidelity.

You don't have enough input voltage 18VDC will not work (especially with the c multiplier) if there is a mains sag and give a blast of hum. I use a 28VCT transformer, 25VCT was not enough for some venues. Others recommend more.
Elliott Sound Products - Power Supply for Preamplifiers is the bare minimum and Power Supply for Preamplifiers (Revision D) uses LM317/LM337 w/full protection diodes.

Note your circuit can blow the SSM2019 if phantom power is on and any shorts (or plugging in wrong mic) at the XLR will dump C4, C5 into the IC's input and it's dead. It's common to have clamp diodes there.

[SuzyC]

Quiescent Current(ground current) changes very little for 78xx regs


78xx regs  have no trouble with dropout voltage with 18V supplies, unlike 317 reg types.

[floobydust]

78xx regs  have no trouble with dropout voltage with 18V supplies, unlike 317 reg types.

No, they both use Darlington pass transistors giving dropout voltage of 1.5-2.3V depending on load, and OP has extra cap multiplier losses.
Having only 1V of headroom means a mains sag of a percent to two gives you a very loud blast of hum out of the mic preamp.

[Benta]

Quiescent Current(ground current) changes very little for 78xx regs

I'd say that 0.5 mA change at 3 mA ground current is significant. Works out at ~17%.

[SuzyC]

Quiescent Current specsheet(above) show geriatric 78xx quies cur to be 8ma, not 3ma.
-------------------------
Nope, specs for Dropout V are better than that.

2 per-cent sag of 18V =.36V      18-.36V= 17.64V   

Dropout V spec=1.1V or 16.9V Vin to Reg   which will dropout at high current output beginning at 16.1V

It is uncommon for a preamp to draw so much current like 1A or so..it is not a P.A. system.

[Tim88]

Some great posts, much appreciated, will address them when I get done cooking supper. One thing of note, the preamp draws less than 80 mA so I knew I had a lot of headroom on the PS. The link to the project is here if anyone is interested: https://www.instructables.com/Build-the-Four-Channel-SSM2019-Phantom-Powered-Mic/

Thanks again!

From the website Q &A:
"DJJules

Answer 1 year ago

Absolutely. The SSM2019's draw at the most 10ma each and the OPA2134's about the same so all in 80ma all in so you are easily covering that with 850ma"

[Benta]

You're showing a LM7905 which is not the subject here.
The LM7815 has a typical drop-out voltage of 2 V. Unfortunately no max. is listed in the data sheet, but thinking 2.5...3 V is not unrealistic.

[SuzyC]

Nope, I first posted above both 78xx and 79xx specs(Sorry, they do not show dropout V, so I post the date 2000 Nat Semi Spec Sheet page showing dropout here)

From looking at the dropout graph at 80mA, the Nat Semi 78xx spec sheet clearly shows less than 1.5V.. but  not 2.5..3V

Notice how the dropout voltage decreases with temperature, but o.k. about 1.5V with  80mA OP's current draw. The hotter the device, the better the dropout voltage spec.
------------
Assume that the OP's idea is to correct the .23V difference in output voltage between the positive and negative regulated output voltage:

I will assume that the quiescent cur for the 78xx series is typically 6mA(I measure this often). 


Assume  .5ma difference in Quies Cur will result in about an output voltage change of 15V +/- (8% of .23V) due to correction circuit change).do the math..not so much is it?

[Tim88]

Like I said, maybe I am just being too picky.  Was under the assumption with opamps circuits, you want the supplies to match. This is the preamp schematic, tiz for a microphone. Thanks!

(Attachment Link)

You don't need the supplies to match. It won't make any difference. The circuit attached to the original post can be trimmed down considerably and will still work perfectly.

This is the first pre-amp I built. I used a PS design like the one you posted (just the + rail part at 12V) and it produced horrible line hum. I added the capacitance multiplier and it cleared most of it up. The mic sounded very high treble and just plain bad. Wasn't sure what was going on with that amp design so went to this other one. Admittedly, I did not test it minus the cap multiplier, maybe it will sound fine, will definitly test it.

[Tim88]

With opamps you do not need exact matching as it was said above.
Anyhow, your above schematics is not good with the 7815 level adjusting - it should be wired as below..
PS: the 7815 output voltage will be always higher (with trimmer resistance set >0 ohm) than without the trimmer, 15V plus 0 .. 1.7V with your values.
You want to add up only 0.32V, therefore I would use 3k3 instead of your 1k and you'll get those 0.32V in aprox. the middle of the 100ohm trimmer setting..

Appreciate the advice and will change that resistor to a 3.3k. I just used a 1k per the spec sheet recommendation. Am gonna test the setup with the voltage inputs unbalanced and see if it sounds any different. Probably doesn't as most have said but it's a learning deal for me. Thanks!

[Tim88]

Rule #1 for beginners:
Do not try trimming 78xx/79xx regulators like that. The ground pin current varies wildly with load current and input voltage.
The LM317/337 are the right parts for doing that.
Live with the 78xx/79xx voltages as they are. They are 1970s devices.

Good info on the 317/337's, much appreciated. This pre amp draws very little current, <80mA, so I figgered I was safe on trimming the 7815. Thanks!

[SuzyC]

You can't trim the 7815 lower..only higher in output voltage.

[Tim88]

For some reason people think exact +/-15V is required and trimming to exact values = better fidelity.

You don't have enough input voltage 18VDC will not work (especially with the c multiplier) if there is a mains sag and give a blast of hum. I use a 28VCT transformer, 25VCT was not enough for some venues. Others recommend more.
Elliott Sound Products - Power Supply for Preamplifiers is the bare minimum and Power Supply for Preamplifiers (Revision D) uses LM317/LM337 w/full protection diodes.

Note your circuit can blow the SSM2019 if phantom power is on and any shorts (or plugging in wrong mic) at the XLR will dump C4, C5 into the IC's input and it's dead. It's common to have clamp diodes there.

Some great info there, thank you. I have been to the sound-au.com site but somehow missed that good article. Will add the clamp diodes. Had seen them in some othere schematics and wasn't quite sure their purpose.