Power Rails voltage drop

From circuitbending to homebrew stompboxes & synths, keep the DIY spirit alive!

Moderators: Kent, Joe., luketeaford, lisa

User avatar
cackland
Super Deluxe Wiggler
Posts: 2211
Joined: Thu Dec 28, 2017 5:42 pm
Location: Los Angeles, California

Power Rails voltage drop

Post by cackland » Mon Aug 24, 2020 10:01 pm

I've just checked the voltage coming from the rails throughout various stages of my circuit and noticed a significant difference from the expected +-12V rails.

So I hooked up my multimeter to the power entry point in my module and noticed the same voltage straight off the rails.
Screen Shot 2020-08-25 at 12.58.24 pm.png
Note: Excuse the test point nodes. These are placed at various points around my circuit.

My multimeter is reading +-9.62V. This means that all the op amps in the circuit will be powered by this voltage, and not the suspected +-12V, which means they have less headroom.

I thought this voltage drop might be because of the demand of my circuit, downstream. However measuring directly after the 10ohm resistor, gives me this reading.

Hoping someone can chime in as to what is going on.

Cheers
You do not have the required permissions to view the files attached to this post.

nigel
Veteran Wiggler
Posts: 730
Joined: Wed Jan 16, 2013 6:49 am
Location: Melbourne, Australia

Re: Power Rails voltage drop

Post by nigel » Tue Aug 25, 2020 12:42 am

cackland wrote:
Mon Aug 24, 2020 10:01 pm
My multimeter is reading +-9.62V. This means that all the op amps in the circuit will be powered by this voltage, and not the suspected +-12V, which means they have less headroom.
If your incoming power is 12V, but your module power is reading 9.6V, then R42 must be dropping about 2V. That's nearly 1/2 Watt, so assuming it's a regular 1/4W resistor it must be smoking by now. Headroom is not what you should be worrying about. Is your main power supply still providing 12V?
cackland wrote:
Mon Aug 24, 2020 10:01 pm
I thought this voltage drop might be because of the demand of my circuit, downstream. However measuring directly after the 10ohm resistor, gives me this reading.
It doesn't matter where you measure it, the voltage on the module power rail should be constant (within a few millivolts)

KSS
Super Deluxe Wiggler
Posts: 3731
Joined: Mon Jan 25, 2016 7:28 am

Re: Power Rails voltage drop

Post by KSS » Tue Aug 25, 2020 3:30 am

check the batteries in your meter.
Treat utility modules as stars instead of backup singers.
Treat power supplies like Rockstars instead of roadies!
Chase magic sound, not magic parts.

User avatar
cackland
Super Deluxe Wiggler
Posts: 2211
Joined: Thu Dec 28, 2017 5:42 pm
Location: Los Angeles, California

Re: Power Rails voltage drop

Post by cackland » Tue Aug 25, 2020 6:38 am

nigel wrote:
Tue Aug 25, 2020 12:42 am
Is your main power supply still providing 12V?

It doesn't matter where you measure it, the voltage on the module power rail should be constant (within a few millivolts)
Interesting.. I'm measuring 11.6V in between the 1N5819 diode and the 10 ohm resistor. So much more than a few millivolts from the rails.

I don't understand how a 10ohm resistor is dropping that much. Many times I thought I had a different multiplier of 10 in there, but nope.. measured a straight 10 ohm.

Doesn't make sense to me.
Last edited by cackland on Tue Aug 25, 2020 6:39 am, edited 1 time in total.

User avatar
cackland
Super Deluxe Wiggler
Posts: 2211
Joined: Thu Dec 28, 2017 5:42 pm
Location: Los Angeles, California

Re: Power Rails voltage drop

Post by cackland » Tue Aug 25, 2020 6:38 am

KSS wrote:
Tue Aug 25, 2020 3:30 am
check the batteries in your meter.
Good call... however multimeter has new batteries. Other voltages measured in the circuit are correct.

User avatar
cygmu
Super Deluxe Wiggler
Posts: 1559
Joined: Mon Feb 15, 2016 3:33 pm
Location: UK

Re: Power Rails voltage drop

Post by cygmu » Tue Aug 25, 2020 7:07 am

If your modules are drawing 200mA then the 10R will drop 2V. That seems plausible to me. What are you powering from this?

User avatar
htor
Wiggling with Experience
Posts: 270
Joined: Sun Dec 01, 2019 6:20 am
Location: Leipzig, Germany

Re: Power Rails voltage drop

Post by htor » Tue Aug 25, 2020 8:08 am

cackland wrote:
Tue Aug 25, 2020 6:38 am
nigel wrote:
Tue Aug 25, 2020 12:42 am
Is your main power supply still providing 12V?

It doesn't matter where you measure it, the voltage on the module power rail should be constant (within a few millivolts)
Interesting.. I'm measuring 11.6V in between the 1N5819 diode and the 10 ohm resistor. So much more than a few millivolts from the rails.
i think he meant measuring the voltage directly on the rail. sounds like your power supply is providing plenty anyway!
if you measure it after the diode its forward voltage drop will eat some hundred millivolts.

User avatar
cackland
Super Deluxe Wiggler
Posts: 2211
Joined: Thu Dec 28, 2017 5:42 pm
Location: Los Angeles, California

Re: Power Rails voltage drop

Post by cackland » Tue Aug 25, 2020 9:15 am

cygmu wrote:
Tue Aug 25, 2020 7:07 am
If your modules are drawing 200mA then the 10R will drop 2V. That seems plausible to me. What are you powering from this?
It’s a multi channel VCA processor module. Haven’t tested the current module draw at the moment...

Back to the idea of headroom. If the module is drawing a certain amount of current, that shouldn’t affect the potential voltage all the ic’s are powered from right?

If the op amps are only being powered by sub -+10v, and if I’m using rail to rail op amps, I’m losing headroom.

Unless I have this concept wrong, I want the op amps powered as close to +-12V as possible.

User avatar
cackland
Super Deluxe Wiggler
Posts: 2211
Joined: Thu Dec 28, 2017 5:42 pm
Location: Los Angeles, California

Re: Power Rails voltage drop

Post by cackland » Tue Aug 25, 2020 9:17 am

htor wrote:
Tue Aug 25, 2020 8:08 am
cackland wrote:
Tue Aug 25, 2020 6:38 am
nigel wrote:
Tue Aug 25, 2020 12:42 am
Is your main power supply still providing 12V?

It doesn't matter where you measure it, the voltage on the module power rail should be constant (within a few millivolts)
Interesting.. I'm measuring 11.6V in between the 1N5819 diode and the 10 ohm resistor. So much more than a few millivolts from the rails.
i think he meant measuring the voltage directly on the rail. sounds like your power supply is providing plenty anyway!
if you measure it after the diode its forward voltage drop will eat some hundred millivolts.
Yup forgot about that forward diode drop... makes me question how to compensate for this and any other losses?

User avatar
neil.johnson
Super Deluxe Wiggler
Posts: 1432
Joined: Thu Apr 08, 2010 3:51 pm
Location: Cambridge, UK
Contact:

Re: Power Rails voltage drop

Post by neil.johnson » Tue Aug 25, 2020 9:55 am

How does the negative rail look? Are you seeing the same voltage drop across R43? Whatever your circuit is doing it is pulling around 200mA from the positive rail = 2.4W. Toasty.
Are there any digital bits that are being driven off a local 5V regulator? Perhaps a 7805?
If your circuit really is pulling that amount of current then (a) what you are measuring is correct, and (b) you might want to rethink your power entry circuit.

It would really help if we knew what your circuit was? Or at least what current you are expecting it to pull. And also what the negative side looks like as well.

Neil
Random ramblings :: http://www.njohnson.co.uk
SSI parts and more :: https://www.cesyg.co.uk

KSS
Super Deluxe Wiggler
Posts: 3731
Joined: Mon Jan 25, 2016 7:28 am

Re: Power Rails voltage drop

Post by KSS » Tue Aug 25, 2020 6:19 pm

Along the lines of what neil brought up.
Is the PSU a half wave wall wart typev supply? Those don't handle uneven rail draw very well.
Treat utility modules as stars instead of backup singers.
Treat power supplies like Rockstars instead of roadies!
Chase magic sound, not magic parts.

nigel
Veteran Wiggler
Posts: 730
Joined: Wed Jan 16, 2013 6:49 am
Location: Melbourne, Australia

Re: Power Rails voltage drop

Post by nigel » Tue Aug 25, 2020 7:55 pm

You have all the information you need. Presumably the input voltage (the line going off the page to the left) is 12V. The 1N5819 is a Schottky diode, so it will reduce the voltage by about 0.3V - you measure 11.6V between the diode and the resistor, so that's about right. You say that the voltage on the module side of the resistor is 9.6V, so the voltage across the resistor is 2V. By Ohm's law, I = V / R = 2 / 10 = 1/5 Amps, or 200 milliAmps. So your module must be using 200 milliAmps from the positive supply. If you want to increase the voltage inside your module (9.6V), you need to either (1) reduce the current it uses, (2) remove the Schottky diode, which will give you another 0.4V or so, or (3) remove the 10 Ohm resistor. Ideally, you should try to reduce the current - as neil.johnson points out, that's a lot of current for a single module.

User avatar
cackland
Super Deluxe Wiggler
Posts: 2211
Joined: Thu Dec 28, 2017 5:42 pm
Location: Los Angeles, California

Re: Power Rails voltage drop

Post by cackland » Tue Aug 25, 2020 9:37 pm

neil.johnson wrote:
Tue Aug 25, 2020 9:55 am
How does the negative rail look? Are you seeing the same voltage drop across R43? Whatever your circuit is doing it is pulling around 200mA from the positive rail = 2.4W. Toasty.
Are there any digital bits that are being driven off a local 5V regulator? Perhaps a 7805?
Yes, there is a digital section. An STM32F3xx, and I do have a various LM4040 series shunt regulators in the circuit.
neil.johnson wrote:
Tue Aug 25, 2020 9:55 am
(b) you might want to rethink your power entry circuit.
Is there any reference material you can recommend as a redesign?
neil.johnson wrote:
Tue Aug 25, 2020 9:55 am
It would really help if we knew what your circuit was? Or at least what current you are expecting it to pull. And also what the negative side looks like as well.

Neil
As mentioned above, its a VCA processing module, with various other digital sub circuits. I'm not sure what current I should be pulling, I'm a novice in the power department so this is part of my learning.

User avatar
cackland
Super Deluxe Wiggler
Posts: 2211
Joined: Thu Dec 28, 2017 5:42 pm
Location: Los Angeles, California

Re: Power Rails voltage drop

Post by cackland » Tue Aug 25, 2020 9:38 pm

KSS wrote:
Tue Aug 25, 2020 6:19 pm
Along the lines of what neil brought up.
Is the PSU a half wave wall wart typev supply? Those don't handle uneven rail draw very well.
The power supply is a mean well GSM90A15-P1M.. The same as Intellijel provides with their 7U case.

User avatar
cackland
Super Deluxe Wiggler
Posts: 2211
Joined: Thu Dec 28, 2017 5:42 pm
Location: Los Angeles, California

Re: Power Rails voltage drop

Post by cackland » Tue Aug 25, 2020 9:40 pm

nigel wrote:
Tue Aug 25, 2020 7:55 pm
You have all the information you need. Presumably the input voltage (the line going off the page to the left) is 12V. The 1N5819 is a Schottky diode, so it will reduce the voltage by about 0.3V - you measure 11.6V between the diode and the resistor, so that's about right. You say that the voltage on the module side of the resistor is 9.6V, so the voltage across the resistor is 2V. By Ohm's law, I = V / R = 2 / 10 = 1/5 Amps, or 200 milliAmps. So your module must be using 200 milliAmps from the positive supply. If you want to increase the voltage inside your module (9.6V), you need to either (1) reduce the current it uses, (2) remove the Schottky diode, which will give you another 0.4V or so, or (3) remove the 10 Ohm resistor. Ideally, you should try to reduce the current - as neil.johnson points out, that's a lot of current for a single module.
This is making me question a lot of things about my circuit, and how to reduce the overall current consumption. A lot to learn, so appreciate the explanation.

I've seen some circuits with these diodes in place of the 1N5819.

https://au.mouser.com/ProductDetail/621-DMP3099L-7
https://au.mouser.com/ProductDetail/621-DMN3404L-7

The data sheets are an overload of information. Would they be an improvement over the 1N5819?

User avatar
cackland
Super Deluxe Wiggler
Posts: 2211
Joined: Thu Dec 28, 2017 5:42 pm
Location: Los Angeles, California

Re: Power Rails voltage drop

Post by cackland » Tue Aug 25, 2020 9:45 pm

I was also thinking about implementing an inline current rail test point, using a 2.54 2 pin header.

Is it as straight forward as this? (With jumpers when not testing)
Screen Shot 2020-08-26 at 12.44.56 pm.png
You do not have the required permissions to view the files attached to this post.

Synthiq
Veteran Wiggler
Posts: 594
Joined: Mon Feb 06, 2017 3:10 pm
Location: California

Re: Power Rails voltage drop

Post by Synthiq » Wed Aug 26, 2020 12:08 am

The processor uses at most 70mA at 72MHz so that doesn't explain the 200mA you see. Even with 70mA current consumption you should consider using a switching voltage regulator for improved efficiency and less current from the +12V supply. Do you need to run it at max frequency or can the application run at a lower frequency to save current?

Does any component get warmer than expected? That's a good indication where the extra current is consumed.

KSS
Super Deluxe Wiggler
Posts: 3731
Joined: Mon Jan 25, 2016 7:28 am

Re: Power Rails voltage drop

Post by KSS » Wed Aug 26, 2020 3:11 am

A current test point is better as part of somethng external you plug between the module and its distribution when needed. You've already got two diodes and two resistors that aren't really doing anything positive for you. I'd remove all six of these in the latest schem.
Treat utility modules as stars instead of backup singers.
Treat power supplies like Rockstars instead of roadies!
Chase magic sound, not magic parts.

User avatar
neil.johnson
Super Deluxe Wiggler
Posts: 1432
Joined: Thu Apr 08, 2010 3:51 pm
Location: Cambridge, UK
Contact:

Re: Power Rails voltage drop

Post by neil.johnson » Wed Aug 26, 2020 3:45 am

I would keep the diodes, then split the circuit into power domains: one digital domain (ideally DC-DC switcher to minimise power losses), then one or more analogue domains each with their own input power filter.

The STM32F3xx is probably running on 3.3V, so if you had a linear regulator that alone would be dissipating 1/2W if fed from 11.6V.

A lot of the design decisions will depend on whether this is going to be a commercial project (mass-produced, BOM cost sensitivity, reliability, etc), or a one-off hobby project. Some context would help.

Neil
Random ramblings :: http://www.njohnson.co.uk
SSI parts and more :: https://www.cesyg.co.uk

KSS
Super Deluxe Wiggler
Posts: 3731
Joined: Mon Jan 25, 2016 7:28 am

Re: Power Rails voltage drop

Post by KSS » Wed Aug 26, 2020 5:02 am

neil.johnson wrote:
Wed Aug 26, 2020 3:45 am
I would keep the diodes,
In series as presently drawn?
Why?
Treat utility modules as stars instead of backup singers.
Treat power supplies like Rockstars instead of roadies!
Chase magic sound, not magic parts.

User avatar
neil.johnson
Super Deluxe Wiggler
Posts: 1432
Joined: Thu Apr 08, 2010 3:51 pm
Location: Cambridge, UK
Contact:

Re: Power Rails voltage drop

Post by neil.johnson » Wed Aug 26, 2020 5:23 am

KSS wrote:
Wed Aug 26, 2020 5:02 am
neil.johnson wrote:
Wed Aug 26, 2020 3:45 am
I would keep the diodes,
In series as presently drawn?
Why?
Reverse polarity protection. Schottky's will drop about 0.3-0.4V, which in the grand scheme of things is nothing to worry about. Any design that depends on the power rails being EXACTLY 12.00000000000000000000V is a poor design IMHO. For bonus points you can add two more diodes and a 10-way header and then it'll work no matter which way round the cable is plugged in.

That's better than the other scheme of reverse diodes, which relies on the diodes being stronger than the power supply, and hoping that the flimsy ribbon cable doesn't melt before the user notices or the power supply shuts down. Too much hope for a power system.

Neil
Random ramblings :: http://www.njohnson.co.uk
SSI parts and more :: https://www.cesyg.co.uk

User avatar
neil.johnson
Super Deluxe Wiggler
Posts: 1432
Joined: Thu Apr 08, 2010 3:51 pm
Location: Cambridge, UK
Contact:

Re: Power Rails voltage drop

Post by neil.johnson » Wed Aug 26, 2020 5:50 am

An alternative to diodes are MOSFETs, which will have a lower voltage drop if you really need to squeeze the last few 100mV or so out of the rails.

Neil
Random ramblings :: http://www.njohnson.co.uk
SSI parts and more :: https://www.cesyg.co.uk

KSS
Super Deluxe Wiggler
Posts: 3731
Joined: Mon Jan 25, 2016 7:28 am

Re: Power Rails voltage drop

Post by KSS » Wed Aug 26, 2020 6:32 am

Thanks for your reasoning.
I worry less about the drop than the variability *of* the drop with current draw.
Treat utility modules as stars instead of backup singers.
Treat power supplies like Rockstars instead of roadies!
Chase magic sound, not magic parts.

User avatar
cackland
Super Deluxe Wiggler
Posts: 2211
Joined: Thu Dec 28, 2017 5:42 pm
Location: Los Angeles, California

Re: Power Rails voltage drop

Post by cackland » Wed Aug 26, 2020 8:34 am

neil.johnson wrote:
Wed Aug 26, 2020 5:50 am
An alternative to diodes are MOSFETs, which will have a lower voltage drop if you really need to squeeze the last few 100mV or so out of the rails.

Neil
Yes, you must of missed this mentioned above:

I've seen some circuits with these mosfet diodes in place of the 1N5819.

https://au.mouser.com/ProductDetail/621-DMP3099L-7
https://au.mouser.com/ProductDetail/621-DMN3404L-7

Assuming these are more efficient?

User avatar
cackland
Super Deluxe Wiggler
Posts: 2211
Joined: Thu Dec 28, 2017 5:42 pm
Location: Los Angeles, California

Re: Power Rails voltage drop

Post by cackland » Wed Aug 26, 2020 8:37 am

Synthiq wrote:
Wed Aug 26, 2020 12:08 am
The processor uses at most 70mA at 72MHz so that doesn't explain the 200mA you see. Even with 70mA current consumption you should consider using a switching voltage regulator for improved efficiency and less current from the +12V supply. Do you need to run it at max frequency or can the application run at a lower frequency to save current?

Does any component get warmer than expected? That's a good indication where the extra current is consumed.
Currently using two LM1117MPX-3.3 voltage regulators for the stm32, one the digital VDD and the other for the VDD_A. This chip wouldn’t work without having power to the ADC and have read that it is good practice to have split power for both normal 3.3 and analog 3.3. Maybe I should go for just one??

I could definitely reduced the frequency and see how low I can go and how that affects the current draw

Post Reply

Return to “Music Tech DIY”