Voltage Regulators, Datasheets, and PSSR

Go To Last Post
17 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Howdy.

I have been looking at voltage regulator datasheets all morning and I am wondering if I am taking a good approach to interpreting datasheets.

On my PCB I have a Mega324p that is reading an analog signal from a gyro IC. The Mega is also controlling a motor through an H-bridge. I recently changed to a different motor that has 1/10th the inductance of the old motor. Now whenever the motor activates am am getting like 20 mV p-p niose (IIRC) on my Gyro signal. I am getting noise from other un known places too.

I think I have a pretty good trace layout with ground planes under the analog portion of the board.

I am currently using two LP8340 v-regs on vcc and avcc which have a PSSR of <10Db at high frequencies. So I am hoping that I will get an improvement if I find a v-reg with higher PSSR. I have attached my excel sheet of the devices I am currently considering.

1) Is PSSR the right spec to be looking at?
2) What other specs should I look at?
3) Do you think that the ADP3335 is a good choice?

Attachment(s): 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have no idea about the regulator, however are your ground planes for analog/digital/power separate and connecting at a single point?

What about the diodes on the H-bridge? Are you relying on the mosfet body diodes?

The other thing to consider is the zobel network across the motor and/or an inductor depending on the motor power...

Can you post the schematic/pcb/layout?

Michael

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

What is PSSR?

I don't think that your problems are anything to do with the choice of regulator, they are probably caused by poor PCB design.

Leon Heller G1HSM

Last Edited: Tue. Nov 10, 2009 - 06:42 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

krazatchu wrote:
I have no idea about the regulator, however are your ground planes for analog/digital/power separate and connecting at a single point?

What about the diodes on the H-bridge? Are you relying on the mosfet body diodes?

The other thing to consider is the zobel network across the motor and/or an inductor depending on the motor power...

Can you post the schematic/pcb/layout?

Michael

Yes the grounds are seperated and connected at a single point. Also I have an iron curtain complete with no-mans-land between analog and digital traces figuratively speaking.

I am using the internal diodes in my MOSFETs. See schematic.

My motor power can spike to over 1 Watt.

Attachment(s): 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Sorry I assumed that if it was in a datasheet it must be common terminology.

PSSR = Power Supply Rejection Ratio

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

leon_heller wrote:
What is PSSR?

I don't think that your problems are anything to do with the choice of regulator, they are probably caused by poor PCB design.

I only have 4 traces that cross my iron curtian: Battery+, GND, and two other traces that do not go anywhere near the gyro or the h-bridge.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I'm just guessing here, someone pls correct me if I'm wrong...

The new lower inductance motor is going to create faster back-EMF..
The body diodes aren't reacting fast enough now that the back-EMF is faster...

So perhaps add an RC network across your motor to slow down the back EMF to the point where the body diodes can react in time...

Also might want to consider adding some fast schottkys in parallel with the body diodes...
Something like the 1n5819 (IIRC)? Or the SMD equiv. ss14?

Just a guess anyhow... maybe I'm totaly wrong, wouldn't be the first time...

Michael

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Well, look at what 10db means. In terms of voltage, it means a reduction by a factor of (about) 3. But, that is the reduction between the input and the output of the voltage regulator. You are quoting signal on the gyro output. There is certainly some degree of PSRR for the gyro that adds to the effective PSRR of the regulator.

So, if regulator PSRR is the issue, you should be able to observe large switcher signals on the gyro supply voltage. But, don't fool yourself. Be sure to measure between the actual ground pin of the gyro and the power supply pin of the gyro. Don't assume a "good ground" and use some other point for your measurement ground.

I am willing to bet that the signal AT the gyro is good and you are seeing the effects of large currents though signal grounds.

Jim

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

dpaulsen wrote:
Sorry I assumed that if it was in a datasheet it must be common terminology.

PSSR = Power Supply Rejection Ratio

That isn't relevant to your problems, as I suggested. You need to look at power and ground connections.

Leon Heller G1HSM

Last Edited: Tue. Nov 10, 2009 - 07:28 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ka7ehk wrote:
Well, look at what 10db means. In terms of voltage, it means a reduction by a factor of (about) 3. But, that is the reduction between the input and the output of the voltage regulator. You are quoting signal on the gyro output. There is certainly some degree of PSRR for the gyro that adds to the effective PSRR of the regulator.

So, if regulator PSRR is the issue, you should be able to observe large switcher signals on the gyro supply voltage. But, don't fool yourself. Be sure to measure between the actual ground pin of the gyro and the power supply pin of the gyro. Don't assume a "good ground" and use some other point for your measurement ground.

I am willing to bet that the signal AT the gyro is good and you are seeing the effects of large currents though signal grounds.

Jim

Thanks for the suggestion. Back when I was troubleshooting this issue I was able to observe noise on the supply pin of the gyro. However, like you said I could have been fooling myself because I was not pulling my ground directly from the gyro.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

My gyro has a spec called "output signal supply drift" and it lists it as 250 mV/V

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

That sounds like a DC (or, at best, a "low frequency") spec, and not indicative of PSRR.

You MIGHT consider adding power supply filtering to the supply that controls the gyro and the a/d converter. But, only if your ground system is proven to be clean.

I would also look at this. It takes some care. Connect scope ground to the actual ground pin of the gyro (or very, very close to it). Put the scope probe on the microprocessor ground. This will show you how much voltage difference there is between these two points. You may be surprised.

If this is small, I would then look at whether or not your 20mV observation is real. It could be that that point you are using as ground isn't. That COULD be giving you an indication of noise that is larger than it really is. Been there, done that. Made incorrect conclusions as a result of it. Embarrassed by it more than once. It can be hard to sort out!

Jim

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Also if you are powering the whole thing off of a wall wart then you may have voltage drops in the ground lead of the wall wart to the board which will vary when heavy loads are switched. The ground itself would sag and rise with loads being turned off and on.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ka7ehk wrote:
That sounds like a DC (or, at best, a "low frequency") spec, and not indicative of PSRR.

You MIGHT consider adding power supply filtering to the supply that controls the gyro and the a/d converter. But, only if your ground system is proven to be clean.

I would also look at this. It takes some care. Connect scope ground to the actual ground pin of the gyro (or very, very close to it). Put the scope probe on the microprocessor ground. This will show you how much voltage difference there is between these two points. You may be surprised.

If this is small, I would then look at whether or not your 20mV observation is real. It could be that that point you are using as ground isn't. That COULD be giving you an indication of noise that is larger than it really is. Been there, done that. Made incorrect conclusions as a result of it. Embarrassed by it more than once. It can be hard to sort out!

Jim

Thanks Jim. I think your comments have been very effective in helping my think about the issue in a realistic and methodical way. Sometimes its easy to jump to a conclusion without methodically eliminating all possibilities.

I will go back and reprobe the problem.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

alwelch wrote:
Also if you are powering the whole thing off of a wall wart then you may have voltage drops in the ground lead of the wall wart to the board which will vary when heavy loads are switched. The ground itself would sag and rise with loads being turned off and on.
Thanks for the suggestion, but it is battery powered.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

How long are the leads from the battery to the board?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

alwelch wrote:
How long are the leads from the battery to the board?
~6 INCH