Today I got 4 LT3045 but the noise is the same. What can I do wrong?
I have the 3.34V but it is still noisy
I have the 3.34V but it is still noisy
Frankly, I'm amazed you managed to get something as small as an LT3045 to work with clip-on leads!
Depending on how exactly you laid out the ground wires, it could be related to the self-inductance of the ground wires. The regulator produces a voltage of 3.32 V with respect to the ground point of the 33.2 kohm in parallel with 0.47 uF. That ground point should be connected to the ground of the board with the microphone and to the analogue ground of the microcontroller board with wires that don't conduct any of the digital ground current of the microcontroller. (At least that is one strategy to minimize coupling via ground wire inductance, another is to use a ground plane, as is often done on printed circuit boards.)
Depending on how exactly you laid out the ground wires, it could be related to the self-inductance of the ground wires. The regulator produces a voltage of 3.32 V with respect to the ground point of the 33.2 kohm in parallel with 0.47 uF. That ground point should be connected to the ground of the board with the microphone and to the analogue ground of the microcontroller board with wires that don't conduct any of the digital ground current of the microcontroller. (At least that is one strategy to minimize coupling via ground wire inductance, another is to use a ground plane, as is often done on printed circuit boards.)
First of all I don't think your noise issue has anything to do with the regulator. TPS736xx has low enough noise for this application.
Your LT3045 implementation while inventive is an abuse to the expensive regulator and I would be surprised if it does not oscillate. Regarding the output capacitor the datasheet states this:
"For stability, use a minimum 10μF output capacitor with an ESR below 20mΩ and an ESL below 2nH."
Electrolytic capacitor connected with 10cm wires hardly meets the above criteria.
Your LT3045 implementation while inventive is an abuse to the expensive regulator and I would be surprised if it does not oscillate. Regarding the output capacitor the datasheet states this:
"For stability, use a minimum 10μF output capacitor with an ESR below 20mΩ and an ESL below 2nH."
Electrolytic capacitor connected with 10cm wires hardly meets the above criteria.
bohrok2610 could have phrased it a bit more friendly, but has a point about possible instability. A well-known rule of thumb says that wires at a large distance from a ground plane have a self-inductance of the order of 1 nH/mm, so the wires to the output decoupling capacitor have to be very short indeed. The capacitor itself also has to have very little inductance, which pretty much means that you have to use a ceramic multilayer SMD capacitor.
It hasn't, actually. The datasheet shows noise levels of 100 nV/√Hz to 300 nV/√Hz. A typical two-pin electret microphone with built-in FET has an A-weighted average output noise voltage density of 55 nV/√Hz and practically no power supply rejection.
More importantly, if I understand it correctly, the supply current of a microcontroller is flowing through it.
Most lab supplies have higher noise than TPS736xx. As the noise is in LF I would suspect some grounding related issue.
And to get lower LF noise than TPS736xx the Cset in LT3045 should be at least 4.7uF (not 470nF).
And to get lower LF noise than TPS736xx the Cset in LT3045 should be at least 4.7uF (not 470nF).
Didn't I do that?
I can't test it any better. I cannot construct a circuit board. I am a little sad that it does not improve.
So while the LT3045 is better than the TPS736xx. But I can't use the LT3045 because I can't connect it correctly? And then I have the problem that Vdda and Gnda are already hardwired.
What can I improve about the power supply? Or is there nothing I can do to improve the voltage supply?
I have attached a WAV of the background noise.
What type of 230V/5V power supply did you use? Have you tried powering the MCU with 5V from your lab power supply or USB from your PC?
I don't know. In the schematic you posted, all grounds are connected via ground symbols and the photo only shows the regulator. No idea what your ground wires look like and whether there is a wire somewhere of which the inductance or resistance causes crosstalk.
https://www.meanwell.com/productPdf.aspx?i=410#1
Mean Well RS 100-5
The LED Stripes need much current.
Not yet, but i make it now.
LT3045 cannot be used without a proper PCB. There are prebuilt boards with LT3045 available. But I would first make sure that the regulator on the MCU board needs improving.
That may have quite a lot of noise that gets into your MCU board through the ground wire.
I now know what I need to do to make the noise go away.
But I need this. When I disconnect the data line to the strips.
Is the noise gone.
But I need this. When I disconnect the data line to the strips.
Is the noise gone.
Are your pullup resistors connected to raw 5V or 3.3V? Can't you use internal pullups on these GPIOs?
Connected to 5V because the stripes need 5V.
And configured as open drain
Can't you use internal pullups on these GPIOs?
Yes but i need 5V
If these are just some LED stripes I would use a separate power supply for those.
BTW have you tested what happens if you use 3.3V for the pullups or internal GPIO pullups? Your LED strips may work with that setup as well.
BTW have you tested what happens if you use 3.3V for the pullups or internal GPIO pullups? Your LED strips may work with that setup as well.
Last edited:
your are right. it work with the 3.3v from microcontroller.
the rms noise dropped from -28.86dB to -36.62dB connected with labor power supply
i use now a 2.2k resistor between the gpio and data signal. to protect the first led.
here are the settings are possible:
//*****************************************************************************
//
// Values that can be passed to GPIOPadConfigSet as the ui32Strength parameter,
// and returned by GPIOPadConfigGet in the *pui32Strength parameter.
//
//*****************************************************************************
#define GPIO_STRENGTH_2MA 0x00000001 // 2mA drive strength
#define GPIO_STRENGTH_4MA 0x00000002 // 4mA drive strength
#define GPIO_STRENGTH_6MA 0x00000065 // 6mA drive strength
#define GPIO_STRENGTH_8MA 0x00000066 // 8mA drive strength
#define GPIO_STRENGTH_8MA_SC 0x0000006E // 8mA drive with slew rate control
#define GPIO_STRENGTH_10MA 0x00000075 // 10mA drive strength
#define GPIO_STRENGTH_12MA 0x00000077 // 12mA drive strength
//*****************************************************************************
//
// Values that can be passed to GPIOPadConfigSet as the ui32PadType parameter,
// and returned by GPIOPadConfigGet in the *pui32PadType parameter.
//
//*****************************************************************************
#define GPIO_PIN_TYPE_STD 0x00000008 // Push-pull
#define GPIO_PIN_TYPE_STD_WPU 0x0000000A // Push-pull with weak pull-up
#define GPIO_PIN_TYPE_STD_WPD 0x0000000C // Push-pull with weak pull-down
#define GPIO_PIN_TYPE_OD 0x00000009 // Open-drain
#define GPIO_PIN_TYPE_ANALOG 0x00000000 // Analog comparator
#define GPIO_PIN_TYPE_WAKE_HIGH 0x00000208 // Hibernate wake, high
#define GPIO_PIN_TYPE_WAKE_LOW 0x00000108 // Hibernate wake, low
Last edited:
I could reduce the noise even further by not sending data over SPI so to the stripes.
But that I no solution.
Does anyone have an idea how I can improve this?
But that I no solution.
Does anyone have an idea how I can improve this?
Do you mean too much with Meanwell? I would use that only for the LED strips and something less noisy for the MCU board.