suggestion of a low noise reference voltages (converter) for the 4 input pins of image sensor

Thread Starter

somaye2022

Joined Mar 23, 2022
99
Hi all,

I need a low noise DCDC converter for the 4 input pins of an image sensor. The input voltage is 4.7V and I need three different outputs Voltage of 2.5V, 2.6V, and 2.2Volt, and max.50mA output current.

The sensor needs to obtain low noise relative to its GND or VCC with respect to the reference voltage!

Could anyone tell me how could I find a low noise DCDC, when it is not too much in Datasheets of lots of Converter about Noise?

Thank you in advance with best regards
 

Attachments

Last edited:

Papabravo

Joined Feb 24, 2006
18,431
You need to be a great deal more specific about your requirements. Do you consider the voltage and current ripple that are part of the DC-DC conversion process to be noise? Perhaps you are referring to external sources of noise from either the input source or the sensor environment. You can choose components with the express purpose of limiting the amount of ripple and conducted noise in the system, but you cannot get rid of it entirely.

Edit: After reviewing the datasheet you need some voltage references that can supply 10-20 mA, not DC-DC converters. This is not quite what your original post suggested. What did you plan to do for the several hundred milliamps at 3.3 Volts?
 
Last edited:

Thread Starter

somaye2022

Joined Mar 23, 2022
99
You need to be a great deal more specific about your requirements. Do you consider the voltage and current ripple that are part of the DC-DC conversion process to be noise? Perhaps you are referring to external sources of noise from either the input source or the sensor environment. You can choose components with the express purpose of limiting the amount of ripple and conducted noise in the system, but you cannot get rid of it entirely.

Edit: After reviewing the datasheet you need some voltage references that can supply 10-20 mA, not DC-DC converters. This is not quite what your original post suggested. What did you plan to do for the several hundred milliamps at 3.3 Volts?
Thank you for your reply. Exactly, you are right. I require some voltage references that can supply 20mA. Before I thought about a low noise converter, the voltage reference of the sensor board was designed with an impedance divider and an OP(LMC6482), and an RC filter; for this purpose, there is a LineNoise problem that you could see in my post on the subject "Line frequency noise". I tried to solve the problem with the help of changing the resistor and cap, but I couldn't solve decrease the noise well enough, and The sensor obtains the low noise relative to its GND or VCC with respect to the reference voltage! Because of that, I decided to redesign the circuits with a step-down converter. I thought maybe this Idea could reduce the line noise of the sensor boards well enough. Do you have any other Ideas for preparing such a voltage reference for a sensor Board with high sensitivity to noise?

Thank you again for your help.
With best regards
I'm looking forward to hearing from you
 

DickCappels

Joined Aug 21, 2008
8,710
Maybe this will give you an idea:

On one project 60 Hz hum was getting into the output of the product. It was a very noise sensitive imaging tube deflection circuit and the smallest amount of noise resulted is a large effect in the resulting video signal. A little looking around showed the LM7805 regulator had insufficient input ripple rejection. After I put an LM7812 ahead of it as a pre-regulator the problem went away.

You might find that your best solution is to make your own regulator using a low noise voltage reference and some low noise amplifiers.

Be careful of ground loops
 

Papabravo

Joined Feb 24, 2006
18,431
Thank you for your reply. Exactly, you are right. I require some voltage references that can supply 20mA. Before I thought about a low noise converter, the voltage reference of the sensor board was designed with an impedance divider and an OP(LMC6482), and an RC filter; for this purpose, there is a LineNoise problem that you could see in my post on the subject "Line frequency noise". I tried to solve the problem with the help of changing the resistor and cap, but I couldn't solve decrease the noise well enough, and The sensor obtains the low noise relative to its GND or VCC with respect to the reference voltage! Because of that, I decided to redesign the circuits with a step-down converter. I thought maybe this Idea could reduce the line noise of the sensor boards well enough. Do you have any other Ideas for preparing such a voltage reference for a sensor Board with high sensitivity to noise?

Thank you again for your help.
With best regards
I'm looking forward to hearing from you
It would help greatly if we had more information and schematics of the complete system. If a part of your system has to be isolated from external noise sources you should consider the use of common mode and differential mode chokes in series. This can reduce the contributions of those sources to sub milli-volt levels
 

Thread Starter

somaye2022

Joined Mar 23, 2022
99
Thank you for your reply, that's right. I attached a block diagram of the whole board, I hope it helps to clarify the problem further. Generally, I need a smooth reference voltage of 2.2V, 2.5V, and 2.6V with a maximum 20mA output current that could support the image sensor reference voltages to give me clean images of the 4 Output of the sensors, and in all the testing and investigation of the board, it was discovered that the problem comes from a sensor reference voltage. Because of that, the sensor needs low noise relative to its GND or VCC concerning the reference voltage!
The circuit of the noisy reference voltage used is also attached.

with best regards
 

Attachments

Last edited:

Thread Starter

somaye2022

Joined Mar 23, 2022
99
Hi everyone,
Ideally, from a battery voltage (~4.7V nom.) I want to generate voltages of 2.2, 2.5 and 2.6V with an output current of max 20mA. Since these 2.2, 2.5, and 2.6V are intended as a reference voltage for a sensor board (in Block Diagram attached), a normal voltage divider with two resistors with some Capacitor and resistor as filter alone would not be sufficient(had noise) and a buffer was added. As buffer LTM6482 was chosen(is attached). My goal is to get these voltages as "clean" as possible, without any noise. But this developed Circuit had Noise.

as another Idea, I search a lot for a step-down converter for my case. I haven't found an appropriate one until now.
Because This variant as a step-down converter seems could not to be a reasonable one. Many low voltage step-down regulators with high 95% efficiency have been looked at so far, e.g., TPS62200 is one of them. The output voltage of this regulator has, at least, a 1% ripple concerning the datasheet. With, e.g., Vout=2.6V, one has a 26mV ripple, which should not happen with the sensor board of the camera. This ripple is not more than 0.5mV before the amplifier, with a gain of 62!(was also attached the simulation)

Can anyone suggest a solution that would give me in this case or an idea of how could develop a circuit from 4.7 Volt supply to 4 low noise output Voltages of 2.2V, 2.5V, 2.6V with max 20mA output current?
Thank you in advance
 

Attachments

Last edited:

Papabravo

Joined Feb 24, 2006
18,431
Since you already have an LTspice simulation, I suggest you do a noise analysis to determine where it is all coming from. Likely sources are the opamp and the resistors. The opamp you have chosen is not known for its low noise characteristics.
 
Last edited:

Thread Starter

somaye2022

Joined Mar 23, 2022
99
Since you already have an LTspice simulation, I suggest you do a noise analysis to determine where it is all coming from. Likely sources are the opamp and the resistors. the opamp you have chosen is not known for its low noise characteristics.
Thank you for your reply.
You are right, now I decided to choose another one https://www.mouser.de/ProductDetail/Analog-Devices/ADA4661-2ACPZ-R7?qs=pqQDoYZ2quYvk8FtGfF/fg== , and it has better noise characteristics. I'm not sure if there is a better choice for this case?
Regarding noise analyses, I haven't yet done a noise analysis, but my colleagues tell me the noise in the Test came from resistors, and cap without OP was the result in the Test also not better, but the problem is with changing the circuit with res and cap also come out not enough good results. Because of that I think when I use another OP it will help maybe with the noise problem, am I right? or do you have another Idea?
 

Papabravo

Joined Feb 24, 2006
18,431
I suggest that the use of a precision reference like the LT1389 as in the following. R1 can be much higher if you need to reduce the current in the reference device U1.

1653063850548.png
 

Thread Starter

somaye2022

Joined Mar 23, 2022
99
I suggest the use of a precision reference like the LT1389 as in the following. R1 can be much higher if you need to reduce the current in the reference device U1.

View attachment 267717
Thank you very much for your suggestion. I have compared the Suggestion with the last design regarding Noise characteristics. I have attached the result. I think, regarding the simulation result there isn't a big difference between the two Results when I design the circuit for the frequency range of 1KHz to 40KHz. what do you think about that?

Thanks a lot with best regards
 

Attachments

Thread Starter

somaye2022

Joined Mar 23, 2022
99
Hi everyone, I have a question regarding the output noise of AP2127. I have already done a noise analysis for that. As sources, I have chosen the supply voltage, and I see Vout and V(R1, R11, R12) as the sum of the Vrms(noise) of all resistors is 6uVrms. But inside the datasheet has been written the Vrms noise of LDO AP2127 is about 60uVrms. It means that there is a big difference between my noise analysis(Vnoise LOD0=0?) and the one from the datasheet. Could you please give me some hints/solutions for these differences? Respectively, How would I calculate the Total Noise of the System for 10kHz-100kHz to reach about a noise amount of about 60uVrms?

Thank you in advance with best regards.
 

Attachments

Thread Starter

somaye2022

Joined Mar 23, 2022
99
I suggest using a precision reference like the LT1389 as in the following. R1 can be much higher if you need to reduce the current in the reference device U1.

View attachment 267717
I saw by LT1389 that there is a long factory lead time, and because of that, I tried to choose another one that is fast the same as it, namely LT1634-1.25. Honestly, I will order some samples to see if it will help clear the line noise problem on the Sensor-Head-Board. Shall LT1634 be a good choice instead of LT1389?
Thank you in advance, with best regards!
 

Papabravo

Joined Feb 24, 2006
18,431
I saw by LT1389 that there is a long factory lead time, and because of that, I tried to choose another one that is fast the same as it, namely LT1634-1.25. Honestly, I will order some samples to see if it will help clear the line noise problem on the Sensor-Head-Board. Shall LT1634 be a good choice instead of LT1389?
Thank you in advance, with best regards!
I chose a part that was available for simulation. If it provides a stable 1.25 volts, it seems like it will do the job. When you get the parts it will be up to you to make that determination in the lab. It is OK to believe in simulation results, but consider Ronald Reagan's sage advice: "trust, but verify".
 

Thread Starter

somaye2022

Joined Mar 23, 2022
99
I chose a part that was available for simulation. If it provides a stable 1.25 volts, it seems like it will do the job. When you get the parts it will be up to you to make that determination in the lab. It is OK to believe in simulation results, but consider Ronald Reagan's sage advice: "trust, but verify".
Thank you very much for your hint.
 
Top