Author Topic: Low ripple low noise small power supply design  (Read 24936 times)

0 Members and 1 Guest are viewing this topic.

Offline VEGETA

  • Super Contributor
  • ***
  • Posts: 1475
  • Country: jo
  • I am the cult of personality
    • Thundertronics
Re: Low ripple low noise small power supply design
« Reply #375 on: March 18, 2021, 07:41:51 am »
Ok I will try soldering the switchers next Monday and do more detailed tests.

Quote
Is it another power supply? a battery?

it is an adjustable wall power supply which can deliver 12v @ 3amps, not bad quality.

Quote
Your 1.6 ohm load will draw over 3 amps on a 5V source.

the switchers are supposed to have 4 amps of maximum current, so 3.2 amps should be fine. of course the final output needed for 5v is less than 2 amps. the dreamcast won't take more than 1 amps actually. However, we wanted to have 2 amps for the 5v rail just in case mods are installed. in summery, the circuit should output 2-3 amps and more. However, the mosfets will get hot without heatsink so a heatsink is required. I didn't install one yet since I don't have one but as you said, more tests will be done.




Quote
It's going to be very hard to diagnose a problem that makes a switcher droop.  First thing to try is other loads.  Find out if it is drooping at lower currents like 2 amps, 2.5 amps, 1 amp...  If it is only drooping above some current the current limit must be kicking in.  If the droop depends on the current over a wide range, then something is wrong that is interfering with the regulation.

yes I will have other loads once I make a complete board with 2 switchers. it is very hard to solder them! last night I barely could solder one switcher using the "scrap the solder mask and tin it with solder then put ic and heat using iron technique".

If I get one board complete and did all these tests, I will send it to you via EMS if you accept.



On a side note, the TI design is about to be complete. I am just figuring out the inductor choice now.

Offline VEGETA

  • Super Contributor
  • ***
  • Posts: 1475
  • Country: jo
  • I am the cult of personality
    • Thundertronics
Re: Low ripple low noise small power supply design
« Reply #376 on: March 23, 2021, 08:36:34 pm »
I have managed to solder the 3.3v switcher and mosfet, and it works.

It outputs the following:

without elec. cap: 3.4v (+some 10s of mV, didn't remember).
with elec. cap: 3.360v

with elec. cap + 1.3R load: 3.294v (starts at about 3.3v but slowly decreases as it gets hot, no heatsink).

However I noticed something very weird, that is the switcher doesn't start at all until I momentarily short the pin #1 which is the output with pin #9 which is the input of 12.6v. I noticed this very clearly.

After this moment of touch, it starts working as described.


On a side note, I finished the design using the TI part TPS62913 and sent it to PCBway and AllPCB. Got a final offer from PCBway for 5 completely assembled boards of 181$ including shipping and still waiting for AllPCB offer. 181$ is so expensive since the TI part from PCBway is about 3.5$ each = 35$ total for 5 boards. Not gonna be economical for production with this price but it gets decreased with quantity + it is still pre-production unit. If you can help, I can send you the schematic + layout to check. I followed it as best as I can according to datasheet and app note.


Offline gnuarm

  • Frequent Contributor
  • **
  • Posts: 941
  • Country: aq
Re: Low ripple low noise small power supply design
« Reply #377 on: March 23, 2021, 09:11:48 pm »
Not sure why the cap would impact the output voltage.  A 6 mV droop with temperature is good.  Very few applications are actually concerned with the exact voltage.  Noise on the rail is more important if you are running analog circuitry from it. 

I forget, what problems are caused by the ripple in the supply?  Does it mess up the video sync?  It seems like that could be resolved with some filtering capacitors on the supply into the impacted circuitry.  I can't remember why you need to get sub millivolt ripple levels.

Why not post your schematics here?  It is helpful if you can print them to a PDF with the text intact so part numbers can be selected and copied and ref des can be searched. 
Rick C.
  - Get 1,000 miles of free Supercharging
  - Tesla referral code - https://ts.la/richard11209
 

Offline VEGETA

  • Super Contributor
  • ***
  • Posts: 1475
  • Country: jo
  • I am the cult of personality
    • Thundertronics
Re: Low ripple low noise small power supply design
« Reply #378 on: March 23, 2021, 09:22:33 pm »
the cap at the final output pin is 220v elec. cap which is necessary for linear circuit stability, this was our conclusion.

I want low ripple since I really like the idea of designing a quality PSU, plus ripple and noise affects output video signal in analog video game consoles. so why not make a very good psu once and for all.

I attached the schematic of this design and also "low_noise_PSU" which is the new one.


Offline gnuarm

  • Frequent Contributor
  • **
  • Posts: 941
  • Country: aq
Re: Low ripple low noise small power supply design
« Reply #379 on: March 23, 2021, 11:06:37 pm »
Is this schematic "as built"?  I see C38 and C39 which need to be removed. 

I see three separate filtering circuits.  I don't understand why you need them to feed the on board switchers.  Filtering the input to U1 is useful, but not going to the switchers.  Does Vcc go on to feed other sensitive circuits?  If so, then the filters prevent noise from the switcher inputs from feeding to the other circuits.  You do need the larger cap and the smaller cap to smooth the current pulses. 

It is important to route the paths around the input, through the input caps back to ground very short.  This path is an important source of noise.  The current through the output inductor and cap are not.  The current through the inductor is pretty constant... that's what inductors do.  The diodes D1 and D2 need to be as direct as possible between LX and ground. 

I don't see a reason to not increase R1 from 100 ohms to 1K.  Worse case draw on the output of that regulator is 2 mA, so dropping 2 volts won't harm a thing.  Higher resistance gives higher attenuation of noise.

Who said the output cap needs to be 220V???  That is an absurdly high voltage.

This is your original design from weeks ago, right?
Rick C.
  - Get 1,000 miles of free Supercharging
  - Tesla referral code - https://ts.la/richard11209
 

Offline VEGETA

  • Super Contributor
  • ***
  • Posts: 1475
  • Country: jo
  • I am the cult of personality
    • Thundertronics
Re: Low ripple low noise small power supply design
« Reply #380 on: March 23, 2021, 11:21:39 pm »
Quote
Is this schematic "as built"?  I see C38 and C39 which need to be removed. 

yes as built. I tried removing these 2 caps previously and it didn't affect the circuit. adding an elect. cap made it stable even with these 2 caps.

Quote
I see three separate filtering circuits.  I don't understand why you need them to feed the on board switchers.  Filtering the input to U1 is useful, but not going to the switchers.  Does Vcc go on to feed other sensitive circuits?  If so, then the filters prevent noise from the switcher inputs from feeding to the other circuits.  You do need the larger cap and the smaller cap to smooth the current pulses. 

one for each switcher, that makes it 2 filtering circuits. and another one to deliver 12v input to output connector which goes to dreamcast and it is used to drive gd-rom drive motor... no need for it to be low noise at all, but also it shouldn't add the noise to the system as it uses the same ground. filtering it this way was the solution. this rail also is the input to the 5v reg and linear stage op-amp as you see, which need to have low noise.

filtering switchers input is not really necessary but it would be a problem if the 12v input itself is very bad noise and ripple, I thought it might affect the switchers so i added a small inductor with caps.

Quote
It is important to route the paths around the input, through the input caps back to ground very short.  This path is an important source of noise.  The current through the output inductor and cap are not.  The current through the inductor is pretty constant... that's what inductors do.  The diodes D1 and D2 need to be as direct as possible between LX and ground. 

the layout is done properly and took a lot of time. very good layout as datasheet and better.

Quote
I don't see a reason to not increase R1 from 100 ohms to 1K.  Worse case draw on the output of that regulator is 2 mA, so dropping 2 volts won't harm a thing.  Higher resistance gives higher attenuation of noise.

this may aid in making it more low noise, but the problem was that the circuit wasn't functioning to begin with.

Quote
Who said the output cap needs to be 220V???  That is an absurdly high voltage.

actually it is 220uF 16v, that was a typo.

Quote
This is your original design from weeks ago, right?

this is the design that was produced, yes.

and the "low_noise_psu.pdf" is a totally new design based on new TI switcher which promises <1mV of total ripple and noise. I spent some time with this to make it very good layout. please check it and see if you have a comment. the boards are not manufactured yet, only quotation.

Offline gnuarm

  • Frequent Contributor
  • **
  • Posts: 941
  • Country: aq
Re: Low ripple low noise small power supply design
« Reply #381 on: March 24, 2021, 01:37:30 am »
Quote
Is this schematic "as built"?  I see C38 and C39 which need to be removed. 

yes as built. I tried removing these 2 caps previously and it didn't affect the circuit. adding an elect. cap made it stable even with these 2 caps.

Zero reason to add them back.  They do nothing but slow the feedback which can increase the noise.  Feedback is what reduces noise.  Slow it and the noise is not properly mitigated.  Get rid of them permanently.


Quote
Quote
I see three separate filtering circuits.  I don't understand why you need them to feed the on board switchers.  Filtering the input to U1 is useful, but not going to the switchers.  Does Vcc go on to feed other sensitive circuits?  If so, then the filters prevent noise from the switcher inputs from feeding to the other circuits.  You do need the larger cap and the smaller cap to smooth the current pulses. 

one for each switcher, that makes it 2 filtering circuits. and another one to deliver 12v input to output connector which goes to dreamcast and it is used to drive gd-rom drive motor... no need for it to be low noise at all, but also it shouldn't add the noise to the system as it uses the same ground. filtering it this way was the solution. this rail also is the input to the 5v reg and linear stage op-amp as you see, which need to have low noise.

You seem to want to add things without understanding how they work.  I assume the dreamcast circuitry is fed through the output connector on the pin labeled "12V"?  That pin also feeds the 5V regulator, so any noise coming back from that is fed into your voltage reference.  Why not feed that from the Vcc input before the filter? 

Remove the two filters feeding the two switchers leaving a 22uF cap and a 100nF cap.  With a 3 amp load, you might want to increase the capacitance from 22 uF to 100 uF. 

C * dV = I * dT, dV and dT are deltas

I don't know the frequency your switcher is running, but assuming 1 MHz, that gives...

dV = I * dT / C = 3 A * 1 us / 100 uF = 0.030V

So 100 uF will give 30 mV ripple going into the switcher which should be fine.  22 uF will be 136 mV ripple. 

Also, the ground circulation paths should be isolated from the rest of the design if you can.  A very short path between the high current/high frequency loops on a separate area of copper, like an island.  Connect that area to the rest of the ground plane via a single connection.  Remember the thermal breaks I was complaining about in one of your layouts?  One connection like your thermal break.  This prevents voltage differentials from interfering with the rest of the circuit. 


Quote
filtering switchers input is not really necessary but it would be a problem if the 12v input itself is very bad noise and ripple, I thought it might affect the switchers so i added a small inductor with caps.

That is not of use.  The place to add filtering is between the switcher and the linear.


Quote
Quote
It is important to route the paths around the input, through the input caps back to ground very short.  This path is an important source of noise.  The current through the output inductor and cap are not.  The current through the inductor is pretty constant... that's what inductors do.  The diodes D1 and D2 need to be as direct as possible between LX and ground. 

the layout is done properly and took a lot of time. very good layout as datasheet and better.

Quote
I don't see a reason to not increase R1 from 100 ohms to 1K.  Worse case draw on the output of that regulator is 2 mA, so dropping 2 volts won't harm a thing.  Higher resistance gives higher attenuation of noise.

this may aid in making it more low noise, but the problem was that the circuit wasn't functioning to begin with.

You asked for advice.  This is my advice.  If you want to have noise in your circuit, the reference will cause it more than anything else. 

Much of your original design was either adding ineffective components like the filtering before the switchers or even counter productive like the cap on the FET gate.  The reference is the gold standard for the entire rest of the design.  Keep that clean or the output will be hard to clean up.


Quote
Quote
Who said the output cap needs to be 220V???  That is an absurdly high voltage.

actually it is 220uF 16v, that was a typo.

Quote
This is your original design from weeks ago, right?

this is the design that was produced, yes.

and the "low_noise_psu.pdf" is a totally new design based on new TI switcher which promises <1mV of total ripple and noise. I spent some time with this to make it very good layout. please check it and see if you have a comment. the boards are not manufactured yet, only quotation.

I don't see a layout, only schematics.
Rick C.
  - Get 1,000 miles of free Supercharging
  - Tesla referral code - https://ts.la/richard11209
 

Offline VEGETA

  • Super Contributor
  • ***
  • Posts: 1475
  • Country: jo
  • I am the cult of personality
    • Thundertronics
Re: Low ripple low noise small power supply design
« Reply #382 on: March 24, 2021, 06:44:13 am »
Quote
Zero reason to add them back.  They do nothing but slow the feedback which can increase the noise.  Feedback is what reduces noise.  Slow it and the noise is not properly mitigated.  Get rid of them permanently.

I didn't add them, the boards are already manufactured (125 boards, was my mistake to make many boards as mentioned).

Quote
You seem to want to add things without understanding how they work.  I assume the dreamcast circuitry is fed through the output connector on the pin labeled "12V"?  That pin also feeds the 5V regulator, so any noise coming back from that is fed into your voltage reference.  Why not feed that from the Vcc input before the filter?

yes the output to dreamcast is fed through that 6 pin connector, it has ground, 12v, 3.3v, and 5v.

Quote
That is not of use.  The place to add filtering is between the switcher and the linear.

Noted.

I actually made another iteration doing just that but as I mentioned, I went for a completely new design using TI part. So all new design is going to be the TI part, not this circuit.

I may use this circuit as side project for the future, just to get it working properly.

Quote
You asked for advice.  This is my advice.  If you want to have noise in your circuit, the reference will cause it more than anything else.

Much of your original design was either adding ineffective components like the filtering before the switchers or even counter productive like the cap on the FET gate.  The reference is the gold standard for the entire rest of the design.  Keep that clean or the output will be hard to clean up.

So you mean I feed 12v output pin directly from 12v input without filtering from the 1000uF of reference? I could just put caps on the output pin instead and keep the reference alone.

Quote
I don't see a layout, only schematics.


I provided the layout as images in previous posts, kindly refer to them now since I am away from my laptop for now. when I return, I will upload them again.

________

for now I will focus on the TI part, kindly refer to the attached design "low_noise_psu.pdf" and see it if you have any note.

Offline VEGETA

  • Super Contributor
  • ***
  • Posts: 1475
  • Country: jo
  • I am the cult of personality
    • Thundertronics
Re: Low ripple low noise small power supply design
« Reply #383 on: May 31, 2021, 10:17:40 am »
Hello, if anyone still here.

I have got the design I done with TPS62913, and it worked perfectly fine first go, no problem what so ever. I still didn't measure total output ripple due to no scope, but I noticed something indicative.

I have a 12v brick which is very bad and noisy, when i connect it to the power supply I have (called re-dream, which promises 30mv ripple, best in market) it results in a lot and a lot of noise and garbage in picture. When I use this bad source with my design, it produces noticeably less noise and garbage.

Anyway, I adjusted the design slightly to move some components for better arrangement, and added 2 more 22uF input caps per TPS62913 which now has 4 caps instead of 2, datasheet doesn't put a limit on input capacitance but just output capacitance. I also added 5 22uF caps to 12v rail itself (which also feeds the 2 rails). Plus, I added another ground pour on bottom layer and connected it to top layer ground, separated nicely from 12v one.

I noticed 5v rail goes to 5.13v on load, so I don't know if this is dangerous or not.
« Last Edit: May 31, 2021, 10:21:27 am by VEGETA »
 

Offline gnuarm

  • Frequent Contributor
  • **
  • Posts: 941
  • Country: aq
Re: Low ripple low noise small power supply design
« Reply #384 on: June 10, 2021, 08:30:33 pm »
I seem to recall the TPS62913 is a low noise switcher.  Did you follow that with a linear?  Not that I am encouraging that, just asking the question since that was your original concept. 

I don't recall the details, but I think you are outputting 5V and 3.3V at maybe 3 amps.  That should leave you lots of headroom, 5V perhaps.  Add up your max load on the 12V input (after the voltage conversion which will reduce the input current) and consider adding a resistor between the input and the caps.  The caps can only filter input noise as part of an RC filter (or LC).  If you don't add an R or L to the input filter design the cap is working with the low impedance of the 12V supply.  By adding a resistor you likely get a much better filter with the same size input cap.  It does dissipate power, which can make an L a lot more attractive. 
Rick C.
  - Get 1,000 miles of free Supercharging
  - Tesla referral code - https://ts.la/richard11209
 

Offline VEGETA

  • Super Contributor
  • ***
  • Posts: 1475
  • Country: jo
  • I am the cult of personality
    • Thundertronics
Re: Low ripple low noise small power supply design
« Reply #385 on: June 10, 2021, 09:12:55 pm »
I seem to recall the TPS62913 is a low noise switcher.  Did you follow that with a linear?  Not that I am encouraging that, just asking the question since that was your original concept. 

I don't recall the details, but I think you are outputting 5V and 3.3V at maybe 3 amps.  That should leave you lots of headroom, 5V perhaps.  Add up your max load on the 12V input (after the voltage conversion which will reduce the input current) and consider adding a resistor between the input and the caps.  The caps can only filter input noise as part of an RC filter (or LC).  If you don't add an R or L to the input filter design the cap is working with the low impedance of the 12V supply.  By adding a resistor you likely get a much better filter with the same size input cap.  It does dissipate power, which can make an L a lot more attractive.

Yes I got TPS62913, 2 of it. One for 3.3v and one for 5v, from 12v input, no LDO after it.

I have an L in input plus many caps... but no R since it will get really hot... about 25W is delivered from 12v source to the design + 12v rail. Do you mean a 0R resistor?

Offline VEGETA

  • Super Contributor
  • ***
  • Posts: 1475
  • Country: jo
  • I am the cult of personality
    • Thundertronics
Re: Low ripple low noise small power supply design
« Reply #386 on: June 13, 2021, 06:16:31 pm »
Kindly see the schematic.

these are the bulk caps added on 12v rail + input of each dc-dc regulator.

I assume you mean I add 0.05R series resistor before the 1000uF cap? or multiple 0.05R resistors between the caps to act as multistage?


Offline gnuarm

  • Frequent Contributor
  • **
  • Posts: 941
  • Country: aq
Re: Low ripple low noise small power supply design
« Reply #387 on: June 13, 2021, 09:04:09 pm »
You have inductors, so forget the resistor.  I don't understand all the same value caps.  You have six 22 uF caps at the 1000 uF cap, then four more at he regulator input.  The image is clipped, so I can't tell about the output, but you have three before the ferrite bead (which is not really an inductor, it's a high frequency resistor) and at least three more after. 

You need to read the data sheet carefully.  While total capacitance can be an issue for some regulators, the ESR of the capacitors can also be an issue, too much or too little.  Be sure to read all the details on this in the data sheet.  Double check the value of C23 as well.  That is specifically for compensation and needs to be adjusted when R7 and R8 are changed.  It doesn't need a high degree of accuracy, but might be different for the 5V and 3.3V circuits or adjust the R's to work with the same value of C23.
Rick C.
  - Get 1,000 miles of free Supercharging
  - Tesla referral code - https://ts.la/richard11209
 

Offline VEGETA

  • Super Contributor
  • ***
  • Posts: 1475
  • Country: jo
  • I am the cult of personality
    • Thundertronics
Re: Low ripple low noise small power supply design
« Reply #388 on: June 14, 2021, 09:19:05 am »
You have inductors, so forget the resistor.  I don't understand all the same value caps.  You have six 22 uF caps at the 1000 uF cap, then four more at he regulator input.  The image is clipped, so I can't tell about the output, but you have three before the ferrite bead (which is not really an inductor, it's a high frequency resistor) and at least three more after. 

You need to read the data sheet carefully.  While total capacitance can be an issue for some regulators, the ESR of the capacitors can also be an issue, too much or too little.  Be sure to read all the details on this in the data sheet.  Double check the value of C23 as well.  That is specifically for compensation and needs to be adjusted when R7 and R8 are changed.  It doesn't need a high degree of accuracy, but might be different for the 5V and 3.3V circuits or adjust the R's to work with the same value of C23.

I used 22uF in the design as it was recommended in the datasheet, now I just increased the number of them to get more filtering.

TPS62913 requires a certain amount of output capacitance, not more. Therefore I chose a suitable number of output caps. All other values are verified from datasheet such as C23 and R7\R8. the boards I have now works perfectly fine using them... I just slightly modified the feedback resistors value to get exact 3.3 and 5v since the current one is a bit more.

Ferrite beads are a part of feedback resistor for the switchers since this is what the switcher's design is, to achieve low noise and low ripple performance.

So the new modification is just better and more precise feedback resistors + more input caps.
« Last Edit: June 14, 2021, 09:22:04 am by VEGETA »
 

Offline gnuarm

  • Frequent Contributor
  • **
  • Posts: 941
  • Country: aq
Re: Low ripple low noise small power supply design
« Reply #389 on: June 19, 2021, 04:58:42 pm »
You have inductors, so forget the resistor.  I don't understand all the same value caps.  You have six 22 uF caps at the 1000 uF cap, then four more at he regulator input.  The image is clipped, so I can't tell about the output, but you have three before the ferrite bead (which is not really an inductor, it's a high frequency resistor) and at least three more after. 

You need to read the data sheet carefully.  While total capacitance can be an issue for some regulators, the ESR of the capacitors can also be an issue, too much or too little.  Be sure to read all the details on this in the data sheet.  Double check the value of C23 as well.  That is specifically for compensation and needs to be adjusted when R7 and R8 are changed.  It doesn't need a high degree of accuracy, but might be different for the 5V and 3.3V circuits or adjust the R's to work with the same value of C23.

I used 22uF in the design as it was recommended in the datasheet, now I just increased the number of them to get more filtering.

I hope you realize the data sheet is referring to the total capacitance, not the value of any components.  Using multiple caps means you need to add up the total to compare to the data sheet values.  I can see in your drawing 132 uF of capacitance on the output.  Is that what the device is specified for?


Quote
TPS62913 requires a certain amount of output capacitance, not more. Therefore I chose a suitable number of output caps. All other values are verified from datasheet such as C23 and R7\R8. the boards I have now works perfectly fine using them... I just slightly modified the feedback resistors value to get exact 3.3 and 5v since the current one is a bit more.

If the data sheet is specifying a total capacitance maximum, then that is what you design to with the sum of all output capacitance including what is on the load boards.


Quote
Ferrite beads are a part of feedback resistor for the switchers since this is what the switcher's design is, to achieve low noise and low ripple performance.

So the new modification is just better and more precise feedback resistors + more input caps.

Not sure what your point is.  I mentioned to you about the feedback capacitor.  If you are using the exact same values as recommended in the data sheet in both the 3.3V and 5V circuits fine.  But the resistor values can't be the same between the two. 
Rick C.
  - Get 1,000 miles of free Supercharging
  - Tesla referral code - https://ts.la/richard11209
 

Offline VEGETA

  • Super Contributor
  • ***
  • Posts: 1475
  • Country: jo
  • I am the cult of personality
    • Thundertronics
Re: Low ripple low noise small power supply design
« Reply #390 on: June 20, 2021, 05:55:36 am »
Quote
I hope you realize the data sheet is referring to the total capacitance, not the value of any components.  Using multiple caps means you need to add up the total to compare to the data sheet values.  I can see in your drawing 132 uF of capacitance on the output.  Is that what the device is specified for?

Quote
If the data sheet is specifying a total capacitance maximum, then that is what you design to with the sum of all output capacitance including what is on the load boards.

this IC is new and unique since it uses 2 stages, the 2nd stage being a ferrite bead. the unique thing is that it requires that ferrite bead to enter the feedback loop itself as part of its low noise features. So you have 2 options, either do single stage like any other dc-dc converter or use 2 stages like explained. achieving that very low noise and ripple requires 2 stages.

Datasheet specifies total effective output capacitance including both stages to be 200uF, which I didn't exceed for sure, one of them is a 47u elec. cap as final bulk capacitor preceded with 22uf ceramic ones. 22uf based on datasheet but i can get anything else.

why you asked about 22u? I used them in the design output stage and therefore thought about using them in input filtering to avoid using another part number for no use. space is very tight so I can't use big elec. caps as input bulk filtering caps... therefore using more 22uF seems like the best idea.

You mentioned the target board capacitance but it is very hard to know, the voltage will go to several paths for sure like LDOs, ICs, etc... which I guess won't add capacitance, at least not in bulk

Quote
Not sure what your point is.  I mentioned to you about the feedback capacitor.

you mean the one across one of feedback resistors? that one is based upon datasheet recommendation to eliminate as much noise as possible but still very much within range. I tested this design on practical use and it works perfectly fine. this "mod" is just adding more input filtering, that is it. so it won't change the functionality of the device.

Quote
If you are using the exact same values as recommended in the data sheet in both the 3.3V and 5V circuits fine.  But the resistor values can't be the same between the two.

Feedback resistors are chosen as needed, and yes they are not the same. in the new mod I changed them to have a more precise value + used 0.5% resistors for extra precision since the 5v rail was outputting 5.15v which is not as I wanted. Now they should be spot on... we'll see.

I will do a final check and order 10 of it.
« Last Edit: June 20, 2021, 05:57:56 am by VEGETA »
 

Offline gnuarm

  • Frequent Contributor
  • **
  • Posts: 941
  • Country: aq
Re: Low ripple low noise small power supply design
« Reply #391 on: June 24, 2021, 12:21:16 am »
Quote
I hope you realize the data sheet is referring to the total capacitance, not the value of any components.  Using multiple caps means you need to add up the total to compare to the data sheet values.  I can see in your drawing 132 uF of capacitance on the output.  Is that what the device is specified for?

Quote
If the data sheet is specifying a total capacitance maximum, then that is what you design to with the sum of all output capacitance including what is on the load boards.

this IC is new and unique since it uses 2 stages, the 2nd stage being a ferrite bead. the unique thing is that it requires that ferrite bead to enter the feedback loop itself as part of its low noise features. So you have 2 options, either do single stage like any other dc-dc converter or use 2 stages like explained. achieving that very low noise and ripple requires 2 stages.

Datasheet specifies total effective output capacitance including both stages to be 200uF, which I didn't exceed for sure, one of them is a 47u elec. cap as final bulk capacitor preceded with 22uf ceramic ones. 22uf based on datasheet but i can get anything else.

why you asked about 22u? I used them in the design output stage and therefore thought about using them in input filtering to avoid using another part number for no use. space is very tight so I can't use big elec. caps as input bulk filtering caps... therefore using more 22uF seems like the best idea.

You need to count all capacitance when meeting the spec for the data sheet.  Too much capacitance can cause oscillation, too little capacitance can cause oscillation.  You need to match the requirements in the data sheet whatever they are.


Quote
You mentioned the target board capacitance but it is very hard to know, the voltage will go to several paths for sure like LDOs, ICs, etc... which I guess won't add capacitance, at least not in bulk

Quote
Not sure what your point is.  I mentioned to you about the feedback capacitor.

you mean the one across one of feedback resistors? that one is based upon datasheet recommendation to eliminate as much noise as possible but still very much within range. I tested this design on practical use and it works perfectly fine. this "mod" is just adding more input filtering, that is it. so it won't change the functionality of the device.

This capacitor does NOT eliminate noise.  It is a high pass filter.  It passes high frequency noise past the feedback resistor so the active regulator can cancel it out.  The resistors have to match the capacitor to get the right characteristics.  Both regulators can't have the same resistors, so the capacitor has to be tuned to the resistors.  Read the data sheet.


Quote
Quote
If you are using the exact same values as recommended in the data sheet in both the 3.3V and 5V circuits fine.  But the resistor values can't be the same between the two.

Feedback resistors are chosen as needed, and yes they are not the same. in the new mod I changed them to have a more precise value + used 0.5% resistors for extra precision since the 5v rail was outputting 5.15v which is not as I wanted. Now they should be spot on... we'll see.

The point is the capacitor needs to match the different resistor values.


Quote
I will do a final check and order 10 of it.
Rick C.
  - Get 1,000 miles of free Supercharging
  - Tesla referral code - https://ts.la/richard11209
 

Offline VEGETA

  • Super Contributor
  • ***
  • Posts: 1475
  • Country: jo
  • I am the cult of personality
    • Thundertronics
Re: Low ripple low noise small power supply design
« Reply #392 on: June 24, 2021, 05:34:43 am »
Quote
You need to count all capacitance when meeting the spec for the data sheet.  Too much capacitance can cause oscillation, too little capacitance can cause oscillation.  You need to match the requirements in the data sheet whatever they are.

well, I can't measure Dreamcast capacitance very accurately but as I told you, I actually tried the design on Dreamcast and it works perfectly fine. therefore, capacitance is ok. I may just use a multimeter to measure capacitance across 5v and 3.3v rails... but will this be accurate?

Quote
This capacitor does NOT eliminate noise.  It is a high pass filter.  It passes high frequency noise past the feedback resistor so the active regulator can cancel it out.
the datasheet specifies that this capacitor helps eliminate more noise, yes it is not a traditional low pass filter but the final functionality is this.

The datasheet doesn't specify this cap value to be dependent on feedback resistor but rather a range of its own. I picked the proper value so that it is not too high or low.

here is what datasheet says:

Quote
A feedforward capacitor (CFF) is not required for proper operation, but can further improve output noise.
However, care must be taken in choosing the CFF, since the power good (PG) function may not be valid with
a large CFF during start-up, and can cause spurious triggering of the PG pin during a large load transient. The
noise performance with various CFF is shown in Figure 6-31.

and since I practically tested it and it works, then no startup issue is there. Plus, there is no load transient since load is fairly constant. No cons mentioned in the app note (mentioned in datasheet after quoted paragraph) applicable here.
_______


On the other side of the ocean, I added 2 small 1uH inductors on 12v line between the many 22uF capacitors to further enhance noise reduction just before the switchers.

Also, I sent the gerbers and bom to allpcb to manufacture 10 boards.

Offline gnuarm

  • Frequent Contributor
  • **
  • Posts: 941
  • Country: aq
Re: Low ripple low noise small power supply design
« Reply #393 on: June 25, 2021, 12:38:19 pm »
Quote
You need to count all capacitance when meeting the spec for the data sheet.  Too much capacitance can cause oscillation, too little capacitance can cause oscillation.  You need to match the requirements in the data sheet whatever they are.

well, I can't measure Dreamcast capacitance very accurately but as I told you, I actually tried the design on Dreamcast and it works perfectly fine. therefore, capacitance is ok. I may just use a multimeter to measure capacitance across 5v and 3.3v rails... but will this be accurate?

Can't say, depends on the meter.  The issue is that the actual values of capacitance vary widely.  +80/-20% from rated value is not unusual for electrolytic caps.  So you may be testing with a unit in the -20% range and will have a problem with a unit in the +80% range... or the other way around.   
 

Quote
Quote
This capacitor does NOT eliminate noise.  It is a high pass filter.  It passes high frequency noise past the feedback resistor so the active regulator can cancel it out.
the datasheet specifies that this capacitor helps eliminate more noise, yes it is not a traditional low pass filter but the final functionality is this.

The datasheet doesn't specify this cap value to be dependent on feedback resistor but rather a range of its own. I picked the proper value so that it is not too high or low.

here is what datasheet says:

Quote
A feedforward capacitor (CFF) is not required for proper operation, but can further improve output noise.
However, care must be taken in choosing the CFF, since the power good (PG) function may not be valid with
a large CFF during start-up, and can cause spurious triggering of the PG pin during a large load transient. The
noise performance with various CFF is shown in Figure 6-31.

and since I practically tested it and it works, then no startup issue is there. Plus, there is no load transient since load is fairly constant. No cons mentioned in the app note (mentioned in datasheet after quoted paragraph) applicable here.

Uh, the load is anything but constant.  That's why regulators are needed. 

If you are happy with the value chosen, fine.  But the capacitor is a filter when combined with the resistors and its operation depends on the corner frequency of the combination even if they don't explain this in the data sheet.  They make some assumptions about the knowledge and experience of the designer. 

We always perform testing to find problems, but you need to understand that testing does not show the absence of problems, only that they have not been uncovered.  That's why there can be field recalls of nearly any equipment.  You can't test quality into a product.  It has to be designed in by understanding what is being designed.

Quote
_______


On the other side of the ocean, I added 2 small 1uH inductors on 12v line between the many 22uF capacitors to further enhance noise reduction just before the switchers.

Also, I sent the gerbers and bom to allpcb to manufacture 10 boards.

Good luck.
Rick C.
  - Get 1,000 miles of free Supercharging
  - Tesla referral code - https://ts.la/richard11209
 

Offline VEGETA

  • Super Contributor
  • ***
  • Posts: 1475
  • Country: jo
  • I am the cult of personality
    • Thundertronics
Re: Low ripple low noise small power supply design
« Reply #394 on: June 25, 2021, 05:02:58 pm »
Quote
Can't say, depends on the meter.

Mine is aneng 8009, it displayed open circuit for 5v rail and 2.3mF for 3.3v rail. I guess it is not reliable since the design worked perfectly well and for a long duration without problems. if this reading was correct then it wouldn't even start.

Quote
Uh, the load is anything but constant.  That's why regulators are needed.

I meant it is a retro gaming console with 22w rated PSU, it does nothing but one task. therefore its load is fairly constant. it doesn't swing or so..

Quote
If you are happy with the value chosen, fine.  But the capacitor is a filter when combined with the resistors and its operation depends on the corner frequency of the combination even if they don't explain this in the data sheet.  They make some assumptions about the knowledge and experience of the designer.

they gave a limited range you choose from but they relied on you not going to top range since it will affect stability, or you can ditch it all together and still work. I added a reasonable amount which shouldn't make any problem.

I suggest you go through the datasheet to see what it is all about. this IC uses new technology such as including ferrite bead inside the loop itself and so on.... I hope the result is as good as they claim.

Offline gnuarm

  • Frequent Contributor
  • **
  • Posts: 941
  • Country: aq
Re: Low ripple low noise small power supply design
« Reply #395 on: June 25, 2021, 05:23:27 pm »
Quote
If you are happy with the value chosen, fine.  But the capacitor is a filter when combined with the resistors and its operation depends on the corner frequency of the combination even if they don't explain this in the data sheet.  They make some assumptions about the knowledge and experience of the designer.

they gave a limited range you choose from but they relied on you not going to top range since it will affect stability, or you can ditch it all together and still work. I added a reasonable amount which shouldn't make any problem.

I suggest you go through the datasheet to see what it is all about. this IC uses new technology such as including ferrite bead inside the loop itself and so on.... I hope the result is as good as they claim.

If it were my design I'd be all over the data sheet.  But it's not, so I won't.  Far too many activities demanding my time.

BTW, adding the ferrite bead in the control loop is not "new technology".  It is just part of the frequency response of the control loop.  At very high frequencies the ferrite material becomes lossy and the characteristic changes from inductor to resistor.  At those frequencies the control loop has zero gain so the output circuit is just passive.  At the frequencies the control loop is effective the ferrite bead has so little inductance it is just a small addition to the larger inductor.  I'm not sure adding it to the control loop has much effect compared to adding it to the output after the control loop.
Rick C.
  - Get 1,000 miles of free Supercharging
  - Tesla referral code - https://ts.la/richard11209
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf