I apologize if this topic is covered already. I can't find a definite answer that applies to my project. I have read that either way power consumption is about equal but nothing about noise.
Also I apologize if this is posted in the wrong section. I'm assuming this is a "Power" topic.
Question: Should I power my +3.3V out regulator (HT7333A) from the +5V (AMS1117-5) or from the power source (USB or batteries)? Basically feed one regulator to the other, or feed the 3.3 regulator directly from either power source. See attached schematic.
I've got a project where I've spec'd 5V for the microcontroller and IC and 3.3V for the NRF24 radio. I've also got USB power for programming/bench testing and battery power for general operation. Take a look at the schematic if you will. I've employed several Schottky diodes to manage USB/battery power at the +5V level and at the feed to the HT7333-A. Not even sure if that will work.
My concern is noise. I watched the video from EEVblog #859 - Bypass Capacitor Tutorial and so I'm worried that from supply to NRF24 there is more likelihood of noise issues by misuse of bypass capacitors if I feed the 3.3 regulator from the output of the 5.0 rather than directly from the batteries or USB. I've also read that NRF24 are notoriously finicky and although a 10uF capacitor at NRF24 Vin is blind general practice to newbs like me, how can I be sure it will work in all cases?
The schematic is very good. Sorry to say you have major problems with the regulators. No high frequency bypass on either one, these are mandatory, it will show them in the data sheet, the 100nF are OK. Extra bypass capacitors do no harm but can help. Note all liner regulators have what is called head room, this is the minimum voltage needed to keep it in regulation. If there is ripple on the input it is measured from the minimum voltage point. I would consider using several different values because of the RF component on in your system Also the use of bulk capacitors on inputs to regulators, you have one on the 3.3 but not on the 5. You have no bypass capacitors on either of the LSI parts or radio, use one at each supply pin. The 3,3 regulator is only a 500mW device. The nRF24L01 has 125 channels and uses about ~14 mA per channel. Look at how many channels you are using to help with your power calculations. On the AVCC I would isolate with a resistor in the 1K range and bypass it just in case. I do not have the time to read the data sheet but I remember something about being able to connect one of the internal references to the Aref pin, the resistor should save the micro if this happens because of a software mistook. Do not forget you have made a great start and get credit for asking for help in the problem areas. This response is to help you get started in solving your problem, not solve it for you.
Good Luck & Have Fun!
Gil
bigbangus:
Question: Should I power my +3.3V out regulator (HT7333A) from the +5V (AMS1117-5) or from the power source (USB or batteries)? Basically feed one regulator to the other, or feed the 3.3 regulator directly from either power source. See attached schematic.
Since the HT7333A is very low input voltage (12V absolute max), you definitely want to protect it by powering it from the 5V rail. Its very low dropout so there's no issue there
I've got a project where I've spec'd 5V for the microcontroller and IC and 3.3V for the NRF24 radio. I've also got USB power for programming/bench testing and battery power for general operation. Take a look at the schematic if you will. I've employed several Schottky diodes to manage USB/battery power at the +5V level and at the feed to the HT7333-A. Not even sure if that will work.
Worth breadboarding it to make sure before committing to a PCB?
My concern is noise. I watched the video from EEVblog #859 - Bypass Capacitor Tutorial and so I'm worried that from supply to NRF24 there is more likelihood of noise issues by misuse of bypass capacitors if I feed the 3.3 regulator from the output of the 5.0 rather than directly from the batteries or USB.
Why? The output noise of a regulator will be in the datasheet if it matters (usually not, the decoupling caps do the work)
I've also read that NRF24 are notoriously finicky and although a 10uF capacitor at NRF24 Vin is blind general practice to newbs like me, how can I be sure it will work in all cases?
Nothing special about the NRF24 to my knowledge, give it regulated power with enough current, it works. If you're using a breakout with all the RF section on it its simple to use, the breakout should have its own decoupling, it won't hurt to add more. I think some people had issues using the built-in 3.3V Arduino supply which is very low current only. Read the datasheet, it will tell you how much current is needed. The one with the PA and LNA needs more power (and can work with 5V directly I believe)
[/quote]
gilshultz:
The schematic is very good. Sorry to say you have major problems with the regulators.
Hi Gil no just concern. Still in design phase.
gilshultz:
The schematic is very good. Sorry to say you have major problems with the regulators. No high frequency bypass on either one, these are mandatory, it will show them in the data sheet, the 100nF are OK. Extra bypass capacitors do no harm but can help.
I've learned that every IC should have a 0.1uF cap for bypass, but I don't see that in the datasheet. For the HT7333A the datasheet just shows an application with 10uF in and 10uF out. Should I ignore all datasheets and just add 0.1uF capacitors to Vin of every chip on my board?
MarkT:
Since the HT7333A is very low input voltage (12V absolute max), you definitely want to protect it by powering it from the 5V rail...Nothing special about the NRF24 to my knowledge, give it regulated power with enough current, it works.
OK you've convinced me to drop 3 of those diodes and just power the HT7333A from the +5V. Karma added.
MarkT:
If you're using a breakout with all the RF section on it its simple to use, the breakout should have its own decoupling, it won't hurt to add more.
1)What do you mean about a breakout with the RF section? Please forgive my newbness.
2)The HT7333A will be solely feeding the NRF24 is that what you mean?
MarkT:
The one with the PA and LNA needs more power (and can work with 5V directly I believe)
You know I can't find a concrete number. The Nordic datasheet says double digit mA consumption numbers, but some people claim 116mA and Gil from the first reply to this discussion says 14mA*125channel for a possible whopping 1.7A!
bigbangus:
Hi Gil no just concern. Still in design phase.
I've learned that every IC should have a 0.1uF cap for bypass, but I don't see that in the datasheet. For the HT7333A the datasheet just shows an application with 10uF in and 10uF out. Should I ignore all datasheets and just add 0.1uF capacitors to Vin of every chip on my board?
All digital chips always need high-speed decoupling, so include it even if the datasheet neglects the issue.
Linear regulators are not digital. Follow the datasheet - in general analog chip datasheets mention decoupling requirements as they vary more. This incluces ADCs and DACs.
bigbangus:
I've learned that every IC should have a 0.1uF cap for bypass, but I don't see that in the datasheet. For the HT7333A the datasheet just shows an application with 10uF in and 10uF out. Should I ignore all datasheets and just add 0.1uF capacitors to Vin of every chip on my board?
It will never hurt to add a 0.1 µF ceramic cap, but it's probably not necessary.
Mind the capacitor type if mentioned in the data sheet. Many regulators are not stable with ceramic caps; some require tantalum; others will suggest electrolytic. In your schematic there's a 1µF unpolarised on the output of your AMS1117 - that has to be a polarised one (tantalum or electrolytic), can't use ceramic, and it indeed needs at least 10µF on the output and I think also on the input. Have to look that one up.
A single 100 nF is normally enough on the Vcc pin of your ATmega. The cap C1 on the AREF pin won't be doing anything unless you use the internal reference.
You power teh NRF24L01 at 3.3V but there are no level shifters on the SPI lines. Are those inputs 5V tolerant?
Do you really need that FT323 on board? I'd just add an FTDI header and connect FTDI adapter when programming/running off USB power.
I didn't try to analyse your power supply parts, that's too scattered around your schematic.
bigbangus:
1)What do you mean about a breakout with the RF section? Please forgive my newbness.
2)The HT7333A will be solely feeding the NRF24 is that what you mean?
A breakout board with the nRF24L01+ chip and the associated RF matching network to couple it to the antenna. Rather than the nRF24L01+ chip by itself (which would be much trickier - I just wanted to be sure)
MarkT:
A breakout board with the nRF24L01+ chip and the associated RF matching network to couple it to the antenna. Rather than the nRF24L01+ chip by itself (which would be much trickier - I just wanted to be sure)
Ah thanks for explaining this to me. Yeah I'm using a breakout board from Amazon.... yeah definitely not up to the task of integrating that into my PCB!
wvmarle:
Many regulators are not stable with ceramic caps; some require tantalum; others will suggest electrolytic. In your schematic there's a 1µF unpolarised on the output of your AMS1117 - that has to be a polarised one (tantalum or electrolytic), can't use ceramic, and it indeed needs at least 10µF on the output and I think also on the input. Have to look that one up.
This is great information but also confusing as to why the Nano Schematic 3.2V doesn't conform to what you're saying.
wvmarle:
A single 100 nF is normally enough on the Vcc pin of your ATmega. The cap C1 on the AREF pin won't be doing anything unless you use the internal reference.
Good to know. Again... Nano Schematic 3.2V doesn't comply?
wvmarle:
You power teh NRF24L01 at 3.3V but there are no level shifters on the SPI lines. Are those inputs 5V tolerant?
Yes they are 5V tolerant and my prototype seems happy.
wvmarle:
Do you really need that FT323 on board? I'd just add an FTDI header and connect FTDI adapter when programming/running off USB power.
Yes. My enclosure has 8 screws and the original PCB already had a hole for one so I'm planning on leveraging that. Google DumboRC-X6
wvmarle:
I didn't try to analyse your power supply parts, that's too scattered around your schematic.
Thanks for your time anyway with everything else. I'll try to make better schematics in the future, !@#$
That schematic you link to does not use the AMS1117, instead it uses the LM1117. Different regulator - check its data sheet for capacitor requirements.
For the ATmega328p 100 nF is normally enough; more won't make a difference. I don't know why they decided on 2x 1uF and 1x4.7uF (though it does have two Vcc pins, so a second ceramic decoupling cap is needed: one for each Vcc pin; the 4.7uF is bulk capacitance, electrolytics including tantalum have too high ESR for decoupling of digital circuits which is why ceramic is used there).
wvmarle:
That schematic you link to does not use the AMS1117, instead it uses the LM1117. Different regulator - check its data sheet for capacitor requirements.
Ah totally missed that! Yeah I'll use the LM1117-5 with 10uF in and 100uF out as per the datasheet! Thanks!
wvmarle:
For the ATmega328p 100 nF is normally enough; more won't make a difference. I don't know why they decided on 2x 1uF and 1x4.7uF (though it does have two Vcc pins, so a second ceramic decoupling cap is needed: one for each Vcc pin; the 4.7uF is bulk capacitance, electrolytics including tantalum have too high ESR for decoupling of digital circuits which is why ceramic is used there).
Gotcha that makes sense. Thanks now I have a lot more confidence in what I'm using.