Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Recommended boost regulator/dc dc converter for a 9v battery? (continuous conduction)

Status
Not open for further replies.

fuji

New Member
I find it difficult to pick the right boost regulator or DC DC converter that is continuous conduction for a 9v battery. Lowering the output current (5mA to 10mA) is important for me to save battery life. Anyone recommend a boost regulator/dc dc converter for a 9v battery where I can set the current output as low as possible and is accurate. Some boost regulators are accurate on high or low frequencies and ripples, just can't decide which regulator is better than the other.

Secondly, I need this regulator to be accurate enough at perhaps around 3%, 2% reference on the voltage output.

I have a voltage converter which can convert the 9v battery to a higher voltage, but its not BOOST to keep the voltage and current constant (continuous conduction) on the output before the battery dies out.
 
What voltage? Load current and leakage drain?
 
This seems to fit your description: **broken link removed**

I have not used one.
 
it is unclear what you want. Do you want a boost converter that is capable of outputting 5 - 10mA whether the load demands it or not, or do you want something like an LED driver where the output current is regulated and not the output voltage? The first part regulates output voltage, regardless of the current (load). The second regulates the output current, regardless of how much voltage is needed to drive that current.

Please clarify.

The output ripple is dependent on the load current, the amount of output capacitance, the capacitance ESR and the switching frequency. It is largely independent of the chip you use and is determined by the external components. The absolute output voltage accuracy is determined by the accuracy of the internal reference and the feedback resistors

If you want to know more about boost converters, I have written a tutorial on them. See the link below
 
it is unclear what you want. Do you want a boost converter that is capable of outputting 5 - 10mA whether the load demands it or not, or do you want something like an LED driver where the output current is regulated and not the output voltage? The first part regulates output voltage, regardless of the current (load). The second regulates the output current, regardless of how much voltage is needed to drive that current.

Please clarify.

The output ripple is dependent on the load current, the amount of output capacitance, the capacitance ESR and the switching frequency. It is largely independent of the chip you use and is determined by the external components. The absolute output voltage accuracy is determined by the accuracy of the internal reference and the feedback resistors

If you want to know more about boost converters, I have written a tutorial on them. See the link below

Awesome I'll definitely take a look at the link. I am still new to learning buck/boost converters.

For my circuitry, I need to boost (step-up) voltage from 9v to 12v and regulate the current (10mA) to be demanded by the load. I'm using the 1n5817 to reduce the voltage loss on the output as well. If there is no load connected, I still want the output current to be at 10mA despite the disconnected load. So you are correct, I need the output at 10mA whether the load demands it or not, but is on continuous conduction mode. So basically I need to regulate both voltage and current...if this is what a boost/step-up converter is suppose to do in general in continuous mode.

I forgot to mention that my first old dc/dc converter was the mc34063. Problem is, it doesn't work
well on low output currents and the accuracy gets worse. The IC has a maximum of 100Khz frequency. It works as a step-up discontinuous conduction mode, but not step-up continuous conduction mode. Unless there is a way to step-up the mc34063 on continuous conduction mode? The datasheet has no information on stepping up the converter as continuous conduction mode. Is there a tutorial for the mc34063 to be on step-up continuous conduction mode?

I'm sure you'll agree as well with everyone else that this cheap dc/dc converter is not that accurate. I agree as well, but its pretty darn cheap.
 
Last edited:
All that any power source can do is to limit it's output current. They cannot force output current to flow through a disconnected load.

If you need a minimum load current on the output, you will need to provide an alternate path for it to follow. The simplest way would be to put a 10mA current sink across the output. But that means that, when the load is connected, your output current will be the total of the load and the sink, or 20mA.

If you need the load current to always be 10mA, then you need additional circuitry to measure the load current and adjust the current sink to make up the difference.

But, do you really need a 10mA minimum load? Or is that just what your existing supply needs to prevent the output voltage from rising when unloaded? If so, then fixing the source of the problem would be a better choice. Few PWM controllers can maintain a constant voltage with absolutely no load, but many can do so with the small load that the feedback voltage divider resistors provide. It's basically how close to zero% duty cycle that the converter can run.

As for accuracy, what do you really need? If you use a trim pot in your feedback network, you can trim the output to whatever accuracy your voltmeter can guarantee. (Note) But, do you just need accuracy at room temperature? Or do you need accuracy across a wide operating temperature? And for how long? Even the best voltage reverences drift over time.

The stated initial voltage accuracy of the MC34063 is 2%. But the accuracy of a complete circuit is also dependent on other factors as well.


(Note) Don't expect to get an accurate measurement from a cheap voltmeter. And don't confuse resolution for accuracy.
 
All that any power source can do is to limit it's output current. They cannot force output current to flow through a disconnected load.

If you need a minimum load current on the output, you will need to provide an alternate path for it to follow. The simplest way would be to put a 10mA current sink across the output. But that means that, when the load is connected, your output current will be the total of the load and the sink, or 20mA.

If you need the load current to always be 10mA, then you need additional circuitry to measure the load current and adjust the current sink to make up the difference.

But, do you really need a 10mA minimum load? Or is that just what your existing supply needs to prevent the output voltage from rising when unloaded? If so, then fixing the source of the problem would be a better choice. Few PWM controllers can maintain a constant voltage with absolutely no load, but many can do so with the small load that the feedback voltage divider resistors provide. It's basically how close to zero% duty cycle that the converter can run.

As for accuracy, what do you really need? If you use a trim pot in your feedback network, you can trim the output to whatever accuracy your voltmeter can guarantee. (Note) But, do you just need accuracy at room temperature? Or do you need accuracy across a wide operating temperature? And for how long? Even the best voltage reverences drift over time.

The stated initial voltage accuracy of the MC34063 is 2%. But the accuracy of a complete circuit is also dependent on other factors as well.


(Note) Don't expect to get an accurate measurement from a cheap voltmeter. And don't confuse resolution for accuracy.

Thanks for the reply. This is good info to learn from.

If you need a minimum load current on the output, you will need to provide an alternate path for it to follow. The simplest way would be to put a 10mA current sink across the output. But that means that, when the load is connected, your output current will be the total of the load and the sink, or 20mA.

So basically both currents from the sink and the load will double. Then there is an extra circuit to bring back down the current from 20mA to 10mA again.

But, do you really need a 10mA minimum load? Or is that just what your existing supply needs to prevent the output voltage from rising when unloaded? If so, then fixing the source of the problem would be a better choice. Few PWM controllers can maintain a constant voltage with absolutely no load, but many can do so with the small load that the feedback voltage divider resistors provide. It's basically how close to zero% duty cycle that the converter can run.

My idea was that I only need enough current to power the rest of the components on breadboard, so I chose 10mA. They are all low voltage and low current run. I noticed through experience that when connecting a load to a boost converter, their is a slight voltage drop (possibility the ESR of the Inductor). So what your saying is, a PWM along with the load can prevent this voltage drop from occurring before entering the load?

Can a PWM like the TL494CDR be used between a boost converter and the load for regulation?

OR would this be better:

SC2604
Simple PWM Boost Controller
with Input Disconnect FET Drive

As for accuracy, what do you really need? If you use a trim pot in your feedback network, you can trim the output to whatever accuracy your voltmeter can guarantee. (Note) But, do you just need accuracy at room temperature? Or do you need accuracy across a wide operating temperature? And for how long? Even the best voltage reverences drift over time.

Accuracy across a wide operating temperature instead of room temperature. Time duration would be at least 30 minutes, if not, 20 minutes.

If you need the load current to always be 10mA, then you need additional circuitry to measure the load current and adjust the current sink to make up the difference.

Thats right. The load current will probably still be double that of 10mA. I'm only dropping lets say 20mA down to 10mA but I'm still using 20mA at the load current.
 
Last edited:
Your 9v battery is not going to last very long.

Worked pretty well on buck mode with the mc34063. After keeping the power on for two days with no load, it was still above 8v. Problem is, its not continuous conduction on boost mode, where I can almost juice out the entire battery instead of having to use buck where I got 5 volts left in the battery and became unusable.
 
Worked pretty well on buck mode with the mc34063. After keeping the power on for two days with no load, it was still above 8v. Problem is, its not continuous conduction on boost mode, where I can almost juice out the entire battery instead of having to use buck where I got 5 volts left in the battery and became unusable.

I'm not sure I understand quite what you're saying here.

In a switch mode power supply, continuous conduction mode means that the current in the inductor does not fall to zero during the switch off time before the switch turns on again. In discontinuous mode the inductor current does fall to zero during each switch cycle. Neither is inherently good or bad, just different. Many converter circuits operate in both modes when their operating conditions vary.

Why do you feel that discontinuous mode is bad in your application?

Also, PWM (Pulse Width Modulation) is how nearly all switch mode converters, including the MC34063, work. The make the pulse width wider or narrower in order to regulate the output.
 
Sounds like fuji wants the SMPS to boost from fairly weak batteries. perhaps he should look at a joule thief arrangement.
Anyway, I did a lot of research on the mc33063 and NCP3063 over the last few days and the NCP3063 with a feed forward loop seems to provide minimal ripple and no audio level frequencies under low loads.


The TL494 series has its uses (not only SMPS) as it very configurable via twin available comparators and a separate DTC over ride along with push pull outputs. For certain applications it can work well but it's top end frequency is not very high for reducing component sizes. From a SMPS cost and space perspective the feed forward NCP3063 is a better solution.
 
I'm not sure I understand quite what you're saying here.

In a switch mode power supply, continuous conduction mode means that the current in the inductor does not fall to zero during the switch off time before the switch turns on again. In discontinuous mode the inductor current does fall to zero during each switch cycle. Neither is inherently good or bad, just different. Many converter circuits operate in both modes when their operating conditions vary.

Why do you feel that discontinuous mode is bad in your application?

Also, PWM (Pulse Width Modulation) is how nearly all switch mode converters, including the MC34063, work. The make the pulse width wider or narrower in order to regulate the output.

Thanks for the reply.

I learned that boost mode in discontinuous mode, when the battery loses power and drops in voltage, the voltage would stay the same on the output, but the current drops gradually as the voltage drops from the battery.
 
Last edited:
Sounds like fuji wants the SMPS to boost from fairly weak batteries. perhaps he should look at a joule thief arrangement.
Anyway, I did a lot of research on the mc33063 and NCP3063 over the last few days and the NCP3063 with a feed forward loop seems to provide minimal ripple and no audio level frequencies under low loads.


The TL494 series has its uses (not only SMPS) as it very configurable via twin available comparators and a separate DTC over ride along with push pull outputs. For certain applications it can work well but it's top end frequency is not very high for reducing component sizes. From a SMPS cost and space perspective the feed forward NCP3063 is a better solution.

Thanks for the reply I'll definitely take a look at the NCP3063 . The frequency has an extra 50khz as compared to the mc34063, which is 100khz. I like that the precision reference dropped to 1.5% as it is 2% for the mc34063. Still worth it.
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top