Charge pump 5V to 15V (voltage tripler)

anishkgt

Joined Mar 21, 2017
549
Hello All,

Searching for a charge pump to boost a 5V to 15V. The source to the charge pump comes from a Buck converter with wide input range from 4.5V to 30V. The 15V powers a MOSFET driver driving 6 N-FETs. I have not calculated the current required.

The MAX1683 has a app note showing it can be used as a voltage tripler by adding two schottky diodes in series but the Iout is about 30mA. I wish to know if i can use the same configuration with another charge pump to get a higher Iout or is there a better way to get 15V from the 5V rail. Off the shelf ones have 5V to12V but are expensive.

anishkgt

Joined Mar 21, 2017
549
The application is for a spot-welder so switching frequency is about once every 1 or 2seconds with Ton between 10ms to 100ms.

ericgibbs

Joined Jan 29, 2010
19,106
hi anish.
What current will the 15Vdc have to supply to load.?
E

anishkgt

Joined Mar 21, 2017
549
Just realised i had to figure that out first before i came this path. Since the switch-on time is quite low i am guessing couple of mA but with six mosfets it would take a couple more, again unsure.

What is the max current that these charge pumps can deliver ?

ericgibbs

Joined Jan 29, 2010
19,106
hi a,
If you are considering the maximum possible achievable output current design, that will cost more.
You said you are on a budget, so use a15Vdc PSU and measure the actual required current and design to give a 20% overhead.

Lets know what you measure and we can take it from there/
E

anishkgt

Joined Mar 21, 2017
549
A quick calculation
Qg = 145nC (from the mosfet datasheet)
fsw = 1Hz (considering 1sec interval per/pulse)
Converting 145nC to coulomb = 0.000000145
Amps = 0.000000145/1Hz, which is 0.145uA
Would this be correct to estimate the approximate current required to charge the gate for a second? A second would be too much for the application. So it would be less 0.145uA.

ericgibbs

Joined Jan 29, 2010
19,106
hi a,
Check the d/s for the MCP1415 which you have posted as the driver, Page #6, shows the Iq.
Your figure of 145uA Gate current does not take into consideration the operating current of the driver.

Post the design circuit for the project.
E

Attachments

• 39.4 KB Views: 5
• 840.2 KB Views: 4
• 7.9 KB Views: 4

anishkgt

Joined Mar 21, 2017
549
i maybe stupid to see beyond what is required here but the .6mA to 1.1mA is the supply current to power the driver. The MAX1683 can supply up to 45mA by default and 30mA in a voltage tripler configuration. Would that not be enough ? correct me if i am worng.

ericgibbs

Joined Jan 29, 2010
19,106
hi,
Post the design circuit for the project, so that we can discuss the project.

E

BobTPH

Joined Jun 5, 2013
9,272
A quick calculation
Qg = 145nC (from the mosfet datasheet)
fsw = 1Hz (considering 1sec interval per/pulse)
Converting 145nC to coulomb = 0.000000145
Amps = 0.000000145/1Hz, which is 0.145uA
Would this be correct to estimate the approximate current required to charge the gate for a second? A second would be too much for the application. So it would be less 0.145uA.
You have calculated the current to switch the MOSFET on over a period of 1 second. That may give you the average current needed, but that is certainly not what you want to do. You want to supply a high current pulse to turn it on quickly.

The good news is that a capacitor can store up this energy over a second the release it quickly when needed. In any case, 30mA should be plenty if switching once per second.

Bob

anishkgt

Joined Mar 21, 2017
549
Is the rise time and fall time of the MOSFET part of determining the max current ? How is it determined ?