TPS60400: charge pump doubler voltage issue

Hi Viola,

Could you share your schematics?

Best regards,

Varun

My setup is refernce from EVM board like this.

I test on EVM board.

BR,

Viola

Hi Viola,

Did you have this problem, only after replacing the diodes on the EVM with the 1N5819? Or did you see this problem with the default assembled diodes on the EVM?

The part should be able to handle 30mA max output current on the VPOS. However the wrong VPOS at 11.4V instead of around 10V will reduce the max output current.

Best regards,

Varun

 around 10V will reduce the max output current.=>how to calculate the maximum current?
and did you know why the VPOS is 11.4V not 10V?

Hi Viola,

I'm not sure of what the exact formula will be. However from the information in the datasheet it seems like the max current should be =(0.03×10)/11.4 =~26mA

I think it would make sense to first figure out the wrong regulation of 11.4V instead of 10V. 

• Could you help measure the differential voltage individually across caps C21 and C22? Both these caps in series is what contributes to the double voltage.
• Can you check your input voltage is stable at 5V? There are no transient spikes? If there is a higher VIN then this can also increase the output
• Can you check if the output regulates back to 10V if you add a load like 2m to 5mA? It could be the output overshot the target 10V due to some transient input spike and is unable to discharge back to 10V as the diodes are reverse biased.
• Do you see the problem with the default diodes on the EVM? Or was this regulating correctly?

Best regards,

Varun

Hi Varun,

below is my reply.

• no load => differential voltage of C21=5.003V, C22=5.76V, Vout= 12.09V
• load: 35mA => differential voltage of C21=4.997, C22=4.505V, Vout= 8.86V
• yes, I check my input voltage is 5V, which is from power supply.
• I add a load around 4.75mA and Vout=9.44V
• EVM board didn't have the default diodes, it should solder componenet by customer.

Best regards,

Viola

Hi Viola,

Thanks for the info.

C22 charges to VIN when the C1+ pin of the IC connects to GND and diode D20 gets forward biased. The fact that C22 is higher than 5V, indicates to me, there was a transient overshoot of VIN. This could have happened during startup itself, as the chances of overshoot are higher while turning on the power supply. If C22 gets higher than 5V, the diode D20 gets reverse biased and there is no path to discharge C22 in the absence of a load current. You can see that the output reduces to 9.4V when you have some load current and have allowed C22 and C24 to discharge.

Could you take a scope shot of the startup when VIN is turned on from 0 to 5V? Please probe VIN, VPOS and the voltages across C22 and C21.

Best regards,

Varun

Hi Varun,

So for this charge pump application, is it means I can not get a stable voltage when I have a variable load?

Or how could I get the stable output voltage, when I have the load?

Best regards,

Viola

Hi Viola,

The charge pump has a certain output resistance. For 5V VIN, this is around 12Ohms. So the output voltage will reduce slightly at high load currents.

I think the voltage drop you see is combination of this drop across the charge pump output resistance + the forward voltage drop across the diode D20. The forward voltage drop across the D20 seems to be around 500mV, as the voltage across C22 is around 4.505V with the 4.75mA load.

The initial overshoot you see, I believe comes from an overshoot in your input voltage during turn on.

Best regards,

Varun