Stealing charge pump power from RS232 driver IC

[Jester]

I can eliminate several components from a circuit if I can "steal" a couple of mA from the charge pump circuits of an existing ST3232BDR (RS232 driver)

Does anyone have real world experience pulling an extra 1-2 ma from the charge pump output pins to power an IC external to the actual ST3232?

https://www.st.com/web/en/resource/technical/document/datasheet/CD00002253.pdf

 

[esp1]

hi,
I have connected the +/-9v from a MAX232 RS232 driver to power OPA's with no problems.
Low current drain of 2mA to 3mA should not cause a problem.
E

 

[schmitt trigger]

My advice to anyone who does something different, is to validate your assumptions by "pushing the envelope" a little.

For instance, let's assume you measure or calculate a maximum load of 2 mA. Then test your circuit with an additional 1 mA load (can be a resistor) to make it 3 mA.

If everything continues to work fine, then you can safely assume that you have sufficient margin to account for component tolerances, temperature effects, etc.

 

[D.A.(Tony)Stewart]

Without getting into too much detail , understand that your Tx is bipolar signals but this chip uses an Rx threshold of +1.6 +/- 0.8 with hysteresis , approximately and can only supply 2mA if the RS-232 load is zero.

The source and load impedances of RS-232 are designed often with 1k source and 10k loads. So 2mA load will cause transient dips in voltage that are caused by data patterns affecting both reliability of charge pump voltage and RS232 data.

it might work for one chip without external noise, but worst case? Who knows with your input signal patter, toggle rate and loading.

 

[Jester]

Without getting into too much detail , understand that your Tx is bipolar signals but this chip uses an Rx threshold of +1.6 +/- 0.8 with hysteresis , approximately and can only supply 2mA if the RS-232 load is zero.

The source and load impedances of RS-232 are designed often with 1k source and 10k loads. So 2mA load will cause transient dips in voltage that are caused by data patterns affecting both reliability of charge pump voltage and RS232 data.

it might work for one chip without external noise, but worst case? Who knows with your input signal patter, toggle rate and loading.

In this application, I'm not using the RS232 driver for RS232. I'm using it instead:
1) to drive an ultrasonic transducer in a 3.7V system
2) to power the mux and op amp

The power supply aspect seems to be working fine, breadboard reality test substituting a LT1014 for U3 not so great.

 

[KlausST]

Hi,

just out of curiosity:
The vertical R = 100k, is it for impedance matching?
I wonder if it makes signal better or worse compared to a 0R ?

Klaus

 

[Jester]

Hi,

just out of curiosity:
The vertical R = 100k, is it for impedance matching?
I wonder if it makes signal better or worse compared to a 0R ?

Klaus

I will test and see, the idea was that the Mux internal resistance (closely matched at about 100Ω) would make this a differential configuration. I don't have a MCP6021 to test with today, but I did find a TL084 (GBW=4MHz) so I replaced the LT1014 (GBW=1MHz), and the response is somewhat improved.

 

[KlausST]

Hi,

I see,

my thought was, it is to be symmetric to the introduced charge by the analog MUX...

For lower frequencies i think it doesn´t matter, but it will have influence when the MUX switches...postivie or negative, i don´t know...

Klaus

 

[Jester]

Hi,

I see,

my thought was, it is to be symmetric to the introduced charge by the analog MUX...

For lower frequencies i think it doesn´t matter, but it will have influence when the MUX switches...postivie or negative, i don´t know...

Klaus

Correct, however the timing of the transmit burst as well as mux switching can be controlled tightly by the uC, so I think that aspect should be okay. I'm going to start a new post regarding the receive signal, the power supply aspect looks to be okay.

Thanks for your thoughts (much appreciated).