ACF2101 Switched Integrator, HOLD pin not functioning in TINA simulation

T.M.,

Thanks for using the texas Instruments AFC2101.

I would like to take a look at your problem with the AFC2101 over the week-end and I will get back to you no later than Monday afternoon.

Thank you for the quick reply.

I did investigate whether my method of driving the RESET port could be the problem due to it injecting currents into the subcircuit however was unable to find anything. Hope you will be able to provide some insight into the behaviour.

Looking forward to your findings.

T.J.

Thank you for giving me the week-end to look into this. I spent some time looking at the code and your output. Where is the switch that creates Vswin? Or should I say, were does the signal Vswin come from?

B.B.

The signal for the voltage sensor Vswin (Voltage SWitched Input) is generated by the piecewise linear current source Iin across the Com and SwitchIn port of the ACF2101 sub-circuit. This sensor is a leftover from my investigation of the circuit.

To try and figure out why the HOLD was not working I examining the ACF2101 netlist I found that the switch associated with the HOLD port connects or disconnects the SwitchIn and In+ ports. However the circuit also defines a second switch across these ports which is dependant on the voltage between SwitchIn and GND. If this switch would have been closed (low resistance) during a HOLD condition it would explain a lot. Vswin is used to ascertain the state of the second switch during simulations.

When I run the simulation I find the voltage between SwitchIn and GND to be around 475mV, which means that the switch is in transition from the 1000G to 10 ohm as defined by the model (see below). So the switch is not closed during the simulation however resistance should still be quite high if I'm not mistaking.

Hope this answers your question. If not don't hesitate to ask, I would very much like to figure out what is wrong with the circuit.

-------------------------------------------------------------

Relevant code from the sub-circuit describing the switches in question. The code is found on lines 117, 118, 128 and 130.

SH1 31 2 34 30 SWH                 <-----                  Switch between SwitchIn and In+ controlled by HOLD and GND
SH2 31 2 31 30 SWT                  <-----                  Switch between SwitchIn and In+ controlled by SwitchIn and GND

.MODEL SWH VSWITCH(RON=1000G ROFF=1.5K VON=2 VOFF=0.8)
.MODEL SWT VSWITCH(RON=1000G ROFF=10 VON=0.45 VOFF=0.55)

T.M.,

There is nothing wrong with the switches in the ACF2101 macro model. I think that your input source is causing the problem. If you place a voltage probe on SwIn you will discover that, when in hold, the voltage will go to approximately 10 volts. This is well above the positive supply voltage and the ACF2010 model is not operating correctly. You are overdriving the switch. You can reduce the impact of this problem by placing parrallel diodes across the current source. When that is done you will see that the voltage across the current source is limited to +/- 6 mV. You will also see the hold function working properly.  

In terms of the actual device, if you place a photodiode on the front end this phenomena will not happen. The device will work properly.

Hi B.B,

I'm not seeing the phenomenon you describe so I would like to make sure I understand correctly and that we are using the same circuit and model for the ACF2101. When I run a transient simulation of the schematic I provided using Tina 9.3.50.40 I find the voltage between SwitchIn and GND (using the volt meter named Vswin or a probe on the output pin of the current source) to be 475.57mV max, not the 10V you find.

Could you perhaps provide me with the circuit you are using so I can compare the two and figure out why we are measuring different values.

To further investigate I made sure the initialization demand for the model (HOLD and RESET should be equal to com) is still valid (it is). I have also investigated if the switch times where to long (changed from 5V/1ms to 5V/1us) and changed some of the switch times to make sure the HOLD, RESET and SELECT switches do not change at the same time. None of the changes has any effect on the circuit behaviour.

P.S. I have updated the schematic I use and added the ACF2101 sub-circuit to the attached zip file to make sure we are using the same code.

Hi B.B.

Thank you for sharing the file. With it I've been able to figure out that the problem was and why we where getting different results. As you stated earlier the problem was in the method I use to excite the circuit. I'll quickly describe my findings in below so others may be able to use the insight it provided in the future.

First of the reason you where measuring 10V on Switchin while I got around 500mV.  The reason was that switch SH2 in the ACF2102 spice model was commented out. I suspect that was due to my investigation and I had forgotten to uncomment the line when I posted the file. After uncommenting the line both simulations gave the exact same results.

As Bonnie stated the problem was in the excitation source I'm using i.e. the current source Iin. Without the two diodes it will provide 10nA to the circuit no matter the resistance of the switches in the spice model. So even into 1000G ohm the source would push 10nA through the "open" switch, an obvious error in hindsight. The addition of the diodes eliminates this.

Thank you for you help.