Description of the: RESET, SETUP, and Special Topics in the Quantum Design 2000/2010 SQUID Controller ------------------------------------------------------- Rev. 12-Oct-2020 RESET Function: --------------- In this SQUID Controller a RESET can be triggered by either: - the front panel RESET push button, - a RESET input signal to the ACU box back panel 15 pin D-Sub connector that is optically coupled before being used in the rest of the controller's electronics, - or triggered by the RESET Discriminator in the ACU box going over threshold. Note I do not think that we ever use this Discriminator triggered form of RESET in the Quick Dipper system. The only way to block this source of RESETs is to set the Discriminator Threshold control on the back panel of the ACU box to its maximum level. See page 26 of the QD Operating Manual. Note that during normal RUN Mode operation this Discriminator is watching the output of the Integrator in the SQUID Amplifier. The RESET function in the SQUID Controller is very simple - all that it does is to discharge the capacitor in the Integrator circuit of the SQUID Amplifier's Feedback Loop. The RESET is accomplished by just shorting out this capacitor with an analog switch while at the same time the input to the Integrator is disconnected by another analog switch. Once the RESET is triggered the short across the Integrator's capacitor is maintained for at least 100 usec which is long enough to completely discharge this capacitor through the one hundred Ohm or so resistance of the analog switch. While the RESET function in the controller itself is simple to understand, how this is used in an overall measurement system is more complicated. As the SQUID controller comes out of RESET the feedback servo loop will lock onto the closest minimum in the periodic flux through SQUID vs RF Voltage across the RF Coil transfer function. Thus unless it just happens that the magnetic environment of the SQUID puts it at a minimum on this periodic transfer function - the servo loop will ramp to provide a small amount of feedback current (typically called a "pull in" current of +- 500 nAmp or less for our SQUIDs) to force the flux through the SQUID to the closest minimum on the transfer function. That is to say - immediately after a RESET the output of the Integrator will be small but it is not likely to be zero and its value is not predictable. Thus the rational way to perform a RESET in the context of an overall measurement system is to: - Sweep the User's current to zero - Assert RESET for a minimum of 100 usec - Allow time for the feedback servo loop to stabilize - Read the output of the Integrator with the DVM, this is the "tare weight" for this measurement - The output voltage from the Integrator can be read from the "OUTPUT" connector on the front of the ACU - Turn back On the User's current and allow it time to sweep up to the desired value and stabilize - Allow time for the feedback servo loop to stabilize - Read the output from the Integrator again - Your measurement is the difference between these two readings of the Integrator output. Note that there is a special Mode of operation called Auto-Zero. Operating in the Auto-Zero Mode is selected by turning the MODE switch on the ACU to its RUN/AZ position. The Auto-Zero Mode effects only how the SQUID ACU box operates. The SQUID Amplifier only knows about SETUP and RUN Modes. In the Auto-Zero Mode the ACU box can be told to start a search for just the right amount of extra DC current to send through the RF Coil so that when the Integrator comes out of RESET the flux through the SQUID will be at a minimum on the periodic transfer function. When told to do this, ACU box finds "just right amount of extra DC current" by sweeping around the input to a 12 bit Digital-to-Analog converter until the output of the Integrator stays right at zero at the end of a RESET. - Yes this is yet another current flowing through the RF Coil. The currents in the RF Coil are: the actual RF current, the flux dither modulation current, the Feedback current, and now the Auto-Zero current, and something I called the SETUP flux sweep current (see below) - The voltage that makes this Auto-Zero current is carried from the ACU box to the SQUID Amplifier on the "External Input" wires in the cable between these two boxes. - The only way to tell the ACU box to search around and find the correct value of Auto-Zero current is to set the MODE Control to RUN/AZ and then press the RESET push button on the front panel of the ACU box. The other 2 ways of triggering a RESET listed above do not cause a re-calculation of the correct Auto-Zero current. - If the MODE Control is set to the normal RUN position then no Auto-Zero current is sent to the RF Coil. I do not think that we use Auto-Zero Mode in the Quick-Dipper system. Note that in general the correct Auto-Zero current needs to be re-calculated any time the magnetic environment of the SQUID sensor changes. Without Auto-Zero the maximum servo loop "pull in" current to get to the closest minimum on the periodic transfer function is about +- 500 nAmp. With a 12 bit Auto-Zero DAC, in theory, this may be reduced to about +- 250 pAmp which I think is about the noise limit of this SQUID system. SETUP Function: --------------- Use of the SETUP Mode to adjust the ACU box TUNE and DRIVE controls and to adjust the SQUID Amp GAIN and MODULATION controls is described on pages 15 and 29 of the QD Operating Manual. Selecting SETUP Mode uses two additional blocks in the SQUID Amplifier that I did not include in its block diagram Figure 17 on page 13 of these notes. The SETUP Mode does the following: - It turns on a SETUP Flux Sweep Current Generator in the SQUID Amp. This is a linear saw-tooth generator running at about 500 Hz that sends a saw-tooth current to the RF Coil that has enough amplitude to sweep back and forth through about 10 flux quanta. - It turns off the feedback loop in the SQUID Amp. - It changes the "Output Signal" from the SQUID Amp from being the output of the Integrator to being the output of an Audio Frequency Detector circuit. The input to this Audio Frequency Detector is the output of the RF Envelope Detector. - It forces the Gain in the ACU box to 1 so that the LED Bar Graph display on the ACU box will have a scale of 0 to +10 Volts and will show the output of the Audio Frequency Detector in the SQUID Amp. Recall that the output of the RF Envelope Detector changes so that it tracks the peak amplitude of the RF waveform. If the amplitude of the RF envelope is increasing and then decreasing at 500 Hz or 100 kHz then the output of the RF Envelope Detector will track these changes. The Audio Frequency Detector output is a DC or slowly changing voltage that tracks the amplitude of the envelope of any audio frequency AC signal coming into this detector. By a slowly changing DC output voltage I mean for example changing at the rate that one might turn the dial on a ten-turn pot. Using the SETUP Mode to adjust the controller to work with a given SQUID sensor is described in the QD Operating Manual. From the SETUP Mode electronics layout I assume: - The 500 Hz saw-tooth modulation current of about 10 uAmp (10 flux quanta) flowing in the RF Coil will NOT modulate the Amplitude of the RF voltage across the RF Coil unless there is something non-linear in the 1 turn SQUID Coil, i.e. unless the SIS junction is doing something. If everything in the 1 turn SQUID Coil is Ohmic or superconducting then there will be no 500 Hz audio frequency modulation of the RF Amplitude across the RF Coil, i.e. the output of the Audio Frequency Detector will be zero and that's what will be shown on the ACU box LED Bar Graph display. - For the SQUID to be useful the SIS junction needs to be doing something non-linear and the SETUP process is how one adjusts for the best non-linearity in the 1 turn SQUID Coil. - You starts with the DRIVE level set low to make certain that you can adjust the TUNE, i.e. the frequency of the RF Generator, to match the LC resonant frequency of the SQUID transformer primary. If the DRIVE is initially set too high then you will not be able to see the effect of your TUNE control adjustments. If you live next door to a 50 kWatt radio station transmitter then it does not matter what frequency you tune your radio receiver to. - Once the TUNE (frequency) of the RF Generator is set correctly then you adjust the DRIVE control to get the optimum non-linearity from the SIS junction. I assume this is the optimum mix of superconducting vs voltage across the SIS junction. The DRIVE control lets you adjust the level of the RF current in the 1 turn SQUID Coil to the value that works best with the critical current of your SIS junction. - Next you adjust the amplitude of the Flux Dither Modulation to get the biggest signal out of the Audio Frequency Detector. This will probably be with about +- 1/4 flux quanta of square wave modulation at 100 kHz. - Finally you adjust the GAIN of the Phase Sensitive Amplifier so that it and the other analog circuits on the SQUID AMP are operating in the middle of their dynamic range. This is my current understanding of what SETUP Modes does and how to use it. Special Topics: --------------- - GAIN Controls: -------------- Note that the QD 2000/2010 system has two controls labeled "GAIN" - one on the SQUID AMP and one on the ACU box. The GAIN control on the SQUID Amplifier is shown in Figure 17 on page 13 of these notes. This GAIN control is under the hatch cover on the Amplifier. This GAIN control adjusts the gain of the AC amplifier that follows the Envelop Detector over a range of 1 to 5. The main point of this gain adjustment is to keep the analog signals in the rest of the Phase Sensitive Amplifier and the input to the Integrator in the middle of their dynamic range. This gain control does not effect the calibration of this system but it does effect the dynamic response of the feedback loop. It does not effect the calibration because the Integrator will still ramp until its output balances or cancels out any flux from the User's Input Coil, i.e. the Integrator will ramp until the output from the Phase Sensitive Detector goes to zero. This Gain Control is adjusted during the SETUP process as described on page 15 and 29 of the QD Operating Manual. The GAIN control on the ACU box is a four position switch giving gains of: 1, 2, 5, 10. During normal RUN Mode operation the Integrator output "Signal" that is received by the ACU box is given this much gain before it is sent out the "OUTPUT" connector on the ACU box front panel. In some sense I guess that the point of this control is to let the user fine tune the RANGE control. Note that it is easy to saturate things with this Gain control, e.g. if the Integrator output is 2 Volts and you set the ACU box for a Gain of 10 then the "Output" signal will just saturate. Note that normally this ACU box GAIN control has nothing to do with the Feedback Loop - normally this GAIN control only effects the signal that comes out of the ACU box's "OUTPUT" connector. But in the Quick Dipper system I believe that this ACU "OUTPUT" BNC signal is the source of the Feedback for servo loop. In that case I expect that this GAIN Control must be kept at "1" to keep the servo loop stable. - Feedback Servo Loop Stability: ------------------------------ Although servo loops are ubiquitous in our Physics lab equipment (e.g. power supplies and temperature controllers) and in everyday life (e.g. the "cruse control" on your car) analysis of their operation is not covered in any Physics classes or books. The block diagrams of an oscillator and of a stable well behaved servo loop are the same. They differ only in the details of the gain and phase vs frequency of their feedback signal. I believe that the Physicist Barkhausen got part way to understand these details and that Bode and Nyquist at Bell Labs made the first complete description. This is the same Nyquist as the Nyquist Sampling theorem. This is really all quiet interesting and gets you to fundamental ideas like the Minimum Phase Shift theorem. Think of all the interacting servo loops at work to get a SpaceX rocket to land right side up. In the QD Operating Manual Figures 1-3 and 1-3A on pages 6 and 7 tell you the details of how the folks at QD designed this controller. This is information that one needs to understand to optimally design an External Feedback setup that uses their controller. The block diagram of the feedback system, Figure 17 on page 13 of these notes, could/should include another block between the Integrator and the VCCS. This block would be called the "loop filter" or sometimes called a "compensator". This is the block that controls the gain and phase of the feedback signal vs frequency to provide a stable well behaved servo loop. In the QD controller the components that control this gain roll off vs frequency are distributed in three places. They are included in the low pass filter at the output of the Phase Sensitive Amplifier, in the Integrator, and in the VCCS. An interesting feature of the way that the feedback signal in the Quick Dipper system operates is that the loop filter elements in the VCCS are skipped over by taking the feedback signal from the output of the Integrator. One could analyze the effect of this choice.