Elektor Project: Cracle-free volume control.
Nederlands
I'm sorry, but I cannot support the other languages in which Elektor is
published. You may write e-mail to me in German, but I'll answer in English or
Dutch to your choice.
Additional information
There are some minor changes w.r.t. what was published in the Elektor article.
A
resistor R5 of 3.3 kOhm is added in the connection to the /CS pin 2 of the PGA2311.
Without this resistor the in-circuit (re)programming of the microprocessor
will not work.
The
resistors Rx, Ry, en Rz are omitted in my design
The stringent separation of analogue and digital groundplanes is less
relevant in this design, because if you do'nt touch the controls for a
second or so all digital activity outside the microprocessor chip will fall
dead. Measurements showed that even while the PGA is actively updated no
noise was measurable with a measuring limit of -130 dB Volt (0.3 uV)
With groundplanes as suggested in fig. 3 of the article or in the directions
in the datasheet a board with this size could not be realized in 2
layers.
The schematic now looks like this: (Drawn somewhat different from what is in the Elektor article )
Fig. 1
Schema.
Top
The PCB I can provide measures approximately 39 x 32 mm.
It is double sided, has metalized holes, goldplated pads, solder masks on both sides and component print on one side.
IC1 and
IC3 (if used) are to be mounted on the bottom side.
All components except IC1, IC2 and IC3 are through-hole.
There is one mounting hole suitable for an M3 column.
Fig. 2 The bare PCB, viewed on the component side.
Fig. 3 Component placing
Fig 4 Solderside (Bottom layer)
Fig 5. Component side (Top layer)
The programming instructions in the Elektor article are meant for the DIP-8 package of the microprocessor. To program this SMD version the best way is to mount it on the PCB and to connect some wires. The pins 1 through 4 of IC1 must be connected to the corresponding pins on the Spyder Stick.
Use a DIL-8 IC socket and solder the wires as shown below.
Fig. 6 Connections to the Spyder stick.
Solder the wires to the PCB on the following points:
Pin 1 of IC1
The proper side of R3 or R5
The proper side of C1, pin 4 of JP1 or the cathode of D1
A GND connection.
During programming the PCB must not be supplied in the normal way. Also the connections to JP1 must be detached. The Spyder stick must have complete control over the microprocessor, including supplying the 3V3 voltage..
For
testing with audio signals disconnect the DIP-socket. After that you may switch
on the normal supply voltages. If you are satisfied with the result you can
remove the wires from the PCB.
The
input signal should not contain a noticeable DC component. You will hear
that while contolling the volume.
The PCB I can provide has no provision for DC blocking capacitors.
Seen the advice in the datasheet
to maintain a source impedance of less than 600 Ohms I advice a DC blocking
capacitor of at least 25 uF.
The
input signal must not be greater than 2.5 V RMS or 7 V pp. Trespassing
this limit will immediately result in serious distortion, even when the
volume is set to zero.
If
you (re)program the CPU in- or external from the circuit with the Spyder
Stick you will get an error message "communication with the target
lost" after downloading te code (F5) and attempting to run it (F5) from
the stick. This is normal behaviour. Because the BKGD pin is used as an
output, diagnostic communication is not possible during run time.
For fast testing with the CPU running on the Stick-power you may connect the
potmeter(s) and the LED. You should see the led going off around one second
after leaving the potmeter alone. This will give the proof that the software
has been downloaded properly
If you do
not have a negative supply voltage available and use the MAX660 a small amount
of crosstalk at the working frequency of ~40 kHz of the MAX660 can be expected.
My measurements indicate a level of -95 dBvolts, that is around 20 uV, or around
the digitization error of the CD-system.
If you use the PGA2311 with a gain of up to +31.5 dB the crosstalk may increase
to 0.5 mV.
In any case I have not been able to detect any intermodulation (sum / difference
frequencies) between this unaudible frequency and any possible music frequency
component, including a 0 dB 44.1 kHz signal.
My measuring limit was around -130 dBVolt, that is 0.3 uVolt, and ranged to 48 kHz. (96 kHz sample frequency)
Below pictures of the spectral analisys.
Fig. 7. Spectrum on the outputs left and right. 0 dB is 1 Volt rms. -100 dB is 10 uVolt.
Fig. 8. Here the gain was set to +31.5 dB. We now see -66 dB, that is 0.5 mVolt crosstalk.
The signals at 11000 Hz and 24000 Hz are from an unknown origin. The are there also when the volumecontrol is not powered
These spectra have been acquired using a Creative Labs Audigy NX2 external USB audio device and the program SpectrumLab. The sample frequency was 96 kHz thus frequencies up to 48 kHz could be captured. As far as I know Spectrum Lab supports most sound devices.
You may download SpectrumLab from here. It is freeware. But use the .INI file in the download package, put it in the program-files directory after installation. The standard settings of SpecLab are for a completely different type of sports.