Pico Technology PicoScope 3206

From sigrok
Jump to: navigation, search
PicoScope 3206

The Pico Technology PicoScope 3206 is a 200MHz analog bandwidth 200MS/s sampling USB oscilloscope.

See Pico Technology PicoScope 3206/Info for some more details (such as lsusb -v output) on the device.

See Pico Technology PicoScope 3000 series for information common to all devices in this series.

Hardware

The brain of the device sits in the FX2. It does the USB communication, programs two DACs and the waveform generator, uploads the firmware into the FPGA, forwards some settings to the FPGA, and controls the seven relays, the input channel crossbar, and the power nets enable signals. The FPGA fetches samples from the ADCs, stores them in the SRAM and waits for the trigger signal. When done, it fetches the samples from the SRAM and gives them to the FX2.

I2C

An ST24C02 EEPROM is directly connected to the FX2. It stores the USB ID and probably some calibration data. All other chips are behind an I2C gate (74VHC4052). The other I2C chips are

  • PCA9554A - 8 bit GPIO controller
    • IO0 controls the relay for the 20:1 attenuator on channel B
    • IO1 controls the relay for the AC high-pass filter on channel B
    • IO2 controls the relay for the 20:1 attenuator on channel A
    • IO3 controls the relay for the AC high-pass filter on channel A
    • IO4 drives the LED
    • IO5 decides which channel is routed to the AD8132+AD9215 behind the channel A connector
    • IO6 decides which channel is routed to the AD8132+AD9215 behind the channel B connector and the LT1715 for triggering
    • IO7 drives the relay to route the waveform generator to the EXT connector
  • MAX5822 - 2 channel DAC
    • Sets the trigger levels
  • AD5315 - 4 channel DAC
    • One output to each AD8132
    • One output to the VREF pin of each AD9215

SPI

There is no SPI controller in the FX2. You have to do bit banging.

  • AD9833 - waveform generator
    • SDATA/SCLK/FSYNC is PD4/PD3/PD2 on the FX2
  • XC2S30 - FPGA
    • DIN/CCLK/PROGRAM is PA1/PA7/PA0 on the FX2

Power Nets

  • The voltage from the optional power supply is converted to 4.9V by an LM3578. This is combined with the nominal 5V from USB. This power rail sources the SP6650 for 3.3V and the secondary 5V power rail.
  • The 3.3V power rail is for the FX2, the EEPROM, the I2C gate, the MAX6424UK29, and the 555 timer. It also sources the secondary 3.3V power rail. It is always on.
  • The secondary 3.3V power rail is for the AD9833, the 25MHz oscillator, both AD9215s, and the LT1715. It is enabled while FX2 PD7 is low.
  • The secondary 5V power rail sources the LM2611 for -4.9V, the SP6650 for 2.5V, the MAX1896 for 6V, and both LP2981AIM5 for 4V. It is limited to about 7mA unless FX2 PD7 is low.
  • The 6V power rail is for the two SST5912s. It is enabled while FX2 PD5 is high.
  • The 2.5V power rail is for the FPGA and the SRAM. It is enabled while FX2 PD6 is high.
  • The -4.9V power rail is for the input channel and trigger signal conditioning, the waveform generator, and FPGA pin 102/103. It is enabled while FX2 PD5 is high.
  • One 4V power rail is for the input channel and trigger signal conditioning, the waveform generator. It is enabled while FX2 PD5 is high.
  • The other 4V power rail is for the AD5315, the MAX5822, the PCA9554, the LED and the relays. This one is always on.
  • FX2 PD5 also controls the I2C gate.
  • PD5/PD6/PD7 can be overridden by dip switches.

FPGA

Thankfully this is only a two layer design without BGA components. On the top layer the 2.5V are distributed below the FPGA to all VCCINT and VCCO pins. Ground is fetched from the other side with vias.

FPGA IOs
Pin Connected to Pin Connected to Pin Connected to
3 LT1715 out A (Trigger A or B) 49 SRAM /WE 94 SRAM D
4 LT1715 out B (Trigger EXT) 50 SRAM A19 95 SRAM D
6 25MHz oscillator 51 SRAM A 96 SRAM D
7 FX2 PD1 54 SRAM A 99 SRAM D
10 FX2 PD0 56 SRAM A 100 Test point
11 FX2 RDY1 57 SRAM A 102 RC network @ -4.9V
12 FX2 WAKEUP 58 SRAM D 103 RC network @ -4.9V
13 FX2 PKTEND 59 SRAM D 112 AD9215-1 D9
18 FX2 FLAGA 60 SRAM D 113 AD9215-1 D8
19 FX2 FD4 62 SRAM D 114 AD9215-1 D7
20 FX2 FD5 63 SRAM D 115 AD9215-1 D6
21 FX2 FD6 65 SRAM D 117 AD9215-1 D5
22 FX2 FD7 66 SRAM D 118 AD9215-1 D4
23 FX2 FD0 67 SRAM D 120 AD9215-1 D3
26 FX2 FD1 68 SRAM A 121 AD9215-1 D2
27 FX2 FD2 69 FX2 PA0 122 AD9215-1 D1
28 FX2 FD3 72 Test point (DONE) 123 AD9215-1 D0
29 FX2 IFCLK 74 SRAM A 126 AD9215-1 CLK
31 555 out 75 SRAM A 129 AD9215-2 D9
37 FX2 PA7 76 SRAM A 130 AD9215-2 D8
38 10kOhm to CCLK? 77 SRAM A 131 AD9215-2 D7
39 FX2 PA1 79 SRAM A 132 AD9215-2 D6
40 SRAM D 80 SRAM A 133 AD9215-2 D5
41 SRAM D 83 SRAM A 134 AD9215-2 D4
43 SRAM D 84 SRAM A 136 AD9215-2 D3
44 SRAM D 85 SRAM A 137 AD9215-2 D2
46 SRAM A 86 SRAM A 139 AD9215-2 D1
47 SRAM A 87 SRAM A 140 AD9215-2 D0
48 SRAM CLK 93 SRAM A 141 AD9215-2 CLK
  • As you can see only 16 data lines of the SRAM are connected. Two data pins are unused. This probably also means that the FPGA discards two of the data bits from the ADCs depending on the selected voltage range.
  • Pin 50 is an address pin only if the board is equipped with a bigger SRAM (K7N161831B).
  • I don't know what the 555 is used for. It looks like it in an astable configuration.
  • Have a good look at pin 102 and 103. They are connected to caps via 110 Ohm resistors. These caps are supplied by the -4.9V power rail via 100kOhm if both pins are inputs. Contrary to what was said here before, this will not kill the FPGA. According to the IBIS data the internal protection diodes will limit the voltage to a safe value and the negative voltage is in fact applied while the FPGA bitstream is uploaded. Anyone knows if these pins serve any purpose?

EEPROM

The I2C EEPROM is used mainly to hold calibration data. Multi-byte values are little endian. There is a checksum, but is very weak and overly complicated. The last two bytes contain the number of iterations that have to be done to a 14 bit LFSR to get a value that equals the sum of the preceding data when interpreted as signed bytes. The weakness comes from their method of reducing the sum to 14 bits: During summing they reset the intermediate value to zero as soon as it uses more than 14 bits. As the bytes added are signed, this happens very often.

Allocation of EEPROM bytes
00-07 see FX2 docs
08 unknown (0x01)
09-0A Hardware version
0B-0C Flags?, bit 0 makes FW disable USB high speed
0D-12 unknown (all 0x00)
13-1C Batch and serial as ASCIIZ string
1D-24 Calibration date as ASCIIZ string
25-26 AD5315 channel 2 value for 0.5V and 1V ranges, trigger not on channel A
27-28 AD5315 channel 2 value for 0.2V range, trigger not on channel A
29-2A AD5315 channel 2 value for 0.1V range, trigger not on channel A
2B-2C AD5315 channel 2 value for 5V, 10V, and 20V ranges, trigger not on channel A
2D-2E AD5315 channel 2 value for 2V range, trigger not on channel A
2F-30 AD5315 channel 1 value for 0.5V and 1V ranges, trigger on channel A
31-32 AD5315 channel 1 value for 0.2V range, trigger on channel A
33-34 AD5315 channel 1 value for 0.1V range, trigger on channel A
35-36 AD5315 channel 1 value for 5V, 10V, and 20V ranges, trigger on channel A
37-38 AD5315 channel 1 value for 2V range, trigger on channel A
39-3A AD5315 channel 1 value for 0.5V and 1V ranges, trigger not on channel A
3B-3C AD5315 channel 1 value for 0.2V range, trigger not on channel A
3D-3E AD5315 channel 1 value for 0.1V range, trigger not on channel A
3F-40 AD5315 channel 1 value for 5V, 10V, and 20V ranges, trigger not on channel A
41-42 AD5315 channel 1 value for 2V range, trigger not on channel A
43-44 AD5315 channel 2 value for 0.5V and 1V ranges, trigger on channel A
45-46 AD5315 channel 2 value for 0.2V range, trigger on channel A
47-48 AD5315 channel 2 value for 0.1V range, trigger on channel A
49-4A AD5315 channel 2 value for 5V, 10V, and 20V ranges, trigger on channel A
4B-4C AD5315 channel 2 value for 2V range, trigger on channel A
4D-74 space for channel C & D? (repeats 0x0200 0x019A 0x00CD 0x0200 0x00CD)
75-76 AD5315 channel 3 value up to 100Msps, trigger not on channel A
77-78 AD5315 channel 0 value up to 100Msps, trigger on channel A
79-7A AD5315 channel 3 value sampling channel A at 200Msps
7B-7C AD5315 channel 0 value sampling channel A at 200Msps
7D-7E AD5315 channel 0 value up to 100Msps, trigger not on channel A
7F-80 AD5315 channel 3 value up to 100Msps, trigger on channel A
81-82 AD5315 channel 0 value sampling channel B at 200Msps
83-84 AD5315 channel 3 value sampling channel B at 200Msps
85-94 space for channel C & D? (repeats 0x0200)
95-96 MAX5822 DAC B scale factor, lower four ranges, trigger on channel A
97-98 MAX5822 DAC B scale factor, upper four ranges, trigger on channel A
99-9A MAX5822 DAC B scale factor, lower four ranges, trigger on channel B
9B-9C MAX5822 DAC B scale factor, upper four ranges, trigger on channel B
9D-A4 space for channel C & D? (repeats 0x7FFF)
A5-A6 MAX5822 DAC A scale factor, trigger on EXT
A7-A8 MAX5822 DAC B offset, lower four ranges, trigger on rising channel A
A9-AA MAX5822 DAC B offset, lower four ranges, trigger on falling channel A
AB-AC MAX5822 DAC B offset, upper four ranges, trigger on rising channel A
AD-AE MAX5822 DAC B offset, upper four ranges, trigger on falling channel A
AF-B0 MAX5822 DAC B offset, lower four ranges, trigger on rising channel B
B1-B2 MAX5822 DAC B offset, lower four ranges, trigger on falling channel B
B3-B4 MAX5822 DAC B offset, upper four ranges, trigger on rising channel B
B5-B6 MAX5822 DAC B offset, upper four ranges, trigger on falling channel B
B7-C6 space for channel C & D? (all 0x00)
C7-C8 MAX5822 DAC A offset, trigger on rising EXT
C9-CA MAX5822 DAC A offset, trigger on falling EXT
CB-FD unused? (all 0x00)
FE-FF checksum

As you can see, there are 8 ranges: 100mV, 200mV, 500mV, 1V, 2V, 5V, 10V, and 20V

  • The upper four ranges use the 20:1 attenuator.
  • 100mV, 200mV, 2V, and 5V use the lowest 8 bits of the 10 bit ADC.
  • 500mV and 10V use the middle 8 bits of the 10 bit ADC.
  • 1V and 20V use the upper 8 bits of the 10 bit ADC.

To use the calibration data to calculate a MAXS822 value for a specific trigger voltage, use the following formula: (voltage*0x40000000+offset*0x40000+0x27330000)/(scalefactor*10)

Photos

Protocol

See Pico_Technology_PicoScope_3000_series#Protocol.

Resources