PE43703 Attenuator

[pa2mrx]

Hi all,

I have a problem with the PE43703, I recently bought some from the bay.

setup: portsdown 2018+ 8 band switch/decoder. software: 201912100

Connected everything and the signal from the FM board is attenuated approx 30dB.
Using the sig.generator the signal is also attenuated and can`t be set.

All working fine with about +9.6dB signal from the FM board, on all bands.

I prepared the PE43703 as per instructions on the wiki.
Data & clk coming from the ADF4351 board.
LE coming from digithin no.10 connector
All three through a 1K resistor.

I see a quick response on the scope on the data & clk connector to the attenuator.
clk: signal around 80µS. data: signal around 231µS (3.2volts) nothing on LE.
What kind of signal do have to see from the digithin no.10 wire? There`s nothing to see, also
nothing on LE to adf4351.

edit: connected everything to the adf4351, I also got signal on the LE, two pulses to low on tx, and two pulses low going
to rx again. But nothing the like on LE on the attenuator.

any suggestions.
Thanks
Frank

[g4hiz]

Hi Frank,

I was testing the same type of attenuator (PE43703) myself recently as it seemed not to respond to programming. I monitored the signals arriving at the serial inputs and all seemed well, there were 16 clock pulses to clock in 16 bits, followed by a short '1' pulse to provide the LE signal and all of the external pins connected to the chip ok.

Note that the attenuator LE signal is different to that for the ADF4351, which goes low for the duration of data being clocked-in, then rises at the end to latch-in the data.

If you look at the data sheet for the PE43703 (or PE43713 - essentially the same) it says that the address in the 16-bit data (000) should match that for the device. However for the attenuator I had, the three address pins were not connected to anything, which implies 'undefined' or floating inputs - so I presume the address in the data couldn't match the device and so wasn't loaded.

Maybe the pcb layout is different in some cases, connecting these pins to 0V and so matching the address, as I have successfully used one of these attenuators already. Needless to say, the attenuator worked fine in parallel mode when programmed by the DIP switches.

One other thing to check, is that you have not accidentally blown up the schottky diode between the 5V input pin and Vdd line for the device. I guess that this was included for polarity protection, but if you have both the serial AND parallel 'blobs' partially in place, this will kill the diode as I did myself trying to fix the above problem. The high attenuation that you have experienced looks like the 'unpowered' attenuation for the device.

Good luck.

73
Jen G4HIZ

[G8GKQ]

Hi Frank

The first thing to check is if you can see the LE signal at the Raspberry Pi end of the 1K resistor. It could be a short-circuit at the input to the attenuator.

Dave, G8GKQ

[g4hiz]

Hi Frank,

I have included a couple of oscilloscope screen shots from my PE43703 testing. The first shows data with clock, in this case 7 ones. The second shows the LE with the clock. The LE is easy to miss. The 'scope was triggered on the clock signal, with horizontal time-base set to 1ms/cm and vertical gain set to 1V/cm.

As a preliminary, check first for continuity of wires to pins and and shorts between pins and of course, that the 5V power supply is present.

73
Jen G4HIZ

[pa2mrx]

Thanks all,
I`ll check things out.

A good time to see (learn) how the scope works using the memory option, didn`t use that a lot, only the analog side of it.

73
Frank

[g4hiz]

Good luck Frank, I hope it works out.

Just to point out that according to the 43703 datasheet, serial address 000 should work with the address pins floating, implying internal pull-down resistors. A check on these pins indicated 455kOhms to ground (probes both ways round, device unpowered). Why my attenuator doesn't work in serial mode shall have to remain a mystery; I've ordered another one from a different supplier!

73
Jen G4HIZ

[pa2mrx]

checked all connections on the attenuator for shorts, chip get power, also high on pin 3.
cleaned everything with board cleaner.
pins are around 455K to ground.
data/clk arrives at chip, but not LE.
Put all outputs on the scope, data/clk/LE/ LE from Pi no 10 pin.
took clk/data on attenuator from board feedthrough pads near chip.

All signals OK, except from PI datapin no.10 (digithin 10) directly.
Nothing there.

took another attenuater same type from another vendor and modded it to serial.
Same problem.
Both work fine using the dip switches before modded.

Put the att under the magnifying camera. The three adress pins A0/A1/A2 on the chip are also not connected to anything.
And if i look carefully, the one on the picture on the batc wiki aren`t connected also.

I can try another raspberry I have here. with the limesdr mini on it.



Here a screenshot of the scope, the bottom line LC3 is connected straight to the output of the PI, pin 10 on the GPIO board, not on pin 10
of the digithin connector. LC 2 is the LE to the ADF board. LC0/LC1 are data and clock from the GPIO adf connector.
What you see is going to TX, att. set to 0dB.

73
Frank