Portsdown debugging, trials and tribulations etc....

[radiogareth]

So I've ordered the box and have all the bits and pieces, time to progress from UGLY mode up the scale a bit.
Made up leads to connect to everything (LO Filter board, Output select, Portsdown Filter modulator.
First error, getting the address lines back to front to the modulator board....second error getting the I-Q and ground leads in the wrong order. My excuse it I used Molex at the destination but screw connectors at the source so clearly guessed the '0' end wrong!
Fired that up and it works nicely as per the test instructions I found here. https://wiki.batc.org.uk/Filter_Modulat ... Parts_List
Having bought the FMB ready built and tested I can't take much of a bow....
Next I checked the voltages on the LO filter board (fed with 3.3 from the RPi) and I see 2.75 on the logic chip test points. Checking the address for 71,146, 437 and 1255 shows the correct binary word present, but putting it in line results in no lockable signals.
So time to get the magnifier on the board, although it was built with care, methodically and I'n not a stranger to SM components.
How best to test it, other than the obvious voltage and no shorts. Is there a DC path in/out of the chip or should I perhaps use the signal generator function and a handy SDR to look at levels in/out? (I shall pursue the data sheet)
SImilar applies to the output switch board, putting it in circuit in inconclusive, probably too much signal leakage across the shack. Very little to get wrong on that one....
I get very quick lock when using the animated test card as a source, nothing from the Pi-camera and a black screen from the C920, although the surround lights up blue when I select it and go to TX. Best I RTFM I guess

Gareth

[radiogareth]

Reply to self.... I think there was a bad joint under pin 19 of the switch chip. Logic that was on pin 8 of the logic gate wasn't present on pin 19. Resoldering that seems to have effected a cure. Putting it in-line slightly reduces the indicated signal strength in MT but in's not scientifically tested. No such problem with the LO filter board, although putting that back in line significantly reduces the output level and drops the MER by 10dB at least. Only a spectrum analyser is going to show me whats going on there I suspect. Be easy enough to rig a 2 bit binary address and see what gives.

Gareth

[g4pdf]

Following my decision to take a working RPi B3 (no +) and reload with the June18 Stretch OS, and following the directions in the GIThub script, I appear to have a functioning unit, but there is no ugly output on 437. I have repeated the build, from the same image file, with the same result. I forced a software update to no avail.
With this problem I'm now off the script. Are there any diagnostics for a RPi with only power, Pi camera and Ethernet attached?

[G8GKQ]

Hi

I take it that you do not have a touchscreen attached and are working from the console menu over ssh?

Please confirm that you haven't changed any of the settings since the initial build?

What does MiniTioune show when you select transmit? Any increase in signal strength or change in behaviour?

Dave, G8GKQ

[g4pdf]

Hi Dave,
I have the Waveshare touchscreen attached too ... functionality seems ok; after the initial build the putty session terminates when the RPi reboots and the BATC logo appears on the touchscreen. I haven't changed any of the initial settings. The Console is working over SSH connection. Callsign and locator adjusted. The Minituine fails to find anything much (no MER), and my 1.5GHz Spec Annie confirms there is no signal or change in the ambient noise floor, wide or narrow sweep, when Tx is selected.

[G8GKQ]

OK

Do you see any evidence of the Lightning flash in the top right corner of the touchscreen?

How are you coupling the spectrum analyser to the RPi?

At this stage, my next test would be to login by ssh (which will give you the console menu), then open a second (duplicate) ssh session which should give you a command prompt; maximise this. At the prompt, type

Code: Select all

rpidatv/scripts/a.sh

This will start the RPi trying to transmit. Then you can see if there any obvious error messages. You can stop the transmission by selecting transmit in the console and then selecting receive in the console. You could copy the text from the terminal screen and paste it to me in an e-mail if you want.

I know the build script works, because I did one this morning and tested it before sending it to VK. One final thought - did you rebuild using the "portsdown" script or the "rpidatv" script?

Dave

[G8GKQ]

Please report the version number that is shown on the "info" screen (select from Menu 2). That will give another clue.

Dave

[radiogareth]

Any reason why I shouldn't shorten my panel mount USB leads and solder them direct to the RPi PCB? It would be much tidier than a bundle of surplus wire taking up case-space?
Gareth

[g4pdf]

Please report the version number that is shown on the "info" screen (select from Menu 2). That will give another clue.

Dave

Software Version: Stretch 201809050

[radiogareth]

Been using the Portsdown sig gen today to try and sort out my 4 way switch and band filter, which worked so much better when I soldered in the ADL5602. Months have passed since I was assembling the boards so I must have left that and forgotten about it. Sig gen function is SO handy, a great inclusion. So handy I'm wondering if I should buy another one JUST for use as a sig gen
Closer, but not there yet...case arrived today though, going to look nice with the touch screen angled up slightly in a modified sloping screen mount.
Gareth