DigiLite set up

[G3GJA]

Just finished building the first of three boards (a 23cm version, another 23cm and a 70cm one to go) and it worked first time. I'm getting good pictures and sound with 82% quality showing on the SL55 receiver.

I'm fortunate to have access to a spectrum analyzer and it looks as though it's the same or similar to the one that Rob M0DTS used to picture the spectrum of the DigiLite in use, and in its USB & LSB test modes, in the CQ-TV 235 article photos.

What the CQ-TV 235 article fails to mention is how the DigiLite Transmit test modes can be used to set up the bias and level pots with an analyzer. Can someone enlighten me please?

Thanks

Clive G3GJA / G8EQZ

[G4EWJ]

I have only occasional access to a spectrum analyser and after setting up using an analyser, I found that the bias (average) level on I and Q was about 1.1v (the data sheet suggests 1.2v) and the peak-peak was 770mV, using In Phase test mode. I found that increasing the peak-peak above this brought spurious signals off the noise floor. I would start a bit lower at 700mV.

This is my method:

If you're using 4000 symbol rate, set the analyser span to 5MHz.

Set In Phase test mode which gives both upper and lower sidebands spaced 2MHz from the LO and adjust the bias pots for minimum LO.

Set the analyser to 2.5MHz span.

Note that due to a bug, the USB and LSB test modes are back to front with DigiLite Config v1-40.

Set LSB test mode. The USB spaced 1MHz from the LO should be the most prominent signal.

Adjust one of the level pots to minimise LSB and LO.

Adjust the other level pot to minimise LSB and LO.

You'll probably find that you can't minimise both at once.

It's like tuning an antenna for minimum swr - move one pot and then tune the other to find the best combination for minimum.

Switch sidebands and do the same again.

Keep switching sidebands to achieve the best mutual minimums.

Start again from the top and see if you can improve on it.

You'll probably find that everything affects everything else to some extent. You can get one sideband well suppressed to find that the other has popped up again. I would aim for about 35dB suppression relative to a single sideband for starters.

There are 2 handy vias for a scope probe: I is in line with R17 near pin 1 of the modulator and Q is in line with and near R14.

I found that there was more swing than you can use on the bias pots, so I changed R31 and R35 to 1k (v5.6 pcb).

Brian

[G3GJA]

Thanks for that Brian. I more or less guessed that was going to be the method after fiddling with the settings whilst in test mode and looking at the results on the analyzer; I found I was able to get the carrier down to the noise floor with careful adjustment of the bias pots and the sidebands minimised with tweaking the level pots.

I presume 82% quality is a good reading for the SL55; it seems to hold that even with 49dB attenuation on the antenna input.

I'm getting about -4dBm out of the modulator with carrier only. Does that suggest the LO level is about right?

BTW, I've made the adjustments easier by using eleven turn pots in place of the single turns pots specified so I won't bother to change the 100R to 1K.

Thanks for your help

Clive

[G4EWJ]

82% quality is what I get on the FS4200 and it never goes higher than that. Maybe the SL55 has the same chipset. The Technomate 1500 will go to 98%.

I was only getting -20dBm out of the modulator without the MAR-6. I hope to be testing it with the MAR-6 on an analyser tomorrow.

Brian