Specifying DATV power output.

[G4EWJ]

What is the standard method of specifying power output for QPSK modulation?

It seems that average power is the one that is used mostly. This is convenient when using a power meter and for dissipation in the PA, but wouldn't PEP be more suitable for an amplitude modulated signal where linearity is affected when approaching the P1 compression point?

A single sideband can be generated by applying certain waveforms to I and Q and this can be measured as the PEP by any power meter.

I'm assuming that I know how QPSK modulation works, but I stand to be corrected.

Brian

[G4GUO]

Hello Brian,

I think you have opened a can of worms!

I am no expert but I believe PAR
(peak to average ratio) for QPSK is between
about 3.5 and 6.5 db depending on the roll-off of
the channel filtering the tighter the filter the
higher the PAR. For an alpha of 0.35 (used in DVB-S) it
is at the bottom end of the range approx 3.5 db.
So measure the average power and adjust appropriately.
According to R&S DVB-T can have a PAR of 20 db.

- Charles G4GUO

[G4EWJ]

The average does seem to depend on several factors. I think I'll stick with PEP, measured as a single sideband. At least I'll know what I mean, even if nobody else does.

Brian

[G4GUO]

I did a a quick Google search and the best I could come up with was
http://www.bird-technologies.com/resour ... r_MDBS.pdf
Discuss!!

Bird seem to recommend their APM-16 meter for measuring digital waveforms, it is a 'true'
average power meter.

All I have succeeded in finding out is that neither my Bird or my Gigatronics meter are suitable.

Most of the modern meters seem to be diode detectors with A/D converters and software.
I suspect that would make a good little project for someone to design.

I will stick to using a spectrum analyser to tell me when I am over driving and not worrying
about the actual power.

- Charles

[G4EWJ]

Interesting. Thanks.

Brian

[M0DTS]

Just a few comments from what i have seen...

Thermal power sensors give much more accurate peak power readings from what i have experienced.


I have noted my output on the spectrum analyser and seen around 6dB from peak to average which is not far away from what has already been said.

I basically take my amplifier (max power) minus 6dB and try run it just under that according to my diawa meter, so for my 120W 23cm amplifier i can run about 25-30W as read from the meter while running digital (DVB-S), this usually gives a reasonable looking spectrum on the analyser.

On the Nokia TTRX amplifiers you will need to run it around 40W output on your power meter for about 160W peak, the intermodulation becomes very noticable above that.

It may not be perfect but is going to get most people in the right ball park regarding power output.

If you are building a whole setup you need to make sure the driver stages prior to the final amplifier are not over driven by using the same technique, making sure the output power from each stage is 6dB or more below the device power rating.

I hope that is of use to anyone building up their digital setup!

Rob
M0DTS

[G4EWJ]

>If you are building a whole setup you need to make sure the driver stages prior
>to the final amplifier are not over driven by using the same technique, making
>sure the output power from each stage is 6dB or more below the device power rating.

By the device power rating, do you mean the P1 point? If so, then on the sideband suppression test, would you expect a power meter to register P1 watts? Can you reduce the spectral regrowth any further by reducing the peak below the P1 point?

Brian

[M0DTS]

The more you go below P1dB the less spectral regrowth you will see until it is not measurable on the analyser, the regrowth only seems to effect the signal at the other end when it's worse than -20dBc but this it not good for the rest of the users on the band!

Trying to recreate my old tests i ran into a problem while making the measurements too...
I forgot about the RBW settings i think, if you measure the single tome it does not matter what RBW setting you have on the analyser but for the data it makes a huge difference and you are effectively ignoring part of the signal with lower bandwidth settings so get a lower power reading.

But the strange thing is if i use the single tone to set power to 100mW on my power meter for example and then switch the data on i get pretty much the same reading... the data average is not showing less than the single tone....!
it may have something to do with how we set the modulator up but can't see where yet..

I'l have a further look tomorrow!

Rob

[M0DTS]

Ok..

By seting the transmitter to no modulation you can set the output power to around 6dB below the P1dB point, 1/4 power.
I did this at 100mW (400mW amplifier being used) confirmed with spectrum analyser and power meter.

(Power meters normally use diodes as the detector so do usually give strange results for anything other than a single tone, differnet meters will give totally different readings for multi tone signals.)

I then swithced modulation on and the power meter showed ~80mW
The spectrum analyser shows around 4dB peak to average difference uisng the highest bandwidth (3MHz) and 'sample' mode. I was running 4Ms at the time.

With this setup the digital output showed IMD3 of about -45dBc... god enough for me..

It's surprising how much you can get away with, i've run my 100W amp at 60W on the power meter by accident while running digital and not had noticable degradation at the other end! The outgoing signal would have been very wide though!

I hope that's useful... and hope it's corerect!

Rob
M0DTS

[G4EWJ]

The 4dB peak to average is in the same area as the figure Charles gave.

Should there be much carrier output with no modulation from a DBM? I had people listening for my carrier on SSB receivers and it was very weak. The sideband test produced a much stronger signal.

Incidentally, what do you get on the analyser when you select the 'In-Phase' test option?

Brian