MAR8 prior to AD8307
I am thinking of using the mar8 in series with the ad8307.
Is the preamplifier part ok (bias resistors and choke)?
I put a preamp input 50R shunt to ensure good match at all frequencies, is what I have done ok?
BW=29 MHz
NF=3.1
MDS=-174+3.1+10*log(BW)=-96.xx dBm..
Bandwidth and NF is pretty high to measure -100dBm signal strength.( I didn't include necessary S/N !, it should be around more or less 10-12dB for safety )
It will somehow work but the sensitivity may not be as you desired..
What can I do to push the ad8307 down to -100dbm for the 1-30MHz and maybe better on 1-60MHZ, as I need to also measure harmonics?
Since the Bandwidth is so large as remarked, there are not so many possibilities to down to -100dBm..If you increase the bandwidth , the MDS will be worse..
Modern Power Meters can not measure -100dBm too.This level can be measured with SA only..
The spectrum analyzer does this in the IF only?
In the AD8307 datasheet there is this page 21 which uses an ad603 to increase the dynamic range. But this is only possible with a filter in place as shown in page 22, and not broadband.
Adding the MAR8 will have any benefit or just use the ad8307 alone?
Forget it. I will use the AD8310 instead. It has a -87dBm of minimum level, which is very close to what I need.
And here is the complete meter for anyone interested.
I wonder, can I change the 6.8k near the meter to a higher value one and use a 100uA meter instead?
If you use a 100uA meter you can probably drive it directly from the AD83xx.
Brian.
This one is 20 bucks, it has a digital display, and is working unbelievable well:
https://www.aliexpress.com/item/3288...chweb201603_53
This is down to -75dbm only. I need it to be close to -90dbm and there are not many cheap ones as far as I have searched. I said -100dbm to have a margin, but I saw that I won't have this luxury easily.
I think I will use the AD8310 instead. It has a -87dBm of minimum level (judging from the datasheet graph), which is very close to what I need.
This would be interesting and it will simplify the circuit. Maybe a series potential divider (or rheostat?) with the 100uA meter, to set accurately the maximum point Brian?
It has a built in voltage amplifier so all you have to do is add a series potentiometer to limit the meter current. You would need a reference to adjust calibration accurately or apply a voltage according to the output graph to the resistor/meter to set a known current. I'm guessing you need to make your own meter scale anyway so it shouldn't be difficult. For example 0dbV input gives 2.5V output (AD8310) so to pass 100uA the series resistor would be ~25K. A 22K fixed resistor and 5K pot should do nicely.
Brian.
25k? That's a lot of headroom, it seems it will work fine! I'll test this thanks.
Do you think I should include a 100nf shunt capacitor from the vout to the gnd, or just omit it?
If you add a capacitor at the output of the internal amplifier it may go unstable. Check the data sheet for the AD83xx device you choose, there is provision for a filter capacitor at one of the pins and information in the text on how to use it.
Brian.
Thanks for pointing this out to me!
Since an analogue meter will be used I wouldn't worry about the ripple at twice the sine frequency (the meter can't respond).
So here is the final circuit.
I wonder, if I set the meter at it's max when the input signal is lower (say -30dbm). then will this somehow increase the sensitivity of the meter at the lower end? (more needle movement)
This one may be an alternative for you.. page 10-11
https://www.nxp.com/docs/en/data-sheet/SA604A.pdf
Yes, but with the risk of damaging the meter and possibly the IC at higher signal levels. It isn't a safe strategy.
A better approach if you want to follow that route is to amplify the lower voltage from the AD83xx but still limit the maximum current to the meter. The best approach is to amplify the incoming HF signal though, to make it more sensitive but as pointed out before, you also increase the noise floor which makes detecting very tiny signals less accurate.
Brian.