Bandido II observations

[Goodrat]

I've been doing a lot of tweeking and testing of this circuit and I'd like to point out a few things.
For one, I noticed a 5MHz oscillation at the input of comparator IC8B. This is quite low, and does not affect much in the way of noise, but a 10K to ground cures this. I am trying to reduce the chatter when the sensitivity is turned up. There is a noise floor, probably picking up 60Hz and such. This can be fixed with a 150pF cap on IC8 1 to 2 (inverting to output). But now the unit is less sensitive. So, I can inrease the receive amp a little. This gets me back to the typical sensitivity, but nothing is gained.
What bothers me most is the 1M ohm pull-up at IC8 A and B. If you notice the Amigo schematic, this is 270K and the resistors out of the comparators are not used (4.7K on the Bandido).
I'm going to try 1K OR-ing resisistors and a 100K pull-up and see if I can get this circuit to calm down. The 10K R51 in the lower comparator seems to control how high the output through D10 is. I may incease this to see if I can get a weak target to be a little louder.
It seems to be Tesoro is squeezing out gain where another gain stage would have been a little more stable.
Any comments would be great.

[Goodrat]

Re: IC8A

When I use a chip cap 10pF between IC8 1 and 2, this took care of the noise also and has more sensitivity than using a leaded 150pF. I also shorted R54 and R55, and changed R57 to 100K and removed C34.
Back to 10 inches on a quarter, and no chatter at max sensitivity, but I'm going for 11 or 12 inches. R29/R28 on the receive amp is 130K, going to try 150K again.

[Goodrat]

Re: IC8A

When I use a chip cap 10pF between IC8 1 and 2, this took care of the noise also and has more sensitivity than using a leaded 150pF. I also shorted R54 and R55, and changed R57 to 100K and removed C34.
Back to 10 inches on a quarter, and no chatter at max sensitivity, but I'm going for 11 or 12 inches. R29/R28 on the receive amp is 130K, going to try 150K again.
Killing the 5MHz oscillation with a 10K at IC8 pin 5 to ground cuts down the sensitivity, so I need to look at that more, or live with it.

[Beachcomber]

Re: IC8A

At the rate you are going you are going to know that circuitry inside and out!!!!

HH

Beachcomber

[Goodrat]

Yes,...

I think I have it committed to memory now.
So far I have this:
R28/29 to 131K
R54/55 to 1K
R57 to 100K
C34 moved to across R53
10pF across IC8A pins 1-2
I think the 5MHz oscillation was caused by the scope probe, so the 10K at IC8 pin 5 is out.
All pots and switched are off the front panel, but I don't think this really gives me less noise. But it is easier to work on. I'm still at 10.5 inches on a quarter due to having to turn down the sensitivity a little.

[Beachcomber]

Re: Yes,...

You are getting really close to the 11 to 12 inch mark. You are probably right that it could have used one more stage of amplification rather than taking the other stages all the way up to thier maximum gain. Even so you may be able to tweak an additional 1.5 inches out of it!!!

Good Luck!!!

HH

Beachcomber

[Goodrat]

Any comparator experts?

I'm thinking of making U8 more typical. Besides a pull-up of 100k, and R54/55 to 1K, remove the 10k, and use 1M positive feedback on both comparators.
But Tesoro has had this setup for a long time, there must be a good reason.

[Goodrat]

I give up!

I just cannot improve the signal to noise.
They got this thing tweeked just right for what it is.

[mikael]

Re: Any comparator experts?

I think, the most critical parameter is not the signal to noise, (except for the last stage) because we receive the signal at a good level.
It is the slew rate of the AOP because we compare the phases.
The two pathes have to be very symetrical to obtain the better result.
I am going to optimize it, i just have to finish to mount all the components on my PCB.
I have layouted with eagle a small PCB with SMD components mounted on both sides (4cm x 8cm) and a big one with a lot a space to debug the circuit.

On the schematic, no decoupling capacitors are shown. They have to be mounted on all supplies to the ground (100nF)

Do you know why they put a PNP transistor between R67 and V- ? it creates a bad voltage drop !
with shotky diodes whith 0.2V drop, instead of 0.7V, we can improve V- to - 4.3V.