Baracuda + Micro

[Michaelo]

Assuming you have the correct timing, a good coil and the correct value of damping resistor, you should be able to get decent depth right?

This was pretty much the advice I found after building my first baracuda with poor depth sensing...
My problem was almost certainly the timing circuitry but after looking at the 40106 I was not convinced it was wired correctly...
Suffice to say these issues have been resolved as I'm sure you are all aware...

In the mean time I built a tool using an ATMEGA328 to generate the timing and it worked perfectly...
I then thought why not add the amp and audio sections from the bara to the Micro... could be useful...

That's what this post is about.
I have attached the schematic and a 3D image for the board for any one interested...

I'm just looking for feedback and to see if anyone is interested...

I aims are to start with the timing (currently the micro only does the timing and battery monitoring), and add additional functionality over time...
As I get the micro to do more of the work, some of the other circuitry may disappear...

The schematic is not finalised (needs more testing)...
Everything will be in the public domain including the code...
Mike

[ODM]

You cannot switch Q3/Q4 like that. They will forward bias with a positive voltage. NJFET is used for analog switching with a negative gate-source voltage cutoff. (74HC)4053 logic IC would replace the three analog switch JFETs well, switch signals at +5/-5 supply with a logic control at +5/0 supply.

[6666]

You cannot switch Q3/Q4 like that.

Would it work if they were 2n7000

[ODM]

Possibly, but it's rarely done! With J- and MOS FET alike one runs into the problem of narrow signal range. A n-jfet switch will have a problem at switched voltages approaching negative supply, and n-mosfet with voltages approaching positive supply. Also, njfet will start turning itself on in this application if signal voltage swings low enough, and mosfets have a parasitic diode between drain and source, so they are rarely used in this application at least without a mirror configuration and floating switch supply. But that is usually a bit overkill so people favor integrated analog switches that are built in CMOS. This kind of construction has n- and p-channel mosfet parts built together in a way that enables using them for nearly rail to rail switching at regular logic voltage inputs for control.

One way of dealing with the limited input range of simple discrete switches is clipping the preamplifier's output swing to an agreeable range, but that's still more components for essentially the same reason.

[6666]

The schematic is not finalised (needs more testing)...

Have a close look at the 412.

[6666]

Possibly, but it's rarely done! With J- and MOS FET alike one runs into the problem of narrow signal range. A n-jfet switch will have a problem at switched voltages approaching negative supply, and n-mosfet with voltages approaching positive supply. Also, njfet will start turning itself on in this application if signal voltage swings low enough, and mosfets have a parasitic diode between drain and source, so they are rarely used in this application at least without a mirror configuration and floating switch supply. But that is usually a bit overkill so people favor integrated analog switches that are built in CMOS. This kind of construction has n- and p-channel mosfet parts built together in a way that enables using them for nearly rail to rail switching at regular logic voltage inputs for control.

One way of dealing with the limited input range of simple discrete switches is clipping the preamplifier's output swing to an agreeable range, but that's still more components for essentially the same reason.

Thanks for the reply, might have to modify a project I'm working on.

[eclipse]

Would it work if they were 2n7000

Also J113.

[6666]

Hi ODM in this article it talks about useing a 2n7000 as a switch for a relay driven by a micro, you are saying that the same switch would not work for a PI switch for Q3/Q4 ?
http://www.w9xt.com/page_microdesign...switching.html

[ODM]

Basically, yes; that article is actually a great example. The load is makes for a drain voltage that generally never falls below source voltage. This ensures the mosfet's parasitic diode from source to drain is never turned on. And the fixed gate drive in relation to source ensures the source voltage never rises high enough to start cutting off the mosfet.

From figure 6/7 we see that using 2N7000 as switch, as Vgs is positive supply minus signal voltage, we start running into problems at positive preamp output swing. And the parasitic diode which is a "feature" of mosfets is shown drawn in figure 1. When preamp output goes below diode forward voltage, the diode conducts. This is why jfet are more commonly used for analog switching. But a mosfet makes for a fine power switch, like we see in the coil driver.
http://www.st.com/web/en/resource/te...CD00005134.pdf

The situation is more complex with the way dual input integrator works compared to common single input integrator, as the noninverting input is used as well, and the inverting input tries to follow this voltage.

[Michaelo]

You cannot switch Q3/Q4 like that. They will forward bias with a positive voltage. NJFET is used for analog switching with a negative gate-source voltage cutoff. (74HC)4053 logic IC would replace the three analog switch JFETs well, switch signals at +5/-5 supply with a logic control at +5/0 supply.

I have only tested the micro end, the rest is simply a transposed schematic quickly drawn but, it's probably as well you mentioned it... I had forgotten they require a negative gate voltage.
There are a few way around the problem including the 4066 (I see it's been used by other in place of the 2N5484), but there may be an alternative without adding another chip... it should be fun finding a way around this...

Mike

[Teleno]

I think C13/R20 have to be connected between GND and pin 3 of U2A. You have connected them to the +5V rail so they' re actually doing nothing in the circuit.

It's becoming a too common phenomoenon to have previously good circuits really messed up by careless modifications.

[ODM]

Could use a 4066, but it needs full supply logic input swing. 4051-4053 can use a logic control voltage referenced at a mid supply pin, for example +5/0V if using +5/-5V supply.

I guess that hiccup with the integrator hookup was what 6666 was suggesting at.

Regarding all curious configurations people upload of a previous circuit, good hygiene in revision documentation is something people usually/unfortunately learn after other "more fun" things

[Koala]

Nice thing about the Barracuda is it low component vs performance. A lot of work is going into sorting the Barracuda (why some work and others don't) and a lot of micro PI are already out there some including ground balance and discrimination.

Seems a lot like putting a cast on arm when a plaster would do.

[Michaelo]

I think C13/R20 have to be connected between GND and pin 3 of U2A. You have connected them to the +5V rail so they' re actually doing nothing in the circuit.

It's becoming a too common phenomoenon to have previously good circuits really messed up by careless modifications.

With the greatest respect, your comment

"previously good circuits really messed up by careless modifications"

is uncalled for...

If you read my post you would have noticed I said the circuit was transposed, no testing was done... The very reason I started this thread was to discussion the potential for adding micro control to an existing circuit and in the process to uncover/discover errors...


Having said that, it was a pretty stupid mistake on my part, and apologies for it, but the baracuda rev 2 actually has design errors to begin with so I don't feel too bad...

Back to matter in hand...
I've ordered up some parts to see if the issue with the JFET can be resolved but in the mean time I continue to fix other issues...
This schematic has the correction mentioned... no correction for the JFET has been incorporated as yet...

[Michaelo]

Nice thing about the Barracuda is it low component vs performance. A lot of work is going into sorting the Barracuda (why some work and others don't) and a lot of micro PI are already out there some including ground balance and discrimination.

Seems a lot like putting a cast on arm when a plaster would do.

Why so negative... it's just a discussion, I'm not forcing anyone to read it...
Ironically, my design will reduce the component count, simplify the design and increase performance... so as you were saying...
Mike