Full m/lab SD2000 schematics

[Andrei]

from datasheet

[ZED]

Hello Andrei

The OP249 has some nice spec's but at 10 hertz its to noisy,see page 8 (voltage noise density),the LT1057 at 10 hertz is 25nv as opposed to the OP249 which has almost 60nv at 10 hertz.
For me thats to much noise,but for other detectors it may be fine,i have a total gain stage starting at the integrator and finishing at the ADC of 8645,so an integrator amp producing 60nv of noise at 10 hertz
would give me a noise signal that the ADC would convert.

[Andrei]

Hello ZED! I wanted to ask about electromagnetic interference (EMI).
Is the SD2000 factory board two-layer or more?

[Mechanic]

I'm not Zed, but I'll answer that..

The sd2000 board is only 2 layer. They were all 2 layer up to the gpx, the GPX uses 4 layer due to the layout density.

EMI comes in through the coil.

Cheers Mick

[Anis]

much appreciated for all your support!
Thank you all

[Andrei]

I'm not Zed, but I'll answer that..

The sd2000 board is only 2 layer. They were all 2 layer up to the gpx, the GPX uses 4 layer due to the layout density.

EMI comes in through the coil.


Cheers Mick

Ок, thanks!
my Board turned out to be noisy, I decided to make a different trace, use the top layer of the same for the ground polygon and ground the smd components. and there was a question, whether it is necessary to divide polygons of the earth of different modules: transmit cct, logic cct, receive

[Andrei]

It is interesting to see how minelab did it, but on the Internet there are few photo boards sd2000

[Bayu]

G.s how many pcb size and width please

[GoldenKiwi]

For the record, most of the answers are all over on Doug's forum.
I had to test it out for myself, I tried manipulating the 33k, and found that it needs to be between 30k and 40k so 33k is perhaps optimum.
The gain resistors also appear to be optimum.
That only leaves the emitter tail current. Data sheets suggests a current between 3 to 4mA gives the lowest noise figure.
So this is probably what unregistered was on about. The bias on the current regulator transistor is beyond me and may be better with a simple diode stabilizer.
Also the compensation resistor maybe better around 220 ohms, especially if the tail current is trimmed fro lowest noise, but I've not tried that yet.
I'll leave it there.

Cheeeers

Hi, I'm looking at the SD2000 LM394 into NE5534 based front end and am trying to understand it. I've searched and seen this discussion but I don't recall any specific answers. I would love some help if possible to try and understand.

What is the function of the 150R and 1NF between the legs of the LM394 LTP?
I'm am unsure how to calculate the emitter tail current with the current source arrangement, nor the resulting gain of this first stage. Can anyone suggest a specific link of similar arrangement and calculations or please describe the calculation?
I am trying to understand the collective gain of this composite first stage (LM394+NE5534). Elsewhere the gain was suggested to be 6000-8000. https://www.geotech1.com/forums/show...highlight=6000

I'm trying not to be lazy...I've ordered a copy of "The Art of Electronics" to try and help me.

Any help would be very much appreciated!

[waltr]

That is a very interesting circuit and not the easiest to understand.

The 1nF-150 Ohm are in series. At DC to low frequencies the 1nF cap is high Z (open) but at higher frequencies the cap conducts putting the 150 Ohm across the NE5534 inputs. I call this a 'snubber' as it attenuates high frequencies and Fast impuses (pulse edges).

The LM394 & NE5534 is a basic instrument amp. Check this App Note:
http://bitsavers.trailing-edge.com/c...es/AN-0222.pdf

I also do not know how to calculate the exact gain.

[green]

Is the gain (1000/22)+1=46.5?

[ZED]

Yes green yer right mate the gain is around 47,thats how i would calculate it too.

GoldenKiwi heres a sim of the cct your interested in,just unzip it,hope it helps,with my basic understanding of electronics this preamp cct is way out of my league.
This preamp is only found in the SD2000,the later machines just use a 797 chip with receive fets.
The bias on the constant current circuit can be done in any number of ways,generally you want to minimize drift with temperature and so using the temperature coefficient of a diode is a big help but
a pair of resistors in this instance is as basic as it gets.
Long tail pair preamps like this one are great for minimizing DC offset on the output when used with matching transistors.
The four resistors on the collectors i imagine are to improve the tolerances between them,the tighter the match between the collector resistors the better the DC offset is.
Plus waltr is spot on with regard to the 1n cap and 500r resistor.

Also the spice model for the LM394 isn't available so the SSM2210 was the next best thing.Dont remember where i got the model from though but thanks all the same.

[waltr]

Thanks for sim model.
I did this but got wrong connection (gnd instead of -5V) to current source.
Ran a freq sweep and Gain is 33dB from 0 to 20kHz. That is a gain of 44.
Close enough to your result.

The 1n & 500r rolls off the gain by -6dB at 240kHz.

[ZED]

Thanks waltr ! and so do you think the circuit is fast enough for a P.I. preamp ? is 200kHz fine or is say 5mHz better ? i have read a variety of opinions.


Heres a good read on various parts of a preamp circuit.

http://www.douglas-self.com/ampins/dipa/dipa.htm

[waltr]

Good paper on audio amps.

That SD2000 works so the 200kHz bandwidth seems ok.

We do need Fast >1MHz GBW op-amp to obtain good linearity with a 200kHz cut-off.