Minelab SD2000 front end replacement amp

[jtz54]

It's been several years since some of the threads were started on improving the SD2000. Since then I'm sure some parts are obsolete while others have much improved
replacements available. DOes anyone know of a drop in replacement for the LM394 in the front end that will make a significant improvement ? Are there other components
that should be replaced?

Thanks!

[Davor]

No.

LM394 is one of the sexiest components that we, as a community, are left without. There are only replacements, but hardly any real improvements. For a replacement seek MAT12.

Trouble with disappearing low noise components is that the industry demands far less of real low noise components than before in pre-CD era. There are however some very nice, and costly, opamps designed recently, but it will not help you much with your quest.

The topology that uses LM394 is very nice, and achieves ~1nV/sqrt(Hz) of noise. Going lower than that is possible with hand-picked pairs of transistors that have very low Rbb, such as BC337 (!) that has Rbb about 30 ohm. There is a new topology that delivers even lower noise, but that's about it.
.

[Mechanic]

Hi Jtz,

The question you really need to ask is : Is the preamp producing the dominant noise in the system?

From my observations the answer is no.

The best way to find a noise source or at least the section in the circuit that noise is coming through is to for a start, remove the receive wire and connect it to ground at the coil plug. Then look at the output of the filters using the testpoints on the audio board, the middle is ground, the others are ch1 and 2. Set your scope to 10mv/div DC and .5 seconds sweep or slower. Also set the threshold just so the threshold is breaking and compare both channels, I would be interested to see if ch2 is noisier than ch1 on yours. Then you start going through each stage of the circuit and ground the output of each section and comparing the measured noise. When you get to the area where the noise suddenly disappears you have found the dominant noise source. Then you need to figure out why. I have found that the low pass filter seems to be the dominant noise source in the 2000, if I short the input to gnd(the input of the 43k resistors) I have the same noise, but if I short the output of those resistors the noise is gone...... Perhaps too much gain on the low pass filter? Perhaps the noise level of the 43k resistors is too high? Replacing the tl074 for a lower noise opamp seemed to do very little.

Cheers Mick

[moodz]

No.

LM394 is one of the sexiest components that we, as a community, are left without. There are only replacements, but hardly any real improvements. For a replacement seek MAT12.

Trouble with disappearing low noise components is that the industry demands far less of real low noise components than before in pre-CD era. There are however some very nice, and costly, opamps designed recently, but it will not help you much with your quest.

The topology that uses LM394 is very nice, and achieves ~1nV/sqrt(Hz) of noise. Going lower than that is possible with hand-picked pairs of transistors that have very low Rbb, such as BC337 (!) that has Rbb about 30 ohm. There is a new topology that delivers even lower noise, but that's about it.
.

SSM2212 is an alternative to LM394.

[moodz]

SSM2220 is PNP but better noise figure 0.7 nV

[Davor]

PNPs are generally less noisy than NPNs. In NPN matched pairs choice you'll find SSM2212 at 1nV and declared as 28ohm Rbb transistors. Practically same as BC337. Its predecessor SSM2210 is now extinct same as LM394.
Going lower than 1nV will be possible only in case you use more current, transistors with lower Rbb, or both. BF459, as reported by Uwe Beis, goes below 0.5nV, so why not buying a bunch of these and select a matched pair, screw them together and shove them inside?

[Davor]

On a 2nd thought, it might be a good idea to aid a preamp with a simple transformer to go for real low noise. There is ample of headroom for improvement in PI machines with the coil resistance being rather low. As coil resistance is the ultimate noise barrier, and 1nV/sqrt(Hz) noise equivalent resistance is 60ohm, there is over 10dB of potential improvement if only a real low noise input is provided. A square root of impedance ratio gives you voltage ratio. With all the transformer losses, the input noise equivalent impedance can be below 1ohm.

I'm about to play a bit with transformers for improving input referenced noise of IB machines, but the principle is the same. Good mic transformers are a pain to find, but there may be some alternative routes to get there.

[Bill512]

Hi Jtz,

The question you really need to ask is : Is the preamp producing the dominant noise in the system?

From my observations the answer is no.

The best way to find a noise source or at least the section in the circuit that noise is coming through is to for a start, remove the receive wire and connect it to ground at the coil plug. Then look at the output of the filters using the testpoints on the audio board, the middle is ground, the others are ch1 and 2. Set your scope to 10mv/div DC and .5 seconds sweep or slower. Also set the threshold just so the threshold is breaking and compare both channels, I would be interested to see if ch2 is noisier than ch1 on yours. Then you start going through each stage of the circuit and ground the output of each section and comparing the measured noise. When you get to the area where the noise suddenly disappears you have found the dominant noise source. Then you need to figure out why. I have found that the low pass filter seems to be the dominant noise source in the 2000, if I short the input to gnd(the input of the 43k resistors) I have the same noise, but if I short the output of those resistors the noise is gone...... Perhaps too much gain on the low pass filter? Perhaps the noise level of the 43k resistors is too high? Replacing the tl074 for a lower noise opamp seemed to do very little.

Cheers Mick

I remember myself often to fade, in this "low noise quest". It is a fascinating search, but first we must look at Mechanic's observations.
If these are true (and I don't have any reason to believe the opposite) simply, any tiny improvement in any other block is useless.

[jtz54]

Thanks for the info. I will try this and post the results.