Not so DC waveform of minus 5 volt rail at the ne5534 - could this be cause of noise problems

[Qiaozhi]

Not suggesting a redesign. The question, is the noise level what is expected.

I think this is more or less the expected result from this simple design. It is what it is.

[green]

I think this is more or less the expected result from this simple design. It is what it is.

Your noise level, a few mv or Polymer's noise level, 300mv at U7 pin7? How many mv of p-p noise should be expected at U2 pin6? Seems there should be away to make a measurement and say this is what is expected and there is nothing that can be done to improve it without redesign. There are ways to calculate amplifier out noise knowing noise specs, not sure I'm doing it correctly. Could someone calculate p-p noise at U2 pin6 with coil input shorted? Maybe with some other detectors such as Surf or HH also. It appears noise from amplifier is causing Polymer's noise level. I'm thinking if integrator input is connected to common and the noise is a lot less at post amplifier out then the amplifier is causing the noise problem, maybe I'm wrong.

[green]

http://www.geotech1.com/forums/showt...9120#post39120

thread on noise, now all I have to do is understand it

[6666]

http://www.geotech1.com/forums/showt...9120#post39120

thread on noise, now all I have to do is understand it

Thanks for dredging up that old thread it has some good gear in it, must reread it.

[Polymer]

Does anything alter if you put a 100 PF cap across the preamp input pins ?

Hello 6666,
I have tried this now. I used a 100pF ceramic cap.
There were no changes that I could see on pin 7 LF412.
Maybe to subtle for me to recognize.

Thank you for your input nonetheless.

Perhaps what you've mentioned will carry more when I have gotten rid of the GREAT NOISE.

[Polymer]

Sorta feel lost in the thread I started, geeeez ...... Putting up more ...

My ongoing quest about my threshhold problem found me this tidbit (THANK YOU Teleno):



I disconnected C1 and found what Teleno states has merit.

The threshold jumpiness is approximately halved!

BUT there is a helluva lot more noise on pin 6 now.

dada

nocnoc

Looks like the LT1054 frequency which was not a problem in my battery test before can really be seen now.
Disconnected that again, put in second battery. Bingo, the helluva lotta noise is gone.
Didn't do much more for the threshold jumpiness though.

If anyone has a scope pic of threshold jumpiness of Baracuda, Geotec Baracuda, Surf PI or Hammerhead, PLEASE post.
It would be great to have reference pictures.

To be continued ...

Found this, don't understand it yet, may be interesting for others, see "low frequency leakage noise" part.

[waltr]

Interesting experiments.
Unfortunately, I have an old analog scope so slower time base traces are very hard to see (200ms/div).
I do have some threshold jumpiness but doesn't seem to be a problem in actual use. I adjust the threshold just above the jumpiness and run the Gain in the SAT stage much lower (~30 instead of 100) which reduces the peaks of the jumpiness.

I do run a large value cap on the MOSFET TX and also use JFETs instead of analog switch for the sampling.

[Polymer]

Originally Posted by SVEN1 Recently rec'd some knock off Texas Instruments NE5534's, they give off a rather wonky wave pattern, not even close to what it normally looks like. The detector still functioned.
As soon as I changed out to a real TI NE5534, the wave form was normal.


Thanks Sven!
I will look into that as I have some TI NE5534's as you've mentioned in use.
I didn't get them from the bay or China, but even some of the more reputable dealers sometimes seem to have wonky parts in their supply chain

I got some TI 5534's from another vendor and some old Signetics ones too. What you mention was part of the problem! Good find and info!
Now I can discern external noise, like rotating coil etc., from the "casual" internal noise.

[Polymer]

Here two scope pics of Pin 7 of U7(LF412) @ 200ms.
One with C15 diconnected and one with C15 connected again at same scope range.
From what I can tell the signal without C15 is dead quiet ... the threshhold sounds super solid here.

barageo_pin7_lf412_nocoil_noRdamp_C15no.pngbarageo_pin7_lf412_nocoil_noRdampR_C15yes.png


scope pictures don't look like pictures from reply #21 http://www.geotech1.com/forums/attac...2&d=1523301124 http://www.geotech1.com/forums/attac...4&d=1523301171 what am I missing?

I have been wondering about this too. You're not missing anything! Perhaps the following answers your question.

Out of desperation I have been looking/asking around what might be causing so much interference.
I think I might have found one of the main EMI culprits. Perhaps some seasoned engineers might want to chime in on this one.
Apparently it is not continuously on, which might explain my differing results when looking at the behavior of the PI outputs.

In a nutshell: Transmitters for measuring ocean currents. Roughly 5 to 50 MHz. Pulsed 20us to 150us.

I don't know if you guys in the U.S. have problems with this at some beaches. Anyway, here's more info on that:

1A-SeaSonde_v2_20100331.pdf

The closest transmitter to me is 1.2 miles. When testing away out from town I was even closer to this bugger

[6666]

With reguard to post #51 and the large electro caps, aren't you supposed to have multiple vlaue parallel electro caps to over come that charge up problem IE: 1,000uf, 500uf, 100uf, .1uf, so when the rail takes a big hit the smaller caps recover faster to bring the voltage up fast.

[waltr]

I have been wondering about this too. You're not missing anything! Perhaps the following answers your question.

Out of desperation I have been looking/asking around what might be causing so much interference.
I think I might have found one of the main EMI culprits. Perhaps some seasoned engineers might want to chime in on this one.
Apparently it is not continuously on, which might explain my differing results when looking at the behavior of the PI outputs.

In a nutshell: Transmitters for measuring ocean currents. Roughly 5 to 50 MHz. Pulsed 20us to 150us.

I don't know if you guys in the U.S. have problems with this at some beaches. Anyway, here's more info on that:

[ATTACH]42939[/ATTACH]

The closest transmitter to me is 1.2 miles. When testing away out from town I was even closer to this bugger

It is possible that SeaSond is part of your problem.
If it has a regular periodic pulse then adjust the PPS of your PI Detector to null the interfering pulse.
This would require a pot on the timing circuit that produces the pulsing. A few commercial PI detectors (White's TDI for one) has this control to reduce interference.
I found on my HH2 that I needed to fine adjust the PPS rate to reduce AC Mains interference.

[Polymer]

With reguard to post #51 and the large electro caps, aren't you supposed to have multiple vlaue parallel electro caps to over come that charge up problem IE: 1,000uf, 500uf, 100uf, .1uf, so when the rail takes a big hit the smaller caps recover faster to bring the voltage up fast.

Hmmm, interesting question.

Somewhere I have read about what you are talking about and if I remember right it has to do with resonances.

I think that in this case all capacitors in parallel regardless of size will have the same voltage across them at all times.
So from what I can picture the smaller caps would NOT "bring up" voltage faster as long as there are bigger capacitors still sucking current.

I thought a big fat cap was crucial for coil switching. Teleno's information has helped me reduce the wobble some more. I now have a 10uF in parallel to the coil
which clears up the new noisy stuff pretty much as compared to the coil without a capacitor.

All this has had me thinking about coil "charging". A coil can't be "charged" instantaneously, it initially resists current flow and
takes some time (5 tau) to reach max current flow. Perhaps there is an optimum capacitor size for a given coil LR.

I think the speed at which the coil builds up the magnetic field is not that important as compared to the "kick" at switch off.

I suppose all this also depends on the robustness of the power supply too.

[waltr]

The reason for different value/type caps to by-pass is due to each cap having a different frequency (time) response. Then nearly all frequencies will be by-passed with the caps.

Interesting idea of having other cap value to charge the coil. I use a single 4700uF cap and do have a slight amount of 'wobble' into the audio stage.

Yes, coil charge rate is not as important as the sharp switch off. I found that some resistance in series with the coil/MOSFET can improve stability without decreasing sensitivity.
I run a 10 Ohm series R to get Tau = 33us and then a 100usec TX pulse (3Tau). 3Tau is enough and the current is only very slightly higher if TX pulse is 5Tau.

[Polymer]

Noise issues continued ...

I did implement a 10 Ohm series resistor, like you mentioned waltr. It does help!

Have been playing around a bit and found a noise culprit signal when touching the just unsoldered 1k resistor leg - the other end still connected to the 5534.
Yeah, I know I shouldn't solder when somethings running, but I just couldn't help myself.
Anyway, I got a nice 208kHz sine signal output on my oscilloscope. It was interrupted occasionally.
Searched on the net, it looks like an NDR (non directional radio beacon) for maritime and aviation purposes. There are a few around here as I have found out.

Had a power blackout in my area a few days ago. I did run home from my walk to go test my Surf PI, to see if that horrible low frequency wobble is gone and
hoping power wouldn't be back to soon.

Alas, wobble not gone. Still there as usual. At least I know it is not 50 Hz interference now.
It was a simple test, switch on and look at LED - which was flickering.

One thing that did make things quieter was tossing out the charge pump.
Put in a 9v battery & neg 5v regulator. The -5v part of my circuit draws round 14mA@3kHz.
That's ok for me, fits in easily and weight is not much of a concern since it's for underwater.

End of noise journal for today, have a great weekend!

[waltr]

Guessing the 208kHz beacon was on back-up power.
At lease you know its not AC mains (50Hz).

Fine adjustment of Pulse rate may help. Don't remember if you tried that.