2 questions-Integrators/noise

[Anonymous]

Hi all,
There has been lots of posts on the front end circuit but what about the integrator. Looking in all my reference books I don't see any integrators with differential inputs. Is there any preferred circuits out there? How do you compute the time constants? What is a good place to start for that? What chip are you guys using for the op amp? or what is the criteria for choosing that amp for the integrator?
I have built Eric's front end circuit and have a question maybe someone can answer as well. I have built it on a breadboard and for the most part seems to work. The only thing about it is while looking at the output on the scope, I see a flat line at the top, then a very short one at the bottom (the transmit pulse?) and then it goes to the center and is a long, very noisy line after that. Is this line noise because I built it on a breadboard? Should I build it on a perf board instead or is this normal what I am seeing? As I bring metal close to the coil, the noisy line goes to the bottom and becomes part of the flat strait line. Since I haven't done this before I don't know for sure what I should be seeing. Thanks for any help on these questions, I know there are a lot.
Fred

 

[Anonymous]

Hi Fred
I think you would prefer to get and answer from Eric, but I have a few suggest for you. The noise you see at the output could come from different sources, for instance it could be noise from something else in your lab that is picked up by the coil. The coil work as wideband antenna and pick up everything thats making noise around you, for instance try to switch on and off a computer in your lab. Anything with a uP is making a lot of noise that is picked up by your coil. It is possible that if you moved your breadboard out in the garden and run it on a battery the noise would completely disappear. You could try to shorten the 5534 to ground at the input and maybe see the noise go down. Be aware that another source of noise could be your power supply. Always use a battery when doing those eksperiment because its free of noise, and make sure that you use a capacitor from ground to plus and minus of the 5534 and keep it very close to the device. Of course it would help a little to make a real board with very short connections to replace the breadboard, this can also be testet by placing your breadboard in a metalbox to shield it from airborn noise.
I hope this can give your some idea to trace where the noise source is.
Mark

 

[Anonymous]

Hi Mark,
Thanks so much for your suggestions. I have tried them all except for shorting the input to ground to verify the noise goes away. I'll try that next. I have been running the circuit off of a gell cell battery. I plan to run it from 24V but most of my testing has been on 12V. I believe I turned everything off (except the scope) but it didn't make a difference. Anymore suggestions is greatly appreciated. I really didn't want to build it on a perfboard until I had it working completely on the breadboard.
Thanks,
Fred

 

[Anonymous]

Hi Fred
Another idea is to try moving around the coil try to turn it around both horizontally and vertically. Because the coil will pickup practically everything from DC to MW radio tranmission you could expect noise coming from a lot of different souces. Another good noise generator is neon tubes, so if possible use a standard old bulb
Instead of shorting the input you could insert a resistor with the same ohm value as your coil. If the noise still exist it was not something picked up by the coil.Again try to locate the sources. Have you tried another 5534 ? It might be partly defect.
Hi Eric if you read this keap on writing your book , we need it. I am running out of suggestions I might have to ask you to take over from hear!.
Good luck
Mark
.

 

[Anonymous]

Hi Fred,
Sounds as though you have got it about right. The flat line at the top is the duration of the TX pulse. This causes the RX to saturate in the positive direction. The short line at the bottom is caused by the coil back emf at switch off and the recovery time of the 5534. After recovery, the 5534 then amplifies any noise, or signal, that is picked up by the coil. As Mark points out, the coil will pickup any electromagnetic noise going and doing tests indoors can be pretty horrendous. What I often do for indoor tests is to wind a coil in a figure of 8. The noise picked up on the two halves then cancels out. The only thing to be aware of is that the response to a horizontal coin has a big dip in the middle of the coil. Vertical coin is OK though. If you want to get really technical you could have a second (identical) coil wired into the non- inverting input of the 5534. Use the same value damping and input resistor plus protection diodes and a 1M0 from pin 3 to ground. Space the second RX coil a foot above the normal coil and you have a noise cancelling gradiometer; provided you have the second coil wound in the same sense. Anyway, with your present arrangement, when you bring metal up to the coil, the noisy line goes to the bottom and if you look, you will see that it comes back to the zero line with an exponential shape. This is a typical eddy current decay signal. Compare the decay signal of a nickel to a quarter. I

 

[Anonymous]

Hi Fred,
I thought that I had posted a couple of integrator schematics before. Go to my post on Integration VS Sample and Hold on 14th May and you will find them.
Eric.

 

[Anonymous]

You guys are absolutely teriffic. Thanks so much for the help Mark and Eric. I dug into it last night and first thing I removed the coil from the circuit. It was still noisy so I new that it wasn't coming from the coil. I disconnected the output of the 555 and now it was just a straight line of noise. I went to the power supply where I am using a max660 to generate the 5V and put my scope on 10mv and saw bad switching noise. I put pin 1 to VCC and verified it was the switching noise as it jumped from about 10khz to 80khz. I tried capacitors for filtering and it made no difference, maybe you guys have a suggestion for cleaning it up. It is being amplified be the 5534 dramatically. Its a lot more noise on the max660 output than I would have expected so maybe this is a result of the breadboard, I'm not sure. In order to further prove to myself this was the cause of my problem I disconnected the max660 from the circuit and put a bench power supply on the 5534 VCC. Now I get a much cleaner signal coming out of the op amp. I reconnected the 555 and the coil and saw a much cleaner signal on the output than originally. It was clean enough that I could move the coil around and verify that the noise level would change. Any suggestions on cleaning up the voltage doubler output is welcomed. Thanks for all you guys help.
Fred.

 

[Anonymous]

Hi Fred
Glad you succed to find the noise source. As I told you use batteries and I mean both for the positive and the negative when you eksperimet, later on you can switch to more elegant solutons like generating a negative voltage. I do not know the MAX660 but most of those are types that use PWM (Puls Widht Modulations) and that generate a of noise. Other types change there frequence and thats even worser. Best thing is to use some of those inverters often seen with a 555 using a constant squarewave. Again my advise is build it all with batteries, and do the negative inverter later.
Mark

 

[Anonymous]

Hi Fred,
I'm not quite clear whether you are generating the +5V with the MAX660 or the -5V. The -5V can be derived straight from the 79L05 and you need an inverter to generate +5V. Basically, I use a 555 hung between 0V and -12V generating a 60kHz square wave. I then use a standard capacitor charge pump arrangement referenced to the 0V line. This then gives about +10V, which is then passed to a 78L05 type to give the +5V. The standard 555's are good at generating switching spikes so in practice I use a TLC555 driving a pnp/npn pair across the -12V to get adequate current. It is important to use an LC filter on the -12V into the inverter to prevent interference getting back out on to the TX supply. I use 100uH choke and a 22uF tantalum. I haven't looked into the MAX660, but if it can be used, run at the higher frequency - 80kHz I believe. I agree with Mark in that it is easier to start with batteries, or a good linear bench power supply. By the way, I have had little success in using switched mode psu's on mains powered industrial detectors. Always too much noise however I have tried to filter and screen them.
Eric.

 

[Anonymous]

Hi Forum!
Excellent postings lately. The flat top in your scope postings ,Eric, sure worried me for a while when it first appeared here at home. The eye does deceive - I printed your scope pictures - they overlay exactly - but who would guess that.
I wasted a lot of time in the power section. My mosfet - currently an IRFP 450 (Yes! retrieved from a computer power supply) is driven by complimentary emitter followers across the same supply - and driven by 6 X 4050 buffers. Overkill? I do not know, but it is very clean.
Power? Wasted lots of time here with Maxim chips - not knocking - but I could not get the P.S. clean - tried all the application notes - inductors and low impedance capacitors didn't clean it up.
I am currently enjoying regulated + and - supplies of 7.5v - maximum for CMOS, although a previous effort with a balanced input via a Maxim 412 with up and down supplies of 5v rails was very nice. Decided to pursue low-impedance input circuitry to defeat noise and RFI - still fumbling around here.
I went to Newport converters - have experienced both one and five watt devices. They are very clean and I have had no problems since - though paying the price of increased battery drain - but who wants to fossick beyond four hours a day?
g.

 

[Anonymous]

Thanks Eric,
That is exactly what I did, I started with a 555 and then I moved to the MAX660 to reduce the parts count. I rebuilt my circuit from the breadboard over to a peice of perfboard. In doing so I put the filter you suggested on it and the circuit seems to be much quieter. If I don't build an acceptable detector in all of this, I'm still having fun and learning. BTW, where do I inquire about the deepstar? None of the detectors I own or have built get anywhere near the results I'm reading on this forum in regards to the Deepstar.
Thanks,
Fred

 

[Anonymous]

Hi Fred,
Where are you located?
Eric.

 

[Anonymous]

I am located in the Houston,TX USA area. I would use it to hunt the beaches near Galveston and Padre island.
Fred

 

[Anonymous]

Hi Fred Jones,
Didn't see your email address listed. If you email me at:< wcrabtre@cape.com > I will supply you with the Deepstar information you requested. You can view the Deepstar and it's specifications on my web site also, at: www.surfscanner.com
Thanks,
Mr. Bill