Tweet
After innumeral trials and errors, suddenly I got a quantum leap in the front end amplification of the PI. The coil and preamp setup suddenly produce an amplification of the signal ( at a 6 uS first delay), that is about 10 times stronger than before. This at first happened by a mistake on the bread board, but I have torn the whole thing down 3 times and rebuilt it with the same result. I am aware that, to work on the breadboard at mega hz frequency is at best fickle and totally unreliable and subject to all sorts of parasitic capacitances and inductances, so I dont know the real reason yet for this circuit to function. I intend to make a few more tests and then make a Pcb to see if it still works. In the mean time maybe some experts from the forum would be so kind to give an opinion? Tinkerer
-
-
Once you get past the preamp, the breadboard parasitics should not have much impact. Sounds like you got something in front of the preamp to settle faster. -
Hi Tinkerer,
I am interested in what you found to increase the sensitivity. Would you be willing to share the concept so we can determine what is happening?
Without any information, one can only speculate that you may have added a certain amount of positive feedback and got it to work without any oscillation. However, that is just a guess. If you have been successful, great.
RegComment
-
still not sure myself if this is real
thanks, Carl and Reg for your input.
I will prepare some explanations and a circuit over the next days and post it. As mentioned earlier, this is at the bread board and oscilloscope level, so it may not stand up later down the design.
At one point I did have a lot of oscillation and eventually managed to atenuate them. This makes me think that I might have a resonant circuit.
I will try to run an FFT, maybe this will give me a clue.
TinkererComment
-
Here is the setup
This is the setup. More tomorrow TinkererComment
-
Hi Tinkerer,
You have an interesting design there. I tried something similar using a transistor as a power diode across the opamp. Instead of tying the base to ground, one simply ties the base to the collector. The idea is to limit the opamp for faster recovery.
Unfortunately, such a design is transistor dependent and I didn't have the transistor I should use, so it is one of those things I have on the back burner. I figure one would have to use a very fast transistor for the best results.
RegComment
-
JC1
Hi Tinker,
Looking at your circuit. First I am not sure the 5817s are the way to go. the capacitance looks terrible. maybe try some 4148s, just two. the gain of 2,200 is probably reducing your bandwidth and cutting into the quick return. 500 should be faster, but you will have to add another amp to make up for gain. the transistor "clamp" capacitance maybe slowing down the amp as well, so make sure this circuit is "helping" speed things up, which it may be.
MHOComment
-
-
Diodes
Thanks for the input. I agree with you on the 5817, however, my choice is limited to what I can find. The second diode in parallel serves two functions, one is to take some of the load when I increase the power. I found that of the limitations of the preamp is the very long recovery time of the diodes. Looking at the specs, including the 1448, the recovery time should be no problem. But we really use the diode outside its specifications, with the penalty of a long recovery time.Originally posted by Unregistered
The other use of the diode is that it gives just the right amount of capacitance, sounds silly, but when one has to do with the means at hand, it seems to work.
The 2,200 fedback just happened to be the highest I had. the setup works well anywhere from 200K to 2,2 M
It also works well with a tx pulse of 34uS up to 70uS.
What makes me happy about this circuit, is that there are still many improvements possible and even as is, the results seem to be quite OK.
TinkererComment
-
Transistor
Hi Reg,Originally posted by Reg
thanks for the input. I tried several transistors, unmarked generic NPN with HFE anywhere between 195 and 130, scavanged from broken TV sets etc. The one new transistor I tried, is the 2N2222 and it works fine.
I have no idea how to calculate the actual gain of this preamp. Maybe somebody could help me with that?Comment
-
Diodes
Hi,
The 1N4448 have half the Total capacitance of the 1N4148, not that it's much anyway, but if one were on a misson to reduce it to the pits.
Cheers
Kev.Comment
-
The diodes are crucial
Hi Kev, The more I work with the limiter diodes, the more I find they are of crucial importance in reducing the first delay time. At one time I came to good results, (notwithstanding the noise level) with germanium diodes. Just when I thought I had found the solution, I discovered that my batch 0f 10 diodes had huge variatins every one of them, so I gave up on Germanium diodes and tried to get some Schottky's. the ony ones I managed o get were the 5817, and frankly the results were miserable. I tried 1448 and 914 but in the end went back to the 5817 because of its higher power rating. I will definetly try several other types of diodes as soon as I can get hold of them. Here is one problem with the diodes: When I use a bit more power , like 12 V- 70 uS - 310 uH - 12 ohm, the diodes really get stressed too much. This shows as a prolongued recovery time. This recovery time ends up being the limiting time for the first sample. For now I have tried to counteract the oscillation caused by the diode by applying an oscillation that is 180 different in phase. Endeffect a smooth decay curve. Anyway this is what I think it is, maybe the experts can correct me. Any opinions are welcome. TinkererOriginally posted by KevComment
-
JC1
Hi Tinker,
Troubleshooting things long distance is a bit akward,
But here goes with my "guesses". Isn't the 634 ohm damp
res. a bit high? maybe something 300 - 400 ohm?
Your gain should roughly be the feedback resistor divided by the 1K resistor. You can leave the 2.2 meg and increase the 1k to 4k to get 500 gain.
By the way I have spent some time going after the little nuggets designing a PI, and it is hard to get the little ones.
anyway I felt I needed bandwidth in the front end, which WILL increase noise and the noise received, but needed for fast sampling, and getting the small signal up. The 4148s I used, I don't remember a problem but maybe they were consistant enough not to matter.
You are right, they get banged pretty hard, but some form of switching diode should be faster at the recovery. Though I believe I have seen schematics on this site or somewhere that was using 5817s so maybe. I never tried them so don't know for sure.
Make sure you are getting 100kHz bandwidth or there abouts I would say.Comment
-
JC1
also run about 10kHz pulse rep rate.
I once made the damping resistor very high because I know the thing is also wrecking the receive signal. And of course I got a nice ringing waveform. Waving some metal around and with the scope offset max and gain up, I could see I was getting what I felt at the time a bigger signal.
But how to get this signal out of that waveform, is troublesome. Integrate the whole mess and try to messure the change? I don't know, I didn't persue it.
And of course as ground changes in the field, this ringing is going to change, so point sampling is dangerous.Comment
-
JC1
Hi All,
The data sheet on the 5817 has the cap at 50 - 200 pf. And I know diodes can change like crazy, but times three is 300 - 600 pf of capacitance and I just see the little tiny nuggy signal getting soaked up by this capacitance.
MHOComment
Comment