One ground plane, or two?

[Detectorist#1]

Hello Carl,
Very strange! Must to have some unknown (at this moment) reason. The main parameters of my pinpointer are:
L=18uH
Ipeek=8A
Upeek=950V
Rdump=390R
Rclamp=200R
Non-inverting first amplifier OPA11611AID with Ku=101
MOSFET IPD95R450P7
Additional diode in the Drain of MOSFET is STPS2L25U type. Removing of this diode haves catastrophic influence on the possible first delay.
First delay=4us
Working frequency 50Hz
TX Pulse 25us
9V 6F22 battery. 16mA current consumption.

[Detectorist#1]

Hi Carl,
Sorry for late reaction. I think, in your picture from oscope haves some strange (for me) thinks. First, the saturation voltages on the output of the first amplifier during TX pulse are strange - with 1K clamp resistor the saturation voltage is lower than with 5K clamp resistor. Second, haves strange, more high saturation voltage just after the end of the TX pulse. Do you have some explanation for these thinks?

[Detectorist#1]

Hi Carl,
Also, if red trace is for 1K clamp resistor, white trace - for 2.2K and pink trace is for 5K, the strange things continues. The red trace points under-damped case. The white trace points also under-damped case but more near to critically damped case. The pink trace points also under-damped case but more near to critically damped case in comparison with white trace. Where is my mistake?

[Carl-NC]

The test circuit used an inverting amp, like this:


For Rs = 1k, 2.2k, 4.7k Rfb = 100k, 220k, 470k. This alters the output offset due to the input bias current through Rs. Nominally, for the NE5532 it should be 235mV for Rs=4.7k. This probably explains the varied offsets, although the saturation levels are still curious. Might be due to the load effect of Rfb.

Since Rs contributes to overall damping, as Rs was changed so was Rd. For the Rs=1k case, Rd should have been reduced a little more to get closer to critical damping. However, doing so will not affect the saturation knee to any great amount so the overall benefit would still be about the same.

[Detectorist#1]

Hi Carl,

What confused me. If red trace points under-damped case (1K clamp), critical damping and over-damping cases have to be on the right of the red trace - later , not earlier! DC offset isn't matter.

[Carl-NC]

Yes, critical damping will slightly shift the saturation knee to the right but not by much.

[Detectorist#1]

Hi Carl,
What is the probability for possible earlier sampling of the output voltage of NE5532 with clamp resistor of 4.7K instead of 1K as result of earlier exit from the saturation state of NE5532 - not as result from faster switch -off time of 1N4148 diodes?

[Carl-NC]

I'm not sure it would make much difference. The gain is always 100 so the GBW should be about the same, except that the feedback resistor is higher for the 4.7k option and that should be a slight detriment. Besides that, you have overdrive recovery speed and that should not vary much either as in both cases the input overdrive is about the same.

[Detectorist#1]

Hi Carl,
Maybe the point of this discussion will be make after similar tests with faster OpAmp.

[Carl-NC]

Maybe, but I've done this trick with several different opamps and it always helps. Reducing the diode's forward current reduces the diffusion capacitance and makes turn-off faster.

[Detectorist#1]

Yes, that's true. But please see the attachment to #9 of this thread. That confuses me.

[Carl-NC]

I don't know what you are doing to get those measurements. I've always found it curious that Schottky diodes never seem to work as well as a plain old 1N4148 in this application despite having far less diffusion effect. But the correlation of reduced clamp current to faster settling with the 4148 is pretty consistent. Here is a Spice sim using a MAX410:

[Detectorist#1]

Hi Carl,
As usual, the processes are more complex as it seems at first look. The temperature of the diode haves serious influence on diffusion capacitance according the theory. My measures are with 2us current pulse and 2ms pause. But if the current pulse is 5us (decay time) and the pause is 50-100us, the temperature of the diode will be higher than room temperature and the diffusion capacitance will be seriously higher. This explains my measurements and your observations when you use high frequency (10-20)KHz search frequency. In this case, seriously reduced peek current via 1N4148 diodes will have significant influence . Tomorrow, I will made measurements for dependence of the diffusion capacitance of 1N4148 from the temperature.

[Carl-NC]

My tests were at 1kHz, I think the TX pulse was 100us.