New Version of Bipolar PI MD with GB and DD Coil

[Detectorist#1]

Hi PahWiz,
This boost converter is tested with input voltage 5.8V - 8.4V. When the input voltage is 8.4V, the output voltage is 8.1V. These 0.3V are the voltage on the 1N5819 diode. Please, assembly this simple scheme and measure the voltages with 100mA load. It is not possible to use +9V or +10V. The supply voltages are limited by the limits of TC4428CPA (4.5V -18V). I prefer to not use supply voltage up to the upper limit. In the datasheets for LiPo accus, end (empty) voltage is close to 2.75V, not 3.5V.
For the second remark, I agree with you. Pin1 is -8Vin and pin3 is +8Vin. The labels with red color near to the input power connector have to be removed. Sorry for this mistake but this is the typing error and it is very easy to be recognized even from less experienced people.
Nevertheless, thank you for your so careful analyzing of the scheme!

[PahWiz]

sorry for late response. yeah, i know the boost converter work when the input voltage is 8.4V (but efficiency drops) and the output = input - diode voltage drop but increasing voltage to ±9V solve this without exceeding the voltage limit for TC4428CPA since 18V is the max recommended supply voltage (Absolute Maximum voltage is 22V). But either ±8V or ±9V works fine.
about the Li-po voltage it depends on battery internal resistance and current draw from battery, but 2.75V is completely discharged. to avoid damaging battery it is better to set low battery 20% percent (maybe somewhere around 3.0V or so).
i was looking at schematic and noticed adding parallel Turn-off diode across R3 and R33 (MOSFET Gate) increase turn-off speed.



TC4428CPA (U5) pin5(VDD) connected to +8V through series resistor R1 (1Ω), same thing happened to pin3(VSS). i think it is better to directly connect MOSFET driver supply pins to ±8V and keep the 1Ω resistor for MOSFET source pin, because during transient current peak probably cause voltage drop across U5.

Ho much is the wattage for R18(47R near the coil connector)? Is 0.25W enough?

Thanks for your efforts, nice design!

[Op04]

Hello PahWiz,
I just want to say about your green marks there... Just my opinion, It will make coil TX On a bit more time, we need faster, so we need R1&40

[Detectorist#1]

Hi PahWiz,
1. No problem with +/-8V supply.
2. See the scheme of the Power Board Rev.D. The "Battery Low" voltage is set to 6V (e.g. 3V for cell)
3. Adding of diodes in parallel of gate resistors not improve nothing . Output resistance of TC4428A is 7oHms (not zero) and no improving in OFF timing of the MOSFETs
4. C3 and C23 haves big capacity and they are LOW IMPEDANCE type. This ensures very little change (appr. 100mV) of the voltage during TX pulses

I need to learn more!

[PahWiz]

I don’t think R1 & R40 affect the ON time. The main purpose of the 1Ω resistor is current limiting (limits peak current during coil discharge) and protects MOSFETs from avalanche breakdown during inductive kickback. Also damps LC resonance between supply capacitors and trace inductance and prevents power rail oscillation at MOSFET switching frequency.

[Detectorist#1]

Hi PahWiz,
You are far from the true. R1 and R40 are filters from TX currents to prevent voltage noise on power supply lines. This noise is possible to disturbs the rest parts of the detector. R24 and R25 are the damping resistors and they also protects MOSFETs from avalanche breakdown. For damping of LC resonance on power rails you are right. And I have one question for you -Do you use AI for learning of the PI detectors?

[PahWiz]

Hello Detectorist#1
1) yeah, no problem, just suggestion
2) nice.
3) you are right. i didn't consider the Output resistance of TC4428A. The improvement is 30% but for PI metal detector it's negligible.
4) yeah, main current source for TX pulse is 1000uF/Low ESR electrolytic capacitor (works fine). But it can be avoided by connecting TC4428A supply pins to ±8V directly (need extra 10-22uF caps near TC4428A).

[Detectorist#1]

Hi PahWiz,
Eliminating 100mV of noise on the TC4428 power rails does not improve the performance of this driver, you just add additional capacitors to the board (just in case).

[Laith]

Hi PahWiz,
Eliminating 100mV of noise on the TC4428 power rails does not improve the performance of this driver, you just add additional capacitors to the board (just in case).

Do you make it and test it?

[Detectorist#1]

Hi Laith,
Do you mean the metal detector with this scheme? Please read the start of this tread on page 1

[Laith]

Hi Laith,
Do you mean the metal detector with this scheme? Please read the start of this tread on page 1

Yup

[Laith]

is it ready to get it out in the field ?

[Detectorist#1]

Yup

[Op04]

I don’t think R1 & R40 affect the ON time. The main purpose of the 1Ω resistor is current limiting (limits peak current during coil discharge) and protects MOSFETs from avalanche breakdown during inductive kickback. Also damps LC resonance between supply capacitors and trace inductance and prevents power rail oscillation at MOSFET switching frequency.

Coil tau = L/R

[Laith]

Yup

Can you send Gerber file and hex file