TX Driver

[pito]

DRV8872
You need a capacitor in series with L, without it the IC will die, DC motor has much higher inductance.

[Carl-NC]

Everything I do now includes a micro. As soon as you get past just driving the TX and you also need to drive demods, read in user pots, and generate a PWM for audio then the micro is far easier than designing all that in discrete circuitry. For my money (and time), I would use a DRV8872 and a small micro that can drive it and do everything else I want.

[pito]

So he has two options to choose from.

[Daves]

So he has two options to choose from.

Hi again Pito, Why do we connect 100uf?

[Daves]

Everything I do now includes a micro. As soon as you get past just driving the TX and you also need to drive demods, read in user pots, and generate a PWM for audio then the micro is far easier than designing all that in discrete circuitry. For my money (and time), I would use a DRV8872 and a small micro that can drive it and do everything else I want.

Hi Carl,
There is something I want to ask;
Capacities shown in the picture
1)What is the purpose of use?
2)For what purpose and how is it used in TX?
3)For what purpose and how is it used in RX?
4)What is the type of capacity? What should you pay attention to when choosing and using this capacity?

[Carl-NC]

Without schematics, I have no idea.

[kt315]

ceramic. 1-5 uF

[Carl-NC]

Actually, most look like PPS film caps, probably in the range of 100nF - 1uF. Still can't say what they do without a schematic.

[pito]

Hi again Pito, Why do we connect 100uf?

To prevent DC short circuit, large value of it make amplifier AFC flat. No resonance coil is producing much less power then resonance coil, H bridge is giving 2x more power ( post #14 ) than TC4022 but need software and for that you need to talk to Carl.

[JoyJo]

In the experiments we tried to use the driver DRV8833. The outputs were connected in parallel to achieve the required current margin. BUT in the end, we got a phase jitter of the TX signal at the output relative to the control signal. Therefore, they refused to use ready-made driver chips.

[Carl-NC]

In the experiments we tried to use the driver DRV8833.

I've never used that chip. I'm using the DRV8872 in a variable frequency VLF direct-driving a 1mH TX coil and it works well.

[Davor]

Hi guys,

What do you think the TX driving method should be like?
Garrett uses Lm386 most of the time.

I simulated Garrett LM386 Tx with a control loop, and I don't like how it uses phase to do control. Given proper conditions it breaths at VLF. IMHO amplitude limiter is the way to go about controlling amplitude. Loop methods in oscillators have stability issues.
The way to go is direct digital drive, as there should be no VLF jitter about it, whether from a MCU or a PLL, but not analogue.

[Daves]

I simulated Garrett LM386 Tx with a control loop, and I don't like how it uses phase to do control. Given proper conditions it breaths at VLF. IMHO amplitude limiter is the way to go about controlling amplitude. Loop methods in oscillators have stability issues.
The way to go is direct digital drive, as there should be no VLF jitter about it, whether from a MCU or a PLL, but not analogue.

Hi Davor,

Could you explain your suggested method (digital drive) a little more?
Any scheme etc. that you can suggest to try? Is there anything?
and I don't understand what you don't like about phase control with Lm386, can you explain if possible?

[Davor]

In reverse order...
Using phase to influence amplitude may sound reasonable because you get a change of amplitude in a process. But the relationship between phase and resulting amplitude is somewhat convoluted, and the components required for it to work, say coil, are critical. What you get is basically an off-resonance drive with a control loop that is barely stable. I don't like it.

Digital drive is anything that has an output that toggles between GND and Vcc. Say, a bridge driven by a microcontroller, or any other digital circuitry. It supplies a non-resonant coil directly, and the only extra provision must be something that will prevent DC current to flow for more than, say, a millisecond, through a Tx coil. When you observe the spectrum of such a Tx, you'll notice a peak at a base frequency, and many harmonics. You may filter out those if you want, so that is not an issue. But if you supply that Tx with well stabilised DC, and because there's no meaningful jitter, all amplitude variability is gone, and with it all the "ghost target" responses due to poor amplitude stability. Because in the digital domain you may have also the orthogonal phases prepared for Rx, the phase relationships of the I/Q paths in Rx will be exact, which leads to reliable GB and discrimination.

Resonant Tx coils are a way to recuperate energy, but so is the digital drive. When polarity is toggled, the current first flows back to the battery, and because modern MOSFETs provide very little losses, digital drive is even more efficient.

So apart from a spectrum that is rich in harmonics, which is not a problem, everything else is better than analogue.

[Daves]

I've built and used this circuit and it works fine clocked from an 80MHz Cortex M4. I wouldn't clock it from an excessively slow micro. I also AC-couple the clocks to prevent a stuck state. Most times I now use a DRV8872 which makes for a very compact TX circuit.

Hi Carl,

I buyed DRV8872 and am trying to drive with tx only. The square wave is not very smooth. Is there a setup you can recommend? I plugged the Rx into the circuit and checked if I could see a square wave, but I saw a sawtooth, triangle wave.