Experimenting with Half Bridge Bi-Polar Transmitters

[SaltyDog]

Sorry to bump ... anyone have an answer to my phasing question?

[SaltyDog]

@Altra
Having a go at the coil. I presume you have two windings (two D's) that have a TX and RX wound together.
How are they phased? I presume TX on each coil is connected so that they add in the centre? What about the RX, how are they wired together, cancelling?

Cheers

[Altra]

I used a self made coil. A good pre-made coil is the ML dd commander for gpx or sd. The TX circuit is about the same, I changed the fet driver to a TC4428 which has an inverting channel for the highside mosfet. Also increased the TX capactor to about 4.4uf for a wider pulse. I found everything very flexible, most important is adjustable TX pulse to tune the wave form.

[SaltyDog]

@Altra

Thanks. Have a good break. You are using the same TX circuit as per your PDF document?
And lastly, what brand IB coil did you use (If you can remember)

Cheers

[Altra]

Hi Salty, I am away until the end of January so can't help much. The video is using a induction balanced coil. TX 300uh and Rx about 400uh in a dd config

[SaltyDog]

@Altra

I have built the 1/2 bridge and a preamp, but I don't get the same results as you do in the video https://www.youtube.com/watch?v=M4tSrN7QcU4
I notice that video uses a seperate TX & RX coil, what else is different, in the video, than the setup you showed in the pdf of the 1st post?

Cheers

[green]

Hi All,

Just a question on the timing signals for the two 1/2 bridge FET's.
I take it one is pulsed on, then "shortly" after the other one is pulse on ... Then, I take it the there is a wait for the rest of the time to get your overall frequency?

I.E. I am imagining a 100us pulse to the top FET followed by both FET's being off for 100us, then the bottom FET pulsed for 100us, then both FET's off for 800us, to give an overall 1khz frequency..
Is this what is intended?

I'm thinking the second pulse starts at 500us for 1kHz frequency.

[SaltyDog]

Hi All,

Just a question on the timing signals for the two 1/2 bridge FET's.
I take it one is pulsed on, then "shortly" after the other one is pulse on ... Then, I take it the there is a wait for the rest of the time to get your overall frequency?

I.E. I am imagining a 100us pulse to the top FET followed by both FET's being off for 100us, then the bottom FET pulsed for 100us, then both FET's off for 800us, to give an overall 1khz frequency..
Is this what is intended?

[SaltyDog]

Agree, the OPA847 is not very good re common mode voltage range..

I will order the AD8031 and LMH6624 to try, they both are low noise and good gain/BW ratio devices. (I have a baseline NE5534 as well)

Cheers

[dbanner]

Check with the datasheet. This thing about common mode input voltage range and hopefully, rail to rail performance. You're gonna be really appreciative by the opamp you choose.
Don't be hung up by the opamp's typical characteristics. Go by it's min and Max values. Even so, some are better than others. it's bandwidth product, etc..

[SaltyDog]

You want the clamp to be really fast and I suspect that LEDs will be really slow.



Those 1M resistors form a 180nV/rtHz noise generator. So I would use an external amp stage before the ADC. Also, looks to me that the internal opamp has a range of only 1:4.8 and is mostly for setting the input range.

All-in-all, it will probably work and may have reasonable performance. I'd guess it to be about equivalent to a 12-bit ADC preceded by a G=100 preamp.

Ok, thanks for the response ..

I will add an OPA847 before the ADC with a gain of 10, that should give me 2uV resolution LSB.

Cheers

[Carl-NC]

Yes, but it does not need to be as low as +- 0.6v !! The ADC I specified has inputs that can take +-20V, so it only has to be limited less than that .. it also has input protection diodes and 1M input resistors .. not a problem.
That is why I have used LED's so as to get more dynamic range .. and soft clipping ..

You want the clamp to be really fast and I suspect that LEDs will be really slow.

No, if you had read the data sheet of the ADC specified, you will see it has input resistors of 1M and protection diodes .. AND has an included op-amp to give gain ..

Those 1M resistors form a 180nV/rtHz noise generator. So I would use an external amp stage before the ADC. Also, looks to me that the internal opamp has a range of only 1:4.8 and is mostly for setting the input range.

All-in-all, it will probably work and may have reasonable performance. I'd guess it to be about equivalent to a 12-bit ADC preceded by a G=100 preamp.

[SaltyDog]

To add to Qiaozhi's reply, you will need to limit the current to the ADC to something typically less than 10mA so a series resistor will be required, most likely around 1k ohms, so from my past experience with trying this, there is not much to be gained going direct into the ADC.
I would be using a x10 opamp as ADC driver which serves to protect the ADC and give it better dynamic performance.

cheers

No, if you had read the data sheet of the ADC specified, you will see it has input resistors of 1M and protection diodes .. AND has an included op-amp to give gain ..

[SaltyDog]

The diodes are there to limit the high flyback voltage from the coil and hence prevent it from destroying the opamp. You have to wait for the final portion of the decay before sampling, otherwise the target signal will be swamped by the remnants of the flyback. The signal may only be a few uV for deep or miniscule targets.

Yes, but it does not need to be as low as +- 0.6v !! The ADC I specified has inputs that can take +-20V, so it only has to be limited less than that .. it also has input protection diodes and 1M input resistors .. not a problem.
That is why I have used LED's so as to get more dynamic range .. and soft clipping ..

[Mdtoday]

To add to Qiaozhi's reply, you will need to limit the current to the ADC to something typically less than 10mA so a series resistor will be required, most likely around 1k ohms, so from my past experience with trying this, there is not much to be gained going direct into the ADC.
I would be using a x10 opamp as ADC driver which serves to protect the ADC and give it better dynamic performance.

cheers