New Bipolar Boost TX and Front End

[Mdtoday]

I have had at least one response indicating a preference for a single board version of the Bipolar PI. After some thought, I also see a benefit to a single board. I have included all of the features that we had in the two board version and the use of the ADuM4120 for the isolated MOSFET drivers. The layout is on a 100mm x 100mm 4 layer board. The option is there to not populate the 20 bit ADC circuitry and to use the CMOD-A7 XADC (12 bit) to process the RX. There is a 3mm keepout on all edges of the top and bottom layers. Top and bottom layers are GND filled and stiched with with an array vias. A 10 pin header strip is included to bring out all of the CMOD-A7 timing signals for testing as needed. The power connection, search coil connection, probe coil connection, and RX coil connection are now via JST-VH series connectors. The power connection is a 3x1 connector, while the 3 coil connections are by 2x1 connectors. 3.2MM mounting holes are at all four corners of the PCB. There is only one feature yet to include on the board... an RF shield cage for the analog circuitry.


This layout/design needs extensive vetting/validation. PCB production is not envisioned until we are finished with our proof of concept testing of version 1.

Very nice work JL! I like it a lot.

[KingJL]

Single board version of Bipolar PI version 2

I have had at least one response indicating a preference for a single board version of the Bipolar PI. After some thought, I also see a benefit to a single board. I have included all of the features that we had in the two board version and the use of the ADuM4120 for the isolated MOSFET drivers. The layout is on a 100mm x 100mm 4 layer board. The option is there to not populate the 20 bit ADC circuitry and to use the CMOD-A7 XADC (12 bit) to process the RX. There is a 3mm keepout on all edges of the top and bottom layers. Top and bottom layers are GND filled and stiched with with an array vias. A 10 pin header strip is included to bring out all of the CMOD-A7 timing signals for testing as needed. The power connection, search coil connection, probe coil connection, and RX coil connection are now via JST-VH series connectors. The power connection is a 3x1 connector, while the 3 coil connections are by 2x1 connectors. 3.2MM mounting holes are at all four corners of the PCB. There is only one feature yet to include on the board... an RF shield cage for the analog circuitry.

The 3D veiws of the new board implementation:
Top veiw


Bottom veiw


The schematic:
Bipolar PI ADuM Single Board Schematic.pdf

The BOM:
Bipolar PI Single Board BOM.pdf

Gerbers:
Bipolar-PI_singleboard_Gerbers(1-7-2020).zip

Kicad project archive is TX_ADuM_single-board(1-7-2020).zip

This layout/design needs extensive vetting/validation. PCB production is not envisioned until we are finished with our proof of concept testing of version 1.

[KingJL]

... I will wind another test coil ready for the bench tests , the highest inductance ones I have are 412uH for the Rx of Cc coils I make, the TX is 320uH.

I've got to do that myself. I want to have a DD and a mono with TX at ~320. Haven't decided on the RX for the DD... factors to consider: (1) balanced (RX = TX) should provide the best noise/soil condition performance. (2) RX > TX probably provides slightly better RX signal in ideal conditions (if conditions are ideal, probably shouldn't be using DD).

[Mdtoday]

Looking great!! I don't think there is much to improve on... 16V difference between gate and source during "on" time and 2.41V during off. I will update the Kicad files to match.

Yes, I think it?s a good compromise

I will wind another test coil ready for the bench tests , the highest inductance ones I have are 412uH for the Rx of Cc coils I make, the TX is 320uH.

[KingJL]

Changed back C35/C38 to 2n2F from 4n7F left R35/R40 at 10Meg Ohms results attached

[ATTACH]49065[/ATTACH]

[ATTACH]49066[/ATTACH]

Looking great!! I don't think there is much to improve on... 16V difference between gate and source during "on" time and 2.41V during off. I will update the Kicad files to match.

[Mdtoday]

I'll try and post results of the latest cap (2n2F) change on my test boards later today.

Changed back C35/C38 to 2n2F from 4n7F left R35/R40 at 10Meg Ohms results attached

[Mdtoday]

Only marginally better... drive increase of ~150 mv... off time difference gate/source reduced from 1.689v to 1.609 mv.

Also tried raising 5V to 5.3V. Two reasons for trying this: (1) The differnce between gate and source during off time is caused by insufficient drive from the pulse transformer. At 5V, the EL7202 is at the bottom end of the voltage operational limits... a slight increase in voltage makes a significant (relatively) difference in drive current to the pulse transformer. (2) The plus 5V will probably have to run at about 5.3V to maintain a 5V level being supplied to the CMOD-A7 thru D1 of the I/F board. With 5V at 5.3V, the difference between gate and source during off time is reduced to 0.837 (extremely acceptable) and drive is increased by ~2V.


Meanwhile, I am loading the rest of the TX portion of the board. I am scheduled to get my 2x10 P127 female ribbon cable connectors sometime in the next 2 weeks. Also I am scheduled to receive a Hantek 2D72 about this coming Wednesday.

Looking good. for all my tests I had set the 5v supply to around the same (5.4v) so that there was 5v at the CMOS-A7. So we are at least getting consistent and similar results.

The Hantek 2D72 looks good too.

I'm off to endure some dental pain later this morning so I'll try and post results of the latest cap (2n2F) change on my test boards later today.

[KingJL]

My test board has 0 ohm shunts in the primary so it will be interesting to see what your measurements are.

Only marginally better... drive increase of ~150 mv... off time difference gate/source reduced from 1.689v to 1.609 mv.


Also tried raising 5V to 5.3V. Two reasons for trying this: (1) The differnce between gate and source during off time is caused by insufficient drive from the pulse transformer. At 5V, the EL7202 is at the bottom end of the voltage operational limits... a slight increase in voltage makes a significant (relatively) difference in drive current to the pulse transformer. (2) The plus 5V will probably have to run at about 5.3V to maintain a 5V level being supplied to the CMOD-A7 thru D1 of the I/F board. With 5V at 5.3V, the difference between gate and source during off time is reduced to 0.837 (extremely acceptable) and drive is increased by ~2V.


Meanwhile, I am loading the rest of the TX portion of the board. I am scheduled to get my 2x10 P127 female ribbon cable connectors sometime in the next 2 weeks. Also I am scheduled to receive a Hantek 2D72 about this coming Wednesday.

[Mdtoday]

As much as I hate to do it, this AM I am going to lift the toroids, replace the 1 ohm resistors in the primary circuit with 0 ohm shunts, and then re-install the toroids. Then I am leaving the circuit alone! Indications are that this could increase the primary current by as much as 50 ma and should drive the difference between gate and source to less than 1 V.

My test board has 0 ohm shunts in the primary so it will be interesting to see what your measurements are.

[KingJL]

... Further testing of the MOSFET driver levels will await until the TX components are loaded and I receive some 127mm pitch 20 pin (2x10) female ribbon connecters so I can easily get to test points on the bottom of the TX board...

As much as I hate to do it, this AM I am going to lift the toroids, replace the 1 ohm resistors in the primary circuit with 0 ohm shunts, and then re-install the toroids. Then I am leaving the circuit alone! Indications are that this could increase the primary current by as much as 50 ma and should drive the difference between gate and source to less than 1 V.

[Mdtoday]

Reloaded C29,C32,C35,C38 with 2n2. Measured drive between M1 gate and COIL+ (M1 source). I am satified that the drive from the pulse transformer drive circuitry is adequate. 5V set exactly to 5V... anything above 5V increases the drive output. I will now start loading the rest of the TX components. Further testing of the MOSFET driver levels will await until the TX components are loaded and I receive some 127mm pitch 20 pin (2x10) female ribbon connecters so I can easily get to test points on the bottom of the TX board

Your board is Looking great, I will leave the toroids on the little carrier boards for now and just change the caps back to 2n2F.
I have almost finished loading another set but will wait for further testing before I complete that one.

[KingJL]

. Checked drive to M1 (I can actually get to that signal) and got about 9V. I need to reload C29,C32,C35,C38 with 2n2 ( I had originally loaded with 5n6). Progress... (slowly)!

Reloaded C29,C32,C35,C38 with 2n2. Measured drive between M1 gate and COIL+ (M1 source). I am satified that the drive from the pulse transformer drive circuitry is adequate. 5V set exactly to 5V... anything above 5V increases the drive output. I will now start loading the rest of the TX components. Further testing of the MOSFET driver levels will await until the TX components are loaded and I receive some 127mm pitch 20 pin (2x10) female ribbon connecters so I can easily get to test points on the bottom of the TX board.

Pulse Transformer M1 MOSFET drive output:


Current status of top and bottom of TX/RX board:

[Mdtoday]

Finally got that sorted out... now have solid connection between CMOD and TX/RX board. Checked all logic signals and they are as they should be to the RX/TX board (previously I was missing TXA, BSTB_E, and DAMP_E). Checked drive to M1 (I can actually get to that signal) and got about 9V. I need to reload C29,C32,C35,C38 with 2n2 ( I had originally loaded with 5n6). Progress... (slowly)!

Great news!

[Mdtoday]

Just finished some test plots for the FQS4901... For a typical FQS4901, the value of Vth seems to be about 3.5V... With a typical MQS4901, at 2V the RDS is about 6.5Meg, 3V is about 26.5k, 3.5V is about 410 ohm, 4V is about 23 ohm, 5V is about 3.6 ohm, 6V is about 3.27 ohm, 10V is about 3.23 ohm.

For M6 and M7 the key is a difference during off of less that 3V with a drive of anything above 5V being satisfactory. At 3V during "OFF" total resistance added to the damp parth is over 100k (which is more than sufficient). With an "ON" voltage of 5V total resistance added to the damp path is ~14 ohm (6V is ~.13).

Thanks for the explaination JL, yes I see what you are saying and it makes perfect sense. I was confusing myself with getting the best logic drive signal with minimal droop and not keeping my thinking relative.
I will replace the caps with 2n2F and repost results

[KingJL]

...
I seem to have an issue with J9 or the mating connector on the I/F board... I will be sorting that out.

Finally got that sorted out... now have solid connection between CMOD and TX/RX board. Checked all logic signals and they are as they should be to the RX/TX board (previously I was missing TXA, BSTB_E, and DAMP_E). Checked drive to M1 (I can actually get to that signal) and got about 9V. I need to reload C29,C32,C35,C38 with 2n2 ( I had originally loaded with 5n6). Progress... (slowly)!