New Bipolar Boost TX and Front End

[KingJL]

I soldered some wires to the PCB to connect to a .1" female header so I can probe the relevant areas without slipping the probe off everywhere...

Slick move!!

[KingJL]

... If we have to go that route, I would like a core just a little larger OD (~20 - 24 mm) & ID (~11 - 13 mm) and a ui such that 5 turns would yield between ~70 and 90 uH primary with a 3.5: - 4:1 turns ratio.

I think I have a good candidate. I am placing an order and when I get back from my Thanksgiving day holiday trip, I will wind and test.

Another (probably a better fit as it has a higher ui due to larger cross sectional diameter) id here. In fact just placed an order. I will wind and test. With this last one 5 turns of 22 awg results in between 70uH and 92 uH (range of accuracy of permeability factor). For the secondary I will use 18-20 turns 28 - 32 awg.

[KingJL]

Thanks JL.

The shots in last post were all without coil attached.
All measurements referenced to 0V.
The primary windings did have nearly 5 turns when you see the way I looped the wire to pads, this has been changed back to 4. The wire gauge is 25 AWG. I will also change this next to .63 (22 AWG) next.

I soldered some wires to the PCB to connect to a .1" female header so I can probe the relevant areas without slipping the probe off everywhere.
Here are some shots across TX_COIL+ and TX_COIL- no coil attached...

[ATTACH]48473[/ATTACH]
[ATTACH]48474[/ATTACH]
[ATTACH]48475[/ATTACH]
[ATTACH]48476[/ATTACH]

I will rewind the transformers tomorrow and measure the gate voltages again but I do see an improvement with just reducing the windings back to 4.

I will also measure the coil current once this is done.

Great! The waveform across COIL_ & COIL+ is just about what I would expect. Now the gate drive for M3 (Boost_A_G) to ground will be riding on top of (in reference to) TX_COIL- to gnd and the gate drive for M1 (Boost_B_G) will be riding on top of (in reference to) TX_COIL+. In the mean time I am doing to see what info I can find on Amidn's site to research some aspect of toroid cores available for our purpose, just in case we have to wind our own. If we have to go that route, I would like a core just a little larger OD (~20 - 24 mm) & ID (~11 - 13 mm) and a ui such that 5 turns would yield between ~70 and 90 uH primary with a 3.5: - 4:1 turns ratio.

[Mdtoday]

I am assuming that this is with a live coil. If that assumtion is correct, would it be possible to get the same measurements without a coil connected. I suspect that the gate drive to M1 and M3 are too low. It appears that we are only getting a little over 5V effective pulse between source and gate. By observing the gates without the coil connected, we take the voltage drop across the coil out of the picture and makes things a little less dynamic. Without the coil, I would expect the source for the active boost MOSFET to be a 5V, then we can see the effective voltage between gate and source.

By looking at the pictures of the pulse transformer, it appears that your primaries are 5 turns... I think 4 would work better and give us more drive. Are the primaries 24 AWG or larger? I used 22 awg and 4 turns on the primary to get my 15v on my test jig (secondaries had 14 turns of much lighter wire just like yours). We may have to get more turns on the secondary to achieve what we need... I was hoping that would not be the case. When i was looking for available toroid inductors, I was looking for 16 - 18 turn secondaries and planned on 4 turn larger awg primaries but could not find any... had to settle for the 1.4mH 14 turn. But had in mind that we could rewind them with 16 if needed (still using 4 turn heavy guage primary).

Also if the coil is connected coud we get a shot at R9 or R11 for a guage of coil curren and profile.

I am calling it a night... good night!
Have a great day!!

Oh forgot to mention that your boards look great!

Thanks JL.

The shots in last post were all without coil attached.
All measurements referenced to 0V.
The primary windings did have nearly 5 turns when you see the way I looped the wire to pads, this has been changed back to 4. The wire gauge is 25 AWG. I will also change this next to .63 (22 AWG) next.

I soldered some wires to the PCB to connect to a .1" female header so I can probe the relevant areas without slipping the probe off everywhere.
Here are some shots across TX_COIL+ and TX_COIL- no coil attached...






I will rewind the transformers tomorrow and measure the gate voltages again but I do see an improvement with just reducing the windings back to 4.

I will also measure the coil current once this is done.

[KingJL]

Ok, found the Tek scope probe..

[ATTACH]48469[/ATTACH]
[ATTACH]48470[/ATTACH]
[ATTACH]48471[/ATTACH]
[ATTACH]48472[/ATTACH]

I am assuming that this is with a live coil. If that assumtion is correct, would it be possible to get the same measurements without a coil connected. I suspect that the gate drive to M1 and M3 are too low. It appears that we are only getting a little over 5V effective pulse between source and gate. By observing the gates without the coil connected, we take the voltage drop across the coil out of the picture and makes things a little less dynamic. Without the coil, I would expect the source for the active boost MOSFET to be a 5V, then we can see the effective voltage between gate and source.

By looking at the pictures of the pulse transformer, it appears that your primaries are 5 turns... I think 4 would work better and give us more drive. Are the primaries 24 AWG or larger? I used 22 awg and 4 turns on the primary to get my 15v on my test jig (secondaries had 14 turns of much lighter wire just like yours). We may have to get more turns on the secondary to achieve what we need... I was hoping that would not be the case. When i was looking for available toroid inductors, I was looking for 16 - 18 turn secondaries and planned on 4 turn larger awg primaries but could not find any... had to settle for the 1.4mH 14 turn. But had in mind that we could rewind them with 16 if needed (still using 4 turn heavy guage primary).

Also if the coil is connected coud we get a shot at R9 or R11 for a guage of coil curren and profile.

I am calling it a night... good night!
Have a great day!!

Oh forgot to mention that your boards look great!

[Mdtoday]

Actually the signal going to R36 might be accessable from the top side as the are 2 vias with a bridge on the top side where we had to cross anothe signal line.

Ok, found the Tek scope probe..

[Mdtoday]

You mean other things are going on in the world... wow!

Haha, yes

[KingJL]

Sorry, back at the bench now, had to go out for a while, will set up now.

You mean other things are going on in the world... wow!

[Mdtoday]

Actually the signal going to R36 might be accessable from the top side as the are 2 vias with a bridge on the top side where we had to cross anothe signal line.

Sorry, back at the bench now, had to go out for a while, will set up now.

[KingJL]

I guess I can answer that myself... no, they are on the backside also. I guess that we'll just have to wait till we fire it up and see the TX current and voltage waveforms.

Actually the signal going to R36 might be accessable from the top side as the are 2 vias with a bridge on the top side where we had to cross anothe signal line.

[KingJL]

... Is R36 or R41 accessible? ...

I guess I can answer that myself... no, they are on the backside also. I guess that we'll just have to wait till we fire it up and see the TX current and voltage waveforms.

[KingJL]

No problem... I am somewhat relieved! Can you get a scope pick of the signal at say R27? Then we could see the actual gate drive pulse... need to insure that everything looks good and that we are back on track.

That may extremely difficult, so don't waste too much time on it... Is R36 or R41 accessible? That would show the gate drive output for the damp MOSFETs... which would give high confidence to the driver operation. Those signals should be high for ~440usec and low for ~60 usec. In fact it would be more of a verification of operation than the boost signals as the high has to be maintained for such a relatively long time. I would be interested in droop.

[Mdtoday]

No problem... I am somewhat relieved! Can you get a scope pick of the signal at say R27? Then we could see the actual gate drive pulse... need to insure that everything looks good and that we are back on track.

Ok, will set up for that shortly and post.

[KingJL]

Yes, that looks very familiar, trust me to pick the wrong side!

Thanks JL, apologies for that

No problem... I am somewhat relieved! Can you get a scope pick of the signal at say R27? Then we could see the actual gate drive pulse... need to insure that everything looks good and that we are back on track.

[Mdtoday]

The pic is great... everything looks good.... Lets make sure we have a problem... The waveform pic of Boost_A_G was sampled on the Gate side of the circuit, correct? The reason i ask is if it was sampled on the secondary side of the diode array it would look like this:


Look familiar?

Yes, that looks very familiar, trust me to pick the wrong side!

Thanks JL, apologies for that