Half-bridge/gate drive heeeeelp!

[bitrex]

On 5/2/2020 1:56 PM, jlarkin@highlandsniptechnology.com wrote:

On Sat, 2 May 2020 13:05:54 -0400, bitrex <user@example.net> wrote:

On 5/2/2020 5:59 AM, piglet wrote:
On 01/05/2020 8:08 pm, bitrex wrote:
On 5/1/2020 2:47 PM, John Larkin wrote:
On Fri, 1 May 2020 14:03:15 -0400, bitrex <user@example.net> wrote:

Here's the situation in simplified schematic form, generate a
high-efficiency sine wave at around 3-5MHz putting out about 5 watts
into resistive load:

https://www.dropbox.com/s/tz1zp0d8oy8zkws/IMG_20200501_133832793.jpg?dl=0


Just drive it like two half-bridges with appropriate logic-level
wavaeforms (some dead-time) and if you look at the two switching nodes
together on a plot you get the red/black waveform there under inductor
L1. LC filter it through a coupled inductor. and Bob's your Dad's
uncle.
in theory.

First attempt at it on a protoboard:

https://www.dropbox.com/s/1n69o0yzhd2rbpk/Schem.PNG?dl=0

"it looks great in the sim" and the connection of the boost diodes to
the switching nodes gives a nice ~12-13 volt Vdd to common rail above
the lockout threshold for the flying driver's supply nodes and keeps
all
the component dissipation low.

https://www.dropbox.com/s/sv5fox0ywcmpypv/Sine.PNG?dl=0

Total miserable nightmare to get working right as a prototype, losing
gate drivers for unknown reason, these chips seem way slower than the
~50n prop delay specified in the datasheet, can't get up to speed. I
don't think this 4120 likes being driven the way it's connected here in
reality and they're dying on me, or I've made an error in the
implementation somewhere i can't readily spot.

Anyway, I hate it, looking for any suggestions or alternate gate-drive
schemes for this, I'd like a cheaper faster and more rugged one, as
someone mentioned I don't really need the ADuM4120's internal
transformer isolation if I could find a good way to level-shift the
incoming logic level drive appropriately to the negative rails the
lower
FETs have. Thanks

I can envision a lot of DC circulating in addition to switching
losses.

This will be difficult with mosfets and regular gate drivers, as you
note. You might consider something simpler, like a single mosfet or
SiC fet in a tuned class-C amp. RF stuff. One power part on a heat
sink.

You could probably use some gate drivers, like IXDN602, as the output
device itself if all you need is 5 watts. They are cheap and fast and
handy for a lot more than gate driving.


Huh. That's an interesting chip. Goes up to 35 input, too. Wonder if
they'd like being paralleled...

This is my second idea for KISS-approach:

https://www.dropbox.com/s/dewy6eqpggoagg5/IMG_20200501_145539343.jpg?dl=0


Use two non-isolated gate driver chips with like 10ns propagation
delay. One of two shown. Keep the flying-boostrap connection the same.
And "level shift" simply by capactively coupling the logic inputs.

It needs an extra two regulators to make a Vcc appropriately above the
negative supplies on the primary side. But I think it's better than
messing with four isolated drivers, I don't really need them for
isolation just level-shifting.

Why do you need the funny voltage power rails? If you're making a sine
wave let your output LC do the voltage transformation. Sounds like
you're trying to make a sort of 80meter ham band QRP transmitter but
using an SMPSU approach?

Did you rule out traditional push-pull (two low side n-ch devices) for
some reason?

piglet


Like using a center-tapped transformer, you mean?

Or at least a single positive supply.

If you are going to use discrete fets, it's easy to drive them if the
sources are all ground.

I agree, these negative rails are a pain. My alternative idea is to do
something like this, four FETs and one phase of this:

<https://www.researchgate.net/profile/Yu_Zhang368/publication/224633035/figure/fig1/AS:386501029449729@1469160434606/Three-phase-full-bridge-inverters.png>

I could keep the +25 and 25/2 rails for the top two FETs, and keep the
drive logic the same as I have it I think. Use an off the shelf L and
l:l transformer. The client would like to be able to raise and lower the
output voltage in steps by adjusting the supply voltages in
steps...maybe set a step-up ratio to the midpoint of the desired range,
actually. 50-200V P2P

Sloman will be happy to design a transformer for you.

He likes that circuit very much.

 

On Fri, 1 May 2020 14:03:15 -0400, bitrex <user@example.net> wrote:

Here's the situation in simplified schematic form, generate a
high-efficiency sine wave at around 3-5MHz putting out about 5 watts
into resistive load:

https://www.dropbox.com/s/tz1zp0d8oy8zkws/IMG_20200501_133832793.jpg?dl=0

Just drive it like two half-bridges with appropriate logic-level
wavaeforms (some dead-time) and if you look at the two switching nodes
together on a plot you get the red/black waveform there under inductor
L1. LC filter it through a coupled inductor. and Bob's your Dad's uncle.
in theory.

First attempt at it on a protoboard:

https://www.dropbox.com/s/1n69o0yzhd2rbpk/Schem.PNG?dl=0

"it looks great in the sim" and the connection of the boost diodes to
the switching nodes gives a nice ~12-13 volt Vdd to common rail above
the lockout threshold for the flying driver's supply nodes and keeps all
the component dissipation low.

https://www.dropbox.com/s/sv5fox0ywcmpypv/Sine.PNG?dl=0

Total miserable nightmare to get working right as a prototype, losing
gate drivers for unknown reason, these chips seem way slower than the
~50n prop delay specified in the datasheet, can't get up to speed. I
don't think this 4120 likes being driven the way it's connected here in
reality and they're dying on me, or I've made an error in the
implementation somewhere i can't readily spot.

Anyway, I hate it, looking for any suggestions or alternate gate-drive
schemes for this, I'd like a cheaper faster and more rugged one, as
someone mentioned I don't really need the ADuM4120's internal
transformer isolation if I could find a good way to level-shift the
incoming logic level drive appropriately to the negative rails the lower
FETs have. Thanks

Something like this maybe?

https://www.dropbox.com/s/87fj0uyt8llt6hq/5W_Sine_Waves.JPG?raw=1

If you need more voltage, you could do a Colpitts sort of step-up for
the bandpass case. Takes one more cap.

Aside, I have been burned a couple of times (literally burned) by
selecting inductors that appeared to be within their ratings, but got
hot from core loss and/or skin loss.

That's happening right now on our 300 watt class-D amp. Fortunately,
it wouldn't be awful to hang some toroids on the surface-mount
inductor footprints.





--

John Larkin Highland Technology, Inc

Science teaches us to doubt.

Claude Bernard

 

[bitrex]

On 5/2/2020 3:07 PM, jlarkin@highlandsniptechnology.com wrote:

On Fri, 1 May 2020 14:03:15 -0400, bitrex <user@example.net> wrote:

Here's the situation in simplified schematic form, generate a
high-efficiency sine wave at around 3-5MHz putting out about 5 watts
into resistive load:

https://www.dropbox.com/s/tz1zp0d8oy8zkws/IMG_20200501_133832793.jpg?dl=0

Just drive it like two half-bridges with appropriate logic-level
wavaeforms (some dead-time) and if you look at the two switching nodes
together on a plot you get the red/black waveform there under inductor
L1. LC filter it through a coupled inductor. and Bob's your Dad's uncle.
in theory.

First attempt at it on a protoboard:

https://www.dropbox.com/s/1n69o0yzhd2rbpk/Schem.PNG?dl=0

"it looks great in the sim" and the connection of the boost diodes to
the switching nodes gives a nice ~12-13 volt Vdd to common rail above
the lockout threshold for the flying driver's supply nodes and keeps all
the component dissipation low.

https://www.dropbox.com/s/sv5fox0ywcmpypv/Sine.PNG?dl=0

Total miserable nightmare to get working right as a prototype, losing
gate drivers for unknown reason, these chips seem way slower than the
~50n prop delay specified in the datasheet, can't get up to speed. I
don't think this 4120 likes being driven the way it's connected here in
reality and they're dying on me, or I've made an error in the
implementation somewhere i can't readily spot.

Anyway, I hate it, looking for any suggestions or alternate gate-drive
schemes for this, I'd like a cheaper faster and more rugged one, as
someone mentioned I don't really need the ADuM4120's internal
transformer isolation if I could find a good way to level-shift the
incoming logic level drive appropriately to the negative rails the lower
FETs have. Thanks

Something like this maybe?

https://www.dropbox.com/s/87fj0uyt8llt6hq/5W_Sine_Waves.JPG?raw=1

If you need more voltage, you could do a Colpitts sort of step-up for
the bandpass case. Takes one more cap.

Aside, I have been burned a couple of times (literally burned) by
selecting inductors that appeared to be within their ratings, but got
hot from core loss and/or skin loss.

That's happening right now on our 300 watt class-D amp. Fortunately,
it wouldn't be awful to hang some toroids on the surface-mount
inductor footprints.

Since I have the four-level drive waveforms available already how about
something like this using the IXYS chip:

<https://www.dropbox.com/s/vdsoeigilgxmbrl/IMG_20200502_161215072.jpg?dl=0>

The schottkeys aren't ideal but for low powers should be OK for now I
think. If they need more power at some point I can figure that out later
with a proper half-bridge drivers and external FETs.

 

[bitrex]

On 5/2/2020 4:25 PM, bitrex wrote:

On 5/2/2020 3:07 PM, jlarkin@highlandsniptechnology.com wrote:
On Fri, 1 May 2020 14:03:15 -0400, bitrex <user@example.net> wrote:

Here's the situation in simplified schematic form, generate a
high-efficiency sine wave at around 3-5MHz putting out about 5 watts
into resistive load:

https://www.dropbox.com/s/tz1zp0d8oy8zkws/IMG_20200501_133832793.jpg?dl=0


Just drive it like two half-bridges with appropriate logic-level
wavaeforms (some dead-time) and if you look at the two switching nodes
together on a plot you get the red/black waveform there under inductor
L1. LC filter it through a coupled inductor. and Bob's your Dad's uncle.
in theory.

First attempt at it on a protoboard:

https://www.dropbox.com/s/1n69o0yzhd2rbpk/Schem.PNG?dl=0

"it looks great in the sim" and the connection of the boost diodes to
the switching nodes gives a nice ~12-13 volt Vdd to common rail above
the lockout threshold for the flying driver's supply nodes and keeps all
the component dissipation low.

https://www.dropbox.com/s/sv5fox0ywcmpypv/Sine.PNG?dl=0

Total miserable nightmare to get working right as a prototype, losing
gate drivers for unknown reason, these chips seem way slower than the
~50n prop delay specified in the datasheet, can't get up to speed. I
don't think this 4120 likes being driven the way it's connected here in
reality and they're dying on me, or I've made an error in the
implementation somewhere i can't readily spot.

Anyway, I hate it, looking for any suggestions or alternate gate-drive
schemes for this, I'd like a cheaper faster and more rugged one, as
someone mentioned I don't really need the ADuM4120's internal
transformer isolation if I could find a good way to level-shift the
incoming logic level drive appropriately to the negative rails the lower
FETs have. Thanks

Something like this maybe?

https://www.dropbox.com/s/87fj0uyt8llt6hq/5W_Sine_Waves.JPG?raw=1

If you need more voltage, you could do a Colpitts sort of step-up for
the bandpass case. Takes one more cap.

Aside, I have been burned a couple of times (literally burned) by
selecting inductors that appeared to be within their ratings, but got
hot from core loss and/or skin loss.

That's happening right now on our 300 watt class-D amp. Fortunately,
it wouldn't be awful to hang some toroids on the surface-mount
inductor footprints.



Since I have the four-level drive waveforms available already how about
something like this using the IXYS chip:

https://www.dropbox.com/s/vdsoeigilgxmbrl/IMG_20200502_161215072.jpg?dl=0

The schottkeys aren't ideal but for low powers should be OK for now I
think. If they need more power at some point I can figure that out later
with a proper half-bridge drivers and external FETs.

Oops, those should both be 25 volt rails, sorry

 

On Sat, 2 May 2020 16:25:57 -0400, bitrex <user@example.net> wrote:

On 5/2/2020 3:07 PM, jlarkin@highlandsniptechnology.com wrote:
On Fri, 1 May 2020 14:03:15 -0400, bitrex <user@example.net> wrote:

Here's the situation in simplified schematic form, generate a
high-efficiency sine wave at around 3-5MHz putting out about 5 watts
into resistive load:

https://www.dropbox.com/s/tz1zp0d8oy8zkws/IMG_20200501_133832793.jpg?dl=0

Just drive it like two half-bridges with appropriate logic-level
wavaeforms (some dead-time) and if you look at the two switching nodes
together on a plot you get the red/black waveform there under inductor
L1. LC filter it through a coupled inductor. and Bob's your Dad's uncle.
in theory.

First attempt at it on a protoboard:

https://www.dropbox.com/s/1n69o0yzhd2rbpk/Schem.PNG?dl=0

"it looks great in the sim" and the connection of the boost diodes to
the switching nodes gives a nice ~12-13 volt Vdd to common rail above
the lockout threshold for the flying driver's supply nodes and keeps all
the component dissipation low.

https://www.dropbox.com/s/sv5fox0ywcmpypv/Sine.PNG?dl=0

Total miserable nightmare to get working right as a prototype, losing
gate drivers for unknown reason, these chips seem way slower than the
~50n prop delay specified in the datasheet, can't get up to speed. I
don't think this 4120 likes being driven the way it's connected here in
reality and they're dying on me, or I've made an error in the
implementation somewhere i can't readily spot.

Anyway, I hate it, looking for any suggestions or alternate gate-drive
schemes for this, I'd like a cheaper faster and more rugged one, as
someone mentioned I don't really need the ADuM4120's internal
transformer isolation if I could find a good way to level-shift the
incoming logic level drive appropriately to the negative rails the lower
FETs have. Thanks

Something like this maybe?

https://www.dropbox.com/s/87fj0uyt8llt6hq/5W_Sine_Waves.JPG?raw=1

If you need more voltage, you could do a Colpitts sort of step-up for
the bandpass case. Takes one more cap.

Aside, I have been burned a couple of times (literally burned) by
selecting inductors that appeared to be within their ratings, but got
hot from core loss and/or skin loss.

That's happening right now on our 300 watt class-D amp. Fortunately,
it wouldn't be awful to hang some toroids on the surface-mount
inductor footprints.



Since I have the four-level drive waveforms available already how about
something like this using the IXYS chip:

https://www.dropbox.com/s/vdsoeigilgxmbrl/IMG_20200502_161215072.jpg?dl=0

The schottkeys aren't ideal but for low powers should be OK for now I
think. If they need more power at some point I can figure that out later
with a proper half-bridge drivers and external FETs.

Seems awfully complex to make a sine wave.

The IXYS part claims anti-shoot-thru features, but it might get hot at
high supply voltage and high clock rate. 5 MHz and, say, 25 volts
might not work.

Do you want 100V p-p sine at 5 MHz? Sounds like RF to me. Class C amp
maybe. 50 volt supply, one fet, LC tank.



--

John Larkin Highland Technology, Inc

Science teaches us to doubt.

Claude Bernard

 

[Lasse Langwadt Christense]

lørdag den 2. maj 2020 kl. 22.46.01 UTC+2 skrev jla...@highlandsniptechnology.com:

On Sat, 2 May 2020 16:25:57 -0400, bitrex <user@example.net> wrote:

On 5/2/2020 3:07 PM, jlarkin@highlandsniptechnology.com wrote:
On Fri, 1 May 2020 14:03:15 -0400, bitrex <user@example.net> wrote:

Here's the situation in simplified schematic form, generate a
high-efficiency sine wave at around 3-5MHz putting out about 5 watts
into resistive load:

https://www.dropbox.com/s/tz1zp0d8oy8zkws/IMG_20200501_133832793.jpg?dl=0

Just drive it like two half-bridges with appropriate logic-level
wavaeforms (some dead-time) and if you look at the two switching nodes
together on a plot you get the red/black waveform there under inductor
L1. LC filter it through a coupled inductor. and Bob's your Dad's uncle.
in theory.

First attempt at it on a protoboard:

https://www.dropbox.com/s/1n69o0yzhd2rbpk/Schem.PNG?dl=0

"it looks great in the sim" and the connection of the boost diodes to
the switching nodes gives a nice ~12-13 volt Vdd to common rail above
the lockout threshold for the flying driver's supply nodes and keeps all
the component dissipation low.

https://www.dropbox.com/s/sv5fox0ywcmpypv/Sine.PNG?dl=0

Total miserable nightmare to get working right as a prototype, losing
gate drivers for unknown reason, these chips seem way slower than the
~50n prop delay specified in the datasheet, can't get up to speed. I
don't think this 4120 likes being driven the way it's connected here in
reality and they're dying on me, or I've made an error in the
implementation somewhere i can't readily spot.

Anyway, I hate it, looking for any suggestions or alternate gate-drive
schemes for this, I'd like a cheaper faster and more rugged one, as
someone mentioned I don't really need the ADuM4120's internal
transformer isolation if I could find a good way to level-shift the
incoming logic level drive appropriately to the negative rails the lower
FETs have. Thanks

Something like this maybe?

https://www.dropbox.com/s/87fj0uyt8llt6hq/5W_Sine_Waves.JPG?raw=1

If you need more voltage, you could do a Colpitts sort of step-up for
the bandpass case. Takes one more cap.

Aside, I have been burned a couple of times (literally burned) by
selecting inductors that appeared to be within their ratings, but got
hot from core loss and/or skin loss.

That's happening right now on our 300 watt class-D amp. Fortunately,
it wouldn't be awful to hang some toroids on the surface-mount
inductor footprints.



Since I have the four-level drive waveforms available already how about
something like this using the IXYS chip:

https://www.dropbox.com/s/vdsoeigilgxmbrl/IMG_20200502_161215072.jpg?dl=0

The schottkeys aren't ideal but for low powers should be OK for now I
think. If they need more power at some point I can figure that out later
with a proper half-bridge drivers and external FETs.

Seems awfully complex to make a sine wave.

The IXYS part claims anti-shoot-thru features, but it might get hot at
high supply voltage and high clock rate. 5 MHz and, say, 25 volts
might not work.

Do you want 100V p-p sine at 5 MHz? Sounds like RF to me. Class C amp
maybe. 50 volt supply, one fet, LC tank.

https://www.nxp.com/products/rf/rf-power/rf-ism-and-broadcast/1-600-mhz-broadcast-and-ism/100-w-cw-over-1-8-250-mhz-50-v-wideband-rf-power-ldmos-transistor:MRF101AN

https://www.nxp.com/products/rf/rf-power/rf-ism-and-broadcast/1-600-mhz-broadcast-and-ism/300-w-cw-over-1-8-250-mhz-50-v-wideband-rf-power-ldmos-transistor:MRF300AN

 

[bitrex]

On 5/2/2020 4:45 PM, jlarkin@highlandsniptechnology.com wrote:

On Sat, 2 May 2020 16:25:57 -0400, bitrex <user@example.net> wrote:

On 5/2/2020 3:07 PM, jlarkin@highlandsniptechnology.com wrote:
On Fri, 1 May 2020 14:03:15 -0400, bitrex <user@example.net> wrote:

Here's the situation in simplified schematic form, generate a
high-efficiency sine wave at around 3-5MHz putting out about 5 watts
into resistive load:

https://www.dropbox.com/s/tz1zp0d8oy8zkws/IMG_20200501_133832793.jpg?dl=0

Just drive it like two half-bridges with appropriate logic-level
wavaeforms (some dead-time) and if you look at the two switching nodes
together on a plot you get the red/black waveform there under inductor
L1. LC filter it through a coupled inductor. and Bob's your Dad's uncle.
in theory.

First attempt at it on a protoboard:

https://www.dropbox.com/s/1n69o0yzhd2rbpk/Schem.PNG?dl=0

"it looks great in the sim" and the connection of the boost diodes to
the switching nodes gives a nice ~12-13 volt Vdd to common rail above
the lockout threshold for the flying driver's supply nodes and keeps all
the component dissipation low.

https://www.dropbox.com/s/sv5fox0ywcmpypv/Sine.PNG?dl=0

Total miserable nightmare to get working right as a prototype, losing
gate drivers for unknown reason, these chips seem way slower than the
~50n prop delay specified in the datasheet, can't get up to speed. I
don't think this 4120 likes being driven the way it's connected here in
reality and they're dying on me, or I've made an error in the
implementation somewhere i can't readily spot.

Anyway, I hate it, looking for any suggestions or alternate gate-drive
schemes for this, I'd like a cheaper faster and more rugged one, as
someone mentioned I don't really need the ADuM4120's internal
transformer isolation if I could find a good way to level-shift the
incoming logic level drive appropriately to the negative rails the lower
FETs have. Thanks

Something like this maybe?

https://www.dropbox.com/s/87fj0uyt8llt6hq/5W_Sine_Waves.JPG?raw=1

If you need more voltage, you could do a Colpitts sort of step-up for
the bandpass case. Takes one more cap.

Aside, I have been burned a couple of times (literally burned) by
selecting inductors that appeared to be within their ratings, but got
hot from core loss and/or skin loss.

That's happening right now on our 300 watt class-D amp. Fortunately,
it wouldn't be awful to hang some toroids on the surface-mount
inductor footprints.



Since I have the four-level drive waveforms available already how about
something like this using the IXYS chip:

https://www.dropbox.com/s/vdsoeigilgxmbrl/IMG_20200502_161215072.jpg?dl=0

The schottkeys aren't ideal but for low powers should be OK for now I
think. If they need more power at some point I can figure that out later
with a proper half-bridge drivers and external FETs.

Seems awfully complex to make a sine wave.

The IXYS part claims anti-shoot-thru features, but it might get hot at
high supply voltage and high clock rate. 5 MHz and, say, 25 volts
might not work.

Is there a similar chip with thermal protection? I built in plenty of
dead-time to the drive waveforms. But with a light load that the drivers
will cook off.

Do you want 100V p-p sine at 5 MHz? Sounds like RF to me. Class C amp
maybe. 50 volt supply, one fet, LC tank.

Can't do a narrowband resonant tank, the frequency needs adjusting in
steps, the circuit is clocked.

 

[bitrex]

On 5/2/2020 6:07 PM, bitrex wrote:

On 5/2/2020 4:45 PM, jlarkin@highlandsniptechnology.com wrote:
On Sat, 2 May 2020 16:25:57 -0400, bitrex <user@example.net> wrote:

On 5/2/2020 3:07 PM, jlarkin@highlandsniptechnology.com wrote:
On Fri, 1 May 2020 14:03:15 -0400, bitrex <user@example.net> wrote:

Here's the situation in simplified schematic form, generate a
high-efficiency sine wave at around 3-5MHz putting out about 5 watts
into resistive load:

https://www.dropbox.com/s/tz1zp0d8oy8zkws/IMG_20200501_133832793.jpg?dl=0


Just drive it like two half-bridges with appropriate logic-level
wavaeforms (some dead-time) and if you look at the two switching nodes
together on a plot you get the red/black waveform there under inductor
L1. LC filter it through a coupled inductor. and Bob's your Dad's
uncle.
in theory.

First attempt at it on a protoboard:

https://www.dropbox.com/s/1n69o0yzhd2rbpk/Schem.PNG?dl=0

"it looks great in the sim" and the connection of the boost diodes to
the switching nodes gives a nice ~12-13 volt Vdd to common rail above
the lockout threshold for the flying driver's supply nodes and
keeps all
the component dissipation low.

https://www.dropbox.com/s/sv5fox0ywcmpypv/Sine.PNG?dl=0

Total miserable nightmare to get working right as a prototype, losing
gate drivers for unknown reason, these chips seem way slower than the
~50n prop delay specified in the datasheet, can't get up to speed. I
don't think this 4120 likes being driven the way it's connected
here in
reality and they're dying on me, or I've made an error in the
implementation somewhere i can't readily spot.

Anyway, I hate it, looking for any suggestions or alternate gate-drive
schemes for this, I'd like a cheaper faster and more rugged one, as
someone mentioned I don't really need the ADuM4120's internal
transformer isolation if I could find a good way to level-shift the
incoming logic level drive appropriately to the negative rails the
lower
FETs have. Thanks

Something like this maybe?

https://www.dropbox.com/s/87fj0uyt8llt6hq/5W_Sine_Waves.JPG?raw=1

If you need more voltage, you could do a Colpitts sort of step-up for
the bandpass case. Takes one more cap.

Aside, I have been burned a couple of times (literally burned) by
selecting inductors that appeared to be within their ratings, but got
hot from core loss and/or skin loss.

That's happening right now on our 300 watt class-D amp. Fortunately,
it wouldn't be awful to hang some toroids on the surface-mount
inductor footprints.



Since I have the four-level drive waveforms available already how about
something like this using the IXYS chip:

https://www.dropbox.com/s/vdsoeigilgxmbrl/IMG_20200502_161215072.jpg?dl=0


The schottkeys aren't ideal but for low powers should be OK for now I
think. If they need more power at some point I can figure that out later
with a proper half-bridge drivers and external FETs.

Seems awfully complex to make a sine wave.

The IXYS part claims anti-shoot-thru features, but it might get hot at
high supply voltage and high clock rate. 5 MHz and, say, 25 volts
might not work.

Is there a similar chip with thermal protection? I built in plenty of
dead-time to the drive waveforms. But with a light load that the drivers
will cook off.

Correction, not light load I mean heavy load/short circuit

Do you want 100V p-p sine at 5 MHz? Sounds like RF to me. Class C amp
maybe. 50 volt supply, one fet, LC tank.


Can't do a narrowband resonant tank, the frequency needs adjusting in
steps, the circuit is clocked.

 

[bitrex]

On 5/2/2020 6:11 PM, bitrex wrote:

On 5/2/2020 6:07 PM, bitrex wrote:
On 5/2/2020 4:45 PM, jlarkin@highlandsniptechnology.com wrote:
On Sat, 2 May 2020 16:25:57 -0400, bitrex <user@example.net> wrote:

On 5/2/2020 3:07 PM, jlarkin@highlandsniptechnology.com wrote:
On Fri, 1 May 2020 14:03:15 -0400, bitrex <user@example.net> wrote:

Here's the situation in simplified schematic form, generate a
high-efficiency sine wave at around 3-5MHz putting out about 5 watts
into resistive load:

https://www.dropbox.com/s/tz1zp0d8oy8zkws/IMG_20200501_133832793.jpg?dl=0


Just drive it like two half-bridges with appropriate logic-level
wavaeforms (some dead-time) and if you look at the two switching
nodes
together on a plot you get the red/black waveform there under
inductor
L1. LC filter it through a coupled inductor. and Bob's your Dad's
uncle.
in theory.

First attempt at it on a protoboard:

https://www.dropbox.com/s/1n69o0yzhd2rbpk/Schem.PNG?dl=0

"it looks great in the sim" and the connection of the boost diodes to
the switching nodes gives a nice ~12-13 volt Vdd to common rail above
the lockout threshold for the flying driver's supply nodes and
keeps all
the component dissipation low.

https://www.dropbox.com/s/sv5fox0ywcmpypv/Sine.PNG?dl=0

Total miserable nightmare to get working right as a prototype, losing
gate drivers for unknown reason, these chips seem way slower than the
~50n prop delay specified in the datasheet, can't get up to speed. I
don't think this 4120 likes being driven the way it's connected
here in
reality and they're dying on me, or I've made an error in the
implementation somewhere i can't readily spot.

Anyway, I hate it, looking for any suggestions or alternate
gate-drive
schemes for this, I'd like a cheaper faster and more rugged one, as
someone mentioned I don't really need the ADuM4120's internal
transformer isolation if I could find a good way to level-shift the
incoming logic level drive appropriately to the negative rails the
lower
FETs have. Thanks

Something like this maybe?

https://www.dropbox.com/s/87fj0uyt8llt6hq/5W_Sine_Waves.JPG?raw=1

If you need more voltage, you could do a Colpitts sort of step-up for
the bandpass case. Takes one more cap.

Aside, I have been burned a couple of times (literally burned) by
selecting inductors that appeared to be within their ratings, but got
hot from core loss and/or skin loss.

That's happening right now on our 300 watt class-D amp. Fortunately,
it wouldn't be awful to hang some toroids on the surface-mount
inductor footprints.



Since I have the four-level drive waveforms available already how about
something like this using the IXYS chip:

https://www.dropbox.com/s/vdsoeigilgxmbrl/IMG_20200502_161215072.jpg?dl=0


The schottkeys aren't ideal but for low powers should be OK for now I
think. If they need more power at some point I can figure that out
later
with a proper half-bridge drivers and external FETs.

Seems awfully complex to make a sine wave.

The IXYS part claims anti-shoot-thru features, but it might get hot at
high supply voltage and high clock rate. 5 MHz and, say, 25 volts
might not work.

Is there a similar chip with thermal protection? I built in plenty of
dead-time to the drive waveforms. But with a light load that the
drivers will cook off.

Correction, not light load I mean heavy load/short circuit

Actually there are probs at both the way I've drawn mine. Unloaded it'll
act like a flyback. I'd better put an R in parallel with the C to damp
it if the secondary is unloadded

 

[Jon Elson]

bitrex wrote:

Here's the situation in simplified schematic form, generate a
high-efficiency sine wave at around 3-5MHz putting out about 5 watts
into resistive load:

You want to generate a 5 MHz sine wave with PWM? What kind of PWM frequency
would you need for that? 50 MHz?

Look up the gate driver specs for dissipation, and you will likely see NONE
of them can handle charging and discharging a FET gate at much over 100 KHz
rates without running very hot.

One other thing I ran into while designing full bridge PWM amps was that the
FET body diodes take microseconds to turn on, but when the high-side
transistor cuts off, current flowing out through the inductor needs to be
supplied from somewhere. The body diodes can easily sit there for several
microseconds with a forward bias of 7 - 12 V without conducting at all.
So, I had to add an ultrafast diode across the low-side transistor to
prevent the common node from going too far negative. Many gate drivers can
only handle the common node going so far below ground before they pop.

Jon

 

[Michael Terrell]

On Saturday, May 2, 2020 at 9:37:48 PM UTC-4, Jon Elson wrote:

bitrex wrote:

Here's the situation in simplified schematic form, generate a
high-efficiency sine wave at around 3-5MHz putting out about 5 watts
into resistive load:
You want to generate a 5 MHz sine wave with PWM? What kind of PWM frequency
would you need for that? 50 MHz?

Look up the gate driver specs for dissipation, and you will likely see NONE
of them can handle charging and discharging a FET gate at much over 100 KHz
rates without running very hot.

One other thing I ran into while designing full bridge PWM amps was that the
FET body diodes take microseconds to turn on, but when the high-side
transistor cuts off, current flowing out through the inductor needs to be
supplied from somewhere. The body diodes can easily sit there for several
microseconds with a forward bias of 7 - 12 V without conducting at all.
So, I had to add an ultrafast diode across the low-side transistor to
prevent the common node from going too far negative. Many gate drivers can
only handle the common node going so far below ground before they pop.

Harris Broadcast has built modular AM transmitters that use PWM for modulation, The BCB models that I've seen were about 250 W per module, with 20 modules for the 5KW transmitter used by WQBQ in Leesburg Florida. It was a first generation transmitter digital. It had the optional Microprocessor board that monitored all of the modules, and could email the engineer when there was a problem. I understand it was replaced after 20 years with the latest version, because spare parts were hard to locate. The original RF modules used TO-3 packages that quickly became obsolete, even though Harris Semiconductor had manufactured them. A complex array of dividers and combiners allowed a failed module to be turned off until it could be replaced. That transmitter replaced a worn out Gates 5KW tube based BC5000 series that I removed from the building to make room for them to move their studios back to the transmitter site. It was full of old oil capacitors filled with PCB based transformer oil. It was five times the size of the new transmitter, and it printed out the required FCC data log, so no on site operator was needed. That Gates transmitter went to another local station for spare parts. (WLBE)

 

[bitrex]

On 5/2/2020 9:37 PM, Jon Elson wrote:

bitrex wrote:

Here's the situation in simplified schematic form, generate a
high-efficiency sine wave at around 3-5MHz putting out about 5 watts
into resistive load:
You want to generate a 5 MHz sine wave with PWM? What kind of PWM frequency
would you need for that? 50 MHz?

Look up the gate driver specs for dissipation, and you will likely see NONE
of them can handle charging and discharging a FET gate at much over 100 KHz
rates without running very hot.

You can go much faster than that nowadays, 100kHz isn't even that fast
for a modern buck converter. Pretty slow

One other thing I ran into while designing full bridge PWM amps was that the
FET body diodes take microseconds to turn on, but when the high-side
transistor cuts off, current flowing out through the inductor needs to be
supplied from somewhere. The body diodes can easily sit there for several
microseconds with a forward bias of 7 - 12 V without conducting at all.
So, I had to add an ultrafast diode across the low-side transistor to
prevent the common node from going too far negative. Many gate drivers can
only handle the common node going so far below ground before they pop.

Jon

50MHz PWM? no, they're being driven directly at the lower frequency,
switching the four PSU rails into an LC filter, like firing pistons in a
4-cylinder engine, consider:

<https://www.dropbox.com/s/tz1zp0d8oy8zkws/IMG_20200501_133832793.jpg?dl=0>

(there should be a cap in line with the inductors to eliminate offset
problems in the absence of some kind of active feedback that keeps the
switching nodes locked together)

No PWM here