Weird: GaN-Fet for dc/dc converters with s-parameters

[Gerhard Hoffmann]

<
https://epc-co.com/epc/Portals/0/epc/documents/datasheets/EPC8002_datasheet.pdf
<
https://epc-co.com/epc/Portals/0/epc/documents/product-training/AN015%20eGaN%20FETs%20for%20Multi-Megahertz%20Applications.pdf
>

Way to LTspice model:
< https://epc-co.com/epc/Products/eGaNFETsandICs/EPC8002.aspx >


I'd like to abuse them as analog switches because of their
low resistance (thermal noise) and low capacitance in a
chopper amplifier. Any drawbacks I did not see?
The hope is that they still work at VDS = -100mV as a switch.

The ADG819BRTZ has a pretty minimum of charge injection
when I adjust VCC & VEE carefully but still too much.

cheers,

Gerhard

 

[John Larkin]

On Fri, 2 Aug 2019 16:07:47 +0200, Gerhard Hoffmann <dk4xp@arcor.de>
wrote:

https://epc-co.com/epc/Portals/0/epc/documents/datasheets/EPC8002_datasheet.pdf



https://epc-co.com/epc/Portals/0/epc/documents/product-training/AN015%20eGaN%20FETs%20for%20Multi-Megahertz%20Applications.pdf


Way to LTspice model:
https://epc-co.com/epc/Products/eGaNFETsandICs/EPC8002.aspx


I'd like to abuse them as analog switches because of their
low resistance (thermal noise) and low capacitance in a
chopper amplifier. Any drawbacks I did not see?
The hope is that they still work at VDS = -100mV as a switch.

The ADG819BRTZ has a pretty minimum of charge injection
when I adjust VCC & VEE carefully but still too much.

cheers,

Gerhard

The EPC parts have a sort of substrate diode, which conducts if the
drain gets about 2 volts below the source, which limits voltage range
if used as a series analog switch. The gate also conducts to the
source a fair amount, in both directions. A little negative Vds should
be OK.

It would be OK as an analog switch if the source signal is stiff and
you can tolerate the polarity limits.

Cg-d is microscopic. Cg-s isn't.

The 8002 is a biggish part. Would a smaller one, like EPC2038, work
for you? Much less capacitance.

EPC is cool because they have LT Spice example circuits that just
work, and don't lie very much.

These parts are basically impossible to connect to for testing, and
amazingly EPC doesn't offer adapter boards. We had to make our own to
evaluate the parts.

https://www.dropbox.com/s/iwsfcm5jryuoxev/GaN_Boards.JPG?raw=1




--

John Larkin Highland Technology, Inc

lunatic fringe electronics

 

[John Larkin]

On Fri, 02 Aug 2019 08:32:11 -0700, John Larkin
<jjlarkin@highlandtechnology.com> wrote:

On Fri, 2 Aug 2019 16:07:47 +0200, Gerhard Hoffmann <dk4xp@arcor.de
wrote:



https://epc-co.com/epc/Portals/0/epc/documents/datasheets/EPC8002_datasheet.pdf



https://epc-co.com/epc/Portals/0/epc/documents/product-training/AN015%20eGaN%20FETs%20for%20Multi-Megahertz%20Applications.pdf


Way to LTspice model:
https://epc-co.com/epc/Products/eGaNFETsandICs/EPC8002.aspx


I'd like to abuse them as analog switches because of their
low resistance (thermal noise) and low capacitance in a
chopper amplifier. Any drawbacks I did not see?
The hope is that they still work at VDS = -100mV as a switch.

The ADG819BRTZ has a pretty minimum of charge injection
when I adjust VCC & VEE carefully but still too much.

cheers,

Gerhard

The EPC parts have a sort of substrate diode, which conducts if the
drain gets about 2 volts below the source, which limits voltage range
if used as a series analog switch. The gate also conducts to the
source a fair amount, in both directions. A little negative Vds should
be OK.

It would be OK as an analog switch if the source signal is stiff and
you can tolerate the polarity limits.

Cg-d is microscopic. Cg-s isn't.

The 8002 is a biggish part. Would a smaller one, like EPC2038, work
for you? Much less capacitance.

EPC is cool because they have LT Spice example circuits that just
work, and don't lie very much.

These parts are basically impossible to connect to for testing, and
amazingly EPC doesn't offer adapter boards. We had to make our own to
evaluate the parts.

https://www.dropbox.com/s/iwsfcm5jryuoxev/GaN_Boards.JPG?raw=1

Note that one fet, on the left, is broken. These parts are very
fragile, like little delicate blocks of glass.




--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

[Gerhard Hoffmann]

Am 02.08.19 um 21:12 schrieb John Larkin:

The EPC parts have a sort of substrate diode, which conducts if the
drain gets about 2 volts below the source, which limits voltage range
if used as a series analog switch. The gate also conducts to the
source a fair amount, in both directions. A little negative Vds should
be OK.

The 2 volts should be no problem. It's on the input side of an
amplifier, after all.

It would be OK as an analog switch if the source signal is stiff and
you can tolerate the polarity limits.

Cg-d is microscopic. Cg-s isn't.

The 8002 is a biggish part. Would a smaller one, like EPC2038, work
for you? Much less capacitance.

I've put it on my order list, together with the 2037 (next larger).

thanks!, Gerhard