EMRFD Message Archive 8307

Message Date From Subject
8307 2013-03-01 10:08:30 Ashhar Farhan current commutating mixer's termination senstivity
i am playing around with KISS configuration of current commutating
mixers (like the one described by chris trask in a paper in our file's
area). i am building this with discrete MOSFETs. there is no testing
gear (yet) to test mixer IMD. Which is why I am asking this question :
As I understand it, by comparison, a diode mixer has diodes that carry
the oscillator current as well as the signal current and hence, the
IMD increases if the output ports reflect the signals back into the
mixer and they go through the diodes again. however, using FETs as
switching elements should not cause any non-linear mixing of signals
that can create IMD.
Am I right to assume then, that following up a FET switching mixer
directly with a crystal filter will not degrade the performance too
much as compared to a diode mixer?

- farhan
8308 2013-03-01 12:27:24 Chris Trask Re: current commutating mixer's termination senstivity
>
>i am playing around with KISS configuration of current commutating
>mixers (like the one described by chris trask in a paper in our file's
>area). i am building this with discrete MOSFETs. there is no testing
>gear (yet) to test mixer IMD. Which is why I am asking this question :
>As I understand it, by comparison, a diode mixer has diodes that carry
>the oscillator current as well as the signal current and hence, the
>IMD increases if the output ports reflect the signals back into the
>mixer and they go through the diodes again. however, using FETs as
>switching elements should not cause any non-linear mixing of signals
>that can create IMD.
>
>Am I right to assume then, that following up a FET switching mixer
>directly with a crystal filter will not degrade the performance too
>much as compared to a diode mixer?
>

As long as you remain in the linear range of the FETs (ie - below cutoff and above saturation) you'll find the commutative mixer to be far more linear than a diode ring. You can improve the dynamic range by having the crystal filter shunt the out-of-band signals to ground to forestall the cutoff limitation. It may be easier to use an LC roofing filter to do that.



Chris
8310 2013-03-03 08:44:01 Ashhar Farhan Re: current commutating mixer's termination senstivity
Chris,
Thanks for confirming this. I want to be able to measure it. I will
do it over the next weekend.
- farhan

On 3/2/13, Chris Trask <christrask@earthlink.net> wrote:
>>
>>i am playing around with KISS configuration of current commutating
>>mixers (like the one described by chris trask in a paper in our file's
>>area). i am building this with discrete MOSFETs. there is no testing
>>gear (yet) to test mixer IMD. Which is why I am asking this question :
>>As I understand it, by comparison, a diode mixer has diodes that carry
>>the oscillator current as well as the signal current and hence, the
>>IMD increases if the output ports reflect the signals back into the
>>mixer and they go through the diodes again. however, using FETs as
>>switching elements should not cause any non-linear mixing of signals
>>that can create IMD.
>>
>>Am I right to assume then, that following up a FET switching mixer
>>directly with a crystal filter will not degrade the performance too
>>much as compared to a diode mixer?
>>
>
> As long as you remain in the linear range of the FETs (ie - below
> cutoff and above saturation) you'll find the commutative mixer to be far
> more linear than a diode ring. You can improve the dynamic range by having
> the crystal filter shunt the out-of-band signals to ground to forestall the
> cutoff limitation. It may be easier to use an LC roofing filter to do
> that.
>
>
>
> Chris
>

--
Sent from my mobile device