What kind of device gives a suitable switch for a gated integrator operating in the 100MHz+ bandwidth region ? The to be integrated time window could be as short as 10ns. Common elements such as a mixer or just a diode have charge injection, offset problems, thermal dependencies and such.
Rene
Is the signal bipolar?
John
How about two video switches, making two integrators operating with the offset of 10ns, and subtracting one from the other?
Vladimir Vassilevsky DSP and Mixed Signal Design Consultant
How about:
Can that be treated with medication?
Hot luv (charge) injection.
Jeez, one would think that if they can make nice, clean, crisp company logos that they would have dressed out those data tables to look a bit better.
Everything looks like a copy of a copy of a copy of a copy.
It really gives them a poorer image. That is sad if they actually do have viable, robust products.
impressive speed. Thanks.
Rene
Unfortunately, yes. We're talking about a signal less than a few dozend mV .
Rene
No...it is controlled with medication.
do
The product is reliable and robust, but it is for a niche market. Signetics invented and introduced the lateral MOSFET a very long time ago, and I've used a lot of them over the years, mostly as the SD214.
Sadly, they didn't seem to sell in large enough numbers for Signetics to keep on making them. Philips did take over the technology when they took over Signetics but eventually stopped making the parts in one of their "rationalisastions". Linear System were one of several companies who stepped in to satisfy the small but persistent demand.
-- Bill Sloman, Nijmegen
Impressive indeed, and the various internal capacitances are remarkably low, but not really low enough for 100MHz region.
I'd go for a diode bridge, or convert the signal to a current (plus off-set to keep it unipolar) and steer it through the tail of a long- tailed pair. The Harris (now Intersil) Gilbert cell IC (HFA3101) might be a place to start.
-- Bill Sloman, Nijmegen
You really should amplifiy it all you can before trying to integrate it. Charge injection will be the big problem here. Charge injection in
*both* directions can be a problem.Some sort of current-steering diffamp thing might be an alternate to a series switch.
You could almost filter and burst digitize and sum, but it's borderline at this speed.
Some of the NEC phemts behave like 10 ohm resistive switches with 0.35 pF gate-drain capacitance. The best surface-mount diodes are around
0.2 pF, so that's about a tie.Not an easy one, especially if you need precision.
John
Funny, I was just reading in my semiconductor textbook about "hot holes".
Tim
You can make amazingly strong mixers out of those.
Cheers
Phil Hobbs
Only after it becomes un-bear-able. ;-)
They have another type, the SST824, which has an impressive 5 Ohms Rdson and takes as much as 1 Amp and is to switch within 2ns. I'll be going to try it with a 50 Ohm MMIC as signal driver.
Rene
Thanks for this hint. The 10GHz Gainbandwidth is also impressive. I'll have a look at it.
Rene
I once had to have a switch with ultra low injected charge and found that nice Fairchild's FSA66. Didn't tested it because I had a better solution to my pb so I don't know for the 50fC (yes!) figure but I read something elsewhere that let me think it is for real. Depending on your configuration it might be useful.
Well, it IS an adult book...
Have something to add? Share your thoughts — no account required.
Ask the community — no account required