Simple analog JFET switch - part 1: schema, simulate and test

RE: Simple analog JFET switch - part 1: schema, simulate and test

Jan Cumps — Thu, 25 Jun 2020 17:45:58 GMT

Today I went to the local brick-and-mortar electronic shop.

It's the place where I order my element14 components (in Belgium you can only order at Farnell online directly if you have a business - and I love to support local business - and the markup is often lower than shipping costs).

I got some ammunition to do more discrete semiconductor investigations:

[View:/resized-image/__size/502x627/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-51168a2d-f2a5-435b-b66c-0b04a36cfb30/contentimage_5F00_191746.png:502:627]

Perf board with isolated islands (no patterns. Just holes with a copper pad).

And a set of 100 bipolar transistors.

BC547 / 557 (200 mA), 28 / 28 pieces
BC337 / 327 (800 mA), 12 / 12
BC517 / 516 (darlingtons), 6 / 6
BD139 / 140 (1 A), 4 / 4

I'm not reviewing those transistors - I'm trying to focus on J- and MOSFETs.

But I do need them in the surrounding circuit. As current sources, current mirrors, ... switches, ...

For my last project14 basket (I've been on a roll the last months), I ordered a few pure copper clad boards, and more FETs. They will be the devices under test.

A few things i want to achieve, that were underplayed during my education:

FETs as analog devices. 90% of the theory and labs were focused on bipolar.
P-devices. Anyone and any book shows N-devices almost always. Then they say that P-devices are just the opposite. But my brain does not work like that.

RE: Simple analog JFET switch - part 1: schema, simulate and test

jc2048 — Mon, 22 Jun 2020 06:51:10 GMT

It works in Tina-TI too! This is using the FET model from the library - goodness knows where that came from

or how old it is, but it seems to give sensible results.

[View:/resized-image/__size/620x430/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-51168a2d-f2a5-435b-b66c-0b04a36cfb30/7220.contentimage_5F00_191738.jpg:620:430]

This is a 100mV peak 5kHz sinewave being switched. I've got it running into a 10k load.

[View:/resized-image/__size/620x323/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-51168a2d-f2a5-435b-b66c-0b04a36cfb30/5165.contentimage_5F00_191739.jpg:620:323]

Here I've got the turn-on delay longer than the turn-off [VG2 and the 25R are simulating a 5V processor

output]. That way, two channels could be a break-before-make switch. What I've got here is possibly a

bit abrupt for audio [1ms on time and 0.5ms off], but it would be easy to extend it. You'd need to look

at how much processor noise found its way through to the gate [a small value decoupling cap in parallel

with the 1uF might help]. If the 100k [R4] were taken to a clean supply, that would also help when it

was off.

This is the feedthrough when the FET is off. I make that 160dB of isolation. Not sure you'd see that in

practice.

[View:/resized-image/__size/620x323/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-51168a2d-f2a5-435b-b66c-0b04a36cfb30/7384.contentimage_5F00_191740.jpg:620:323]

Distortion isn't too bad at 100mV peak, though hi-fi buffs might turn their noses up at it

[View:/resized-image/__size/530x448/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-51168a2d-f2a5-435b-b66c-0b04a36cfb30/2526.contentimage_5F00_191741.jpg:530:448]

gets a bit rough with 1V peak [don't think those professional audio people are going to like it]

[View:/resized-image/__size/531x399/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-51168a2d-f2a5-435b-b66c-0b04a36cfb30/1031.contentimage_5F00_191742.jpg:531:399]

RE: Simple analog JFET switch - part 1: schema, simulate and test

DAB — Sun, 21 Jun 2020 18:34:48 GMT

Very good post Jan.

DAB

RE: Simple analog JFET switch - part 1: schema, simulate and test

Jan Cumps — Sun, 21 Jun 2020 15:01:56 GMT

All instruments used in the blog are affordable. I purchased them second hand.

[View:/resized-image/__size/620x826/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-51168a2d-f2a5-435b-b66c-0b04a36cfb30/contentimage_5F00_191737.png:620:826]

The oscilloscope is a Rigol DS1054z. Someone in Belgium bought it as a temporary solution for their lab while they were waiting for another scope to arrive.

But any 2 channel oscillocope that can show 2 V p/p at a few kHz (and survive 5 V on the inputs) will do.

The function generator (could be the audio output of a Pi or PC), DMM (optional), frequency counter (optional) and PSU (can be a fixed low power one between say 5 and 15 V, or a Pi 5 V pi, or a 9 V battery), are a single unit, Metex MS-9150.

All components except the FET were taken off of scrap PCBs.

RE: Simple analog JFET switch - part 1: schema, simulate and test

jw0752 — Sun, 21 Jun 2020 14:48:31 GMT

Hi Jan,

The exploration and sharing of these simple experiments with individual components is one of my favorite types of postings on the forum. They always inspire me to setup and try to duplicate the experiment. Thanks for posting.

John

RE: Simple analog JFET switch - part 1: schema, simulate and test

genebren — Sun, 21 Jun 2020 14:29:16 GMT

Very cool! Thanks for sharing!