Remembering CMOS...

[Svart]

oh great, give him more ideas why don't you??.. :shock:

 

[Swedish Chef]

[quote author="bcarso"]Better than a cleaver...[/quote]

The cleaver is in the shop for a service this week. Still got the tenderiser though... :green:

chef

 

[Swedish Chef]

can anyone suggest a very low leakage FET to try? Preferably DMOS.
I have found one or two but they are SMT. Do they only make them in SMT? Surely not! :grin:
I have a 2SK170 that I'm using in the sidechain anyway, but am wondering if it gets better than that?
It is okay, but I bet there is better...

I'm getting there kids! :thumb:

chef

 

[Svart]

hmm I like the sk170s myself..

not a lot in through hole parts anymore... :sad:

you *might* get away with a 2n5457 depending on what you are looking for..

might I ask what you are building? if it's a secret, that's cool, but we will just bug you until you give up the goods you know..

plus it might help us help you find better parts..


:green:

 

[Swedish Chef]

Yep it's a dual mono VCA comp, feedforward/ back, RMS sensing, linking modes, and a couple of other fun things.
I'll be sure to let you see it when it's done! ;-)

I'll check out the 2n5457.. :green:

Cheers

chef

 

[bcarso]

[quote author="Swedish Chef"]can anyone suggest a very low leakage FET to try? Preferably DMOS.
I have found one or two but they are SMT. Do they only make them in SMT? Surely not! :grin:
I have a 2SK170 that I'm using in the sidechain anyway, but am wondering if it gets better than that?
It is okay, but I bet there is better...

I'm getting there kids! :thumb:

chef[/quote]

Why low leakage? If the other parameters are of no concern then there are smaller geometry devices than the 170, with longer channels so the gate leakage is not so voltage sensitive. The Toshib* process is so clean though that other smaller FETs from somebody else can be quite a bit more leaky despite the geometry.

The low leakage of DMOS is characteristic of unprotected gate parts. But if you are using it as a transmission gate the substrate leakage usually spoils things. And they are so easy to destroy.

 

[Svart]

Bcarso, have you tried the new Linear sk170?

I'd like to know how these stand up to the original.. i'm going to be needing a few soon..

 

[bcarso]

Not yet. They kindly provided me with a couple of the SMD samples iirc but you know how it is....

They are starting to advertize that part specifically I noticed.

 

[Swedish Chef]

[quote author="bcarso"]Why low leakage? If the other parameters are of no concern then there are smaller geometry devices than the 170, with longer channels so the gate leakage is not so voltage sensitive. The Toshib* process is so clean though that other smaller FETs from somebody else can be quite a bit more leaky despite the geometry.

The low leakage of DMOS is characteristic of unprotected gate parts. But if you are using it as a transmission gate the substrate leakage usually spoils things. And they are so easy to destroy.[/quote]

Dagnabit!
I'm still tinkering with the supercap thing and it seems to be the FETs that are the fly in the ointment. They are just leaking too much when off.
Any ideas/ suggestions?

chef

 

[Svart]

Leaking S->D or D->S? or G->S/D?

depending on which, there are some answers for both..

:guinness:

 

[Swedish Chef]

G -> S!!! :sad:

 

[Svart]

so are the parts actually leaking current(blown through?) or are they building charge capacitively? do you have a schemo of the section that we might look at?

I typically work with MOSfets that need the gate charge to be carefully controlled as to not blow through with high currents, however these tend to be even more insulated than what you are working with, but i would think that the same would apply since you are using a much lower current i'm sure.


lets get this figured out! :guinness:

 

[bcarso]

[quote author="Swedish Chef"]Dagnabit!
I'm still tinkering with the supercap thing and it seems to be the FETs that are the fly in the ointment. They are just leaking too much when off.
Any ideas/ suggestions?

chef[/quote]

We need to see a fragment at least of the design I think. But realize that the DMOS has pretty high leakage for reasonable on resistance. If the gate is unprotected the gate leakage should be tiny though.

The 170 JFET has low gate leakage unless the drain-gate voltage is high, but of course it is a depletion mode device so requires a reverse bias to be off.

Most DMOS is enhancement mode so more helpful in that regard, but have the disadvantage of the integral drain-source body diode. You can use parts in reverse connection in some cases and they work fine as switches.

When all else fails (when for example current must be blocked regardless of the direction of flow), back-to-back DMOS will work. If you can switch the low side of a positive supply you can use N channel devices; if not the lower-performance but now readily available P channel devices can be used (they are readily available in large measure because of these very sorts of applications!).

 

[Svart]

yeah we need to see at least this part of the schematic.. is the FET in the audio path? i am assuming it is but if it's not then there may be better choices for this part..

 

[Swedish Chef]

sorry kids! I have had a pretty busy weekend so haven't been around. Apologies for keeping you both hanging...!
The FET/ BJT is not in the audio path just part of a 3 way toggle switching arrangement to fire relays and LEDs. I'll try to get you the schematic for the whole switching gubbins today, but struggle to capture schematics easily from easyPC; don't worry the irony has not passed un-noticed :?

I have changed over to PNP transistors with the relay coils inline with the emmiter/ +V and a 2 way LED strapped across the low side of the coils. In this new config it has remembered it's setting since Friday afternoon...Hurrah!

Will attempt to get a schem up today! :thumb:

cheers

chef x

 

[Svart]

ah ok, i was thinking that the FET was in the audio chain for some reason. this changes a few things.

first off, you would likely benefit from switching from JFETs to MOSFETs in this situation as the gates are usually protected much better from leakage and blowthrough. I've been using some IRF parts, IRLML2502(from memory) as a small switch for relays.

Vdss = 20v
Rds(on)= .045
Id@25c= 4.2A

it's pretty fast but that won't matter for your application. It also has a body diode which will save you some PCB realestate.


(info alert: general MOSFET info follows, please ignore if you already know about MOSFETs..)


keeping the gate charge won't be hard with this part either but you will need to drive the gate both high and low due to that very same charge. remember that (MOS)FET gates are capacitive and will hold a charge for some time even when the voltage is no longer applied. this is great for keeping something ON or OFF for long periods of time but bad when you need to be absolutely positive of the gate's condition at any certain time. usually you have a resistor to pull the gate high or low when there is no control voltage to ensure that it's on or off but this would only drain/keep the gate charge.

:thumb:

 

[Swedish Chef]

[quote author="Swedish Chef"]

Will attempt to get a schem up today! :thumb:

cheers

chef x[/quote]

Soz!

Here's the scrawl...I fought with EasyPC and lost...



chef :thumb:

 

[Svart]

looks good! i would switch the BJTs to logic level MOSFETs(with body diode) myself because of the inductive nature of the relay coils, but at such low levels i wouldn't be too concerned about the BJTs. Is this done in surface mount? you can get dual logic MOSFETs in so-8 packages as well as the others too.


does that 40106 latch to high or low when triggered or does it just follow the input pulse?

I started to think about it and you might be able to get away with rearranging the setup and using MOSFET driver ICs in place of the BJTs AND the 40106 parts since the 4027 has schmitt trigger-like inputs. the mosfet driver ICs come in various standard types of logic inputs and a high/low output that can drive high current (1+amps) which would be more than enough for the relays and LEDs and might reduce your parts count even further!

the boss is sweating me right now so i'll think about it as i do some *real* work..



:thumb:

 

[Swedish Chef]

[quote author="Svart"]looks good! i would switch the BJTs to logic level MOSFETs(with body diode) myself because of the inductive nature of the relay coils, but at such low levels i wouldn't be too concerned about the BJTs. [/quote]
Could you suggest some part nos.? This stuff makes my mind boggle! :?

[quote author="Svart"]Is this done in surface mount? you can get dual logic MOSFETs in so-8 packages as well as the others too. [/quote]
I fear it's through hole. I had a brief experience of trying to troubleshoot SO boards and hated it!


[quote author="Svart"]does that 40106 latch to high or low when triggered or does it just follow the input pulse? [/quote]
It does the opposite of the input; so the input is tied hi and goes Lo when the switch contacts, so the o/p is lo and goes hi when the switch contacts. Teh switch is a momentary though so the states are..well...momentary :grin: Just a pulse for the 4027 really..

[quote author="Svart"]I started to think about it and you might be able to get away with rearranging the setup and using MOSFET driver ICs in place of the BJTs AND the 40106 parts since the 4027 has schmitt trigger-like inputs. the mosfet driver ICs come in various standard types of logic inputs and a high/low output that can drive high current (1+amps) which would be more than enough for the relays and LEDs and might reduce your parts count even further! [/quote]
That sounds VERY cool. Once again though I need a kick in the right direction! :green:

[quote author="Svart"]the boss is sweating me right now so i'll think about it as i do some *real* work..[/quote]



Cheers so much mate! :thumb: :sam: :guinness: :green:

 

[Svart]

still thinking Chef!

i'm trying to finish this little assignment for the boss.


did you want the outputs to latch high/low when the button is pressed multiple times? you have it setup 3 way, so i assume the modes for the relays/LEDs are:

push one: R1 on, R2 off
push two: R1 off, R2 on
push three: R1 on, R2 on

or something to that effect?

sorry i didn't have much time to study the schematic, but will later today.




:guinness: