Piloting the PTR08060WVD with a touchswitch, P mos or N mos? help

[kosliev]

Hi, I can make both a P or N channel mosfet touchswitch to pilot the same regulator from BigBlue's Evercool, except I remember someone had troubles with the N one (which interrupts the ground pin on the regulator) since there was 'runaway' current that still made the regulator have an output, so he used a P mosfet to interrupt the 'positive' pin of the regulator (the pin #2, IN).

Is this true? Since the N mosfet isnt polarized without your fingers on the pads I dont see how the regulator could still go off, i'm attaching my schemes for both in case someone wants to look at them (perhaps theres mistakes?)

Hope the N&P mosfet i picked will do it for the job too

thanks for your time

edit.

Wow the pic came out really small, gonna upload it on imageshack
http://img821.imageshack.us/img821/7949/img009yo.jpg

 

[Phroge]

This is why I'll leave building mods up to someone else. Many of the words in the OP look like real words, but make little sense to me.

I'm sticking to adding a few drops from Bottle A to a little bit from Bottle B to a dash from Bottle C. Then all I have to do is shake it up and put some current to it!

 

[CraigHB]

Use the P-channel. An N-channel switches power on the low side (ground side). It's do-able, but for it to work, you need to make sure that all connections to ground go through the switch. Generally not practical with normal circuit design. There are applications where N-channel FETs are preferable, but this is not one of them.

Better yet, use a touch controller chip to provide the gate input on a P-channel FET. There's a thread around here somewhere on how to do that. Most people don't understand you need to use a protected FET anytime it can be vulnerable to ESD. Using one as a a touch switch makes it highly vulnerable. Even when protected, the protection is only rated for about 2kV using the human body model. A typical static discharge from your finger is in the neighborhood of 15kV. Touch switch controllers usually have a 16kV tolerance to human body ESD.

 

[kosliev]

Thanks for the replies!

CraigHB when you say protected mosfets,you're talking about the built in zener? is that enough to counter ESD's? The Pch I have are these ones
http://docs-europe.electrocomponents.com/webdocs/0478/0900766b80478f5c.pdf

The idea for the mosfet was to forego using switches, saying that they're vulnerable to ESD is heartbreaking to say the least

Electricity why you hate me so

 

[CraigHB]

You're not alone, I've many times been frustrated in designing for ESD tolerance. It's often an afterthought and it's generally a RPITA.

Anyway, protected FETs are quite common and you should easily be able to find one that fits the bill. I've used the Fairchild FDS6679AZ before. You can always tell by the circuit diagram in the data sheet. It has diodes between the gate and source.

 

[breaktru]

I remember someone had troubles with the N one (which interrupts the ground pin on the regulator) since there was 'runaway' current that still made the regulator have an output, so he used a P mosfet to interrupt the 'positive' pin of the regulator (the pin #2, IN).

Is this true? Since the N mosfet isnt polarized without your fingers on the pads I dont see how the regulator could still go off, i'm attaching my schemes for both in case someone wants to look at them (perhaps theres mistakes?)

 

[breaktru]

Like my "Side x Side"

Side x Side Thread

 

[kosliev]

Alright, thanks breaktru, I see you have a 1k resistance at the touchpads, while i saw people use 47k there, I can use 47k, 1k or 10k, which one should i use?

 

[breaktru]

Good question Kosliev.
I've seen dozens of circuits with various gate resistor values. I've seen a web site called Electronic Tutorials with NO resistor and just a tact switch from gate to gnd (P-Ch).

If you take an OHM meter and place it on the highest meg ohm scale and touch the leads to your DRY finger, you will measure several meg ohms. So technically you are not putting a 1k ohm from gate to gnd (P-Ch) but several meg ohms in series with the 1k ohm to gnd. So there fore 1k or 470k ohms at the gate is minimal compared to the resistance of your skin resistance.

I know this not a TECH answer but just my evaluation.

 

[kosliev]

Oh that makes sense to me

Perhaps the 47k or whatever it is, is there incase of misfire with a penny in your pocket? thats not gonna have much resistance at all and the mosfet might get damaged

Electricity is still mostly magic to me

 

[CraigHB]

You can't let the gate "float" on a MOSFET. Gate impedance is typically so high, the wiring can actually pick up stray voltage from the circuit and flicker the MOSFET on and off.

You normally use a pull-up resistor with a P-channel (or pull-down with an N-Channel) so the gate never floats. For a touch switch, you're forming a voltage divider with the pull-up when you touch the contact. Since finger contact is a couple Meg, you want the pull-up 5-10Meg to reduce the effect of the divider.

You always want to maximize the voltage differential between the gate and source on a MOSFET switch to put it fully into the saturated or "on" region. For a touch switch, the higher the pull-up value, the closer it is to the maximum. However, with too high of a pull-up resistance, you can still run into the problem of the gate picking up noise. So, you woudn't want to use something like a 100Meg pull-up resistor.

I've seen schematics showing a MOSFET touch switch with a pull-up of a couple Meg, but that does split the voltage in half. That's usually okay with a mod that uses two batteries is series. In that case, Vgs is around 4V which is fine for many MOSFETS, but for a single battery, cutting Vgs is half would be a problem. You'd definitely need a pull-up with much higher resistance.

I prefer to use a touch controller chip myself, but for a quick and easy touch switch, a MOSFET seems to work pretty good. Use a protected one.

 

[kosliev]

So you're saying that 82M ohm between source and gate is too much, or the 47k ohm in series with the drain is too much?

in that circuit basically lets assume i have 8V total as Vin

since its a divider that'd be Vdgnd=8*47k/47k+82M= ~0,45V

which means Vgs is 8-0,45= 7,55V

is this good enough?

 

[CraigHB]

I don't think I would go that high on the pull-up resistor, probably 10Meg, but if 100Meg is okay, it's okay I guess. With 100Meg, a current flow of 10nA would be enough to start affecting the MOSFET. That's in the neighborhood of what the gate junction can leak. Depends on the FET though. There's also issue with susceptibility to noise as already mentioned.

Unless I'm missing something, I don't see any reason for a resistor between the touch pad and the gate node. It's insignificant compared to skin resistance and it's not going to offer any ESD protection since the human body has about 1.5kΩ when modeling ESD behavior.

"Turn on" voltage can vary quite a bit depending on the FET. You have to look at the data sheet for the FET you are using, but for most FETs, anything over 4V is usually good. Logic level FETs carrying low currents may only require a couple volts. High power FETs carrying large currents may require as much as 10V. However, FETs are not on/off devices like switches. Higher voltage is usually better to reduce "on resistance", but to a marginal limit noted in the data sheet.

 

[breaktru]

I've calculated resistors and also tried what other's have used for circuits ranging from all types of regulators to mosfet switches. I have found that setting it up on a test board and trying out different values always worked best for me... then again I'm not an E.I. I'm a trial an error guy.... LOL

 

[CraigHB]

You would know better than me in practice. I haven't actually built a touch switch using a MOSFET. However, based on general design practice, you would normally never use a 100Meg pull-up resistor for anything. It would still be very unusual to use a 10Meg resistor. The largest pull-up or pull-down resistor I've ever seen used in practice is 1Meg and that was on a demo board, not something intended for every day use. The highest I've used myself is 47K on 5V source and 33K on a 3V source. Of course, there are unusual requirements here which call for unusual components.

 

[kosliev]

Well, after many, many hours of careful soldering and cross wiring, i did get it to work, theoretically.

I measured 5,8V at the connector (the connector being an adapter), plugged the atom in, didnt make contact, took a screwdriver to rise the pin of the adapter, shorted it, saw an electric discharge, pretty much knew it already-

Burnt the mos of the switching regulator, which means, 20$ damage.

Now it just does a clicking sound when I activate it (tick tick tick tick tick), theres 4,2V at the connector but no current (atomizers dont even crackle).


Something else could have fried but I'm fairly positive its the regulator
Cant explain the hate i have right now

 

[kosliev]

Weird, i forgot to solder the (-) of the chip to the main (-) of the circuit, that might be why its not working >_>!

Cant explain the 5,8V before the electric discharge vs the 4,2 after though.
I'll have a coffee, shrug off the excitement of hearing the atom crackling and then i'll see if I can get a voltage reading of the output, also so far, I dont really need to touch the touchpads. I just need to touch the touchpads at the Drain, and the atomizer exterior (or the ego cone), all of which are (+) for it to work

Pretty cool, coffee for now

 

[kosliev]

Alright, it works !

5,7V at Vout, exacly how I wanted it, I have a few things to note, and a question to ask to either breaktru or Craig (hope they're still reading >_>)

to note:

I like HV, the 1/4W trimmer i used hence was always pretty low on the setting (200 total, kept it at 40), hence, it gave out, for this reason (and for lack of components) I decided to use two 470ohm resistors in parallel, meaning I have a 235ohm resistance, meaning the regulator will output roughly 5,7V.

Instead of the 47k resistor, I used a 10k one.

--

Now for questions, in the way I wired things, the outside of the atomizer (or the cone) are (+), inside is (-), if i touch one of the
pads (the one that comes from the drain of the mosfet), and the atomizer/ cone exterior, I'm basically shorting the drain with the pin 4 of the regulator (Vout), which is (+).
This makes me have a 5V output.

There are also other parts exposed I can short it with by just holding it (the + or - of the battery side, for example, which would translate to Vin+, if its Vin- thats GND so its fine that way.

Basically what I'm trying to say is that normally you touch the pads, and connect the the gata-drain to (-), hence polarizing the mosfet and you make the thing fire.

Instead, I can also connect the gate-drain with the pin4 (Vout), which is positive though(via the outside of the atomizer because I wired things the other way around), not GND as it should be. Is this damaging something? If yes I can just rewire it so that the outside of the atomizer is (-) instead, this way I can just use only one pad, the other being the outside of the atom/cone. (though it would be quite a PAIN)

thanks for your help. If its not clear I can just draw it (i.. probably should)

 

[styl3r]

Hi, I can make both a P or N channel mosfet touchswitch to pilot the same regulator from BigBlue's Evercool, except I remember someone had troubles with the N one (which interrupts the ground pin on the regulator) since there was 'runaway' current that still made the regulator have an output, so he used a P mosfet to interrupt the 'positive' pin of the regulator (the pin #2, IN).

Is this true? Since the N mosfet isnt polarized without your fingers on the pads I dont see how the regulator could still go off, i'm attaching my schemes for both in case someone wants to look at them (perhaps theres mistakes?)

Hope the N&P mosfet i picked will do it for the job too

thanks for your time

edit.

Wow the pic came out really small, gonna upload it on imageshack
http://img821.imageshack.us/img821/7949/img009yo.jpg

It was Dave "breaktru" over on his site who had the problem n mosfet