SparkFun Forums

[ttabbal]

FYI, a possible variation on my circuit:
Change the CMOS 555 to the traditional bipolar type (which has much greater drive capability), and let the output of the 555 provide power to the xmtr module *directly*, eliminating both transistors. You get the benefit of the "automatic xmtr power-on/off", you don't need to use an extra PIC pin, and the overall circuit is rather simple.

Pete

That's interesting. How long is that timer running? So when the PIC sends data the low pulse will trigger the 555. I wonder if 5V from the PIC would be enough to keep the 555 from triggering with a 12V VCC. Do you find that the xmtr powers up fast enough or do you lose a byte or two before it powers up? I'm sure I have a 555 around here I could test that with.

[ttabbal]

Hi,
You have Q2 upside-down. As drawn, the FET's body diode will cause the xmtr to be powered up all the time. And you need a resistor (say 10K) between Q2's gate and 5V (in the line directly above Q1). Without it, Q1 will 'pop' when it turns on. And, you want the new resistor to go to VCC-RF, not to 5V.
For Q1, I would use a 2N7000 (N-channel FET). Then you can eliminate R1 and R2.

Pete

Thanks. I have a real problem with wiring FETs backwards on schematics. Good ideas. I'm not sure if I prefer the FETs or the 555. Not needing the pin and software control is nice.

[saipan59]

That's interesting. How long is that timer running? So when the PIC sends data the low pulse will trigger the 555. I wonder if 5V from the PIC would be enough to keep the 555 from triggering with a 12V VCC. Do you find that the xmtr powers up fast enough or do you lose a byte or two before it powers up?

I think I've got it set up for 0.1 seconds.
As for the trigger voltage: That's a good point - there *may* be a problem if the 555 is running from 12V. The 555 datasheet may give the answer. Or, just wire up a 555 by itself on a breadboard, and try it...

I don't know exactly how fast the xmtr powers up. At 4800 baud, it probably loses a couple of bits(?). But any useful protocol should include at least 2 or 3 preamble bytes (i.e. 0xAA) anyway, so losing something at the beginning is not a problem. My protocol starts with AA AA AA AA FF as the first 5 bytes, to ensure that the rcvr is synced up.

Pete

[ttabbal]

That may have been confusing - let me restate my suggestions:

Swap the drain and source on Q2.
Eliminate the connection to 5V.
Add a 10K resistor from the gate of Q2 to VCC-RF. This will keep Q2 'off' when Q1 is 'off', regardless of the voltage at VCC-RF.
Change Q1 to a 2N7000 (with the Source on the ground side), and eliminate R1 and R2.

Pete

I'm still getting used to the terminology in FET datasheets. But I was wondering if I could use a part like the IRF7509 which contains a N and a P MOSFET in a single package to set up the switch? It looks like I could trigger the N from a 5V PIC line and use that to trigger the P channel device to make the high-side switch. Wired as you suggest, do you think that part would work?

I might still do the 555, I just want to understand this stuff a little better.

[saipan59]

Yes, it appears that it would work.
Be aware that it's in a very small SMT package...

Pete

[ttabbal]

Yes, it appears that it would work.
Be aware that it's in a very small SMT package...

Pete

Thanks. I'm not worried about SMT, I've done a lot of soldering on tiny SMT parts and dense ones like TQFPs. No problem there, I just haven't done a lot of design with discrete parts. I'm usually just messing with MCUs and other digital parts.

[bigglez]

Greetings (Still No First Name Supplied),

I'm not worried about SMT, I've done a lot of soldering on tiny SMT parts and dense ones like TQFPs. No problem there, I just haven't done a lot of design with discrete parts. I'm usually just messing with MCUs and other digital parts.

I think the caution has as much to do with
device power rating than handling. Small footprint parts are
not very good at heat transfer. You're using this as a power
switch (albeit for a small load). You may find the larger
single device packages more reliable due to lower self-heating.

Comments Welcome!

[ttabbal]

Thanks for all the info. Here is an updated schematic of the power switch (the 555 I have didn't like a 5V trigger with a 12V Vcc). Did I get the polarity right on the FETs this time? I'm not real worried about self-heating as I'm running it far below its 25C rating and it will be a low duty cycle.

[bigglez]

Greetings Travis,

(the 555 I have didn't like a 5V trigger with a 12V Vcc).

Depends how you triggered it. 5V is a little too
small, as the 555 trig must pass through 1/3Vcc (4V) to trigger.
Also, the trigger must return high before the 555 timout. Usually,
a capacitor coupling and resistor divider bias is placed on the
trig pin, with a steady state bias of 1/2Vcc (i.e. resistors are
equal value).

You can also use the 555 on 5V Vcc and drive a FET high-side
switch from it's output.

Did I get the polarity right on the FETs this time?

Yes, let us know how it works on the prototype hardware.

Comments Welcome!

[mvs sarma]

I tied it to ground to force it low when the PIC wasn't driving it yet (bootup and such). I suppose it doesn't really matter, but it does work on the breadboard. It might not be the best way and could be causing other problems I guess.

Why is high-side switching preferred? Transistors are not my strongest area. I posted a while back about controlling a relay with a PIC and people here suggested the arrangement I used (low-side switch with base pulled low). If it's not the best way, I'd like to understand why so I don't do it wrong again.

You put a 1K pullup resistor to Tx Vcc and then the collector connected to Gnd pin (not actual ground)so that the gnd can be transmitted. This might serve your problem.

Hope to see a reply

[bigglez]

Greetings sarma,

You put a 1K pullup resistor to Tx Vcc and then the collector connected to Gnd pin (not actual ground)so that the gnd can be transmitted. This might serve your problem.

You are reverting to low side switch in the ground?
This was discussed in the thread as a bad idea. I
wonder if you have read (and understood) the
entire thread form the beginning?

Comments Welcome!

[mvs sarma]

It might appear a bad idea. But we used this technique, of course using N-mosfet to switch in and out NE602 based modulator by controlling the ground connection. It was really effective.

This has reference to a Radio receiver design published in Elektor UK Dec. 2006. This mod was discussed and implemented in the forum over there.