Problem applying a mosfet

I think your problem is an inadequate gate-drive voltage for the IRLZ24 mosfet. The IRF datasheet has 4V as the lowest gate voltage with a specified Rds(on), and the output characteristics plots show the mosfet falling well out of its solid turned-on state at Vgs = 3 volts. It may be that the mosfet's Ron is low enough to support the average draw, but not the peak. If this is the case you may be able to restore proper operation with the mosfet switch by adding a large low-esr electrolytic in parallel with the camera.

Is that the "typical transfer characteristics" I ought to be looking at on the data sheet? For some reason I had the idea that is just for a 20 us pulse and not steady state. Ah, but the camera is probably making its decision as to battery voltage in the first instants of operation . . .

If that's the case, running the camera on low voltage with the gate a volt or two higher ought to work - at least as far as determining the cause.

Camera #1 used a step up supply and ordinary high threshold mosfet, I was hoping that using a LL mosfet could eliminate the added supply. I guess there's no free lunch.

What Win means is that you need more than 3V to thoroughly turn this FET on. The guaranteed specs don't go lower than 4V, only the diagram with tyical values does.

Look at figure 1: At 2.5V gate drive the FET isn't able to handle more than 300mA, and only at a considerable voltage drop. Meaning it'll get very hot, very fast. Digital cameras need a whole lot more current. Even the little Nikon Coolpix here does. So you'd have to find a FET that has a guaranteed Rdson at 2V that is low enough to operate the camera.

--
Regards, Joerg

http://www.analogconsultants.com/

Nowhere in your post does it explicitly say what you are doing. Do you have an external battery pack that you are hooking to the camera's external power jack, and switching it with a mosfet? Do you have the batteries inside the camera and are trying to send a logic signal to something on the camera? For that matter, what kind of camera? Are you tooling around inside or outside the camera? And what kind of camera turns on just because there's power to it? Don't you still have to press a button?

The little darling in question is a Saitek "Pocket Cam X" I opened it up soldered wires to the battery tabs and shutter button and ran them out the case to a connector.

Intent is to use a pair of D batteries external to the camera. No external power jack except it will accept 5 V via the USB cable - but that wastes a lot of power.

No, this baby and their megacam power up and initializes when batteries are put in - the reason there's so much radio control, aerial photo, stuff on the web for hacking into it - and the low price.

I'm running a battery life test on it now it has taken 257 pictures and still counting over 2 days on 3 AA NiMH cells. Camera takes ~7.5 seconds to power up and get straight, shutter actuates, 6.5 seconds for pix to go into memory, then every thing goes to sleep until the controller turns the mosfet on again.

Goal is 900-1200 pictures over a week during daylight for time lapse. Time delay per picture adjustable over the range of 1 minute to 8 minutes per pix.

That's the Aiptek PocketCam X, right? Very interesting behavior. What are a few good websites to learn more about it? It sounds quite useful for making a smart uP-controlled time-lapse camera, or a trail camera for shooting wild animals. We have a coyote family active in our backyard that we'd like to know more about. I'd like to understand a bit more about what they're up to.

Here's a table of some of the logic MOSFETs in my stock drawers, with your part at the top for comparison. (Id* is the maximum mosfet current for Id=100C.)

Id* Rds @ Vgs part Vds A m-ohm ------- --- --- --- --- irlz24 55 13 105 4.0 fqa33n10L 100 23 43 5.0 irlz44n 55 33 35 4.0 irl3715L 20 38 15 4.5 irl2505 55 74 13 4.0 fdp7045L 30 75 4.8 4.5 fdp8030L 30 - 3.6 4.5

As you can see, there are far better parts to choose from, than the one you picked. Sorry! However, none of these is specified for operation at 3.0V gate voltage, let alone 2.0V, etc., end-of-life battery voltage. Looking at the fdp7045L's datasheet plots, we can estimate Ron might increase to about 10 m-ohms at Vgs = 3V, and perhaps 20 or 25 m-ohms at Vgs = 2.5V, but below that the on-resistance climbs dramatically.

One good solution for you is to create a small gate-voltage power source, which would say double the battery voltage, for the low-power purpose of running the mosfet's gate, and perhaps a few other things that need more voltage. A part like the 7660 can accomplish that task in one step, from voltages as low as 1.5 volts. From $1.30 at DigiKey.

How 'bout he stick a 9 Volt in there just for the gate for now, and where is his turn on signal coming from anyway? A relay activated by a PIC running off 5 Volts somewhere? That 5 V just may do it- else it wouldn't be called a logic mosfet.

Good question.

Thanks, I'll look into the chip. Camera one - I used a 10 KHZ pulse train out of the picaxe to switch an NPN transistor for a little HV supply to goose the mosfet into conduction. It was a tad unwieldy with about 10 parts just for that function, but it does work well. With the current requirement next to nothing it may be possible to just use the IO pin to toggle the inductor directly (with a diode to protect the 'axe).

Camera one is in a waterproof case and has a mount that slides into a channel on the front of my kayak or motorcycle.

Cameras

The PocketCam X is vastly superior to the smaller camera for resolution, light level correction, current drain, and speed. Code: R-PKX Price:$19.99 +$5 for a 512mb SD card (it has internal memory as well, the SD card boosts the number of pictures you can take from 20 to 640 high res shots) Comes with a tripod, small camera bag, usb cable, software. I think there's a built in microphone and it will take AVI videos. Focus is adjustable but not automatic, but the depth of field is great from about 10' - infinity in the long range mode and down to 9" or so in close up mode.

My first camera is the mini pen cam (they were calling it the mega pen cam but changed the name and color) Code: R-PCM13B Price:$9.99. It comes with a reasonably good belt loop case, microphone and earphone, and a stand that serves as a tripod. The resolution is nowhere near the PCX camera but it is small, very light, and cheap enough to practice on. It says it takes 50 shots - I usually get 60 and even 62 before the memory is full.

Picaxe is the time lapse controller - just BASIC to program costs ~$3

08M and 14M versions cost the same, but the 14 pin has more I/0 and just as easy to apply. You would have all of $20 invested for a "development system" including three controller chips. Programming software is free, you need a PC and serial cable. This thing is a lot of fun . . . for software and forum

for parts in the US - he charges ~$10 for three 08M or 14M parts, includes the programming resistors, shipping is only $1.95.

You don't need all the hoopla - a basic solderless breadboard , serial cable, and holder for 2 or 3, 1.5 V batteries is all it takes. It was harder to figure out how to get the usb to serial cable to work with windows than it was to use the picaxe.

The radio control forums are excellent sources for camera modifications - they send them up with planes, kites, rockets, and balloons.

I got the disassembly and wiring instructions from these sites

for modifying the dolphin jazz camera - identical to the pocketcam X

how to hack the megacam

When I finish the life test of the batteries I plan to try boosting the gate voltage temporarily to see if that works - I figure it will.

With camera one I knew I needed a high gate voltage so just added a little 10 volt boost supply to work the mosfet - I expected switching to a LL mosfet would eliminate the need for 10 volts.

I want to use only one set of batteries and only 3 volts for the camera and controller - which will both work down to around 2 volts. Adding an extra battery is one more thing to check before the camera goes out in the field or, running the whole shebang from 9V wastes a lot of battery power - picaxe wants to see 5 volts or less, camera can also take 5 but wastes power.

One other thing about this PocketCam X - I noticed that the shutter trigger needs a separate NPN to pull it low to snap a picture and it has to be a hard connection not capacitor coupled like Camera One.

While fooling with the shutter, I noticed it would take a picture immediately if the shutter is biased at 1/2 Vcc. That is it initializes and immediately snaps the picture and stores it.

I didn't check to see if it would snap repetitive pix, or if it has enough time to make the light level adjustments, but it may be that the whole time lapse camera will work with only a pair of 555 timers to control it.

I plan to keep using the picaxe controller because I still want to add some logic like increase the delay time after dark and speed it up the next morning.

*Nice* background information on your project - the camera & PICAXE stuff is enticing. :-)

Ed

Thanks. It has been a lot of fun for me. The picaxe is a great toy. Does a lot of neat preprogrammed functions like PWM output, servo output for positioning radio control servos, IR send and receive (uses Sony TV codes), ADC input, plays music or tones to a piezo transducer, etc.. The chips range from 8 to 40 pins for added IO and communications capability.

The active picaxe forum site so you can see what others are doing. The forum is moderated and is monitored by the developers too.

The camera project is also fun. I can leave it at a location and watch what went on all day. boat launch ramp is fun - watch some newbies crashing the docks, parking lots - caught some guy getting arrested, record the scenery as I paddle around the port traffic, etc..

One thing I might try after camera two is in a waterproof box is to add a servo so I can do a time lapse panorama of several shots automatically.

A lot of fun for a very low price.

How about a FET like the IRF7455? They guarantee Rdson to be 20mohm at 2.8V:

A little over two fifty at Digikey in singles.

--
Regards, Joerg

http://www.analogconsultants.com/

One more idea in case even a IRF7455 isn't good enough: You could also use a little single Schmitt Inverter as an oscillator and voltage doubler or tripler. Then you can generate a resonable Vgs even if the battery is down to 2V.

--
Regards, Joerg

http://www.analogconsultants.com/

Whoa, what'd they do to that datasheet? All grey'd out and taking 1026k, compared to only 166k for the dark black 2000 version! It seems adding pbf comes at a severe price!

Anyway, yes there are some sub-4V logic mosfets.

I'll see your irf7455 and raise you a Fairchild fds6894a, only $2.08 at Mouser, and featuring 16m-ohms at 2.5V or even better, 21m at 1.8 volts! But wait, there's more, you also get TWO, yes TWO mosfets in one package, which you can parallel for 10m-ohms at 1.8 volts. Beat that!

I guess we can add a few more entries to the logic-level mosfet table: Id* Rds @ Vgs part Vds A m-ohm ------- --- --- --- --- irlz24 55 13 105 4.0 fqa33n10L 100 23 43 5.0 irlz44n 55 33 35 4.0 irl3715L 20 38 15 4.5 irl2505 55 74 13 4.0 fdp7045L 30 75 4.8 4.5 fdp8030L 30 - 3.6 4.5 irf7455 30 12 10 2.8 fds6894a 20 - 8 2.5 (paralleling two sides) fds6894a 20 - 10 1.8 (paralleling two sides)

That was sort of my first attempt at boosting voltage. In order to use parts already on hand I used 1N4148 diodes in a voltage multiplier

- problem is when you start with only 2 volts (or 1.4 due to the internal transistors) the .6 volt diode drops eat into the multiplying output.

I just used the picaxe to output a 10 KHZ pulse train - it has a totem pole output, and just fed that into the multiplier. The npn/inductor supply used fewer parts.

One of the '7660 parts I suggested will give you 2x voltage without the diode drops (active mosfets are used for charging and rectification), and will draw much less power under the right conditions: icl7660, tc7660s, lmc7660, ht7660, and aic7660, for starts. The latter may be the latest design. I suspect they all do pretty well at low voltages and currents, but unfortunately they fail to use log-plot curves and properly show the data we need.

That looks like it is intended to produce a negative output corresponding to the positive input. The application notes show a positive voltage multiplier added, but it has diodes.

Am I missing something?