Basic problem with transistors

The transistors will have a saturation voltage loss of about 0.4V from collector to emitter.

Why discuss the ridiculous circuit that has three power supply voltages??
Just use a single 5V supply and logic-level Mosfets.

audioguru said:

The transistors will have a saturation voltage loss of about 0.4V from collector to emitter.

Why discuss the ridiculous circuit that has three power supply voltages??
Just use a single 5V supply and logic-level Mosfets.

Click to expand...

Well one is the 3V battery supply which will drive the motor, and the other two are the board parallel port bits which will decide which direction the motor runs in / whether it's running. How could I use a single 5V supply to control the circuits?

Either way, I appreciate that the various problems discussed with the transistors / MOS-FETs make my schematic impossible, but I'm happy I now know that my ground system wasn't impossible.


Thanks for the extra schematic marcbarker, it appears you've got the gist of what I was trying to do now now I just need to think about whether I'll need to find logic-level MOS-FETs to build a similar circuit, build a schematic with the IC suggested by blueroomelectronics earlier, or something else... looks like I'm going to be busy

audioguru said:

The transistors will have a saturation voltage loss of about 0.4V from collector to emitter.

Click to expand...

Unless you provide the sums (or proof) how you arrive at "0.4 V", this is only an opinion. We all know that even the experts get it wrong, that's why we often ask for a 'second opinion' when there's not enough proven data.

Why discuss the ridiculous circuit that has three power supply voltages??

Click to expand...

Your opinion of the posted circuit aside, could you analyse the posted circuit any better than I had done? If you're able to, we're eagerly waiting.

Just use a single 5V supply and logic-level Mosfets.

Click to expand...

Haven't we already covered that already?

In case you (or anyone else) missed what this Thread is about....

It is about a theorectical scheme, posted by Giftiger which potentially reduces the number of signals between a control device and a motor drive circuit, by utilising the fact that between two digital outputs is a 'H-bridge' function already, and that all that needs to be done is the output's current boosting enough so that it can drive a motor directly. I would go so far to say that this scheme is novel and has potential for development. This circuit's mechanical analogy is a hydraulic power assisted steering system, or servo-assisted brakes.

I think we should be encouraging this kind of 'thinking-out-of-the-box' design with all the help we can give. Anything that reduces the amount of signals and circuit complexity is a product is a good thing. Especially for instance in mechantronics for spacecraft/lander design for example. Here there are constraints such as the reduction of mass, the reducing the amount of semiconductor devices, etc.

I didn't know that this project is school homework.

audioguru said:

I didn't know that this project is school homework.

Click to expand...

Our young people are the future of the planet, they inherit it from us after we've finished using it

marcbarker said:

Our young people are the future of the planet, they inherit it from us after we've finished using it

Click to expand...

hi marc,
I agree we should train our replacements, however there is a sub Forum for homework questions.

If the OP posts in the homework forum at least we know its only a tutorial type question.

BTW: do you want a link to a free schematic drawing program.

If you think it's be worth building it, the FETs can be a Fairchild FDFMA3N109 Cellphone RF transistor (Farnell stock code 1324787, £0.20 each) These fets have the schottky diode included for free. https://www.electro-tech-online.com/custompdfs/2009/08/59685.pdf

According to the datasheet (Fig1), and using 2.4 V of drive, the FET is driven well into 'ohmic region', with Rds(on) just over 1 ohm, which is more than good enough.

The data I quote is typical of course, but it's like marriage... it's "better or worse", but nearly always 'typical'.

If you think you'll be unlucky, the FETs are only 20p each and the minimum order charge for Farnell is £30...

This project is not a school homework and while I also wanted to test my design, I am going to ultimately require a functional circuit...

I'm not entirely sure how you could think that this is a homework question. I'd be pretty devoted to my homework if I put in this much time, effort, and money (the rabbit development kit wasn't exactly cheap) into completing a piece of homework. I'm not an electronics student and have never been one.

I am simply an enthusiast who thus far has little electronics experience and am trying to learn for myself (of course with help from the experts in here, and thank you to everyone who has helped so far).

marcbarker said:

If you think you'll be unlucky, the FETs are only 20p each and the minimum order charge for Farnell is £30...

Click to expand...

So replacing the BJTs in my previous schematic with these FETs, it should work alright? I'll post a new schematic when I get chance to draw one up. How much current would be required to drive the FETs? From what I read in an online transistors tutorial a while ago (and this is more or less my only source of knowledge about FETs so bear with me), MOS-FETs are insulated from the rest of the device and therefore require virtually no current, but FETs are slightly different. Though the page did seem to suggest that FETs in general are voltage-operated and don't generally need much current.

What would be the typical gate-source resistance for the device and how much current would be required to drive it? The datasheet mentioned the drains-source resistance at various gate voltages but I didn't see a mention of RGS (if this is the correct term, I'm still learning the notation). Remember that the maximum allowed current through the parallel ports of my board is 8mA, so hopefully it will require less than 4 mA or it'll obviously be unable to drive two of them...

I doubt that you will find Mosfets that will turn on properly with a gate voltage as low as 3.3V.
The gate of a Mosfet has a high capacitance that takes a fairly high current to charge and discharge quickly. If they don't switch quickly then they fry or waste a lot of power making heat if the switching frequency is high.

Switching frequency is unlikely to be high, though I suppose that depends on what 'high' frequency is in this context. The datasheet for the device suggests that the RDS value is less than 200 milliohms when the VGS is 2.5V... it seems unlikely that this will cause a lot of power to be wasted. I don't see a mention of how much current would be required to run the FETs though.

If it is indeed not given in the datasheet, I'm sure your estimate would be better than mine - how much current would likely be required, and if the current simply makes the difference of how long the FET takes to switch on, how much of a delay should be expected if they were run at 4mA? It seems unlikely that they would be damaged in this way at the low switching rate at which the motors are likely to be operated - the motors will be usually held on for a couple of seconds, and then left off for longer.

So far, this seems like it may be a pretty good solution... a FET which may be able to run correctly off of my microprocessor output, at a low cost.

audioguru said:

I doubt that you will find Mosfets that will turn on properly with a gate voltage as low as 3.3V.

Click to expand...

What exactly do you define as "turn on properly"? That's a bit vague.

Perhaps do you mean: "10 V of gate drive, as specified in the datasheet" ?
or: "biased into the ohmic region", ?
or: "may conduct full rated Id without losing saturation"??

The gate of a Mosfet has a high capacitance that takes a fairly high current to charge and discharge quickly. If they don't switch quickly then they fry or waste a lot of power making heat if the switching frequency is high.

Click to expand...

OK then, as you've kindly offered, please calculate for us how of these many Watt / seconds are involved for each motor switch-on (or off)?

Other than that, how does Dynamic performance apply to a DC circuit such as this?

By the way this is a mosfet for a cellphone.

BTW: do you want a link to a free schematic drawing program.

Click to expand...

I have allsorts of schema drawing software but I still prefer the sheer power of creative expression that only paintshop pro provides!

giftiger_wunsch said:

So replacing the BJTs in my previous schematic with these FETs, it should work alright?

How much current would be required to drive the [MOS]FETs?

Click to expand...

The only thing I altered in your schematic was changing the lower BJTs for MOSFETs. I don't know for sure if the circuit will work. I think the IC circuit that was posted earlier was a 'better' solution. However, something about your circuit I like. I like the minimalism of it, and that it's so novel. Is it all your own work? I would like to see it work.

As I said earlier, it's like a power assisted steering, there is pump (motor battery), control valves & torsion bar (transistors) and a driver and steering wheel (your uC IC + software). Two different power sources, independent of each other, only meeting eachother at the control valve.

I chose MOSFETs for the lower transistors, because if it was still a BJT, the base-emitter diode would shunt away the current. When activated, there is a voltage that is 'reflected' back from the motor, towards the controller. This voltage appears across G-S and switches on the appropriate MOSFET.

The upper transistor is switched by the current flowing into it, this current flows through the motor too. The motor and transistor base drive current mingle, as they pass through the motor, then going their seprate ways. The transistor acts (if I analysed the circuit right) like a "emitter follower", and the base current is motor current / hFE.

The MOSFET gate is a capacitor. There is no DC current as such invloved.

marcbarker said:

However, something about your circuit I like. I like the minimalism of it, and that it's so novel. Is it all your own work? I would like to see it work.

Click to expand...

Yeah I came up with the idea early on. It was really just based on a simple concept: if the uC output had matched the requirements of the motors, then I would have simply connected the motors between each pair of parallel port bits; obviously this is not the case which makes the circuitry much more complex, but the idea of running the circuit between the two terminals can still be applied.

With all the other stuff being thrown around in this thread, I still haven't had chance to look into the IC suggested earlier in detail yet, I will do so soon. Right now I haven't looked at it hard enough to grasp more than its basic function so I'll see how well it suits my needs.

Ultimately I would like to stick with my original concept if possible, because if I abandon my own designs then I'm never going to get better at developing and implementing them. If I can use the IC in a similar way I will go with that if it a 'better' H-Bridge circuit; if not I will probably go with my original schematics, obviously with adjustments as discussed in this thread.

So far the biggest changes seem to be that the BJTs should be replaced with logic-level MOSFETs such as the devices you suggested, and that a resistor should be added to the uC circuit to prevent the current exceeding the 8mA limit of the board. A single resistor ~410Ω placed in the ground connection should limit the overall current to ~8mA since other resistances should be low. Though I'm not entirely sure how the MOS-FETs will actually affect this... the gate is insulated, so will current actually flow through the gate at all?

I don't know what your motors will do. Most motors have high frequency pulse-width-modulation for speed control. The high frequency causes the gate driver to use a fairly high pulsing current to charge and discharge the high gate capacitance quickly.

If your motors do not use PWM speed control then the current into the Mosfet gates is momentary.

giftiger_wunsch said:

the BJTs should be replaced with logic-level MOSFETs

a resistor should be added to the uC circuit

will current actually flow through the gate at all?

Click to expand...

Only the 2 lower ones. The fet used needs to be the lowest Vgs(th) voltage you can find easily. If you choose something different you might want to post it here first.

No resistor added to the circuit. That's what's so clever about it. The transistor (if I'd analysed it right) only takes current that's needed.

No DC current in the MOSFET gate.

Audioguru:

There's no PWM speed control for the motors: the motors are only set up to operate at one speed, and it won't be particularly necessary to attempt to change that at any time in the near future. The existing circuitry for the motors is remarkably simple, it literally just involves using the levers in the controller to operate a simple H-Bridge connecting the cells directly to the motors. There are no actual components visible on the two small boards; obviously excluding the switches in the controller. No semiconductors, no resistors, nothing. There is a guide-light white LED mounted in the claw, which would suggest that there is a simple circuit hidden inside this component to limit the current to the LED, but otherwise it's all just conducting battery power directly to motors.


Marcbarker: I'm not really sure I understand why only two of the transistors need to be replaced with MOS-FETs. Could you explain the reason for this?

audioguru said:

I doubt that you will find Mosfets that will turn on properly with a gate voltage as low as 3.3V.

Click to expand...

Try the AO6407 and AO6408.
https://www.electro-tech-online.com/custompdfs/2009/08/AO6407.pdf
https://www.electro-tech-online.com/custompdfs/2009/08/AO6408.pdf

Hero999 said:

Try the AO6407 and AO6408.
https://www.electro-tech-online.com/custompdfs/2009/08/AO6407-1.pdf
https://www.electro-tech-online.com/custompdfs/2009/08/AO6408-1.pdf

Click to expand...

The AO608 MOSFET is interesting. I'd heard much about TrenchFets, but I never did get to use one. At low Vgs & 1 A Id this fet has low typical rds(on). There is a healthy margin too.

**broken link removed**

I think we have all seen manufactured products that are machine-made with such tiny little surface-mount parts. But can we hand-etch a pcb and hand-solder these tiny parts?
Can we buy these little parts in small quantities?