Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

MOSFET use

Status
Not open for further replies.
Please supply a clear image of the current schematic with actual component values...I can simulate it and see what's what.
To figure out the DC% from spec sheets and RC calcs is a bit tedious, also it doesn't consider interactions with the power supply sagging etc that a simulation can help with.
 
Is this good enough? I can't do the nice diagrams that appear here.
CIMG3225.jpg


From Left to Right, the Optosensor is an RS 307913. The SetPot (A) can be regarded as Zero R, the little Cap 15 n.

The 4093 Pins 1, 6, 9, 14 get ~6 v. Pin 2 gets the ragged Low signal from the Optosensor, Pin 3 > Pin 5 via RC. Pin 4 feeds to Pin 8.

Pin 10 is the timed sq wave output, to both FETs (fire and hold) and then to the 4013 (Pin 3, clk).

The 4013 Pins 5, 12 get ~ 6 v.
Pin 1 (Q) goes to FET (hold) only. It is linked to Pin 4 (R) via two caps in parallel (use 15 + 10 n), via a potentiometer (0-100k) which is the "accelerator." Pins 6, 7 go to O v.

The Power resistors for the 15 V circuit are 2R0 in common to the FETs, then 1R3 to FET (fire) and 4R0 to FET (hold); eg 1R3 means 1.3 ohm.

The coil has an Inductance 0.15 mH, impedance 0.15 ohm.

The little diode to the voltage regulator 7806 is an IN4002 after a 1K resistor. The ordinary caps either side of the 7806 are 100n, and the Tantalum cap is 100 microfarads (16 V). (I have no idea what this is for!).
 
It appears that your opto input and output share a common ground and that it is fed by the common 6V supply, which means it is always 'on' and serves no purpose.
It does seem to me that a oneshot 555 timer pulse driving both MOSFETS with an RC on the 'Fire' FET for shorter response might be simpler.

I'll sim the cct and see what it does.
 
The Opto only gets a signal when the light reflector passes. That output causes a momentary drop in voltage in the line, linked to 4093/2
 
Here is what we got using a 555 pulse generator. It was a different set-up but we fed in Low pulses at about 2700 rpm.

The present circuit doesn't use the LM1949 - thats my whole plan.
Pot fully counterclockwise.jpg
 
Well the sim says , it won't work. The HOLD FET is getting no signal.
The fire FET is getting a signal equivalent to the period of the signal from the opto. Kickback is estimated at 4+ V.
 
Your Mosfets are source-followers instead of switches. For the source to be near +15V then the gate must be +25V which is not going to happen with your +6V logic supply.
 
I don't think this design is worth 'fixing' . 4000 series CMOS don't have voltage clamping on their pins and tend to fail in high transient environments.
I suggest a 556 or two 555's as a dual one shot (monostable)....with a small series cap in the input lines to ensure the triggering pulse is of short duration (<700uS).
Setup the first shot to last the 700uS that u want and the second to be 'tunable' between 700uS and 3000 uS for fueling adj.
Have both 555's drive an individual FET that delivers 2.5A to the load. Thus the FETS are paralleled for the 5A/2.5A fire/hold u want making both circuits nearly identical.
**broken link removed**

Download LTspice and design the cct there...it will sim everything for u, and you will have minimal tweaking with the physical build to compensate for any inductance (from the wiring loom) induced pulse delays to the injectors etc. It's not a good idea to copy bits of designs and cobble them together to make critical circuits such as this one.

Also use a unidirectional TVS & switching diode to clamp the kickback to perhaps 12V would speed up the off cycle of the injector. Place this AT the injector terminals if possible. Some snub testing,
**broken link removed**

If you want to go super compact you can switch to SOIC parts and use this dual NFET, with a 556 and a dual NFET your semiconductor parts count is reduced by 50% and reliability improved.
http://www.newark.com/international-rectifier/irf7380trpbf/mosfet-transistor/dp/73R1193

I'd suggest the higher temp. spec 556 as this app. is automotive.
http://www.newark.com/texas-instruments/na556n/ic-precision-timer-dual-16v-14/dp/06M7359
 
The IRF630 is not ideal for this job, since it needs a Vgs of ~10V to get its Rds(on) down to 0.4V. It is likely to run hot with Vgs = 6V and Ids = 5A.
 
All very disappointing. But thank you

We had something that worked but it was thought to be unnecessarily complex. It became unreliable. Its designer said it could be changed (elegant, was his word). But he has never got round (despite being paid).
 
Last edited:
Last edited:
OK - going back to §26 and §29, there seems to be two separate issues. (1) The Timer doesn't work. (2) The power circuit doesn't work (and cannot).
Well, the original timer worked fine. Its only since I modified as advised that its stopped working.
At least I can rely on that.

If these other links help with the power, maybe I'll have another go.

(Looking at the 555 page, I have no idea what units they talk about with frequency - I presume Hz - but its just numbers - where the 1.44 comes from is unclear - and it doesn't tell where to Put In my low signal from the RS optosensor. Is it Pin 7?)
 
Last edited:
The datasheet for a 555 shows the formula for its resistor A and its timing capacitor to be 1.1 x R x C= t (in seconds), for the 555 to be a monostable that you want. The 1.44 number in a formula is used with two resistors and a timing capacitor when it is an oscillator (in Hz) that you do not want.

The monostable will not time out when its trigger input pin 2 is held low too long so a coupling capacitor is in series with pin 2 to apply a short duration trigger then pin 2 goes high and allows a time out.

Here is a description of a coupling capacitor feeding the trigger input pin 2:
 

Attachments

  • 555 proper timeout.png
    555 proper timeout.png
    79.9 KB · Views: 237
§31, MOSAIC can you see if this works in your simulation? You'd have to ignore the RS307913 and just feed in a Low pulse (logic 0) at some frequency eg 40 Hz..
This timer control works on the Original Circuit (vintage 2004).
The 4039 is shown as a single Gate (on the LHS) and the 4013 is used for each Power pulse (one fixed, the other variable). As before, use 10 n for the variable cap, Ch.

On the Original Circuit, there's a damper with a link to ground via a BDA33, base to Pin 12 (Q bar) via 10k resistor. But I haven't drawn that in.

I apologise for not using the 555 but I know the timer is OK, so am more comfortable with it.
CIMG3227.jpg
 
Your new schematic (?) has two rectangles (ICs?) with no part numbers. Another problem is that the base, emitter and collector pins are not shown on the 2N6044 NPN darlington transistors.
Maybe you are using the 2n6044 darlington transistors as emitter followers, instead of as switches. Then when the base is driven up to maybe +5V the emitter is only about +3.5V or a little less.
 

Attachments

  • emitter follower.png
    emitter follower.png
    10.9 KB · Views: 221
Can both terminals of the injector solenoid be isolated from ground (engine block), or is one terminal permanently grounded?
 
§36
With the original set up, the "return" from the inj (the coil works a ferrite plate, so is polarised) goes to the current source O v (as if to engine block, hence to battery, but in our case, not involving engine at all). (In the LM1949 circuit, considered but not used, the coil is "at the beginning of the line" with the power control to ground coming after, ie "downstream" ).
That is true here (except its not built).

We would always keep the circuit in its own discreet loop, away from anything else, electric power source included. So the appropriate terminal is permanently O v but not randomly to "earth".

Of course, the diode is attached (but we found only 6 v kickback and MOSAIC estimated 4 v).
 
§35 Sorry: the rectangles are each side of a 4013, which worked well in our original circuit.
The little gate to the left is a 4093.

As for the Darlingtons, the 15 V source coming in at the RHS in each case (via a power resistor), is to the main Collector. The 6 v source coming in at the top, in each case, is to the Base of the controlling BJT.

In haste I have drawn the Darlingtons the wrong side. It should be as you show in your RH drawing. Thank you!
 
(In the LM1949 circuit, considered but not used, the coil is "at the beginning of the line" with the power control to ground coming after, ie "downstream" ). That is true here (except its not built).
That is the preferred arrangement (for simpler switching) and is why I asked the question in post #36. However, it's not what your schematic shows. You show the injector downstream of the Darlington.
 
#39. Yes, apologies. In truth, I had again copied a published circuit but sketched it wrong (implying ignorance).
This might have been answered before, but for my education, could two MOSFETs replace the Darlington?

You said: The IRF630 is not ideal for this job, since it needs a Vgs of ~10V to get its Rds(on) down to 0.4V. It is likely to run hot with Vgs = 6V and Ids = 5A.

Does this mean: it needs a high Gate voltage (potential) (10V not 6V) to create the Field required to shift electrons (or some other conductor) into the high resistance material, to allow conduction? On the Data sheet, 6V Vgs seemed to permit 5A. If it is important it would be easy to arrange for more volts.
 
Last edited:
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top