How can I build a CMOS current generator ?

[ronv]

Yes, it does slow down the turn off of the coil. Usually a zener diode is used - not a regular diode - to improve the turn off. You need the diode to prevent the voltage from the inductor from rising to the point it breaks down the switching transistor. You can select the zener(s) for a voltage safely below the Vceo of the transistor to optimize it, but remember the zener is dissipating the energy stored in the coil so needs to have a high enough power rating.

 

[malc9141]

I do appreciate this. But I think I'm a bit dim. What you say sounds OK intuitively to me but when I try to imagine what's happening, it doesn't.

If the coil went straight to earth, no "bypass" back to the "top", wouldn't that be the fastest way to collapse the field. Or would it cause something to fizz because of a very voltage? I stlll don't understand why I need the diode at all.

You suggest the beefy Zener. As your argument stands, this would allow more induced current to bypass the diode???

 

[ronv]

Sometimes a picture is worth a thousand words.

The first one is with no clamp at all. The voltage at "Out" goes to several hundred volts which will blow the transistor. This is why you need a clamp.

The second is with just a diode across the coil. Since the diode drop is only about .6 volts the power from the coil is dissipated but slowly. Note the coil current.

The 3rd is with a zener diode to ground. Since it is a 68 volt zener the power from the coil is dissipated faster.

 

[malc9141]

Very many thanks! So if I want a "reasonably" fast drop (driven by 68v), I should use the Zener but not the diode?
Malc

 

[malc9141]

Very many thanks! So if I want a "reasonably" fast drop (driven by 68v), I should use the Zener but not the diode?
Malc

 

[ronv]

Yes, but 2 things to remember (or check):

1- That the breakdown voltage of the switching transistor is higher than the zener voltage.
2- That the power rating of the zener is high enough that it doesn't overheat. - maybe 2 35 volt zeners would be better than 1 68 volt one.

 

[moffy]

If you want max switch off speed then you want to maximise the voltage during switch off. Typically if you place a zener diode in series with the diode it will increase the turn off speed. The zener should be the other way around than the diode so that it is at its breakdown voltage during switch off. Remember to match the current/power rating of the zener to its load/duty cycle.

 

[malc9141]

I completely understand the general point here.

The only thing I'm still not clear about is shown on your circuit diagrams: -
Attachment 3, the coil goes to zero V via the Zener(s). On Attach 2, it goes to "OUT". Is "OUT" 0 V ?

In my actual circuit, which works fine but maybe discharges too slowly - the whole point of this (very welcome) discussion - my circuit has the "bottom" of the coil going straight to 0 V, but there is a bypass back to the "top", via the diode.
In other words, it would be just like your Attach 3 if the diode was substituted for the Zener.

So if I have the diode feeding back (which I have), but a clear run from the "bottom" of the coil to 0 V, is there still a slow discharge????

 

[ronv]

Maybe a drawing with the part numbers is in order.

 

[malc9141]

View attachment 68864

This is sketch which should clarify the question. Sorry it's crude.

A is the circuit I have, at its business end. About 16 A at 25 V from Current Generator for ~2 ms. The diode is shown.

B is what you are suggesting. The Zener is shown.

Question: A gives rather slow discharge ??
B much faster?

 

[ronv]

Assuming the switch is at the bottom of the coil.. B is faster.

 

[malc9141]

The switch is at the Top !!

The Hall Sensor (+ve input) > Inverter circuit (+ve) > Current generator (+ve) > "top" of Coil (+ve).

So the existing circuit, A, simple diode, is optimum ???

 

[ronv]

The Zener hooked up like this is fastest.
Be careful that the voltage rating on the transistor is higher than the supply voltage plus the zener voltage. A large capacitor from the +24 volts to ground close to the diode is also a good idea.

 

[malc9141]

A Year ago.

Thanks if you are there!

I see the coil "wastes" via a Zener. No diode snuffer.

Malc

 

[alec_t]

Here's a mod of the circuit I posted way back, showing how the zener would be connected to shorten the switch-off time. As Ronv said, the voltage rating of the zener (40-70V suggested) should be less than the voltage the switching device (MOSFET in my circuit) can withstand. The zener should also be rated to pass at least ~8A.
View attachment 69005

 

[malc9141]

Thanks. Makes sense & the details are welcome.
Checked, our coil is 4.5 ohms but I don't know the uH. I guess your zener characteristics still stand.

 

[alec_t]

I guess your zener characteristics still stand.

For a 4.5Ω coil and 24V supply the max current is ~5.3A, so the zener should be rated to handle that current.

 

[malc9141]

Getting the hang

Going on from Ronv and Alec's ideas, would this work, in principle?

(The obvious point I had overlooked was that the current in the coil has got to be ~ 5 A (V/R) not what the iGen is capable of delivering! And of course with the time factor for induction, it's less current still.)

Now, if I want everything to go faster (as it does in the car, with this Delphi injector) would the accompanying set-up work? I'm utilising the emf to drive the coil, using the zener to control voltage.

I hope it's not View attachment 69062!

(NB Industrial secrecy ensures that we can't use the car circuitry. When my friend examined a stripped Engine Control Unit, he couldn't see evidence of the coil-energy-harvest that I'd wondered about - and propose here).

 

[alec_t]

Not sure how that's supposed to work. The diode from +24 to the coil top clamps the emitter to 24-0.7, so the base voltage is 24-0.7+0.7 = 24, so no base current flows through R.

 

[malc9141]

I should stick to what I know. I would never have spotted that. Trouble is, someone has to improve this engine's firing!

And I realised there were other stupid oversights.

On quite a different question, in your earlier diagram No 75, what is the V1 circle? Somewhere, you explained the "little circles" but I can't find it now.