555 voltage controlled oscillator question

[Thedrive2112]

Hi again, I have built the circuit using the 4046 and it's working good except it's having trouble when I connect it to the load. I'm taking an output off pin 7 of the 4046 and feeding it into a 393 comparator so I can control duty cycle. The output of the 393 feeds into an op-amp configured as a follower and is used as a buffer. The output of the op-amp feeds the gate of a N channel MOSFET to control the on/off switching of the injector(3 ohm magnetic coil injector). Everything works great until I connect the injector to the MOSFET when connected it causes the frequency to significantly slow down on the 4046 even if I isolate the 4046 and there is nothing connected to the any of the outputs. I have the V+ side of the injector connected directly to the vehicles battery and the GND of the injector connected to my circuit to the DRAIN of the MOSFET. The only thing I can think of is that the injector coil is drawing so much current when connected that it is changing resistance of the resistor connected to pin 11 of the 4046. Anyone have any ideas of what I can do to fix the problem? Do I need to switch the V+ side of the injector rather than the GND side? Any help would be appreciated!

 

[KrisBlueNZ]

If the injectors are isolated from the block (two insulated wires going into them) then you should be able to switch either the supply side or the return side.

Can you draw up a schematic of what you have?

Check that the supply voltage for the 4046 isn't dropping when you connect the injectors. You might need a regulator, e.g. a 78L08.

Check that the control voltage to the 4046 doesn't change when you connect the injectors. You might need to feed the control potentiometer from the regulated rail.

I would power the low-power circuitry from a small battery and only connect the vehicle battery to the injector and the source of the MOSFET. The MOSFET source should be the only place where the two supplies connect to each other. Keep the heavy currents in the injector-MOSFET loop away from the circuitry.

Switching big-a$$ MOSFETs cleanly is not easy. Big MOSFETs draw significant gate current when the gate voltage changes. An op-amp is not a very good way to drive a MOSFET! I wouldn't be surprised if spikes from the MOSFET were coupling back to the comparator and into the 4046! I would use a MOSFET gate driver. The Micrel MIC4422 looks good: http://www.digikey.com/product-detail/en/MIC4422ZN/576-2320-ND/1029862

A gate driver chip needs to be really well decoupled - see the data sheet. It also needs a good solid supply. Ideally this would be the supply you use for your low-power circuitry, because the automotive supply will be very noisy, but it needs to be solid - for example, two 6V lantern batteries, or eight D cells - not some piddly little 9V battery.

The driver itself will put noise onto the power supply, so you might want to isolate the sensitive circuitry using a regulator as I mentioned earlier.

Also, what do the waveforms on pins 6 and 7 look like? Can you display them both on your scope and take a photo?

 

[Thedrive2112]

If the injectors are isolated from the block (two insulated wires going into them) then you should be able to switch either the supply side or the return side.

Got it.

Can you draw up a schematic of what you have?

Yea, I don't really have one drawn up yet mostly because I'm using bits and peaces from other demo circuits I have been trying to make work. I will hand draw one up and post it.

Check that the supply voltage for the 4046 isn't dropping when you connect the injectors. You might need a regulator, e.g. a 78L08.

I do have the circut running through a lm7812c 12v voltage regulator The supply voltage does not drop when injectors are running.

Check that the control voltage to the 4046 doesn't change when you connect the injectors. You might need to feed the control potentiometer from the regulated rail.

Wow,the control voltage does drop significantly (.3-.6 volts) when the injector rail is running. The control voltage is currently being fed from the vehicles pedal position sensor that will put out 0-5 volts depending on how far you push the throttle pedal. I'm feeding this into a op-amp for small amplification and buffering before it hits the 4046's VCO input. The op amp is fed off the regulated supply rail and I figured this would keep the voltage from dropping off but looks like that's not working. This seems like it's the main issue as these changes on the VCO input will change the frequency quite a lot. Any suggestions?

I would power the low-power circuitry from a small battery and only connect the vehicle battery to the injector and the source of the MOSFET. The MOSFET source should be the only place where the two supplies connect to each other. Keep the heavy currents in the injector-MOSFET loop away from the circuitry.

Ok, I will move the MOSFET off the main board and see if that makes a difference. As far as a battery goes I would then need to have some sort of battery charger circuit to charge the battery so I wouldn't need to change batteries all the time?

Switching big-a$$ MOSFETs cleanly is not easy. Big MOSFETs draw significant gate current when the gate voltage changes. An op-amp is not a very good way to drive a MOSFET! I wouldn't be surprised if spikes from the MOSFET were coupling back to the comparator and into the 4046! I would use a MOSFET gate driver. The Micrel MIC4422 looks good: http://www.digikey.com/product-detail/en/MIC4422ZN/576-2320-ND/1029862

http://www.digikey.com/product-detail/en/MIC4422ZN/576-2320-ND/1029862
Is this also the case with logic level MOSFETS?

Also is there a better way than MOSFETS to switch these injectors on and off? I don't need anything high frequency. I thought about relays but they would need to be replaced to often.

Also, what do the waveforms on pins 6 and 7 look like? Can you display them both on your scope and take a photo?

 

[KrisBlueNZ]

I do have the circut running through a lm7812c 12v voltage regulator The supply voltage does not drop when injectors are running.

OK, but 78xx regulators have a dropout voltage of around 2~2.5V. If you use a 12V regulator, you need 14~14.5V at the input, otherwise the regulator will "drop out", i.e. lose regulation. That's why I suggested using a 78L08 or 7808. Or is your automotive supply 24V? If so, then no problem.

Wow,the control voltage does drop significantly (.3-.6 volts) when the injector rail is running. The control voltage is currently being fed from the vehicles pedal position sensor that will put out 0-5 volts depending on how far you push the throttle pedal. I'm feeding this into a op-amp for small amplification and buffering before it hits the 4046's VCO input. The op amp is fed off the regulated supply rail and I figured this would keep the voltage from dropping off but looks like that's not working. This seems like it's the main issue as these changes on the VCO input will change the frequency quite a lot. Any suggestions?

Investigate your ground connections and topology (paths). Heavy current flow will cause voltage drop across wiring; even very thick wire will see a voltage drop if the current is high enough.

Where are you getting the throttle voltage from? From the ECU? What is the relevant 0V reference for the throttle voltage? Can you measure from the throttle voltage to the relevant 0V reference with a multimeter and see whether that voltage changes when you connect the injectors?

If the throttle voltage is actually changing, I don't know how you can fix that. But if you're measuring it relative to the wrong 0V reference, you may be able to fix that. If the throttle voltage isn't changing, but the VCO control voltage is changing, the problem must be in the buffer.

Ok, I will move the MOSFET off the main board and see if that makes a difference.

You don't necessarily have to move the MOSFET off the board. Just make sure that the current paths are controlled - make sure you know what current is flowing where, and make sure that voltage drops due to heavy current flow through wires are kept "outside" the circuit so they won't cause offsets within the circuit.

As far as a battery goes I would then need to have some sort of battery charger circuit to charge the battery so I wouldn't need to change batteries all the time?

I meant that you would use the separate battery as a test. If it solves some problems, this gives you a clue to the cause of those problems.

(use a gate driver because of high gate currents)
Is this also the case with logic level MOSFETS?

Yes. Logic level MOSFETs are like other MOSFETs; they just have a lower gate-source threshold voltage.

Also is there a better way than MOSFETS to switch these injectors on and off? I don't need anything high frequency. I thought about relays but they would need to be replaced to often.

IGBTs, maybe. What's used in commercial injector drivers? I would use MOSFETs or IGBTs. (Search for IGBT on Google, Wikipedia and component suppliers.)

 

[Thedrive2112]

[edit by (*steve*) to try to fix quoting]

OK, but 78xx regulators have a dropout voltage of around 2~2.5V. If you use a 12V regulator, you need 14~14.5V at the input, otherwise the regulator will "drop out", i.e. lose regulation. That's why I suggested using a 78L08 or 7808. Or is your automotive supply 24V? If so, then no problem.

Supply voltage should be 14v best case when the vehicle is running. If I have not seen any dropout on the power rails yet. I will order the regulator you suggest.

Investigate your ground connections and topology (paths). Heavy current flow will cause voltage drop across wiring; even very thick wire will see a voltage drop if the current is high enough.

Got it.

Where are you getting the throttle voltage from? From the ECU? What is the relevant 0V reference for the throttle voltage? Can you measure from the throttle voltage to the relevant 0V reference with a multimeter and see whether that voltage changes when you connect the injectors?

Yes it's from the ECU and there is no voltage change change when the injectors are running when I test directly off the pedal itself.

If the throttle voltage is actually changing, I don't know how you can fix that. But if you're measuring it relative to the wrong 0V reference, you may be able to fix that. If the throttle voltage isn't changing, but the VCO control voltage is changing, the problem must be in the buffer.

You don't necessarily have to move the MOSFET off the board. Just make sure that the current paths are controlled - make sure you know what current is flowing where, and make sure that voltage drops due to heavy current flow through wires are kept "outside" the circuit so they won't cause offsets within the circuit.

So right now the circut is on a breadboard and all grounds are connected to the same rail. Last night I moved the MOSFETs onto it's own breadboard and connected it to the main ground lug in the engine compartment. It seemed to fix the problem with the voltage dropping and freqency changing. However I burned up 2 of MOSFETs in just a few minutes this way. Is it possible now that my MOSFETs are closer to the injector rail I have greater inductive spiking that is overcoming my flyback diode and burning up the MOSFET?

I meant that you would use the separate battery as a test. If it solves some problems, this gives you a clue to the cause of those problems.

Good idea!

Yes. Logic level MOSFETs are like other MOSFETs; they just have a lower gate-source threshold voltage.

Oh ok, I thought that logic level MOSFETs have a gate driver type already built in?

IGBTs, maybe. What's used in commercial injector drivers? I would use MOSFETs or IGBTs. (Search for IGBT on Google, Wikipedia and component suppliers.)

 

[KrisBlueNZ]

Thedrive2112, you need to separate the sections that you quote from your responses. When you're replying to a post from me, your reply has to look like this in the editor:

[QUOTE="KrisBlueNZ, post: 1234567, member: 22166"]This is something I wrote that you have quoted. If you want to respond to it, you need to end the quote first, like this.[/QUOTE]
This is your response to the text that you just quoted. If you want to quote more of my post, you need to put a new opening quote tag at the start of it, as shown below. It doesn't have to have the username, post and member number, the way the first one does.
[QUOTE]Here's some more of my text that you're quoting. If you want to respond to it, you need to end the quote first, like this.[/QUOTE]
Here's your response to the second section of text that you quoted.

... and so on.

I fixed post #23 up for you, but I'll let you fix post #25.

 

[(*steve*)]

oops, sorry, I jumped in there.

 

[KrisBlueNZ]

Supply voltage should be 14v best case when the vehicle is running. If I have not seen any dropout on the power rails yet. I will order the regulator you suggest.

Good. Or you could use a variable regulator such as an LM317 so you could set the output voltage to 10V for example, or a low dropout regulator that doesn't need so much voltage between its input and output. Micrel make a nice range with output currents of 1.5A, 3A, 5A and even 7.5A. See http://www.digikey.com/product-detail/en/MIC29300-12WT/576-1118-ND/771587.

You should also look into load dump: https://en.wikipedia.org/wiki/Load_dump because you should protect your circuit from it. A bit of series resistance and a clamp of some kind is good - a MOV (metal oxide varistor) and/or a big TVS (transient voltage suppressor) diode. This is outside my experience but I suggest you do some Googling for load dump protection in automotive electronics otherwise your circuit might suffer a meltdown one day in the future if you don't hold your tongue at the right angle when you turn the ignition off! I suggest the Littelfuse V18ZA40 MOV with a clamping voltage of 37V at 20A as a starting point.

Yes it's from the ECU and there is no voltage change change when the injectors are running when I test directly off the pedal itself.

OK, good. Then at least you have a chance of avoiding the problem!

So right now the circut is on a breadboard and all grounds are connected to the same rail. Last night I moved the MOSFETs onto it's own breadboard and connected it to the main ground lug in the engine compartment. It seemed to fix the problem with the voltage dropping and freqency changing. However I burned up 2 of MOSFETs in just a few minutes this way. Is it possible now that my MOSFETs are closer to the injector rail I have greater inductive spiking that is overcoming my flyback diode and burning up the MOSFET?

I would definitely not put the MOSFET on a breadboard. The currents are far too high. Here's my suggestion. The diagram shows everything except the insulators.



I'm assuming that the MOSFET has a metal tab. You need to insulate the MOSFET from the sheet using an insulating pad (a thin mica sheet or a silicone-impregnated pad) between the MOSFET tab and the metal, and an insulating collar around the mounting screw.

The tag marked D must make direct contact with the tab of the MOSFET. If the MOSFET has a full plastic package, you'll have to connect to the middle pin, instead of cutting it off as shown in the drawing.

The other tag, connected to the cathode of the big back EMF suppression diode, also needs to be insulated from the aluminium sheet, since it has +12V on it.

I didn't label the back EMF suppression diode, but it's very important. Connecting it next to the MOSFET gives the best protection for the MOSFET.

I've shown the wide copper area on the stripboard being screwed into the aluminium sheet. This is the 0V reference so it's OK to have an electrical connection. In fact it's probably best to have the aluminium sheet connected to 0V. This wide track can be formed by bridging two (or even three) adjacent tracks together using solder or some sort of braided wire.

The gate (left hand pin) connects to a single track on the stripboard, which connects to a 12V 1/2W zener to the source, and a 10 ohm series resistor. This zener protects the gate from voltages that could damage the device.

I've added a big capacitor, marked "DECOUPLING CAP", from +12V to 0V. This should be a good quality ceramic or film capacitor such as http://www.digikey.com/product-detail/en/ECQ-E1106KF/EF1106-ND/56416 (10 µF, 100V). The value isn't important but I suggest at least 1 µF.

You can see how I've kept the high-current paths independent of the low-current paths.

Oh ok, I thought that logic level MOSFETs have a gate driver type already built in?

No! That would be nice, wouldn't it. No; they just have a lower gate-source threshold voltage so you can drive them from a circuit operating from a low voltage, e.g. 5V instead of 12V.

oops, sorry, I jumped in there.

Thanks Steve