For testing I will definitely need to have everything adjustable, but once I've got it all figured out I will want to set the values of everything so that it can't be changed.

I suppose the comparator can be figured out during testing, like you mentioned. I'll look up those other ICs and see if they have schematics that can explain the comparator system to me a little bit better.

For testing in LTSpice, I would just eliminate the TL494, and supply a square wave to the transistor that drives the MOSFET, and 12V to the inductor before the MOSFET?

I expect to have to work through plenty of mistakes, that's how you learn!

The MOSFET I'm planning to use for the solenoid is this one, and it says that it's Vgs is 4.5V. It also has a threshold voltage of 20V, though. I'm not entirely sure how to interpret those numbers, but from everything I've read, 4.5V should be enough to activate it.

If not, I'll just add an NPN like you mentioned.

Regarding the Thai schematic -

The hardest part I've had with all of this is just grasping the insane amounts of information that I've learned in the last 5 days. When you posted that Thai schematic, I couldn't even look at it without being overwhelmed. I had no idea what I was looking at, and pretty much just ignored it.

Now that I've had some practice reading and building schematics, it's not nearly as difficult to understand, and I can actually pull information from it.

It has definitely been fun jumping in feet first into something that I had zero experience with before starting.

 

1. Yes for LTSpice the testing is done without TL494.
It does not produce a square wave, but information about the correctness of the output stage can be obtained using only a square wave.

It is true you won't have voltage regulation in this case.

But if for instance nothing is produced on the output, then there would be a blatant mistake.

2. The MOSFET you want to use just barely turns on at Vgate = 4.5V.
It needs 8V somehow. But most MOSFETs behave that way you can also see that in LTSpice. For the solenoid simply use a 6 Ohms etc. resistor.

If for instance output voltage disappears or goes down dramatically when you supply a square wave, there might be something wrong as well. Simulation is cheap you can change the circuit/try many things at virtually no cost.

 

So I've been playing with LTSpice....aside from the millions of other questions I have, the one thing that stands out is that the lower inductance coil I use, the higher voltages I'm getting out.

The resistances on these coils on Newark are so low that the frequency that I am pulsing at almost doesn't matter.

I'm having a really hard time finding a coil that will give me the results I want. I suppose I could put a resistor in series to bring the voltage down, but that just doesn't seem like the right thing to do here.

 

No this would not be the right thing to do.

I'd say something like 200uH would be the right value, but I might be wrong.

Would it be possible to make the .asc file available?
Or send it to my mail if you don't want to make it visible on a public forum.

Don't take the output voltage too serious.
Because normally there is feedback active, so once it will reach a certain value, the power supply is stopped.

If you get output voltage higher than input voltage, and this keeps stable at 1A or 2A, this is a good sign already.

 

No this would not be the right thing to do.

I'd say something like 200uH would be the right value, but I might be wrong.

Would it be possible to make the .asc file available?
Or send it to my mail if you don't want to make it visible on a public forum.

Don't take the output voltage too serious.
Because normally there is feedback active, so once it will reach a certain value, the power supply is stopped.

If you get output voltage higher than input voltage, and this keeps stable at 1A or 2A, this is a good sign already.

I'll upload the file this afternoon, I want to play with it a bit more. I had it making some really nice curves last night and then I changed something and all of a sudden everything got worse. I want to try to get it back to where it was, and then I'll upload it.

How do I measure current? So far I've only been able to figure out how to measure voltage. Do I have to replace the voltage source with a current source and only monitor current? Or is there some way to overlay the two?

 

I'll upload the file this afternoon, I want to play with it a bit more. I had it making some really nice curves last night and then I changed something and all of a sudden everything got worse. I want to try to get it back to where it was, and then I'll upload it.

How do I measure current? So far I've only been able to figure out how to measure voltage. Do I have to replace the voltage source with a current source and only monitor current? Or is there some way to overlay the two?

Measuring current is easy. If you hover with the pointer over components, when the simulation already was performed, the pointer will change, according to position. If you click wire connections, you get voltage level.

If you click the right position for wires going into components you will get the current. You can display more than one together.

If the display goes off scale, just repeat selecting sources. Once you have understood how it works, it is really quite easy.

Also I suggest to select "Skip initial operation point solution" (or a checkbox with a very similar label). To change this later on, click the text in the schematic which contains a few numbers in a sequence. These are the simulation parameteres. Maybe you figured out already how this works.

Also as I suggested earlier you don't need model data for a solenoind, simply replace it with a 6 Ohms (etc.) resistor. Simulation resistors can't burn out

 

Well, this is the best I could do. I don't know why it surges after the switches at the bottom of the schematic are turned off. And I'm not getting anywhere near the amperage that I need.

 

OK I found some grave mistakes. For instance, not having resistors in the base path of BJT. And the MOSFET gate not grounded.

I have changed the pulse timing to 30uS on/60uS period.

I have changed the coil resistance to 50 mOhm. Which is quite low.
I have added a 10 Ohms load for the time being.

The bottom part of the schematic I have not yet considered.

As well time scale was changed to 10mS.

I am getting a voltage level somewhere around 30V.

But I have not tested just to charge 10000uF, and then to switch on the solenoid, switch off the input pulses at the same time. So that means, only relying on the stored energy.

Maybe the changes I made will make it easier for you to proceed.

That is what makes designing these circuits so difficult. You simply need real-world experience which kind of coil and which kind of transistor drive will work for a certain power level. And then you can reduce components as much as possible to save costs.

But relying on the stored energy to turn on the solenoid in general is a good idea. Having a low overall duty cycle, this circuit should not be too much difficult to implement. A 0.05 Ohms inductor is here even able to supply ~30V permanently...30uSec/60uSec was just chosen by random as a starting point...typical SMPS frequency I think.

 

OK I found some grave mistakes. For instance, not having resistors in the base path of BJT. And the MOSFET gate not grounded.

I have changed the pulse timing to 30uS on/60uS period.

I have changed the coil resistance to 50 mOhm. Which is quite low.
I have added a 10 Ohms load for the time being.

The bottom part of the schematic I have not yet considered.

As well time scale was changed to 10mS.

I am getting a voltage level somewhere around 30V.

But I have not tested just to charge 10000uF, and then to switch on the solenoid, switch off the input pulses at the same time. So that means, only relying on the stored energy.

Maybe the changes I made will make it easier for you to proceed.

That is what makes designing these circuits so difficult. You simply need real-world experience which kind of coil and which kind of transistor drive will work for a certain power level. And then you can reduce components as much as possible to save costs.

But relying on the stored energy to turn on the solenoid in general is a good idea. Having a low overall duty cycle, this circuit should not be too much difficult to implement. A 0.05 Ohms inductor is here even able to supply ~30V permanently...30uSec/60uSec was just chosen by random as a starting point...typical SMPS frequency I think.

I think that frequency is too high for the TL494. According to the spec sheet, the most it can do is 300Hz, which would be a period of 3333uSec.

Finding coils with all of the specs that I need has proven to be fairly difficult, honestly. If it has the inductance I need, it doesn't have the amperage I need, and vice versa.

As far as adding the 10 ohm load goes, I already had a load on the system at R3, it was set to 1000 ohms, but I've been playing with all kinds of different resistances. Does it make a difference if the load is in series (R3) or parallel (R7) with the MOSFET M2?

The graphs that are output by this setup look great, though I don't think the frequency it's operating at is possible, which of course lengthens the charging time.

I've been having a hard time with the MOSFETs. Apparently, to get a certain voltage through them, you have to nearly match that voltage at the gate. Is this true? This would mean that I have to feed the gate from the source, and then switch that using another MOSFET or transistor, but then you have the same problem with that one needing the same input voltage at the gate.

I haven't had much time to play with this new design, I've been pretty busy. I'll try and look at it some more this week. I've found using LTSpice to be pretty frustrating, it seems like none of the changes I make to the schematic do what I want them to do, and I can't figure out a pattern to any of it. It all appears to act randomly, regardless of the changes I make.

 

Just kidding, I reread the user manual, the TL494 operates at a max of 300kHz, not 300Hz. That makes a big difference...

 

I think that frequency is too high for the TL494. According to the spec sheet, the most it can do is 300Hz, which would be a period of 3333uSec.

Finding coils with all of the specs that I need has proven to be fairly difficult, honestly. If it has the inductance I need, it doesn't have the amperage I need, and vice versa.

As far as adding the 10 ohm load goes, I already had a load on the system at R3, it was set to 1000 ohms, but I've been playing with all kinds of different resistances. Does it make a difference if the load is in series (R3) or parallel (R7) with the MOSFET M2?

The graphs that are output by this setup look great, though I don't think the frequency it's operating at is possible, which of course lengthens the charging time.

I've been having a hard time with the MOSFETs. Apparently, to get a certain voltage through them, you have to nearly match that voltage at the gate. Is this true? This would mean that I have to feed the gate from the source, and then switch that using another MOSFET or transistor, but then you have the same problem with that one needing the same input voltage at the gate.

I haven't had much time to play with this new design, I've been pretty busy. I'll try and look at it some more this week. I've found using LTSpice to be pretty frustrating, it seems like none of the changes I make to the schematic do what I want them to do, and I can't figure out a pattern to any of it. It all appears to act randomly, regardless of the changes I make.

To switch a p-CH MOSFET on, you typically need -8V on the gate.

For a coil since you want to use the capacitor charge, you don't need 5A or 10A coil, 1A or 2A would be enough. Let say 200uH (220uH) looks like a good starting point. I don't know if 10000uF are really required. Maybe it can work with less. Buy you need suitable wires from the capacitor to the solenoid.

Yes TL494 = 300 KHz.

I have added 10 Ohms from the output capacitor for simplicity.
If the circuit works (in the simulator), then add the functionality to switch it with the MOSFET.

If you use n-CH MOSFET, the gate voltage needs to be about 8V. They are often used in the GND path, so you need 8V above ground.

The resistors used for the switching MOSFET need to be quite low, and rated for 1W. For 12V maybe also 100 Ohms could be used, and the 220 Ohms resistor left out.