Hi, Gandolf
First as to your questions...
Gandolf wrote:
Thanks, rainbowsally, much appreciated.
I have uploaded the current (in-development) .asc to the TEMP MATURINf folder.
I have also uploaded TL081.301 and OpAmp.asy...I have a few puzzlements about the way LTspice is running this sim....
Is the simulation actually using the TL081.301 model (.subckt)?
Yes. It's an op amp w/ jfet inputs (4 uA input currents measured at the op amp). Probably even a 741 would work though.
Is it OK that the TI model has the filename extension .301 instead of .mod or something like that? (And why the heck did TI give it the .301 extension in the first place?)
Notaclue. ;-) Might be a model revision number?
Why do I have to keep increasing the time step to as high as 1ms (where it is right now) to get the sim to run without a "time step error" message?
Good question.
It may be that you had checked the 'skip initial operating point solution". This forced the circuit into a very non-linear range and the attempt to converge to realistic values for all the components (by iterations using Newton's method).
General solution. Allow it to find the initial operating point, even if it runs a little slower.
Why is it that the comparator is running in the sim even though I didn't include an .include command for its file in the folder?
The comparitor didn't even show up on my schematic.
I plugged in a LT1017 which looks like it's at least pin compatible for now.
So many questions, so few years left....
VB, Maturin
PS: the reason for the various low values of capacitance throughout the circuit is I only use film caps; I refuse to use electrolytics that will render my devices defective in 10 years; this is a real circuit for a real device that will be produced in small but meaningful quantity for people who want it to work a very long time.
Mylar (film) generally has good temperature stability too. Good choice.
PPS: the 555 will actually be the TLC555 from TI, CMOS version, that doesn't spike the supply rail nearly as badly as all the BJT versions, but I didn't bother trying to import a model and .asy for the CMOS 555.
We might want to model it.
Now for some feedback on your circuit.
First of all, very nice schematic!
Now for things that you might not have seen and/or that might be improved. These are general observations and may not be applicable to this exact circuit as I have not run it through anything but a cursory check so far.
1. Availability of parts models.
As mentioned above, I don't think your comparator is an a standard LTSpice d/load. I used any old pin-comptible one that I found. This doesn't apply to the circuit itself, only to the simulation.
2. Dynamic range.
I don't think you want the 10% hysteresis on the comparator (U4). That's the 100K from the output back to the non-inverting input. This will limit your ability to get vary narrow or vary large pulse widths and is unnecessary because you don't have noise on the inputs that could possibly cause a false trigger (seen as spike in the output).
If you do get spikes in the output, you might want to decouple the supply to the comparator itself with something like ten ohm resistors and .1 caps.
3. R9 is unnecessary because the input impedance of the comparators is quite high and the output inpedance of the opamp is quite low. If for some reason the opamp isn't able to drive the tiny capacitance of the comparator input, then no problem but for this unity setup gain that's incredibly unlikely.
4. Your feedback network for the opamp is probably unnecessary. The 1K from out to the inverted input could probably be replaced by a short and the 100K (R8) could probably be eliminated. Basically what you have here is an offset adjustment that is almost certainly unnecessary.
5. Similarly, the 10K (R6) is probably unnecessary. The input current at the opamp is in the range of 37 pico amps (thousandths of milliamps).
In applications where dc offsets are critical at each stage you may need to balance the voltage drops between the two inputs, but almost certainly not here.
This stuff does affect cost and until I can think of something to write about switching power supplies which is presumably the actual topic at "Switcher CAD LTSpice" forum, hopefully no one will be too offended by these notes.
If drilling holes isn't a cost issue, you might want to design them in and just put jumpers in some and leave the others empty so you can optionally use the resistors if you replace the opamps or comparators on your board with cheaper but adequate parts as prices and availability may fluctuate over time.
Now I'll play with your model for a while (make the pots, etc.) and see what you've got going on.
:-)
Stay tuned. We may have a sub-forum coming for discussion of issues anyone thinks are off topic for this forum.
