My Covid-19 project

Falstad's javascript editor has far less features than any Spice program yet far easier and more visual with as many scope probes as you can watch at a time.

There is a learning curve to know how to use it, but I love seeing everything. If you want, you can make anything variable, Beta or gm on FETs, threshold voltage, Supply voltage, Cap ESR using ctrl+ mouse wheel or the component SLIDER menu option.

It beats everything else for me.

Again I think your approach is wrong using a voltage driver with emitter follower and cap load to a valve instead of a PWM voltage controlled current source.

This is my proof of concept

You can specify and adjust some parameters, change options > time sample rate, simulation speed scroll > everything. Just no trace PLL trigger like a scope, but there are event triggers.

We're running up to 15 valves at once, and many need voltages higher than the battery. Furthermore, every valve needs its own constantly-changing power level. The system (hardware + software) continuously, and independently, meters the power to each valve - it's power rather than current that matters in this case.

But you can't use specialized op-amps, transistors or FETs, right?

--
 Thanks, 
    - Win

Yes, I can have a colored font, but it's much more useful to do word or value highlighting. I have the same frustration when using Word. I can highlight a pdf printout of a Word document, after the fact, but I can't insert any highlighting, while composing.

--
 Thanks, 
    - Win

What soft of 'highlighting' are you trying to achieve in Word?

It has myriad options to manipulate the text colour, background, outline etc. Like Adobe it has a (multicoloured) highlighter tool if you want...

--
Cheers, 
Chris.

OK, thanks, you're right. Given your assertion, I looked harder, and found the highlight tool. Thanks again!

--
 Thanks, 
    - Win

You can define simple characteristics like gm, hFE, OA gain, GBW in filter designs built-in but add DCR, ESR, Rs to OA output for current limits.

Your foldback current limit was miss-drawn. But unless the output is shorted, the CL doesn't do much, so you can simply omit it.

After fixing your drawing, it works, but doesn't match my calculations, or SPICE modeling with proper models. For example, on a square-wave input, it's fast, but shows an unrealistic 4A Q4 Q12 charge / discharge current, with resulting slight overshoot. Changing to a LM741 opamp,** and setting the slew rate to 18V/us helped, but it failed to apply the op-amp current limit to the sinking output. So I reduced the PNP beta to 25. Now it comes closer to the real thing.

** With a "real" opamp, it runs slower.

--
 Thanks, 
    - Win

Correction:

--
 Thanks, 
    - Win

Using Falstad requires that you comprehend and specify the source and load impedances of every component, then do your best to add parts to match them . This can be done for any internally compensated Op Amp by adding a serie s R and feedback pF or simply, focus on phase margin effects only when you need to. e.g. adding trace inductance, crosstalk capacitance, DC, ESR.

Modeling of anything comes down to physics and your awareness of assumption s and effects.

So one starts with a * OHm for a speaker for sub-woofer and power amp with damping factor > 1000 or 8 milliohms. In order to reduce resonance effects from back EMF and inertia.

However, for radial or linear motors, you want to control current which is a function of force and back-force for a balanced servo. Since this is a q uasi-static system , the system operating point is analyzed by DC , and Hea viside disturbances.

The "perfect' design from my over 4 decades of bleeding-edge experience wi th Hitachi, NPL, Fujitsu, Toshiba, and many world 1st designs of my own, is SIMPLE, meeting good design & product interface specs. Which are lacking here, so everything could be anecdotal? In general this design looks not b ad, but there is a lot of room for improvement with interface specs. The i nternal document specs annotated is exceptionally well done , unless you co mpare it to a TV service schematic. Readability might be improved , but wh o am I to judge. I can barely sign my own name after right arm ischmemia an d bypass to arm with nerve damage at both ends. (Phrenic and Median)

Whoa, that sounds serious. But at least you can still run a mouse ** and type!

** BTW, I'm a big fan of trackballs, rather than a mouse. You only move your thumb, not whole hand or arm. It turns out thumbs work really well - fast accurate, no carpal tunnel issues, etc. I use a Logitech M570.

--
 Thanks, 
    - Win

Trackballs have (or at least had) the reputation of knackering your thumbs. (Lightpens, your shoulders!)

I've set up my mouse acceleration so that I can move it around almost the entire screen with only my fingers. My wrist doesn't move.

Reminds I was the same until I wore out the neck wires. It was after a 4 Roto-Rooters to the right arm and no right arm pulse they decided to use a leg vein to replace right arm vein from clavicle to elbow. Now good plumbin g but some electrical problems with nerve damage from the surgery on 1st 3 digits so I type with 4th digit on my right dominant side. The phrenetic ne rve damage caused a false sense of drowning just submerged below the surfac e of the Gulf of Mexico last Jan. Freaky minor pressure stimulus to freaked out Phrenic nerves with an aqualung overhead to walk in a coral garden. I had to cop out, immediately.

Yet last year I had the best game of golf in life 1 mos before surprise sur gery on a driving tour through NY, Boston, Cape Breton, PEI, NB and Que.

The CMOS Op Amp is not much different in the output stage from a CD40xx buf fered inverter Zo=300? . You can verify this with current specs ? ? V/I = Zo=RdsOn vs Vdd. This allows wide operation to 16V where Zo decreases with Vdd from std Vgs rules.

The 74HC' series= 50? or 74ALCV'= 25 ? both with 50% toleran ce just as most FETs have RdsOn tolerance and thermal effects.

The critical issue is the interface impedance specs and PWM range.