Solving the Variable Ebrake-Throttle Problem

[Njay]

[*]R100 & R101 become a voltage divider that takes care of a pesky problem with proper Gate Voltage. It turns out that the Gate of U1 activates at about 1.9V, so when Throttle is at 0%/5K Ohms, SP will read about 3.33V which is too high for U1 to shut off… hence the divider. I don’t know enough about circuit design to say where it is a good design, but I managed through it.

You'll need some margin, because this 1.9V threshold (Vgs(th)) will differ from device to device (manufacturing differences; 0.8V to 3V) and conditions like temperature.

[*]Conversely, when Throttle is at 0% the R100 & 101 Voltage Divider drops the Gate voltage below the threshold and cuts off U1 with some small leakage (I think).

I now see that you have a short range to work with, with the throttle voltage approximately in the 2..5V range; cutting let's say 2.5V from this voltage to make sure the MOS turns off leaves only 2.5V for it to use to turn ON. May be short, meaning that the resistance presented in its D-S terminals may be too high to have the desired effect; in that case another FET with "better" characteristics would need to be used (BSN304 is a little better) or the design changed (several possibilities, from adding an extra transistor to using a voltage comparator). Since it's a simple circuit and you'll be building a one off, I would say just build it and measure it

I do not know how to use the Advanced settings and create a slope of the UV voltage, so I can’t tell if it is linear or has other oddities.

PS - I just went back and reworked values on R100 & 101 using 19.6K & 24K Ohms respectively; I fear too high of value would slow the Gate charge too much, but I defer to the ES EE graybeards for the final say.

Unfortunately I'm not one of those graybeards ones but I think it's better, lower values, and would say you should try to use "standard" resistor values as they are cheaper and easier to find (the ones in the sequence 1, 2.2, 3.3, 4.7, 5.6, 6.8, 8.2, ...).

 

[Kingfish]

Njay: Understood and comprehend all that you have said. With the resistor values: That’s was I was doing on the very last iteration, setting to common values available per Digikey. I know with a bit more study this could be accomplished a little easier using local supply as well. I also would like to look at BSN304 for improvement. 8)

With regards to BigMoose (with whom I have absolute respect) and the Current-Shunt Solution, it is not for me to say it is better or worse. Honestly I didn’t study it too closely because the whole shunt business has been difficult for me to work with since Day-One beginning with trying to calibrate the CA-DP. Given the strongly positive reactions to that direction, I owe it to myself to take a closer look.

However – I want to finish what I started first.

Today I wish to focus on adding back the sense circuit for –EBS which shall include the Blocking Diode for 2WD. The primary concern is that R12 (UV) controls both LVC and HVC, and this circuit dorks with both settings when active. We don’t care as much about LVC during Regen, but we certainly must care during normal operations and I wish to attend to that.

Restating what this mod proposes and what it doesn’t:

• R12 (UV) in conjunction with the Controller Firmware controls HVC and LVC (natch).
• If you have modified R12 to change the values of Regen, this mod offers a way to extend that result so that there is range between the high and low limits. In other words, it takes the maximum existing differential and reduces the gap by lurking on the Throttle position.
• With Ebrake enabled at WOT, with proper calibration - Regen is the same as before the mod.
• With Ebrake enabled and at reduced Throttle and going to Zero Throttle, Regen should reduce to a minimum configured value.

Caveats:

• LVC goes higher than pre-configuration when using this circuit however it should not matter if we are in Regen.
• There is the possibility that the Controller only sees the value of R12 (UV) once during Ebrake, or possibly at Boot time; we simply don’t know. If the Controller looks once at Boot, then we’re done and this method won’t work. If it looks once during Ebrake, we could also be shot down.

For the moment, we have to hope for the best and solder on.
Thank you, KF

 

[Kingfish]

BSN304:
Hmmm, appears to not be available. I have tried to source nearest equivalents and came up with this short list:

• VN10LPSTZ – no Spice file.
• ZVN2110A
• ZVN3306A

Suggestions welcomed.
~KF

 

[Alan B]

Need to finish that high-end controller.

Another option I've been considering to get really variable and powerful motor braking is to make a plug-braking type system. It looks to me like a three phase bridge, a PWM'd FET or two and a load resistor would do it. It would take a few more parts to make the circuit regen into the battery, but that is also possible. This would allow the controller to cool of during braking, and I suspect it could provide more braking than is available with the standard regen settings, and continue to low speed where controller regen usually drops out.

 

[Kingfish]

Plug braking sounds interesting, though the way I understood it, there is a lot of heat generated by shorting out the motor. Although I could be wrong.

OK, here is Version 4:

• This one uses 4N25 Optocoupler @ U2 to turn the Mod ON & OFF. It also acts to prevent reverse current when linked to a 2nd controller for 2WD; two birds with one stone.
• Imagine there is a switch just below –EBS; LTSpice doesn’t allow me to draw on in, so to simulate – I cut the trace.
• With –EBS shorted to GND, EBrake = ON, otherwise it is not.
• The difference is that when at WOT, Ebrake = ON, electric braking will be HARD and based upon the values of R12-A & B. When HVC is set at this level, is must be below the Controller’s maximum FET voltage level including a little bit of margin for error.
• When Ebrake = OFF, regardless of Throttle position, the values of all three R12’s A through C are realized and HVC goes to the low setting. It is important that we ensure HVC at this level is about a volt over the highest charge of the Battery Pack.
• Looking at the Resistors, I think that I’d like to have R12-A as a high-turn precision POT. The present schematic is for 15S/63V, but for other systems, likely R100 & R101, as well as R12s will require customization.
• The challenge to keep in mind is that SP must fall between 3 to 5 V when Throttle goes from 0 to 100% (5K to 0 Ohms). UV will depend upon the R12 setting; I suggest using Hyena’s spreadsheet as a guide for the high and low side of HVC and LVC.

Upon buyoff by Njay - my friendly EE, I’ll commit to purchasing parts and spin up a breadboard for validation.

Concerns:
Aside from the obvious, I worry that –EBS will not have enough current to affect U2. I tried this before with a relay and it didn’t work. It may require a little amplifier/driver. I won’t know until I get there.

Thoughts and opinions welcomed, KF

 

[Dlogic]

Electronics is not really my thing. Why do we have to pull the brake lever actuating a switch, or press a regen button? How uncomfortable!!

Some years ago I came up with the idea of a 2 way throttle. When you twist it towards you acceleration occurs and turning it away from you starts the variable regen. The more you twist it in the opposite direction, the harder the ebrake effect. The throttle should have a feelable center point where nothing happens at all. With such a setup you'd barely need the normal mechanic brakes and would enjoy the comfort of an easily controllable ebrake.

We could build 2 pots into the throttle. One for acceleration and one for that variable ebrake.

 

[amberwolf]

That is what I would lke to do as well, but it requires manufacture of a new throttle unit. I have designs in my head for making a durable one, that oculd be made either by CNC machining or printing on those nifty 3D-plastic-printer units recently discussed elsewhere on ES, but I haven't had time to pull it out of my head and into CAD to see if it can actually work as I see it (or be manufactured!).

It doesn't even need two pots, or two halls, though the latter might simplify things.


Still, until Kingfish or someone else comes up with something to make a bog-standard controller operate with variable regen, it doesn't matter if the brake control has an analog output or not.

 

[Njay]

You will need a current limiting resistor in series with the opto's LEd (diode), either at the anode or cathode, or the LED will go pooof.

EBS doesn't provide any current, it's just an input with possibly a weak pull-up (a resistor from EBS to 5V); the 5V power supply must be able to source enough current for the LED. I'm not sure if the transistor part will work ok, I would also have to simulate/test to know.

I'm not sure if you're modeling the throttle correctly; is R8 & R9 inside the controller along with the 5V supply?

I was going to send a hint for variable voltage simulation on the other post but a phone call got in between and I forgot . See the image below; if you need further clarifications go ahead and ask.
You may want to give a value to the "Series resistance" in the "Parasitic properties". This value represents the supply's internal resistance, which may be important to simulate (this resistance defines the voltage sag when current is in demand from the source); for the throttle you can put something like 5K.


.

 

[Kingfish]

Dlogic: That reminds me of how the Throttle works on my tractor 8)
As AW suggested, equipment like that doesn’t exist for us – yet. Although a foot-throttle is interesting and I have seen those on electric motorcycles. Imagine: No wrist fatigue!

- - -


Hi Njay

I have been mulling this over and thinking of an alternate method to trigger the Mod:

First, allow me to attach the working schematic I use for reference with the “Infineon” based controllers; it’s pretty close to the original ebikes.ca 6-FET and not far from the Lyen 12-FET controllers. The basic components are there to support the model circuit. This is where I get the values for R8 through R11.

View attachment 6fet.pdfNote that there are a lot of comments in this file on how to mod this circuit from 36V to 55V (10S to 15S).

• In reference to V4 schematic, everything to the LEFT of the +5V Rail preexists on the Controllers, Resistors 8 through 11 are applied to complete the evaluation of SP and UV in OEM configuration, as well as after modifying R12 with a known (real world) value before divvying it up into three units.
• It crossed my mind to use +12V rail to drive the OptoIsolator, with the GND tied to –EBS after a blocking diode though before The Switch: When shorted, both –EBS and the Opto go to GND. I trust your opinion that we need current limiting on the Opto although I don’t know how to spec it (yet). A little help please
• Another idea I had was to leverage an earlier hint you provided and to use a voltage comparator on –EBS instead of the Opto; Whichever is less complicated works for me.
• Small note: The voltage at SP is also affected by R100 & R101; I had to make those values higher than R8 & R9 to reduce the impact when I was trying to balance U1-Gate voltage.

Advanced Settings:

• I tried this on the +5V rail and had some unexpected results. I don’t think I quite like it… probably cos I am not sure what I’m doing here.
• The sag I experience is from insufficient wire gauge to the controller (I’m in the process of redoing that) although I think that is out of scope for this mod; we have to presume the controller can handle sag, otherwise we’re into modifying the voltage regulators. That said, I don’t have a problem placing a low ESR cap to uphold the supply voltage between the 5V (or 12V) rail & GND before U2.

That's where I'm at so far. Time for another cup o' joe! KF

 

[Kingfish]

Version 5

• Added D1 (1N4007) which is optional, though required for 2WD to prevent the other controller from holding –EBS high during Ebrake. (SPICE Directive given in the comment; this was added to the standard.dio file).
• Added R103 for current limiting. The value might change if D1 is removed.
• Added C1 mainly to stabilize the voltage into U2. The value is TBD.
• With these additions, Current (I) through U2 is now limited to about 75 mA.
• When the Ebrake is applied, -EBS drops to < 1V. Setting R103 higher reduces the voltage, though also the current through U2.

In review, I looked at several other optoisolators and diodes. I don’t have enough experience to rightly choose, although I did spend time reviewing similar controller circuits that are generally within the same power range. What else can I do but to mimic by example?

I think I’ve pushed this about as far as I can go without more help. I can’t think how to make this more simple and clear.

In other news we’re having record-breaking heat in Redmond. I don't have A/C, but I do know of a Bar & Grill named after the town that serves up ice cold Pike’s Naughty Nellie’s on draft and I think I’m going to take a plunge and swim at the bottom of the glass for a while and increase my intelligence by purging the weak brain cells. Happy Hour starts at 3 PM; just enough time for my weekly shave & paste down that cowlick. :wink: ES lurkers welcomed.

Happy feet, KF

 

[Njay]

75mA for the opto's LED is too much, the datasheet mentions 60mA as the "Maximum Absolute Value" (parameter If).

The opto's transistor needs to saturate, because in this case you want it to "turn ON" as much as possible (right?). For it to saturate you need to give a base current that will make it have a collector current higher than the current that will flow through it and R100 and R101 (no current will flow to the MOS). In a quick calculation, we know that the SP node of the circuit will have 3V minimum, and with R100 and R101 this translates to 3V / (21K + 22.1K) ~ 68uA; so we need to set the opto's transistor base current to be comfortably higher than 68uA, it can be much higher. Since this is an opto, we don't deal with base currents; the 4N25 datasheet's tells us that this device has a CTR* >= 20%, which means that the collector current will be 20% or higher of the LED's current (not in all situations but let's simplify here). So for you to get 68uA of collector current you need an LED current of 68 uA / 0.20 = 340uA or higher. Let's set it to 2mA to be on the safe side.
In your circuit you have the 5V rail. When the EBRAKE closes, current will flow from there to GND, going through the opto, R103 and 1N4007 (a 1N4148 will do it and is a tad smaller and cheaper). The opto will take some voltage (the "diode forward voltage drop") out of the 5V, the 1N4007 will take some more and R103 will take the rest. 1N4007 will take out around 0.65V, the opto will take at most 1.5V (value from datasheet, Vf) so we reach 5 - 0.65 - 1.5 = 2.85V that is what R103 will see. Using Ohm's Law (again) and knowing we want 2mA through the opto's transistor (the same current will flow through the opto, R103 and D1 since they are in series) we calculate R103: 2.85V / 2mA = 1425 ~ 1.5K. All of this is an approximation, a starting estimate. As an interesting check, you can see that picking your current R103 value of 39 Ohm, it gives 2.85V / 39 = 73mA pretty close to the 75mA you mention.

The -EBS voltage when the EBREAK is applied is just D1's voltage drop, with the amounts of current we're talking here it shouldn't go over 0.8V.

You can raise R100-101 further to further reduce the impact on SP if needed. The current that the R100-101 branch pulls from SP is coming not from SP (it is an input) but from the 5V rail through the Throttle and R8, with a bit of influence from R9. I see that you mention the variable-resistor throttle but are modeling using the Hall throttle input... shouldn't you be using the throttle connector SLA pin instead of SP?

C1 shouldn't be needed, although it won't hurt.

"Sag" is a generic term, I used it because I though you would understand it better . Let's call it by its EE name, "voltage drop". The drop I'm mentioning is the drop inside the voltage source. In the case of the 5V rail, since it comes from a 7805, it's not important to model it because your circuit is taking very little current and the 7805 has a relatively low internal resistance which won't cause any relevant voltage drop on the 5V rail. 7805 datasheet says ~8mOhm for internal resistance (they call it "output impedance" because it's dynamic and is a resistor in series with the output) so if your circuit draws 20mA (which it doesn't) that represents a drop of 20mA * 8mOhm = 0.16mV on the 5V rail - totally irrelevant.
However, if you modeled your SP point as a voltage source (you would replace R8, R9 the throttle and the 5V rail by a single voltage component that you could configure to give a ramp of voltage simulating a throttle), then the internal resistance wouldn't be irrelevant anymore because it's high compared to the current you draw from it, which you already found out.

As for the advanced settings of the voltage component. You are now simulating 5s of your circuit (.tran 5). Let's say you want to make a ramp on the 5V rail from 0 to 5V, from second 1 to second 4. You would set

Vinitial = 0 (your starting or "base" voltage)
Von = 5 (your final or "top" voltage)
Tdelay = 1 (1s at Vinitial before starting; this delay applies only to the 1st waveform cycle)
Trise = 3 (you have 3s to go from 0V to 5V, since we want to start at time = 1s and end at time = 4s)
Tfall = (empty, in this test you don't care about the fall time because the test will end before reaching it)
Ton = (empty, as Tfall)
Tperiod = (empty, as Tfall)
Ncycles = (empty, as Tfall)

Sorry for the text length, but it seems to me that you're the kind that likes to learn how to fish and not only be given the fish

Have a nice drink !


* CTR = Current Transfer Ratio - transfer ration from "input" (LED) to "output" (transistor)

 

[Njay]

Some years ago I came up with the idea of a 2 way throttle. When you twist it towards you acceleration occurs and turning it away from you starts the variable regen. The more you twist it in the opposite direction, the harder the ebrake effect. The throttle should have a feelable center point where nothing happens at all. With such a setup you'd barely need the normal mechanic brakes and would enjoy the comfort of an easily controllable ebrake.

If I'm not mistaken, that method is actually patented by Vectrix, and that's why you don't see other commercial products using it, as far as I know.

 

[Kingfish]

Version 6

Opto LED current: Understood; Changed R103 to 1.5K, swapped D1 for 1N4148, LED current reset to 2.2 mA, -EBS now at .622mV. I wasn’t quite sure how to interpret the Opto datasheet, but now I get it – thank you!

Throttle R100 SP & SPA: Ahh, that is a conundrum. The short answer is that I use both. On the Road, the HE gave up the ghost and I swapped to a Resistor-Throttle, then later to a new HE-Throttle. Both have particular artifacts depending on the manufacture. Regardless – they both provide feedback via SP, and that’s what I have to use. The more difficult question is which one do I pick to model the circuit? From what I understand, regardless of which type of throttle is use, Voltage-SP will vary between 0-5 as the superset, though from the schematic we can infer it will be 3-5 Volts. I am though at a loss to reconcile the proper circuit model between a POT and HE throttle. :?

Latest and greatest.

Advanced Settings: Ah HAH! That’s awesome! I can now clearly see the ramp of change for UV. And I can also observe the effect of increasing R100 & R101. However there is a new issue: I am measuring the current at the Gate of BN170 and it is between 0 to 30 pA which seems quite small. Does that seem reasonable to you? 8)

Now that we have the slope, I wonder if there is a way to elongate the transition, otherwise it looks like we will have a very short region to feather between High and Low Braking; when SP-out is between 3.6 to 4.2V. I'd like to get that active range closer to 3.2 to 4.8V. Know what I mean Vern? :wink: Maybe I just need to twiddle some more...

ADDENDUM: Changed R100 to 68K and that moved the slope to start at about 3.2V, though ends at 3.6V. Need to figure out a way to stretch the range... hmmm <rub chin>

Fun stuff; really learning a lot here – very much appreciated!
Cheers! KF

 

[Njay]

Regardless – they both provide feedback via SP, and that’s what I have to use. The more difficult question is which one do I pick to model the circuit? From what I understand, regardless of which type of throttle is use, Voltage-SP will vary between 0-5 as the superset, though from the schematic we can infer it will be 3-5 Volts. I am though at a loss to reconcile the proper circuit model between a POT and HE throttle. :?

That's what I would do, concentrate on the voltage sent to the controller (SP).

However there is a new issue: I am measuring the current at the Gate of BN170 and it is between 0 to 30 pA which seems quite small. Does that seem reasonable to you? 8)

Very reasonable . That's some 6 orders of magnitude below the smallest current the rest of the circuit is dealing with, so it's the same as zero. There's always small leaks, everywhere.

Now that we have the slope, I wonder if there is a way to elongate the transition, otherwise it looks like we will have a very short region to feather between High and Low Braking; when SP-out is between 3.6 to 4.2V. I'd like to get that active range closer to 3.2 to 4.8V.

Well, you can try increasing R103, but I assumed from the beginning that you just wanted 2 levels of regen, a hard and a soft. If you want a ramp, then other methods should be pursuit, because all that adjustment you're doing on R100-101 will have to be done for every single BS170 device, due to variations on the characteristics; and the range is subject to moving around a bit depending on temperature (I don't know how to quantify that). An ampop is able to transform or "map" a voltage range into another voltage range (thinking throttle -> UV), but UV also depends on the battery's voltage so it would not be as simple as this.

Fun stuff; really learning a lot here

That is the really important stuff

 

[fechter]

Most hall effect throttles have an output that goes from just below 1v to something around 4v. Most controllers will hit 100% output when the signal is just below 4v. A pot throttle will generally output 0 to 5v.

I'm still a little fuzzy on exactly how those controllers behave in regen mode. The hardware is all there to do variable regen, just not sure about the software. If the circuit gives the throttle input the right voltage when the ebrake is activated, seems like it should work.

One of my scooters had a regen setup that worked off the brake switch. If you let off the throttle, it would just coast. When you barely apply the brakes, the switch comes on and activates regen. The regen was current limited, so provided a nearly constant braking force over a wide speed range. I found this to be very intuitive and had a nice feel to it. Applying more brake of course activates the mechanical braking and gradually increased the braking force.

A bi-directional throttle would be cool, but somewhat hard to make. A hall effect setup in the brake lever has been done before and can give a variable output based on brake lever movement.

 

[Kingfish]

Njay -> OpAmp: Understood. Been looking at examples and models trying to find something reasonable.

Fechter -> R-Throttle verse HE-Throttle: Yes, that seems familiar; I know we had to dork with the Resistor types to create a direct swap; I actually liked the manufacturing quality of the Magura Resistor over the cheapie HE, but not the hassle of tuning them in – and then there were those holes in the range.

OK, I shall keep the circuit reactive on SP to the 3-to-4V range. Kinda frustrating – so close I can taste it. If I could just double that slope width we could be in clover.

On a mission, KF

 

[IBScootn]

Interesting stuff. Good luck KF.

Only successful variable regen I've read about was: http://visforvoltage.org/forum/6520-activating-regen-braking

 

[Kingfish]

Version 7

Nailed it! Right there in front of me the whole time, but didn't get it until I finally found a thread on how to slow down FET switching. There was a discussion which is the better route: add gate resistance to slow the Gate capacitance charge, or to add external Gate capacitance. I couldn’t get the model to accept my resistor, but placing a cap between the Gate and Drain did the trick.

Latest version. Added C2 just above U1.

Big wide ramp!

C2 now controls the slope width. At about 12uf, the width is doubled to about 0.8V rise. R100 controls the beginning of the progression, with less ohms moving toward lower SP voltage. I twiddled with C2 a bit and setting it to 47uf provides a broad 2-volt range. One small caveat: R100 and C2 need to be set together cos changing one will shift the starting voltage of the ramp.

Restating the Component Features:

• R12-A & -B set the HARD Braking voltage of UV (match a bit lower than maximum Phase FET voltage)
• R12-C, when added to R12-A & -B sets the SOFT Braking voltage (match just above Battery pack voltage)
• R100 & R101 set the starting Throttle voltage (about 3 to 3.2V)
• C2 sets the width or range of progression: 12uf for about 0.8 to 1 V, up to 47uf for about 2 volts.

If my EE Consultant Njay buys off on it and there are no other issues, I’d like to get a move onto ordering up parts

Dang, it’s just about Happy Hour again! I don’t think my body could take another night of abuse… though it’s tempting, yes?
Building a thirst, KF

 

[Alan B]

.

 

[Njay]

Sorry but I don't think so, KF. What you're seeing is a transient state, the MOS is just taking longer to respond to changes in "the input". Try stopping the raise of V2 at 3.5V (while still simulating up to 5s).

You should simulate the throttle as a voltage going from 2 to 4V (or whatever range you choose), with a parasitic series resistance of some 10K. This would be your SP-OUT.


You're forgetting a 1.2K resistor from UV to ground inside the controller.

 

[Kingfish]

Sorry but I don't think so, KF. What you're seeing is a transient state, the MOS is just taking longer to respond to changes in "the input". Try stopping the raise of V2 at 3.5V (while still simulating up to 5s).

You should simulate the throttle as a voltage going from 2 to 4V (or whatever range you choose), with a parasitic series resistance of some 10K. This would be your SP-OUT.


You're forgetting a 1.2K resistor from UV to ground inside the controller.

Yeah, I got that after I extended the time to 12s

The 1.2K resistor in the controller is R12; we're replacing that with R12 A-C.

At present I am looking at one of the Linear Technologies op amps (cos they're already built into LTSpice) and trying to figure out how to invert the output between the two voltage limits.

No beer for KF

 

[Njay]

Here's a starting point for you. Has 2 difficult to adjust points, but is scarce on hw. An analog guru will probably make the same with 2 or 3 transistors

"base" is a voltage to adjust; in practice you generate that voltage with a 2K pot from the 5V rail.
R2 is a resistor to adjust; in practice use maybe a 3K3 to 5K pot.
I made a rough approximation to the desired range using common resistor values.

V2 (SP) is simulating the throttle by simulating it's voltage (used 2V to 4V).

First ampop makes the range conversion, the second arranges for UV to be the same voltage as desiredUV. I used the internal R12. LM358 is an old and very cheap dual ampop (I buy mine for ~$0.15 in small quantities) in an 8 pin package.

SP is transformed in SPin by a resistor divider between SP and a voltage. Then SPin goes to the 1st ampop which is a non-inverting amplifier, with gain given by 1+R2/R1 resulting in desiredUV. These 2 transformations convert the input range to the output range.
Then the 2nd ampop "makes its magic" to keep UV equal to desiredUV. An ampop acts on its output in order to make both its inputs the same; this is the property we explore here. There's an output voltage that will make the inputs the same; this output voltage acts by pulling current through the 100 Ohm resistor and thus through R10 and R11 (I modeled them as R10) lowering the voltage at UV.

This is pure theory, not tested in practice. You've been warned

.

 

[Kingfish]

Thanks Njay - this will take time for me to absorb, and I appreciate the effort.

There's just one tiny itty bitty thing that needs correcting: As SP rises, UV decreases. You have the values correct, but the ramps go in opposite direction. This is why I've been looking at inverting the output using a pair of OpAmps. It's driving my batty :? :lol:

~KF

 

[Njay]

oops

No problem, just turn the input non-inverting amplifier into an inverting one, see attach.
Took the opportunity to add a stabilizing cap to the "voltage reference".

.

 

[Kingfish]

OK, I see this working now

Let me see if I understand…

• SP voltage set to emulate throttle – regardless of POT or HE type.
• “base” is a 10K POT, presently at 2.8V. This sets our bottom of desiredUV.
• R1 & R2 form a voltage divider which affects the voltage of UV
• I see now how to correctly invert the signal, however I am at a loss to why we need the second OpAmp.

You have definitely captured my attention with your schematic my friend
Now I'm trying to figure out how to spoon it into other schematic with the optocoupler.

Very intriguing! KF