Throttle Behavior

fechter said:

I'm nearly certain the PAS sensors are something like the motor halls or a SS41, which has an open collector output.

Click to expand...

The one in one of the Fusin PAS sensors I have is an SS41 or similar cllone, IIRC (been a long time since I opened it up), but it operates the same way as a motor hall does (except I dont think it latches like the motor halls, just pulses on (grounding the pullup) whenever its in a strong enough field).

Okay, I attached wires to the grey and black PAS wires and the blue throttle signal wire, all within the controller, with leads coming out for testing purposes. Before further testing I checked the voltage of the throttle signal. With the throttle resting it was about 0.9V and the motor started running at about 1.1V when the throttle was advanced, topping out at about 4.0V at full throttle.

Then I verified that everything worked normally, that is normally for a 2017 controller that has the throttle/pedal issue.

Then I shorted the grey and black PAS wires together and retested. As expected the PAS system no longer functioned, and when set to PAS 0 the throttle had full range while pedaling.

So success for the initial testing! Just for disclosure, these tests were done in the house with the controller wires all hooked up properly but the controller was not physically installed. Also the chain and chain ring were not installed so the back wheel did not turn, and hence no speedometer, so the Bafang did show the expected error 21.

Now I'm ready to complete the project by installing a transistor between the grey and black PAS wires and a resister between the blue throttle signal wire and the transistor. Problem is, I have no idea what to purchase (I have no experience or knowledge in this area). Fechter, can you provide me with the details of exactly what to purchase, and where? Will I be able to just take the part specs in to a local Radio Shack or do they need to be ordered from someplace like Mouser?

Thanks in advance for any additional help. I am excited to be able to make this controller work the way it should, at least the way I think it should.

OK, good news on the testing.

In reviewing the datasheet for the transistor, it seems the super simple circuit with one resistor might have issues with temperature drift. The first one with two resistors will be better in this respect. The easiest thing would be to use a 20k trimmer pot in place of the two resistors, then you could dial it in so the PAS disables when you barely move the throttle. This way you could use practically any NPN transistor also. Otherwise it will take some experimentation to find the right values for fixed resistors and it will depend on which transistor you use.

I like simple, so here are some thoughts. The start voltage for the throttle is currently set at 1.1, but that can be reset to a higher value using the Bafang programming cable and Penoff software. Also, it doesn't seem that having the PAS shut down extremely close to the throttle start voltage would be critical as long as it was always above the throttle idle voltage. I.e. plus or minus 0.2V probably wouldn't even be noticed. Resistor drift caused by heat was mentioned, but how much drift would occur?

Easy to see where I'm going here. Is there a single resistor and transistor combination that would short the PAS signal at about 1.2V with just a reasonable variance due to temperature? Then fine tuning could be done at the throttle start voltage value if necessary.

About five years ago dnmun helped me tame an over aggressive throttle by using a variable resistor he sold me. It was quite sensitive and took many tries to get it to function properly. Plus, once it was housed inside the controller compartment it would not be accessible for adjustment anyway.

I used a 2N5550 I just happened to have lying around and simulated that part of the circuit on a breadboard. The single 100k base resistor will be pretty close, but I don't know the exact threshold where the PAS will inhibit.

I know what a pain it can be to dial in a pot. One reason I like the 10 turn trimmers. The process would be to increase the setting a little at a time until the PAS inhibits without touching the throttle, then back off a little. If you used a pot for initial adjustment, you could disconnect and measure the resistance of it, then use a fixed resistor.

fechter said:

The process would be to increase the setting a little at a time until the PAS inhibits without touching the throttle, then back off a little. If you used a pot for initial adjustment, you could disconnect and measure the resistance of it, then use a fixed resistor.

Click to expand...

Sorry to digress, but the pot is supposed to go in series into the negative supply to the throttle?

Okay, I've been trying to learn a little and it doesn't look like a trimmer resistor is very large and it would allow some trial and error to take place. You mentioned a 20k trimmer resistor earlier, but 100k as being about the right resistance, so does that mean I should get a 100k+ trimmer resistor?

When I look at Mouser they have half a dozen different 2N5550 transistors. For this application are the differences of any importance?

I should really try to do more testing, but I think a 100K trimmer would be better.
Temperature drift might still be an issue since the controller heats up quite a bit as you ride, but I think it will be OK. When building, I recommend attaching the CCW leg of the pot to the ground side so that turning CW increases the voltage where it turns on.

Any of the transistors will work. The differences in those parts are often just the packaging, the way the legs are bent, or the temperature range. There may be something better than a 2N5550 also. It was just a random choice because I had one stuck to my breadboard. I think it will be OK for this application. I looked at some other transistors, like 2N2222 and 2N4401 and they will work about the same as the 2N5550.

Another thing I looked at was using a Darlington transistor. These have much higher gain, so temperature induced gain changes won't have much effect. Downside is I think the lowest you could get it to turn on is about 1.1v. This would allow the super simple single 100k resistor layout but you couldn't get it to turn on at less than about 1.1v. This might actually work OK in practice. Darlington examples: KSP13, MPSA28 (TO-92 package).
I'll dig around my junk and see if I can find a Darlington to test. If it works, it might simplify things.

Here's a sample layout. Top view given, so when wiring, keep that in mind. What I do with these kind of circuits is have all the wires coming out the end opposite the adjuster screw and use heat shrink tubing to cover the whole thing so just the adjuster screw shows on one end and the wires stick out the other end. I think getting it dialed in won't be too hard.

Thanks for this diagram and detailed info, it looks like I can make a good Mouser order for enough stuff to have some extras and to try different approaches. So in about a week I should be back, hopefully with some good test results.

I managed to find a small Darlington transistor and tested it. 2N5308. With a 100k base resistor, it turns on right around 1.15v. This is pretty close to the desired point. I think ideally we want it to turn on at 0.95v to 1.0v.

One other thought I has was if you are pedaling and hit the throttle, the PAS might still be 'engaged' when the throttle signal appears and you might still get the jerking problem for a second or so until PAS turns off. If you use the throttle first, then start pedaling, that should be no problem. If you are on throttle and pedaling and release the throttle, it should go back to PAS no problem.

fechter said:

I'll dig around my junk and see if I can find a Darlington to test.

Click to expand...

If you cant find one, you can just hook up two regular ones to make one.

I don't think throttle jerk will be a problem even if there is a small overlap with the Bafang Start Throttle voltage. I had the throttle set for a nice slow ramp when the other "bad" controller was installed (LUNA eventually exchanged that controller for a good used 2016 controller), and I was able to make a smooth transition at any speed from pedaling with PAS to throttle only when I stopped pedaling. There didn't seem to be any difference if I stopped pedaling before using the throttle or stopped pedaling after applying partial throttle.

I can also increase the Bafang Start Throttle voltage a bit to eliminate the overlap if necessary.

If any fine tuning is needed it probably won't show up until the system is road tested.

Unfortunately, I did not have time to test the circuit over the weekend.

I looked around to see if there was a voltage detector that might work, and they do exist. A voltage detector will switch at a preset level and should be pretty immune to temperature changes. I found some that will trigger at 0.9v and some for 1.0v. The outputs don't seem to be strong enough to pull down the PAS signal directly so would need a transistor to boost things, and to invert the signal. A low threshold MOSFET could be driven directly from the output. A regular NPN transistor would work too, but needs a resistor in series with the base.

A big downside to this approach is the detector chips and the MOSFETs are really tiny surface mount parts, so you would need a PCB to mount them on and hand soldering is a challenge. Plus side is you don't need to adjust anything and it should work over a wide temperature range.

Thanks fechter. I have a Mouser order that should arrive tomorrow afternoon. It includes a couple of straight 100k resistors and a couple of adjustable 100k resistors, and a couple types of transistors.

My first test will probably be with a 100k resistor in series with a KSP13 transistor. If that doesn't work I'll probably try using an adjustable 100k resistor.

When and if I have something that seems to work properly when tested on my no load setup I'll make all the connections permanent using silver conductive epoxy with liquid electrical tape for insulation and then slip some shrink tubing over the whole thing. Then put the BBS02 back together and road test, probably on Wednesday.

Good deal. You get to be the guinea pig..

What works nice is to have the pot hanging out on long wires so you can adjust while testing, then once you get it all dialed in, carefully disconnect the pot and measure the value with an ohmmeter. Then you know which fixed resistors to use in the future. It's a lot of work and only worthwhile if you want to duplicate the setup.

With the single resistor version, you could also use a pot instead of the fixed resistor and dial it in.

Once you get it all installed and working, you may (or may not) find the trigger point drifts with temperature. For sure it will drift, but whether it will be a problem or not I can't tell. Hopefully it is so small you don't notice it.

One more thing to pay attention to is too small of a resistor value might limit the maximum throttle. Just make sure you can still get full speed when the throttle is 100%.

Okay, I finally got success on a no load test.

I used a trimmer resistor, Mouser P/N 652-3286W-1-104LF and a Central Semi Bipolar Transistor, Mouser P/N 610-2N5550. All hooked up like fechter's red and orange diagram a few posts earlier.

The 100K resistor was turned down under 20K when I installed it, and adjusting it up or down one full turn made no noticeable difference. From previous tests I have made It seemed that with a straight 100K resistor there was no change to the controller functionality and with 0 resistance neither PAS or throttle worked at all.

So I will continue some no load tests, but I don't think they will be conclusive until some road tests are done. It seems logical that the throttle is more apt to work best with more resistance, but the interface between PAS and throttle usage may be smoother with less resistance. No way to tell without the load so what I need to do is keep everything exposed so I can adjust the resistor between riding tests without having to remove the controller. I think I can secure things well enough to allow for this.

Working with these little wires and pins and electronic parts was a challenge for my skill and experience level, which isn't going to improve at age 78, so keeping everything solid for road testing will take some additional effort.

Rassy said:

The 100K resistor was turned down under 20K when I installed it, and adjusting it up or down one full turn made no noticeable difference. From previous tests I have made It seemed that with a straight 100K resistor there was no change to the controller functionality and with 0 resistance neither PAS or throttle worked at all.

Click to expand...

This sounds right. If the resistor is near zero, it will be loading down the throttle signal and PAS will be always disabled. If it's too high, the transistor never comes on and it behaves like there was no circuit.

One way to get it in the ballpark would be to measure the grey wire voltage while adjusting (zero throttle, power on). Let's call CCW "low" and CW "high". When it's turned low, the grey wire should have 5v. Adjust until the grey wire comes down to about 2v. This should be pretty close to where you want it. You can do this without pedaling or running the motor.

Then test whether PAS is active or not. Turn the pot higher until it just disables PAS, then back off a little. It might be handy to know what the grey wire reads when it's dialed in correctly.

I know how it is working with tiny little wires. My eyes aren't getting any better with age and the blue throttle wire is really skinny.

Good luck and thanks for testing. If I didn't have a day job, I could have had this installed and tested by now.

Using the pot in series may exacerbate the temperature sensitivity, and make adjusting it very difficult and insensitive to changes. Better to use it as a voltage divider as was shown in one of the early sketches, one end of the pot to the throttle signal, one to ground, transistor fed from the central wiper.

Alan, back up a few more posts to fechter's red and orange diagram. In between he posted a way to use a voltage detecter, but that's not the diagram I was referring to.

This is the voltage divider setup (from above) which should provide better adjustment range. Perhaps you are already using this arrangement, but I would not call this series. The pot forms a voltage divider across the throttle output. Unfortunately when the pot is turned near the CW end the base will load the throttle line, but it should be set farther away from that end anyway. I would turn it down (CCW) till it did not work, then advance it (CW) till it works. That should minimize the throttle loading.

Thanks Alan, I think we're on the same page. I did try a "series" arrangement first using just a straight 100K resistor between the throttle signal and the transistor, but that did not help at all. My no load tests with the adjustable resistor show everything starts to work properly at about 25K, which is where I plan to have it for my first road test tomorrow.

No road test coming! Finished putting liquid tape on everything last night. Did one final no load test and everything working as desired. Wrap everything with electrical Tape this morning and tested again. System was working in the "faulty" mode. Temperature was down 15 degrees so I turned the pot down one turn and retested. Everything worked as desired. Stuffed the wires in and installed the controller and tested again. PAS worked fine but the throttle did not work at all. Pulled the controller off and tested again, but still no throttle. Adjusted the pot both ways, but still no throttle. Installed my good controller and everything works properly.

So perhaps the throttle power supply gave out, or perhaps a short was introduced within my wires, or perhaps the throttle wire I tapped into broke, although just the insulation was cut and the actual throttle signal wire wasn't severed at all. Anyway, I want to ride my trike for a while and may revisit the bad controller when I feel like it and see if the problem is in the wiring I added or if I have actually damaged the basic controller.

It was a challenge that I wasn't quite up to, although it looked like success yesterday.

Anyone else trying a similar experiment, it was mentioned, I believe by Alan B and fechter, that the throttle power supply could be sensitive to being tapped into for this purpose.

EDIT: Mystery solved. The little blue throttle sense wire snapped where I tapped into it. It was already stretched pretty tight so I might just splice in a little extra wire and retest this thing one of these days, but not right now, gotta catch up on some other stuff.

It might be worthwhile to add another resistor to the circuit to insure that it can't draw down the throttle voltage too far, even if the pot it too far toward one end.

Glad you figured out what went wrong there.

Good it was just a broken wire. It does look pretty tight around there. Need to think about a better way to tap into skinny wires.

I'm hoping to test an alternate circuit using the Darlington transistor. This one will have temperature sensitivity also.

There are many, many other ways to do this, but I wanted to try the simplest one first.

Need to think about a better way to tap into skinny wires.

Click to expand...

Here are some mini taps for small wires. May be too large for the wires we are working with:

https://www.crutchfield.com/S-vkLakULdBwn/p_669PP644/Posi-Products-Wire-Connectors.html