power loss and backward rolling residence. HELP!

Edit: So I fixed the 3-speed switch and noted a new symptom; backward rolling resistence (a lot) when the controller is powered up and none when it is powered down. This was definitely not the case before this problem started. I have checked the function of the hall sensor and the mosfets and found them to be functioning properly per the guides on the Grin site. Please help if you can

So my bike has been running perfectly for many months now and I just love it. Build here
http://endless-sphere.com/forums/viewtopic.php?f=2&t=55605


So yesterday I first noticed that the power level felt like I had the 3-speed switch set to level 1 (slowest). And there was no change in the power level in any of the three positions of the switch. I have checked the plug and it has continuity throught the plug. My first thought was that the 3-speed switch had a problem. But then I noted on the CA that when I ran the bike at WOT the battery voltage was sagging from full charge 74v briefly all the way down to 65v. Now I confess that I don't regularly watch for this but that seems like a lot...too much. So, any thoughts on how to proceed? How can I test the 3-speed switch for proper function? What other things might cause these symptoms? Any thoughts from the collected wisdom? BTW my set-up is 18s2p of 8000mAh HK Zippy lipo, running a V2 Cromotor through an 18 fet Lyen controller and a V3 CA. HELP! My baby is sick.

Are you drawing beyond your batteries C rating? Voltage sag is what it is, there is little mystery to it. There is a good chance you've beaten up your pack pretty good, have you checked the internal resistance of your cells? How many cycles has your pack endured?

In summary, it sounds likely that nothing is wrong, I could be wrong though.

Hey bowl, I have not beaten these cells. They are 30C and I have been very cautious to neither overcharge nor over discharge them. Likewise my use of the upper end of the power curve has been infrequent and of short duration. I guess checking the IR would be a good idea I think my charger has this capability. I'll give it a go.

Oh yeah, in answer to your question, I am sure this pack has less than 50 cycles.

If the 3 speed switch makes no changes, then it's broke or disconnected. Even with a weak attery it should show a change.

Dan

In C rating, the C stands for capacity. C rating is capacity times whatever value comes after before C. Lets pretend you have 1000 watt hours of capacity, and take your 30C figure. To find, in this theoretical scenario, what you can continually draw from this pack without harming the cells or seeing very extreme voltage sag, you multiply 1000 watt hours by 30, which comes out to be 30,000 watts. Very much heat and cold increases internal resistance. You should probably know your capacity and C rating, and actively think about that and how to use it on your electric bicycle setup. With a cycle analyst, you could actually do a bit of very simple experimentation using an amp/watt limit to find where a certain amount of draw is considered too much in your opinion. Lets pretend with your capacity and C rating you can continually draw 1500w without causing issues. With your cycle analyst, you could probably configure a limit of around 1400w and then be set for a while, until capacity loss increases (it will).

The more times you cycle the pack, the higher the resistance and lower the capacity your cells are going to become. If you have abused (drawn beyond C rating, over heated, or you ride in extremely bumpy conditions and your pack isn't well protected against that) your pack, these are things that will likely increase internal resistance.

A simple way to test the 3 speed switch theory is to simply unplug it and ride normally. One thing to keep in mind though, some controllers default to a specific throttle level limit without their three speed switch plugged in, your 3 speed switchless mode may be 75%, I say this to prevent you from seeing a false positive. However, 3 speed switches don't change your voltage (at least none that I've ever seen, used, or heard of), they simply change the throttle maximum (in most cases).

The sag is worth investigating but it sounds like a separate issue. There is no way that amount of sag could account for a drastic power reduction or a failure of the switch to function. No point in wandering off the target problem unless a clear connection develops - so don't worry about it for now.

Your build thread doesn't specify (dynamite build BTW!) - but I'm guessing this is a CA 3-speed switch, yes? If so, go to the Aux Pot Setup Section Title screen and look at the percentages there as you flip the switch. The HIGH switch setting should show 99%, etc.

If that didn't turn up anything - Assuming you are routing the throttle through the V3 (you running in 'Normal Mode'), switch to the Diagnostic screen and watch the OUT voltage during a test run. This should rise to your configured ThrO->MaxOut voltage. If it does, then it's not the CA and may be the CA connection or something wonky with the controller or battery wiring (hence the sag). If it doesn't, then you definitely have a V3 issue. While you are on that screen, check to see if a limit flag goes Upper Case during the tests - this will tell you which parameter is limiting your operation. Revisit the configuration of the associated MaxXXX limit parameter according to the table on page 27 of the Guide. Remember that if you are using Current or Power Throttle that it will be normal for the A or W to be Upper Case even at standstill so those cases may be harder to diagnose.

If this doesn't reveal anything, there are quite a few interactions that might account for the V3 squelching performance. If you are using Current or Power Throttle, switch to PassThru mode temporarily so you can repeat the Limit Flag test above without the 'always on' A or W limit flag issue. Restore your previous throttle mode.

Failing that - you can disable the Aux Pot functionality which should give you full power. If this doesn't work, then it's not the switch or Aux Pot configuration. Next I would look at your configuration for MaxCurrent and MaxPower. Finally, re-examining the Throttle In/Out Min/Max settings.

That should pretty much cover everything obvious....

Also - you may find it useful to examine the V3 RBatt value on the CA status screens after every ride. This is the computed battery resistance. You should expect this to increase with age and also with cold weather.

The value assumes good wiring and so a crappy connector can artificially raise the apparent 'battery' resistance. In any case, any sudden jump is an easy indication of battery or wiring issues.

Wow! Thanks for all the great input. So, Tek it is not a CA 3-speed switch. It is a standard style that is plugged into the controller directly. Also it does not have the auxiliary "green" button. So what is a good way to test one of these stand-alone cheapies? I don't know if you recall Tek, but this is the build that you helped me integrate the non-CA cruise control about a year ago (working great BTW). It has been since then that I messed around with the CA settings so I will have to go through and re-learn the details of the settings. I did go through and do a careful visual check of all the connections in the whole bike today and everything looked fine. Unfortunately, I have to work at the fire house tomorrow and the next day so I won't be able to do any diagnostics until after that. Please keep the ideas comming though. Oh and, Go Giants!

I went through and conducted the tests recommended below. Here are the results:

In the diagnostic screen at WOT the voltage rose to 3.77 exactly matching the ThrO->MaxOut.
Since I am not using a CA 3-speed switch (but rather, a 3-speed switch connected to the controller and configured through the controller settings), my AuxPot is currently disabled.
The CA is normally configured for PassThru mode, and no limit flags went to upper case during the test.

Max Current is set at 70A
Max Power is set at 9900
So I don't see how these setting could be limiting performance.

ThrI->Min= .65V
ThrI->Max= 3.10V

ThrO->Min= 1.4V
ThrO->Min= 3.77V

The Rbatt reading in the CA = .149 and 240 aH total on the batts.

Couple other observations; the throttle ( Magura) has a small dead spot at the beginning, smooth increase throuout the range, and a similar small dead spot at the top. This feels like the same throttle response that I have had since I initially set it up. Real-world riding on flat asphalt at WOT results in smooth acceleration but considerably less torque than my normal riding performance (at switch setting II) and reaches 30mph with the noted battery sag from 71.6V down to 65.8V which stays pegged at that 65.8V number and does not drift back up as cruising speed is approached. On the test stand, the top speed is 35mph with about .3V of battery sag total. Interestingly, on this last test ride I noted a few instances of very minor and brief motor stuttering on acceleration. Any thoughts? It has been my experience that motor stuttering is often related hall sensor problems. At this point I am thinking that the 3-speed switch might be bad but that does not account for the performance being unchanged with the switch disconnected nor does it account for the battery sag issue. Any help would be greatly appreciated.

Obiwan007 said:

On the test stand, the top speed is 35mph with about .3V of battery sag total. Interestingly, on this last test ride I noted a few instances of very minor and brief motor stuttering on acceleration. Any thoughts? It has been my experience that motor stuttering is often related hall sensor problems. At this point I am thinking that the 3-speed switch might be bad but that does not account for the performance being unchanged with the switch disconnected nor does it account for the battery sag issue. Any help would be greatly appreciated.

Click to expand...

No load speed and power consumption is not the same as loaded speed and power consumption. Here is a simple way to gain an understanding on what I am talking about. In some way, applying braking forces to the wheel is probably the most safe way to do this, observe the watts on your cycle analyst when the motor is unloaded and then some (not full) braking force is applied. What you should see is an increase in watt usage, and while doing so, the voltage sag you mentioned should also appear. Remember voltage sag is, typically related to power consumption, as in drawing more power (watts) then the pack can easily sustain.

It's hard to know the cause of stuttering sometimes, but you are right in that potential theory relating to halls sensor. The first thing I'd do is make sure the connections for the phase and hall sensors are solid, no corrosion, nothing potentially loose and so on.

Do you have a temperature sensor in your motor?

Yep, I understand the concept and by showing the difference between the riding performance and the no-load perfomance I think it definately showing that a problem exists. When I rode it and got the battery sag way down to 65.8v it was drawing as much as 50A I think. That level of current would normally cause this bike to go like a rocket. The Rbatt readin off the CA is a calculated IR reading I think but I'm not sure if it is in mOhms or Ohms. So I independently checked each battery with my charger which has this function. Most of the batteries came back as 2 mOhms and the rest as 1 mOhm. So battery deterioration does not seem to be the problem if I am reading this right. The brake test would be difficult for me with regen hooked up, I guess I could unhook it but I already know that the bike is drawing excessive power somehow under load. Just got to figure out why. I do have heat sensor hooked to the motor and I double checked the setting and it set at 120 Celsius as the upper limit and it starts cutting back the power at 110c.

Obiwan007 said:

I don't know if you recall Tek, but this is the build that you helped me integrate the non-CA cruise control about a year ago (working great BTW).

Click to expand...

Ya - I remember it well - an interesting little side project. Very glad it's still working as planned.

Obiwan007 said:

I went through and conducted the tests recommended below. Here are the results:
...
The Rbatt reading in the CA = .149 and 240 aH total on the batts.

Click to expand...

Thanks for the detailed results -

So - it looks like the CA is putting out the full WOT voltage and there is no apparent limiting in play. PassThru mode is pretty straightforward and it looks to be operating as intended as well. This pretty much takes the V3 off the table as a prime suspect. Ideally, we might like to run a throttle directly to the controller and disconnect the V3, but your results are pretty convincing.

That said - the reported RBatt looks quite large - particularly for LiPo. This is the V/I value that the CA computes to correct the 'loaded' battery voltage for the Gas Gauge display when underway where 'open circuit' voltage is not available.

Obiwan007 said:

... 30mph with the noted battery sag from 71.6V down to 65.8V which stays pegged at that 65.8V number and does not drift back up as cruising speed is approached.
...When I rode it and got the battery sag way down to 65.8v it was drawing as much as 50A...

Click to expand...

So If we look at RBatt and your questimate of Max Amps we get 50A * 0.15ohm = 7.5V which compares favorably with your observed sag of 71.6V-65.8V = 5.8V. These don't match exactly but they're close which suggests the RBatt figure is roughly accurate and suggests that the CA is probably seeing the same issue that you report.

Since RBatt is based on measured voltage/current, it doesn't seem that the controller three speed switch can be responsible - since that would reduce the current appropriately and would not affect RBatt as we see here. This doesn't explain the 3-speed switch issue - and I'm not a big fan of coincidental failures...

So - following the 'RBatt breadcrumbs' - I would be looking for something distinct from your charger-measured battery resistance that is introducing extra resistance into your power or phase wiring -- a damaged wire (probably at a connector crimp) that is holding on by a few strands, a damaged connector contact (misaligned or badly pitted), or a faulty solder joint where wires join to the controller board.

Obiwan007 said:

...on this last test ride I noted a few instances of very minor and brief motor stuttering on acceleration.
... It has been my experience that motor stuttering is often related hall sensor problems.

Click to expand...

My goodness, you seem to have a symptom from every column! Since it's on acceleration, I might suspect this is a V3 DSGain issue which can induce speed limiting without an accompanying indication in the Limit Flags. You might try the adjustments in section "4.9.2 Speed Gain (PSGain, IntSGain, DSGain)" of the Guide - although the fact his hasn't occurred previously makes this an unlikely cause. The adjustment is otherwise harmless, so...

The 'halls analysis' may well be true, but going after the low-hanging fruit of simple CA adjustments is a simple first step if it's not convenient to disconnect the CA.

Obiwan007 said:

At this point I am thinking that the 3-speed switch might be bad but that does not account for the performance being unchanged with the switch disconnected nor does it account for the battery sag issue.

Click to expand...

Dunno, but easy to test. The switch is a conventional SPDT center-off switch (ON-OFF-ON). I don't know the connector wiring, but you can test the unplugged unit at the connector with a DMM on 'continuity'. In the center, the three wires will be disconnected from on another - at each 'end' switch setting, two wires will be shorted and the other disconnected. When the switch is unplugged, the controller defaults to Speed 2 of the three speeds because that corresponds to the center position where none of the wires are connected.

That's it for me. I'm flat out of ideas for this bewildering bunch of symptoms. Hope one of these helps a bit.

Thanks so much for the comprehensive reply Tek. You always get me thinking in directions I had not previously considered. So that is an easy way to test the functionality of the switch (I knew there had to be one). I'll check it when I get off duty. Next, since I did a thorough inspection of all of the connectors. I doubt that the problem is with the contacts. So how to test if the crimp or solder joint is the problem without stripping heat shrink off of everything. Can I use the DMM to test the continuity quantitatively?, (The current carrying capacity I guess you would say) without exposing the crimp or solder joint. Also you mentioned pitting which got me thinking. The main plug that I use for charging is a six-gang anderson set-up so that I can charge in parllel and discharge in series. I have an inrush control on it so that it does not spark when I connect it in series. But when I charge I have noticed a very small spark each time I connect it to the charger. I just recently read that this may possibly be eliminated by pluging in the charger before connecting to the battery. That said, perhaps I have a bad connector in this plug. This does not make sense though now that I think about it because the bad connector would have to be on the battery side of the plug since that side is the only one involved in the charging circuit. So it follows that the controller side of the plug encounters no spark since when it interfaces the battery side plug it is protected by the in-rush control. So, and I'm speculating here, if there is a poor connection on the battery side of the plug then the batteries would not be charging correctly downstream of that connector (and they seem to be charging fine). Now I'm going to really speculate; if one of those connectors (individual anderson in the six-gang plug) were bad, it should prevent the bike from functioning at all since they are connected in series for discharge. I guess if the connection were marginal it could cause excessive draw simply from a resistance standpoint. If it were bad enough to cause fluctuating current, maybe that would cause a hall signal problem. ....I know that I am wildly speculating now to try to mash together my dispirate symptoms into a cohesive cause. Better to just check the connector when I get home.

So I have done some more testing for issues related to this problem and discover an additional symptom (like I need another one) I think. First the test results; checked the hall sensors via the Grin directions and all appear to be functioning properly. I then checked the fets, also via the Grin method and found significant resistence between each phase wire and the positive and negative power leads off the controller. So no problem with the fets or the halls...right? Now onto the symptom. When the bike is powered off, there is little rolling resistence either forward or backward. When powered on, there is a LOT of rolling resistence backward. I am virtually sure this is new as I would have noticed this much resistence before now. Before all of this testing I also tested the 3-speed switch to find that it had no continuity between any combination of wires. After opening the switch housing I found that the black wire had intermittent contact at the back of the switch. After re-soldering all was well with the switch and on testing there is now a small discernible difference between the 3 power settings but the overall power is very low still and the battery sag issue is still significant. Any thoughts?

Now thinking back I am SURE the backward rolling resistence did not exist before this problem started. I have tested the halls and the fets per the Grin site and all appear to be functioning properly. Please help if you can.