Power "Sag" when going up a long hill. But not overheating

rberger

10 W
Joined
Mar 9, 2015
Messages
76
Location
Saratoga, CA
I resolved the Cycle Analyst intermittent shutdown issue described in and earlier post on this forum: Cycle Analyst shuts down randomly and infrequently.

But I continue to have an issue that I'm pretty sure is not related to the Cycle Analyst, where the power delivered by the motor suddenly drops in half while climbing a 2 mile steep (1500 ft rise) hill. If I stop for a few minutes and restart climbing (I don't reset anything, it stays on the whole time), it works at full power for a little while (several minutes).

It seems to only happen while doing such a long continuous climb. It doesn't show up when I do a bunch of shorter (but equally steep) climbs followed by downhill or level travel.

I did check all the Phaserunner v2.0 settings and they seem correct based on the V2.0 Manual and some earlier feedback from Grin Tech. I will post those at the end of this.

I did check the CA diagnostic view while the sag was happening and:
  • None of the Limit Flags are set (other than the expected intermittent W flag)
  • The Throttle out is pinned to 4.00V while I'm pedaling or throttle is on full during the power sag)
  • It does this with PAS or Full Throttle (i.e. its not a PAS issue)
  • Motor temperature has never been more that 70 to 80 degrees Celsius during the climb.

Now when this sag happens (power drops to 600 to 800 Watts or so, from the peak of 1500 to 1750 watts when there is no sag) I'm going up a twisty narrow steep hill. I suddenly have to pedal at a level that is pretty strenuous and pushing me beyond what is my theoretical max heart rate. I also have to pedal to at least get myself to a safe place on the road so when a car does come by I'm not a sitting duck. Also I don't have the wherewithal to do much else than just try to get to a safe spot. So I'm not able to do much more checking of the CA than glancing at the diagnostic view.

The setup is:
  • Recent Crystalyte H3540 with statorade, 35mm laminations and aluminum stator core (should be able to do 1500W continuous)
  • Cycle Analyst v3
  • Phaserunner v2
  • Two Grin Tech B5213LiM-DT 52v 13Ah 25A Continuous Output Batteries in Parallel (so 50A Continuous Output total)
  • Thun Bottom Bracket Torque Sensor for PAS
  • Thumb Throttle
  • ICE Adventure Recumbent Trike
  • Total Weight (Human, Trike, Motor, Battery, Misc Stuff): ~325 lbs

My theory is that there is something hitting a limit in the Phaserunner settings that is cutting in at the point of Sag. Which mean something is not set right as I'm pretty sure I'm not at the limits of the Motor, Phaserunner or Batteries. But of course I could be wrong...

Would love to hear some suggestions on how to debug this or how people may have fixed something like this before. Below is the Phaserunner setup.
Thanks!

Phaserunner Basic Setup - Click for large view

Phaserunner Advanced (All Defaults) Setup - Click for large view
 
I don't know how to tell the temperature of the Phaserunner.
 
With a thermometer? ;)

Lots of possible ways.

Stick-on strip thermometer on the side of the controller, and stop and glance at it whenever you want a reading.

You could mount the wired (or wireless) outdoor sensor of any indoor/outdoor thermometer against the heatsink side of the controller (tape, glue, etc). Then mount (tape, zipties, etc) the display on the handlebars.

Tape or glue a glass tube thermometer to the side of the controller, stop and glance at it as needed.

tape, glue, ziptie, etc the metal probe of a bbq or candy thermometer to the controller, and stop and glance at it, etc.

Etc.
 
amberwolf said:
With a thermometer? ;)

Lots of possible ways.

Stick-on strip thermometer on the side of the controller, and stop and glance at it whenever you want a reading.

You could mount the wired (or wireless) outdoor sensor of any indoor/outdoor thermometer against the heatsink side of the controller (tape, glue, etc). Then mount (tape, zipties, etc) the display on the handlebars.

Tape or glue a glass tube thermometer to the side of the controller, stop and glance at it as needed.

tape, glue, ziptie, etc the metal probe of a bbq or candy thermometer to the controller, and stop and glance at it, etc.

Etc.

I use an infrared thermometer to check motor and charger temps.
 
rberger said:
My theory is that there is something hitting a limit in the Phaserunner settings that is cutting in at the point of Sag. Which mean something is not set right as I'm pretty sure I'm not at the limits of the Motor, Phaserunner or Batteries.
If you have a laptop and saddlebag or backpack, you might set the Phaserunner Utility to 'Dashboard' and configure it to dump the battery and motor Amps to the left axis and the controller temp to the right. Start the logging, load the laptop into the backpack, and run the course where the power sag occurs. If you stop when it sags and whip out the laptop, you should see the current and temp plots which should help clear up what the PR thinks it's seeing. With that in hand at least you may have a target problem to attack.

How do you have the PR mounted? This does smack of an overheating problem... If you have a thermistor on the shelf, you could stick it on the PR aluminum base and plug it into your CA instead of the motor, or pick up a cheapie BBQ thermometer for a test and just tape the probe along the side of the PR base.

taylorDigitalThermometer.jpg

  • FWIW: I'd start by loading the default XML file from the PR thread and reconfiguring things from scratch without any missteps exactly as the menual recommends - no embellishments or extra settings.. The PR has a nasty habit of going permanently loopy once parameters get too far afield from the defaults. Reloading the defaults seems like a shotgun approach but it's really the only way to get all the parameters back into proper ranges so the controller algorithms will work. Your settings look like you may done this already, but just thought I'd mention it....
 
I do have one of those infrared thermometer devices. I can try that.

My main problem is that I can end up being in a potentially dangerous position if the power sags while I'm on a blind curve and I don't have enough human watts available to move out of danger. In any case that will be what I'll be focused on when it sags and hard to take a temperature or whip out a laptop.

Does the Phaserunner have a way to log its data other than the dashboard? I could at least try to capture that log with a laptop in a bag and then review it after a run.

I did do the reset from default XML, set only a few things specific to my setup. But I could do that again to be sure.

The phase runner is mounted under the front boom and should be getting plenty of air:

Phaserunner under front boom

Click for large view
 
Have you run it through the Autotune feature in setup?

http://www.ebikes.ca/documents/Phaserunner_Manual_Rev2.0.pdf

Mine is mounted to a steel shoe the battery slides onto but barely gets warm under hard use like extended hill climbs; + 1500 feet elevation rise.
Phaserunner_Battery.jpg
 
I did do the autotune process and selected the option to reset things to defaults at the end.

Maybe I'll do a complete restore from defaults XML, autotune and only change the wheel size to be correct and see what happens with that. And I'll bring my infrared thermometer. Also sounds like a good excuse to buy a remote BBQ thermometer. I've been looking for an exuse for that to use on the BBQ :D

But I still would be interested to know if there is a way to log from the Phaserunner so that it could be reviewed later...
 
rberger said:
In any case that will be what I'll be focused on when it sags and hard to take a temperature or whip out a laptop.
Yep - that's why I though the CA or BBQ thermometer might be good (remote view)...

Anyhow, looking at your mount and the not-too-worrisome 80 phase amps max, I'm less convinced it's temp. My PR is configured almost identically and it hardly gets warm -- but I have a heatsink on it and don't really flog it, so no directly applicable first hand knowledge for comparison...

rberger said:
Does the Phaserunner have a way to log its data other than the dashboard?
No.

rberger said:
I did do the reset from default XML, set only a few things specific to my setup. But I could do that again to be sure.
If you are getting desperate, but otherwise should be good to go. (That said, I've reset/reloaded my PR many times 'just to sure' before getting to the end of the Yellow Brick Road...)

Looks like time for an email to Robbie at Grin....
 
I'm going to do the full reset, autotune, minimal change. And go for another ride up the hill, with a thermometer just to be sure before I email the Wizard of Grintech.

I agree that its a low probability that this setup is capable of causing a normally functioning Phaserunner to overheat even without a Heatsink.

Thanks for all the input! I'll let the thread know the results.
 
I reinitialized the Phaserunner with the Default Parameter XML.

Then did an autotune, did not accept its setting things to defaults like I did before with the assumption that the Default Parameter XML is more valid.

Then changed a few specific things:
  • Wheel Diameter to be the actual Wheel diameter
  • Max Current to 40A instead of 30A (Batteries are supposed to be capable of 50A)
  • Regen Voltage to match the battery

I then took it out for the same Big Hill (Strava Activity with Map, elevation and heart rate) and it failed with power sag the same way as the earlier times.

When I recovered enough to pull out the Infrared Temperature sensor, I measured the temp of the Phaserunner at its heatsink and the hottest components and at that point it was down to 57 degrees Celsius. So it was probably a few degrees hotter at the peak. Motor topped out around 70 degrees Celsius (according to the CA).

TempSensorSmall.jpg (After I got home, just showing how I got the temp. That point around the components was the hottest point, hotter than the heatsink portion of the Phaserunner)

I don't know if that is approaching the limits of the Phaserunner or not. It took at least 5 minutes to cool down the Phaserunner to around 47 degrees at which point it worked fine. But would sag after another few minutes of going up the hill. I can get up the hill with about 3 or 4 stops.

I didn't get a chance to try pulling out the laptop and checking the Dashboard.

Does anyone know if 57 degrees Celsius is near where the Phaserunner will start cutting back power?

I will send this to Grin Tech to see if they have any thoughts as well.

Here are the final values I used for the Phaserunner:

Basic__Bat_Max_Current_40___Default_XML___Wheel___Autotune___Manual_Max_Regen_small.png
Advanced__Bat_Max_Current_40___Default_XML___Wheel___Autotune___Manual_Max_Regen_small.png
 
Justin at Grin Tech agreed that the problem is the Phaserunner is probably hitting thermal limit that causes it to cutback power by half or so as I use full power continuously to go up the hill.

I am going to try McGyvering a heatsink that I lucky had around that kind of fits and see if that is enough heat sink to get me up the hill.
The found funky heatsink sm.png

Here is how I plan to mount it on the heat sink
Mounting on Heatsink sm.png

And how it will fit onto the Trike boom
How it will fit on the boom sm.png
 
FWIW, you might get better heat dissipation if you mount the PR against the boom and leave the fins all in the air.

I also recommend a *very small* amount of heatsink paste between the PR and heatsink. Put it on the two surfaces, then use a credit card or similar stiff flat bit of plastic, and basically remove all but what fills the scratches in the surfaces. Then bolt the two together.

If you put a bunch of paste in there, it'll make the heat transfer worse than without any paste at all.
 
amberwolf said:
FWIW, you might get better heat dissipation if you mount the PR against the boom and leave the fins all in the air.

This is why fans were invented. Unless he's going 20mph uphill, I'd suggest that the OP figure out a way to get a fan to blow air across the heatsink fins.
 
I don't see anything in either thread about this trike that says how fast he's going, but a fan could be done easily enough if he has a 12v lighting system, using that to power a computer-case fan.

Even if there's a 5v power available (other than the throttle/etc supply, which probably won't have enough current available) a 12v fan can be powered off of 5V, and it'll run really slow but still move air; if it's a big fan it'd be as good as a small fan spinning faster, and be virtually silent.
 
That heatsink should shed heat pretty well so long as it clears your shoes. I got the pipe clamps in the picture above from McMaster Carr FWIW.

It surprised me that it gets that hot on 1500 watts. I can run mine at 1500 Watts on a similar climb for over 15 minutes and it's only warm to touch but maybe I don't give the thin steel plate enough credit for quenching heat.
 
wturber said:
amberwolf said:
FWIW, you might get better heat dissipation if you mount the PR against the boom and leave the fins all in the air.

This is why fans were invented. Unless he's going 20mph uphill, I'd suggest that the OP figure out a way to get a fan to blow air across the heatsink fins.

Fans are my step 2. Want to see if I can get away with just a heatsink based on Triketech's experience. Just got back from HSC Supply in Sunnyvale (famous surplus place here) and found a couple of interesting heatsinks that are better than the one I showed earlier. Picked up a few different fans too.

I found a rather huge one (~5 inch diagonal) that will operate directly off my main battery voltage. Plus a couple small 12v ones. But I don't have 12volts on board and I don't yet have the ability to properly crimp new anderson connectors (ordering yet another tool, my attempts at doing it without the proper crimper ended in tears).

Going to see if I can put together the heatsink today and get a ride in before sunset.
 
Here are the heat sinks I found at HSC. The most interesting one is this nice copper one that is very compact but has a lot of cooling. I have to figure out how to mount the Phaserunner to the heatsink since I can't really screw it together since there is no surface to screw into. I'm thinking of starting with thermal adhesive and tie wraps. and then Tie Wrapping the whole thing to the boom.
PR on Copper Front View sm.jpg
PR on Copper w Fan Front View sm.jpg

Then there is this monster. Its really a bit too big to fit onto the boom as is. If I could shorten it somehow (Don't know how to cut this, don't have the right tools, but I could go to the Tech Shop or something) I think it could fit in the location I originally planned
PR on bottom of big HS front view sm.jpg

I'm wondering if its viable to put these assemblies behind the seat. Main problem is even less airflow but if I'm going to have a fan....

Open to suggestions and input!

Collection of Heatsinks top view sm.jpg
 
Personally, I'd go with the copper-based one, as it has a lot of surface area for it's size. YOu can even bend the thin fins out into a radial array if you need to, but I suspect just having it on there will help.

If the fins are in the way of drilling holes in teh copper to match the PR's holes, you can remove a fin or two from the spot.

But I'd just drill the holes from the open flat side, after having marked them based on the PR's holes/measurements. Then if the fins are in teh way of the holes, you can remove teh sections needed, if the drill isn't long enough to have done this for you already. :)
 
amberwolf said:
Personally, I'd go with the copper-based one, as it has a lot of surface area for it's size. YOu can even bend the thin fins out into a radial array if you need to, but I suspect just having it on there will help.

Yeah, I'm going to give that a try. I ordered some of those Vibration-Damping U-Bolts that Triketech used to mount it as well.
 
Here's a thread with some heatsink info that you may find interesting. There's a post in there where JLE runs down the supportable phase amp levels. I used these when I mounted up my PR some time back but seem to have 'mis-remembered' them in my post above (oops) - not as much current capacity as I recollected.

https://endless-sphere.com/forums/viewtopic.php?f=3&t=82196#p1220423

Also FWIW - I think for first cut on this and owing to the fat boom, I would reverse the mounting approach you show here - flip the heatsink over with the flat side against the boom and mount the PR between the fins on the down/forward side.

How it will fit on the boom sm.png
This will get the fins out in the airflow while letting you dump a bit of heat to the boom. That heatsink is moderately robust and something of that ilk would be my choice for an exposed location like the underside of the boom - not the fragile CPU heatsink. Also - you don't need to worry about vibration with the PR - the vib-dampening U-bolt will just thermally insulate the heatsink from the boom (not that there's a big contact patch there anyway) - plain vanilla hardware store or antenna-mount u-bolts would be better IMO. I like these saddle clamps from DX Engineering - with a little refinishing the aluminum saddles might up the heat transfer to the boom if you can get a matching diameter:

dxe-sad-200b_ml.jpg
 
teklektik said:
Here's a thread with some heatsink info that you may find interesting. There's a post in there where JLE runs down the supportable phase amp levels. I used these when I mounted up my PR some time back but seem to have 'mis-remembered' them in my post above (oops) - not as much current capacity as I recollected.

https://endless-sphere.com/forums/viewtopic.php?f=3&t=82196#p1220423

Also FWIW - I think for first cut on this and owing to the fat boom, I would reverse the mounting approach you show here - flip the heatsink over with the flat side against the boom and mount the PR between the fins on the down/forward side.


This will get the fins out in the airflow while letting you dump a bit of heat to the boom. That heatsink is moderately robust and something of that ilk would be my choice for an exposed location like the underside of the boom - not the fragile CPU heatsink. Also - you don't need to worry about vibration with the PR - the vib-dampening U-bolt will just thermally insulate the heatsink from the boom (not that there's a big contact patch there anyway) - plain vanilla hardware store or antenna-mount u-bolts would be better IMO. I like these saddle clamps from DX Engineering - with a little refinishing the aluminum saddles might up the heat transfer to the boom if you can get a matching diameter:

View attachment 1
Those clamps will "oval" the thinwall tube that has the boom tube fitted inside.
 
Triketech said:
Those clamps will "oval" the thinwall tube that has the boom tube fitted inside.
In this case the u-bolts are way overkill and only need to hold up a PR and aluminum heatsink. There's no need to torque the heck out of them to the extent of deforming a steel tube - zip ties can do the same job, so over-tightening beyond holding the PR assembly in place makes no sense. Just slip some heat shrink on the yoke to protect the paint.

If you really can't control your wrenching, slip a couple of small compression springs on the arms of the yoke (Ace Hardware) and run down a couple of Nylocks to partially compress the springs and hold the heatsink snugly w/o crushing anything.
 
teklektik said:
There's no need to torque the heck out of them to the extent of deforming a steel tube - zip ties can do the same job, so over-tightening beyond holding the PR assembly in place makes no sense.

They need to be snugged up enough to keep it from swinging under turns. I've played with something similar with a saddle I machined from 6160 with polyurethane pads. PU has a high coefficient of friction against the powder coat. 1/4 turn past snug distorted the tube.

He's already ordered the Pipe hangers anyway, they have about 357° of compression coverage and draw down without distorting the tube. Also been playing with split collars on a PD pad. If the pad is lubed with ethanol just prior to snug & torque the distortion is minimal. No lube and it begins to oval on tightening up. Like this image (click on the link the picture is too big for the page).

http://www.triketech.com/image/Drivetrain/BatteryISO_V1.jpg
 
Back
Top