24S Lipo - the Good, Bad and Ugly

[Alan B]

I see a number of folks here on ES running 24S Lipo on 100 volt FETs.

24 * 4.2 = 100.8 volts hot off the charger

This would seem to be pushing it, though it will drop a bit pretty fast.

So what is folks experience "pushing it" with 24S?
'
Is it reliable?

Is it worthwhile?

Is it a good idea?

Has it caused any problems? Failures?

Would you do it again?

 

[spinningmagnets]

Because the initial surface charge of LiPo drops off rapidly, I am of the opinion that a LiPo battery pack is better served by limiting the charge to 4.10V per cell, and that would make a 24S pack 98.4V (4.20V is harder on the chemicals, adds only a sliver of extra range). That being said, its my understanding that a system can experience voltage spikes, and adding one or two capacitors to the red/black power input wires to smooth out spikes doesn't hurt anything (even if it ends up not 'helping' anything).

So far I have not used more than 37V, so take my opinion for whats its worth. I believe if you are sourcing a controller from Lyen, a capacitor upgrade is an easy and cheap request.

 

[Jeremy Harris]

Very much down to your particular installation and controller/motor combination as to whether it'll be reliable. All the evidence that's been gathered over many years shows that power semiconductor voltage ratings need to be derated a fair bit for reliability, as do capacitor ratings, but it depends on the definition of reliability. An ebike might not need thousands of hours of reliable operation, as it might well suffer from enough non-electrical wear and tear as to make it's life just a few hundred hours or so.

If you get lucky and have a motor/controller combination that keeps power line ripple and spikes at the FETs and capacitors to a volt or two, then you may find it's reasonable reliable, provided the FETs and capacitors are of decent quality. On the other hand, if you have a low inductance motor and capacitors with a high ESR then you may well find the risk of FET or capacitor failure has increased pretty dramatically by running close to the absolute voltage maximum rating.

Some have got away with doing this, some haven't, but it's very hard to determine with any degree of confidence as to which failures reported here and elsewhere are directly due to over-voltage and which are just failures that may have occurred if the voltage was well within limits.

If not being let down by a failure is important to you, then it might be an idea to carry a spare controller................

 

[hydro-one]

My buddy is running a lyen 4110 mosfet 18fet at 103v hot , he keeps asking me to turn up his charger just a bit more, every volt sees a large increase in range, due to his packs being 25s lipo......Personally i think the fets can do it (if real fets, they are overspeced by IR to be reliable for military uses), im much more worried about caps....if a cap blows you can kiss alot of stuff goodby if not the whole board. the inside of your controller is shorted by tons of foil fragments spraying everywhere....I had this happen at around a hundred volts, however i actually think this failure was caused by a bad battery connector causing ripple?

Overall, 100v max , seems ok got a few guys doing that

 

[Jeremy Harris]

My buddy is running a lyen 4110 mosfet 18fet at 103v hot , he keeps asking me to turn up his charger just a bit more, every volt sees a large increase in range, due to his packs being 25s lipo......Personally i think the fets can do it (if real fets, they are overspeced by IR to be reliable for military uses), im much more worried about caps....if a cap blows you can kiss alot of stuff goodby if not the whole board. the inside of your controller is shorted by tons of foil fragments spraying everywhere....I had this happen at around a hundred volts, however i actually think this failure was caused by a bad battery connector causing ripple?

Overall, 100v max , seems ok got a few guys doing that

Sorry, but it's a complete myth that "if real fets, they are overspeced by IR to be reliable for military uses". Mil Spec components are actually massively derated in use, so a 100V component would have maybe a 50 or 60V Mil Spec rating. When the manufacturer publishes a data sheet that says "Absolute Maximum Rating" that's exactly what they mean. Manufacturing tolerances will mean that most of the components may take a small amount more, but it's not something that can be relied on, as semiconductors aren't 100% tested, just a few are sample tested from each batch and a batch would pass if the samples *just* met the absolute maximum rating.

The real issue is the voltage spikes and ripple right at the FETs and capacitors, as this will add to the battery voltage in terms of the voltage those components see. Measuring this isn't easy, especially if you have a low inductance motor, as the spikes are likely to be of short duration and only exist fairly close to the FET pins on the board.

If you want to improve reliability whilst running right at the edge of the ratings, then I'd suggest making sure that the commutation capacitors are up to the job. Too often controllers seem to be fitted with pretty poor capacitors, in terms of ESR, and decent ones can really make life a lot easier for the FETs.

 

[spinningmagnets]

"Updated cap choices for high power controllers" (posted by LFP)
http://endless-sphere.com/forums/viewtopic.php?f=30&t=22194

This is the cap stage of a Honda Insight motor controller. Notice it uses 350vdc rated caps? Pack voltage is always under 175vdc on an insight, but they knew to leave a good deal of margin to keep the caps from being damaged by inductive spikes.

Controllers with 100V caps are recommended for 72V systems, Arlo used 160V capacitors on a controller he intended to use a 100V system. I have no experience with electronics, but for a 24S battery, I would recommend asking Lyen for his upgraded controller cap advice.

 

[Trackman417]

I wouldn't run 24S for an everyday commuter. It's just too much. I ran 24S once, just to see what it was like. Will I do it again? Sure, but only of someone schedules a race, I'll charge up all my battery's, put them on the bike and race.
Keep in mind I am 19 years old, I live for adrenaline and stupid things, but it gets to the point where you can't finess the throttle without being thrown back a little bit. At 100V, forget about touring and sight seeig the twitchy mess of the throttle will ruin it for you.
I would go for 20S. I run at 20S all the time, and I see 43mph. At 100V I see 45mph and a lot of extra heat. No use in the extra charge time, just for more torque. Forget if the components are up for the job, safety is a major concern to.

 

[Jeremy Harris]

big moose wrote a useful post about derating and included this link: http://snebulos.mit.edu/projects/crater/docs/engineering/EEE-parts-derating.pdf

Worth a read, but bear in mind that your own personal definition of reliability may well be lower than the expectation in that document, as I mentioned previously in this thread.

 

[wesnewell]

Here's my thought on this. A component manufacturer that doesn't under spec his components by at least 5% will find himself out of business in no time due to failures, so I have no problem going .8% over rated voltages. More damage will be caused to the caps by not using a pre-charge resistor than going over rated max voltage by .8%. For the fets, I'd be more concerned about the amperage than the voltage as the voltage will likely never hit a fet at more than 100V for a fraction of a second and any decent fet will take .8% over voltage. Obviously if you run the controller at lessor voltages and amps it will last longer because it won't generate as much heat which is the real killer. So that's why I've been using 24s on my cheapass controller since the beginng in June 2011, but only full time for the last few months. I bought 2 controllers so I'd have a spare anyway. At an average of $35 each shipped, I don't really care if one of them blows, but it hasn't. And I always charge to close to 100% charge. so to answer your questions.
Yes, it's been reliable so far.
Worthwhile? Only if you want the extra power/speed.
Probably not the best idea, but I don't care.
No failures, no problems.
Yes. I'd do it again, but mostly did it because I wanted to. Not because I needed it. 99% of my riding is under 15mph. I could get almost twice that with 12s. But that 1% is more fun.

 

[Alan B]

Good discussion folks.

What voltage are the caps in Lyen's 24S controllers rated at? Methods? Steveo's? Others??

 

[itchynackers]

I ran my 12 fet lyen 4110 controller at 25s (102.5V) for over a year without failure. It was (and still is) a daily commuter for my kid, however he now runs at 24s.

 

[Jeremy Harris]

Good discussion folks.

What voltage are the caps in Lyen's 24S controllers rated at? Methods? Steveo's? Others??

The voltage rating isn't really the most important parameter for commutation caps, it's the ESR (equivalent series resistance) together with the ripple current rating. If you used 100V rated capacitors at 100V supply, but they had a poor ESR, so that they allowed ripple and spikes of a few volts on the power rail close to the FETs/caps, then that creates a potential problem for both the caps and the FETs, by creating repeated voltage transients that will be well over the max rating. There is a great deal of evidence from reliability testing that proves, beyond any doubt, that over-stress like this reduces reliability. Even cheap commercial consumer products derate component absolute maximum voltage ratings by 10 to 20%, and manufacturers don't do things like this that add cost if they haven't been found to be essential.

BTW, wesnewell, inrush current isn't harmful to capacitors, neither is very fast discharge current (as witnessed by all those of us with capacitive discharge welders, photographic flash guns, strobe lights etc). As long as the mean current over time remains below the ripple current rating then the capacitor won't be harmed at all. The risk is only one of internal heating, from resistive loss, causing the electrolyte to expand or boil, and this will only usually happen if the capacitor is subjected to continuous high current, not a one-off, short duration event.

 

[cal3thousand]

I wouldn't run 24S for an everyday commuter. It's just too much. I ran 24S once, just to see what it was like. Will I do it again? Sure, but only of someone schedules a race, I'll charge up all my battery's, put them on the bike and race.
Keep in mind I am 19 years old, I live for adrenaline and stupid things, but it gets to the point where you can't finess the throttle without being thrown back a little bit. At 100V, forget about touring and sight seeig the twitchy mess of the throttle will ruin it for you.
I would go for 20S. I run at 20S all the time, and I see 43mph. At 100V I see 45mph and a lot of extra heat. No use in the extra charge time, just for more torque. Forget if the components are up for the job, safety is a major concern to.

Your HS3540 puts you in a very high speed at 24S. Some people have slower winds and 24S is the only way to get up to a good top speed.

And 3 speed switches help too.

 

[Blackssr]

It has been my daily commute of 14 miles a day.

 

[cal3thousand]

I have been running 24s3p on an EV Global Ebike with a Yescomusa motor for the last 3 weeks in the South Florida heat. It has been my daily commute of 14 miles a day. I have a Lyen 12 fet and CA DP. I pull about 40 amps and at wot can hit 48 -53 mph in the bike lane . I usually average 35 mph roundtrip and use about 5 -6 ah. I do not pedal at all. I have a 3 speed switch, also from Lyen and have it set like the this:
Position 1 (35%) - 20 mph
Position 2 - (60%)- 35 mph
Position 3 - (118%) - 48-53 mph

Position 1 is useless, Position 2 is very smooth and my favorite ride. Position 3 is very sensitive at low speed and only use this, on the way home, on a 2 mile stretch of newly paved road with no cross streets. The EV Global ebike handles well at high speed and the body panels keep everything dry during the Florida rain showers. I have regen and front and rear disc brakes to aide in stopping at high speeds. I installed 2.3" Michelin Pilot tires front and rear. They handle well in wet or dry conditions. I installed a high speed 12 volt cooling fan blowing directly on the Lyen controller and it stays cool even at wot. The warm air is exhausted at the back of the body panels below the seat post. I installed bright 10 LED headlights and 12 LED tail lights as well as working brake lights. These are powered by a 110 volt AC adapter for the 12 volt output required for horn and lights.

I feel 24S in the best way to go because the power is there if you need it and the efficiency is there at lower speeds for daily commute. I own an SRT8 challenger and a C6 Corvette convertible and have not put gas in either car it since April and realy enjoy this bike on my daily commute. I also have a Prodeco Phantom X that is collecting dust in the garage. It is just boring and the brake issue is too annoying to enjoy it.

Great detailed post! That's a very nice looking EVG too; I haven't seen one that looks metallic (maybe the photo conditions?)

I haven't used a drop of gas this week, and it feels amazing!

 

[Alan B]

For those people who run 24S and have had failures, are they voltage or current and heat related?

It may be hard to tell, but if the failure occurs when connecting a fully charged battery and prior to applying power, it might be voltage related.

If it occurs when the controller is HOT then it might be current/heat related. 24S systems tend to have a lot of current available.

How many have suffered failures that are likely voltage related?

Is there a good way to tell that the failure is voltage related?

 

[Arlo1]

I run 24s and 100.8 off the charger. My motor is an x5 and its a setup that is super easy on the controller even though I have had my 24 fet running over 250 battery amps and my 18 fet over 200 battery amps! I tried it once with lower amp settings and max voltage on Colossus and popped 3 controllers then tried it again at 84v off the charger and had it last a little long but still blew all the rest of my china controllers up.

So if you are going to run a controller at its max voltage then make sure you have an easy to run motor. My X5 has 255uh inductance and this is why the controller can handle it.

As for any failures mine are usually over currents but never any failures on the X5!

 

[ohzee]

I had a mxus before my HT that I ran at 24s on the same controller. My commute is about 30 minutes each way and as long as I babied the motor
I was fine.. But if I was to go WOT more then 20 minutes that baby would be smelling.. eventually I toasted the motor or at least I got it so hot the
halls failed. I fixed them , but had some issues and now Amber has that motor.

Only failure I have had so far on a 12 fet at 24S

 

[Doctorbass]

Never use 24s lipo with 5303 and 100V fets controller! :lol:
but with the 5304 or 5305 it's a totally different game!.. less high speed as well, but way mre reliable !

Doc

 

[cal3thousand]

Never use 24s lipo with 5303 and 100V fets controller! :lol:
but with the 5304 or 5305 it's a totally different game!.. less high speed as well, but way mre reliable !

Doc

What's better suited to handle 24S on a 100V FET controller? a HT3525 or a HS3540?

 

[Doctorbass]

Never use 24s lipo with 5303 and 100V fets controller! :lol:
but with the 5304 or 5305 it's a totally different game!.. less high speed as well, but way mre reliable !

Doc

What's better suited to handle 24S on a 100V FET controller? a HT3525 or a HS3540?

Use the HT3525.. wich is high torque and not high speed. It will be easy on your controller.

or use the HS3540 with 22s and enjoy!

Doc

 

[Alan B]

So the Cromotor with the standard winding would be a better choice for 24S on a 100V controller than a faster wind.

 

[cal3thousand]

Never use 24s lipo with 5303 and 100V fets controller! :lol:
but with the 5304 or 5305 it's a totally different game!.. less high speed as well, but way mre reliable !

Doc

What's better suited to handle 24S on a 100V FET controller? a HT3525 or a HS3540?

Use the HT3525.. wich is high torque and not high speed. It will be easy on your controller.

or use the HS3540 with 22s and enjoy!

Doc

Thanks for clearing that up!

 

[cal3thousand]

So the Cromotor with the standard winding would be a better choice for 24S on a 100V controller than a faster wind.

That's the logic I got out of it.