PowerVelocity controller review

Get real world experience and user feedback on the electric bicycle products.
User avatar
fechter
100 GW
100 GW
Posts: 11822
Joined: Dec 31, 2006 3:23 pm
Location: California Bay Area, USA

Re: PowerVelocity controller review

Post by fechter » Nov 28, 2017 10:44 pm

Alan B wrote:
Nov 28, 2017 8:14 pm
What is the indicator of synchronous rectification on the PV controller?
I used a scope to look at the gates on both low and high side FETs while running. Both were switching at the PWM frequency. In non-synchronous controllers, the low side just stays off and the diodes do all the work.
"One test is worth a thousand opinions"

User avatar
Alan B
100 GW
100 GW
Posts: 7026
Joined: Sep 11, 2010 7:43 am
Location: San Francisco Bay Area, USA
Contact:

Re: PowerVelocity controller review

Post by Alan B » Nov 28, 2017 10:49 pm

fechter wrote:
Nov 28, 2017 10:44 pm
Alan B wrote:
Nov 28, 2017 8:14 pm
What is the indicator of synchronous rectification on the PV controller?
I used a scope to look at the gates on both low and high side FETs while running. Both were switching at the PWM frequency. In non-synchronous controllers, the low side just stays off and the diodes do all the work.
Excellent. Very definitive.

User avatar
ElectricGod
100 kW
100 kW
Posts: 1350
Joined: Nov 01, 2015 5:24 pm

Re: PowerVelocity controller review

Post by ElectricGod » Nov 29, 2017 2:28 am

fechter wrote:
Nov 28, 2017 10:44 pm
Alan B wrote:
Nov 28, 2017 8:14 pm
What is the indicator of synchronous rectification on the PV controller?
I used a scope to look at the gates on both low and high side FETs while running. Both were switching at the PWM frequency. In non-synchronous controllers, the low side just stays off and the diodes do all the work.
Does this mean I can use mosfets that have no back diode?

User avatar
ElectricGod
100 kW
100 kW
Posts: 1350
Joined: Nov 01, 2015 5:24 pm

Re: PowerVelocity controller review

Post by ElectricGod » Nov 29, 2017 2:39 am

Addy wrote:
Nov 28, 2017 10:05 pm
ElectricGod wrote:
Nov 28, 2017 7:54 pm
LOL...it reveals I made a mistake and nothing else. It looks more like you are looking for something to pick at than anything else. Do yourself a favor and just keep that sort of thing to yourself or state it as a question like..."Did you mean milli ohms? You said Micro ohms and I think you meant milli ohms." That shows a very different perspective and humility on your part. If you knew anything about me, you'd get this one detail...I'm human and I make mistakes from time to time. No big deal and I'm not the least bit bothered by getting things wrong or have an ego I have to protect when I do.
I apologize for being rude, I could have worded that differently. How about we change the subject then? Could you explain why you think some of these FETs are off spec? What was your measurement technique?
Hey...not a problem and thank you for apologizing.

The first test I do is use my component tester on all the mosfets. It's a quick test to determine the VGS(th) voltage. I want all of the mosfets I use to have the same voltage. The AOT290's I've gotten most recently are nearly all 4.2 volts...not the typical spec of 3.5 volts AND 4.2 volts is .1 volts over the max spec. This in and of itself isn't good, but if all the mosfets are of the same VGS(th), then this isn't significant enough to return the mosfets...if they are all the same value. However, the first batch I bought, most are 3.6 to 3.9 volts with a few above 4 volts. The second batch had 1 mosfet that was close to the typical spec and the rest were all 4.1 or 4.2 volts.

User avatar
fechter
100 GW
100 GW
Posts: 11822
Joined: Dec 31, 2006 3:23 pm
Location: California Bay Area, USA

Re: PowerVelocity controller review

Post by fechter » Nov 29, 2017 9:03 am

ElectricGod wrote:
Nov 29, 2017 2:28 am

Does this mean I can use mosfets that have no back diode?
The body diode in a MOSFET is intrinsic. You can't find ones without it. But you would want it there anyway because during synchronous rectification, you need a little dead time to prevent shoot-through. During the dead time, the diode might conduct.
"One test is worth a thousand opinions"

Addy
10 W
10 W
Posts: 66
Joined: Aug 11, 2017 11:15 am

Re: PowerVelocity controller review

Post by Addy » Nov 29, 2017 12:24 pm

ElectricGod wrote:
Nov 29, 2017 2:39 am
Hey...not a problem and thank you for apologizing.

The first test I do is use my component tester on all the mosfets. It's a quick test to determine the VGS(th) voltage. I want all of the mosfets I use to have the same voltage. The AOT290's I've gotten most recently are nearly all 4.2 volts...not the typical spec of 3.5 volts AND 4.2 volts is .1 volts over the max spec. This in and of itself isn't good, but if all the mosfets are of the same VGS(th), then this isn't significant enough to return the mosfets...if they are all the same value. However, the first batch I bought, most are 3.6 to 3.9 volts with a few above 4 volts. The second batch had 1 mosfet that was close to the typical spec and the rest were all 4.1 or 4.2 volts.
That doesn't sound too bad. 0.1V over the max spec isn't much, the component tester could just be inaccurate. Did you do any measurement for RDS? How do you handle and store these mosfets?

User avatar
ElectricGod
100 kW
100 kW
Posts: 1350
Joined: Nov 01, 2015 5:24 pm

Re: PowerVelocity controller review

Post by ElectricGod » Nov 29, 2017 6:52 pm

Addy wrote:
Nov 29, 2017 12:24 pm
ElectricGod wrote:
Nov 29, 2017 2:39 am
Hey...not a problem and thank you for apologizing.

The first test I do is use my component tester on all the mosfets. It's a quick test to determine the VGS(th) voltage. I want all of the mosfets I use to have the same voltage. The AOT290's I've gotten most recently are nearly all 4.2 volts...not the typical spec of 3.5 volts AND 4.2 volts is .1 volts over the max spec. This in and of itself isn't good, but if all the mosfets are of the same VGS(th), then this isn't significant enough to return the mosfets...if they are all the same value. However, the first batch I bought, most are 3.6 to 3.9 volts with a few above 4 volts. The second batch had 1 mosfet that was close to the typical spec and the rest were all 4.1 or 4.2 volts.
That doesn't sound too bad. 0.1V over the max spec isn't much, the component tester could just be inaccurate. Did you do any measurement for RDS? How do you handle and store these mosfets?
.1v over MAX spec is bad when the typical spec is significantly LESS than that.

Testing Rds is pretty easy...turn on the mosfet via 5 volts at the gate and measure the voltage drop across source to drain. A known series resistor gives you the last bit of data...current and then it's just V/I=R to determine Rds. Maybe this is an overly simplistic approach, but it works well enough. Realistically I should feed the mosfet a 5 volt square wave to simulate the signal it gets in a controller and measure the AC voltage drop, but I think a DC voltage gets close enough.

Addy
10 W
10 W
Posts: 66
Joined: Aug 11, 2017 11:15 am

Re: PowerVelocity controller review

Post by Addy » Nov 29, 2017 9:04 pm

ElectricGod wrote:
Nov 29, 2017 6:52 pm
.1v over MAX spec is bad when the typical spec is significantly LESS than that.

Testing Rds is pretty easy...turn on the mosfet via 5 volts at the gate and measure the voltage drop across source to drain. A known series resistor gives you the last bit of data...current and then it's just V/I=R to determine Rds. Maybe this is an overly simplistic approach, but it works well enough. Realistically I should feed the mosfet a 5 volt square wave to simulate the signal it gets in a controller and measure the AC voltage drop, but I think a DC voltage gets close enough.

Well, for your purposes that method might be adequate to point out bad mosfets, but that's not going to accurately measure down to 3.5 mOhm. This is why I doubted that you were measuring it properly. To properly test the RDS you need to do it in pulses so that the mosfet doesn't heat up significantly during the measurement, because this will change the RDS. You'd also want to have very good contact with the mosfet - holding multimeter leads by hand is probably not good enough. You should also apply something more like 10V to the gate - 5V is hardly sufficient and I doubt any decent controller would be driving the mosfets with such a low voltage. It really does make a difference. I've had mosfets with 4V thresholds that nearly melted when driven by 5V because they weren't being turned on well enough.

Anyways, you didn't answer the question about handling and storage. Are you wearing an ESD wriststrap while testing these mosfets? What kind of container do you put them in?

User avatar
ElectricGod
100 kW
100 kW
Posts: 1350
Joined: Nov 01, 2015 5:24 pm

Re: PowerVelocity controller review

Post by ElectricGod » Nov 29, 2017 10:12 pm

Addy wrote:
Nov 29, 2017 9:04 pm
ElectricGod wrote:
Nov 29, 2017 6:52 pm
.1v over MAX spec is bad when the typical spec is significantly LESS than that.

Testing Rds is pretty easy...turn on the mosfet via 5 volts at the gate and measure the voltage drop across source to drain. A known series resistor gives you the last bit of data...current and then it's just V/I=R to determine Rds. Maybe this is an overly simplistic approach, but it works well enough. Realistically I should feed the mosfet a 5 volt square wave to simulate the signal it gets in a controller and measure the AC voltage drop, but I think a DC voltage gets close enough.

Well, for your purposes that method might be adequate to point out bad mosfets, but that's not going to accurately measure down to 3.5 mOhm. This is why I doubted that you were measuring it properly. To properly test the RDS you need to do it in pulses so that the mosfet doesn't heat up significantly during the measurement, because this will change the RDS. You'd also want to have very good contact with the mosfet - holding multimeter leads by hand is probably not good enough. You should also apply something more like 10V to the gate - 5V is hardly sufficient and I doubt any decent controller would be driving the mosfets with such a low voltage. It really does make a difference. I've had mosfets with 4V thresholds that nearly melted when driven by 5V because they weren't being turned on well enough.

Anyways, you didn't answer the question about handling and storage. Are you wearing an ESD wriststrap while testing these mosfets? What kind of container do you put them in?
so then...still nit picking...I see...and only being marginally helpful.

How about you tell us how you would do it instead of constantly finding fault in what I say? Actually present real live help instead of finding fault all the time. I'm sure you are an intelligent person however it comes out as arrogance. I'm sure you know something I don't so just go ahead and tell us all how much you know and POST REAL INFORMATION INSTEAD OF NIT PICKING! If all you can accomplish is to find fault, then just go away!!!

Addy
10 W
10 W
Posts: 66
Joined: Aug 11, 2017 11:15 am

Re: PowerVelocity controller review

Post by Addy » Nov 30, 2017 10:44 am

Ok, there's no need to be so dramatic. The whole reason I was saying these things is because you posted:
ElectricGod wrote:
Nov 21, 2017 5:27 pm
In the product summary it says: RDS(ON) (at VGS=10V) < 3.5 micro ohms. further down in the actual spec section, it says typical is 2.5 and max is 3.2, but I've had quite a few that are well over 3.5 micro ohms. The specs for the AOT290 look great, but consistency and actual product matching up with the specs has been less than spectacular.

I did provide some information on how to measure it more accurately. I'm not saying you need to accurately measure the RDS of your mosfets - just refuting your claim that these mosfets are out of spec. I also asked you multiple times how you handle and store the mosfets, but it seems you're more interested in ranting. If you do have out of spec mosfets, it's probably caused by improper handling and storage.

For proper storage, you should keep the mosfets in an antistatic bag or container and only handle them when you're at an ESD safe workstation. The best thing would be to wear an ESD wrist strap which is grounded to prevent buildup of static electricity. I have an ESD mat at my electronics workbench, so I usually keep one hand touching that so I stay grounded while I'm handling sensitive things like mosfets. When I don't have a mat like that available, I touch something grounded to dissipate static charge before I handle anything sensitive.

User avatar
ElectricGod
100 kW
100 kW
Posts: 1350
Joined: Nov 01, 2015 5:24 pm

Re: PowerVelocity controller review

Post by ElectricGod » Nov 30, 2017 12:07 pm

Addy wrote:
Nov 30, 2017 10:44 am
Ok, there's no need to be so dramatic. The whole reason I was saying these things is because you posted:
ElectricGod wrote:
Nov 21, 2017 5:27 pm
In the product summary it says: RDS(ON) (at VGS=10V) < 3.5 micro ohms. further down in the actual spec section, it says typical is 2.5 and max is 3.2, but I've had quite a few that are well over 3.5 micro ohms. The specs for the AOT290 look great, but consistency and actual product matching up with the specs has been less than spectacular.

I did provide some information on how to measure it more accurately. I'm not saying you need to accurately measure the RDS of your mosfets - just refuting your claim that these mosfets are out of spec. I also asked you multiple times how you handle and store the mosfets, but it seems you're more interested in ranting. If you do have out of spec mosfets, it's probably caused by improper handling and storage.

For proper storage, you should keep the mosfets in an antistatic bag or container and only handle them when you're at an ESD safe workstation. The best thing would be to wear an ESD wrist strap which is grounded to prevent buildup of static electricity. I have an ESD mat at my electronics workbench, so I usually keep one hand touching that so I stay grounded while I'm handling sensitive things like mosfets. When I don't have a mat like that available, I touch something grounded to dissipate static charge before I handle anything sensitive.
Dude...just impart your vast wisdom and get on with it. You obviously think you know more than me so just spill it already. Also, do the world a favor and keep the posturing and fault finding to yourself. We don't care.

How would you test Rds...spell it out in simple steps for the weak minded and inferior folks that post here. Have you built a test rig for doing this test? Post pictures! Do you have a schematic? Post that too! So far you've done lots of nit picking and very little actual beneficial action.

So then...regarding mosfets...I store them in plastic wrap commonly used for covering food. I do all my work at a folding plastic table. It's super cool to see stuff cling to the desk when it picks up a good static charge! I deliberately scuff my feet across super staticy carpet over and over so that I can zap the mosfets as much as possible when I pick them up. Ground strap? What's that? Oh wait...I think my breaker panel has one of those in it for connecting it to the water pipe. Anti-static bag? I thought mosfets ran on electricity so why would I worry about them needing an ESD bag? And...to top it all off my soldering iron cost me $12 from radio shack and one of the wires from the power chord is shorted to the metal shell. I have to be careful to not touch it or I'll get shocked. If I touch the tip to something grounded, there's a giant spark. Is that bad?

Just impart your vast wisdom upon us lowly dim wits so that we can all be impressed with you.
Last edited by ElectricGod on Nov 30, 2017 2:09 pm, edited 1 time in total.

User avatar
madin88
10 MW
10 MW
Posts: 2332
Joined: May 27, 2013 2:02 am
Location: Austria

Re: PowerVelocity controller review

Post by madin88 » Nov 30, 2017 12:26 pm

Addy wrote:
Nov 29, 2017 9:04 pm
Well, for your purposes that method might be adequate to point out bad mosfets, but that's not going to accurately measure down to 3.5 mOhm. This is why I doubted that you were measuring it properly. To properly test the RDS you need to do it in pulses so that the mosfet doesn't heat up significantly during the measurement, because this will change the RDS. You'd also want to have very good contact with the mosfet - holding multimeter leads by hand is probably not good enough. You should also apply something more like 10V to the gate - 5V is hardly sufficient and I doubt any decent controller would be driving the mosfets with such a low voltage. It really does make a difference. I've had mosfets with 4V thresholds that nearly melted when driven by 5V because they weren't being turned on well enough.

Anyways, you didn't answer the question about handling and storage. Are you wearing an ESD wriststrap while testing these mosfets? What kind of container do you put them in?
You should buy one of those "ebay IR meter" and try it out. You will be really surprised how accurate they are.
With the 4-wire method and a calibrated DMM (V and A DC) you can measure it very precise with U = R x I
Yes you have right that temperature is important, but it will not heat up if you pass only some mA (which is enough) through it for measuring the voltage drop.

User avatar
ElectricGod
100 kW
100 kW
Posts: 1350
Joined: Nov 01, 2015 5:24 pm

Re: PowerVelocity controller review

Post by ElectricGod » Nov 30, 2017 2:11 pm

madin88 wrote:
Nov 30, 2017 12:26 pm
You should buy one of those "ebay IR meter" and try it out. You will be really surprised how accurate they are.
With the 4-wire method and a calibrated DMM (V and A DC) you can measure it very precise with U = R x I
Yes you have right that temperature is important, but it will not heat up if you pass only some mA (which is enough) through it for measuring the voltage drop.
Do you have a favorite unit? I'm open to options! These little devices get better and better...I just went to look at the current batch of them and noticed that they now show IR signals from your TV remote. It would be cool to find one that has a basic function generator, does Rds in mosfets, does IR, temp sensor and hall testing and of course all the other stuff that I already have.

User avatar
Alan B
100 GW
100 GW
Posts: 7026
Joined: Sep 11, 2010 7:43 am
Location: San Francisco Bay Area, USA
Contact:

Re: PowerVelocity controller review

Post by Alan B » Nov 30, 2017 7:24 pm

The manufacturer's data sheet usually shows test circuits. Start there. The little testers are useful but don't create the proper conditions for a real measurement that you could compare to the data sheet.

Static electricity will damage these FETs, and they can be partially damaged so they still work but don't meet specifications. Proper handling is essential.

User avatar
ElectricGod
100 kW
100 kW
Posts: 1350
Joined: Nov 01, 2015 5:24 pm

Re: PowerVelocity controller review

Post by ElectricGod » Dec 01, 2017 12:03 am

Alan B wrote:
Nov 30, 2017 7:24 pm
The manufacturer's data sheet usually shows test circuits. Start there. The little testers are useful but don't create the proper conditions for a real measurement that you could compare to the data sheet.

Static electricity will damage these FETs, and they can be partially damaged so they still work but don't meet specifications. Proper handling is essential.
I was tired of Addy's ongoing nit picking and general unhelpfulness so I decided to be sarcastic. You do realize that is what I was doing right?
Addy wouldn't post real information on how to test for Rds and other specs. All he is good for is finding fault and being otherwise useless.

Care to post real information on HOW TO TEST MOSFETS? It would be of actual use to people.

User avatar
ElectricGod
100 kW
100 kW
Posts: 1350
Joined: Nov 01, 2015 5:24 pm

Re: PowerVelocity controller review

Post by ElectricGod » Dec 01, 2017 12:10 am

madin88 wrote:
Nov 30, 2017 12:26 pm
Addy wrote:
Nov 29, 2017 9:04 pm
Well, for your purposes that method might be adequate to point out bad mosfets, but that's not going to accurately measure down to 3.5 mOhm. This is why I doubted that you were measuring it properly. To properly test the RDS you need to do it in pulses so that the mosfet doesn't heat up significantly during the measurement, because this will change the RDS. You'd also want to have very good contact with the mosfet - holding multimeter leads by hand is probably not good enough. You should also apply something more like 10V to the gate - 5V is hardly sufficient and I doubt any decent controller would be driving the mosfets with such a low voltage. It really does make a difference. I've had mosfets with 4V thresholds that nearly melted when driven by 5V because they weren't being turned on well enough.

Anyways, you didn't answer the question about handling and storage. Are you wearing an ESD wriststrap while testing these mosfets? What kind of container do you put them in?
You should buy one of those "ebay IR meter" and try it out. You will be really surprised how accurate they are.
With the 4-wire method and a calibrated DMM (V and A DC) you can measure it very precise with U = R x I
Yes you have right that temperature is important, but it will not heat up if you pass only some mA (which is enough) through it for measuring the voltage drop.
This unit claims to do Rds. I bought one...I'll post how well it works. That's actually helpful!

https://www.ebay.com/itm/Transistor-Tes ... 37idjAYHXQ

User avatar
Alan B
100 GW
100 GW
Posts: 7026
Joined: Sep 11, 2010 7:43 am
Location: San Francisco Bay Area, USA
Contact:

Re: PowerVelocity controller review

Post by Alan B » Dec 01, 2017 1:01 am

It will give you a value which may be useful, but probably not comparable to the specifications test conditions, (one thing you'll probably want to figure out is at what test current). For the spec sheet they probably do it at fairly significant current, and a cheap meter probably does not. As was already covered, 5V on the gate is not a good value to use. If you want to compare to the data sheet you have to use the same values they did.

I just glanced at the AOT290L datasheet, and they give the test circuits as I mentioned before. No need to ask for them. You already have them.

User avatar
ElectricGod
100 kW
100 kW
Posts: 1350
Joined: Nov 01, 2015 5:24 pm

Re: PowerVelocity controller review

Post by ElectricGod » Dec 01, 2017 12:47 pm

Alan B wrote:
Dec 01, 2017 1:01 am
It will give you a value which may be useful, but probably not comparable to the specifications test conditions, (one thing you'll probably want to figure out is at what test current). For the spec sheet they probably do it at fairly significant current, and a cheap meter probably does not. As was already covered, 5V on the gate is not a good value to use. If you want to compare to the data sheet you have to use the same values they did.

I just glanced at the AOT290L datasheet, and they give the test circuits as I mentioned before. No need to ask for them. You already have them.
You are probably dead on...if the tester runs a mosfet at 200mA and more than 5 volts I would be surprised. BUT is this limited test "good enough"?

This is the test schematic for Rds presented in the AOT datasheet. RL would be a watt meter and shunt to show circuit amperage. Rg can be a 1k resistor and doesn't really matter what the value is. Vdd isn't stated in this schematic, but in the one above it looks to be about 12 volts. Vgs also was shown at about 12 volts. Feed the mosfet gate a 12 volt trap wave and watch the voltage drop across source to drain on an o-scope. Do a little math to calculate Rds and you have Rds the "real" way. If this produces similar results to a component tester that does Rds...then this setup is irrelevant since the component tester is producing "good enough" test measurements.

Image

1. I bet egregious mosfets are going to be off the mark regardless of the voltage and current they are run at. IE: An inexpensive component tester capable of Rds at a minimum ought to find those components easily and quickly.

2. A "simple" rig that's essentially the above schematic will provide the "real" test set up. I have everything needed to make that happen.

My hope is that the "quicky check" with the component tester does will result in the same information as actually testing the "real" way produces. OR at least is close enough that it doesn't matter very much to need further testing. My theory is that there is a level of mosfet testing that is "good enough" and that cheap a component tester will provide that. We will see! I do understand finding matched mosfets...hence my level of mosfet testing that I have already been doing. How close to "perfectly matched" is good enough? I guess that depends on the person and their expectations. Will a $30 component tester purchased on ebay provide "good enough" results? My suspicion is that it will, but the proof is testing a mosfet with a component tester and then comparing those results with what the test rig produces. If the results are similar and therefore "good enough" then a mosfet testing rig is a mute point.

I don't know this for a fact, but I suspect that most controller manufacturers assume that the mosfets they bought from whatever supplier are all good and at most, put them through a very quick test to make sure they turn on and off correctly. Beyond that, they put them in a controller and test the controller as a complete unit. If it runs a motor and passes the various system tests they do, then the mosfets are considered "good enough". It's only people like me that might want another level of precision in finding the best pairing of mosfets.

Since no one has provided any actual demonstrations of such work...only opinions, I'm suspicious that they have not done any testing at all. In that case, opinions about what is or isn't going to work to produce "good enough" test results are purely supposition and probably not based on actual experience. I'm not imputing anything, just stating my observations and suspicions. It seems that my basic testing with a component tester and a few steps beyond that are far more than anyone else typically does. Assuming mosfets are "in spec" and NOT testing them in any way at all seems highly probable in most cases. At least I do some basic testing. Which is more than anyone else has shown demonstrable results for. LOL!

When I say, that the AOT290's I have received don't match spec or vary widely, it's because I've done actual testing of those mosfets. This is infinitely more than anything that anyone else has presented, since no one else has presented anything at all...just opinions. I have actual testing on my side. What does anyone else have? Nothing, zero, nada, just opinions and suppositions!

I have a thread that I started recently. It's far from done yet and I have hardly put any content in it. It's almost entirely place holders at the moment. See below. I intend to put a post in there about mosfet testing. The results from various ebay component testers and a "real" test set up all on the same mosfet will be presented. THAT will settle this and then people can nit pick or state opinions all they want, but the test results will be the actual facts. I don't know when I will get to the mosfet testing specifically, but hopefully sooner rather than later. I'll wait at least until I have the component tester I just bought.

"Random tools and EV related projects"
viewtopic.php?f=10&t=91589&p=1337793#p1337793

User avatar
Alan B
100 GW
100 GW
Posts: 7026
Joined: Sep 11, 2010 7:43 am
Location: San Francisco Bay Area, USA
Contact:

Re: PowerVelocity controller review

Post by Alan B » Dec 02, 2017 1:48 am

5V on the gate is not sufficient for a saturation test.

I don't know where you get these ideas. This is not hobby stuff, this is business. There are big dollars riding on this stuff.

The manufacturer is liable for the specs. If they are found to be shipping bad parts they could be in for lawsuits and product liability. It could lead to fires and vehicle crashes, and huge loss of business.

You don't have to test the parts properly. But you should not claim they are out of spec unless you have tested them properly. You are defaming their brand and their product without proper testing. I don't know how good their parts are, but if you are finding parts that don't meet spec in the small quantity you are seeing that would imply to me that they're junk. I don't think Justin and ASI would be able to use them in their products at all if that were true. They would have way too many failed products coming back, and at least Justin's is potted so not easy to repair. It just doesn't add up.

They state 10V on the gate for most of their values. They also state a max gate turn on of 4.1. That's way too close to 5V to expect full saturation and minimum on resistance to have developed. Looks like 6.5 volts is starting to get there on their graph, but that is a typical and not a guarantee. You can see that there is improvement even between 8V and 10V on the gate. If you don't test with 10V on the gate you cannot compare your results to theirs, and you cannot claim "out of spec".

Occasionally there are errors in datasheets. Usually they are just confusing and most often interpreted incorrectly, but sometimes there are real errors, and they won't fix the parts, they might fix the datasheet, but more likely there will be some errata somewhere that you might not easily find.

User avatar
madin88
10 MW
10 MW
Posts: 2332
Joined: May 27, 2013 2:02 am
Location: Austria

Re: PowerVelocity controller review

Post by madin88 » Dec 02, 2017 6:22 am

I just want to bring up that powervelocity mentioned that there is the risk that FET's could die because of bad / inconsistent supply of his AOT290.
My 12F controller died withoun any reason at only 48V in combination with a 9C hub motor.

Even if the method of measuring the FETs from ElectricGod isn't exactly similar to the test conditions of the manufacturer, there are still differences so it should be at least a little bit meaningful.
Yes i agree that antistatic handling is very important and that partial damage could happen, but i assume that everyone who is working with such sensitive parts knows and sticks to those conditions.

Powervelocity.com
10 kW
10 kW
Posts: 617
Joined: Mar 27, 2015 2:15 pm
Location: California
Contact:

Re: PowerVelocity controller review

Post by Powervelocity.com » Dec 02, 2017 4:57 pm

I never mentioned the risk of FETs failing due to inconsistent supply. It's rather consistent and comes form a single supplier.
I did mention that there may be just bad mosfets in batches and that's why I warrant controllers against defects. I have repaired a few (not really many) controllers free of charge even though in some cases customer admitted it was their fault.

Anyhow, I am switching to different mosfets that will be coming directly from the manufacturer bypassing distributors. I expect this will minimize the possibility of failures. Not that it happens very often today but I am willing to invest a bit more in reliability upfront.
madin88 wrote:
Dec 02, 2017 6:22 am
I just want to bring up that powervelocity mentioned that there is the risk that FET's could die because of bad / inconsistent supply of his AOT290.
My 12F controller died withoun any reason at only 48V in combination with a 9C hub motor.

Even if the method of measuring the FETs from ElectricGod isn't exactly similar to the test conditions of the manufacturer, there are still differences so it should be at least a little bit meaningful.
Yes i agree that antistatic handling is very important and that partial damage could happen, but i assume that everyone who is working with such sensitive parts knows and sticks to those conditions.

User avatar
Alan B
100 GW
100 GW
Posts: 7026
Joined: Sep 11, 2010 7:43 am
Location: San Francisco Bay Area, USA
Contact:

Re: PowerVelocity controller review

Post by Alan B » Dec 02, 2017 8:31 pm

Which FETs are you going to use then?

User avatar
ElectricGod
100 kW
100 kW
Posts: 1350
Joined: Nov 01, 2015 5:24 pm

Re: PowerVelocity controller review

Post by ElectricGod » Dec 02, 2017 9:50 pm

Alan B wrote:
Dec 02, 2017 1:48 am
5V on the gate is not sufficient for a saturation test.

I don't know where you get these ideas. This is not hobby stuff, this is business. There are big dollars riding on this stuff.

The manufacturer is liable for the specs. If they are found to be shipping bad parts they could be in for lawsuits and product liability. It could lead to fires and vehicle crashes, and huge loss of business.

You don't have to test the parts properly. But you should not claim they are out of spec unless you have tested them properly. You are defaming their brand and their product without proper testing. I don't know how good their parts are, but if you are finding parts that don't meet spec in the small quantity you are seeing that would imply to me that they're junk. I don't think Justin and ASI would be able to use them in their products at all if that were true. They would have way too many failed products coming back, and at least Justin's is potted so not easy to repair. It just doesn't add up.

They state 10V on the gate for most of their values. They also state a max gate turn on of 4.1. That's way too close to 5V to expect full saturation and minimum on resistance to have developed. Looks like 6.5 volts is starting to get there on their graph, but that is a typical and not a guarantee. You can see that there is improvement even between 8V and 10V on the gate. If you don't test with 10V on the gate you cannot compare your results to theirs, and you cannot claim "out of spec".

Occasionally there are errors in datasheets. Usually they are just confusing and most often interpreted incorrectly, but sometimes there are real errors, and they won't fix the parts, they might fix the datasheet, but more likely there will be some errata somewhere that you might not easily find.
I DID say I would test according to the data sheet method. I even posted a screenshot of what is in the AOT datasheet. Where's the problem here? However if a Chinese component tester can produce the same results, why is that a bad thing? I'd say that's a VERY good thing in fact. It means anyone can spend $30 for a component tester and confirm they are getting good components. THAT is VERY worth while! If it actually pans out. We will see soon enough.

STILL...NO ONE except me has done any actual testing in ANY way at all to prove or disprove anything I have said. Opinions and suppositions mean nothing if not backed by proof! I have proof...what does anyone else have? Nothing at all except their faith, hopes and desires that what the manufacturers do is always legitimate. I personally trust testing. Faith without proof is worthless faith. I have proof that AOT doesn't meet their specs consistently.

IF AOT or anyone is manufacturing low quality components, then yeah, they deserve to get sued. If it's a few bad parts out of thousands of good ones, that's not what I'm referring to. However, if they are producing loads of crap...or ANYONE DOES, then expect lots of people to have no respect for your products since you are NOT keeping to your specs! I'm not imputing anything against AOT, but my testing and their specs commonly DON'T match up.

I've bought 2 batches of AOT290's and tested EVERY SINGLE ONE OF THEM on my component testers. That means 100% testing...not spot testing. I have seen with my own 2 eyes that they are NOT all the same or even close. Those are the facts. Believe them or not. The same mosfet tested repeatedly produces the same results so clearly the component tester consistently tests the same every time. I'm definitely going to ALWAYS test EVERY mosfet I use since results vary widely from the manufacturer specs or from each other in the same batch.

If a Chinese component tester gives me the same results as testing according to the data sheet method, well I'm not ashamed of accepting there is an easy way to find out what I need to know.

User avatar
ElectricGod
100 kW
100 kW
Posts: 1350
Joined: Nov 01, 2015 5:24 pm

Re: PowerVelocity controller review

Post by ElectricGod » Dec 02, 2017 10:06 pm

Powervelocity.com wrote:
Dec 02, 2017 4:57 pm
I never mentioned the risk of FETs failing due to inconsistent supply. It's rather consistent and comes form a single supplier.
I did mention that there may be just bad mosfets in batches and that's why I warrant controllers against defects. I have repaired a few (not really many) controllers free of charge even though in some cases customer admitted it was their fault.

Anyhow, I am switching to different mosfets that will be coming directly from the manufacturer bypassing distributors. I expect this will minimize the possibility of failures. Not that it happens very often today but I am willing to invest a bit more in reliability upfront.
madin88 wrote:
Dec 02, 2017 6:22 am
I just want to bring up that powervelocity mentioned that there is the risk that FET's could die because of bad / inconsistent supply of his AOT290.
My 12F controller died withoun any reason at only 48V in combination with a 9C hub motor.

Even if the method of measuring the FETs from ElectricGod isn't exactly similar to the test conditions of the manufacturer, there are still differences so it should be at least a little bit meaningful.
Yes i agree that antistatic handling is very important and that partial damage could happen, but i assume that everyone who is working with such sensitive parts knows and sticks to those conditions.
Vadym,

I don't hold you at fault or the PV controllers in any way. As you know, I've fried a controller and that's 100% on me. LOL! I'd still buy more controllers from you and have.


To everyone else,

Something that would be good to do is to perform the exact same test series on IRF4110's, AOT290's and the IXYS IXFP180N10T2. The proof is in the pudding! Whichever one produces the most consistent results that match up with the manufacturer spec sheets wins! I'll buy some of the IXYS mosfets soon. They will all get tested and then the results will be whatever the results are. I already have 30+ AOT290's and easily 50 IRF4110's. These mosfets are from several batches made over a good bit of time. If the manufacturers are failing to make consistent components, then that will be proven by the many mosfets I have that have never been used. Realistically a second batch of the IXFP180N10T2 mosfets from something like a year from now would be evidence that they are or are NOT producing consistent components. No imputing intended...just letting the chips fall as they may. I trust testing...that's proof! Everything and anything else is wishful thinking.

User avatar
fechter
100 GW
100 GW
Posts: 11822
Joined: Dec 31, 2006 3:23 pm
Location: California Bay Area, USA

Re: PowerVelocity controller review

Post by fechter » Dec 02, 2017 10:54 pm

I'll be interested in the testing. I've used lots of 4110s with good results.
"One test is worth a thousand opinions"

Post Reply