PowerVelocity controller review

Status
Not open for further replies.
ElectricGod said:
tanstaafl said:
ElectricGod said:
I'd like to find some TO-220 IXYS mosfets with low Rds and high wattage.

This is the best I have found...
IXFP180N10T2
480 watts is comparable to the AOT290 at 500 watts.
.006 ohms vs .0035 ohms for the AOT290...not so great.
180 amps vs 140 amps, but both are leg limited to 75 amps so pretty much a mute point.
21nS vs 31nS on time is impressive.

Is the IXYS a better mosfet than the AOT290?
I guess it has to be a TO-220? I only see some of their better ones in other packages.

I wan't to try some of their mosfets in a 12 or 18 fet PV controller and see how it goes.
You have a source that would be affordable for 12x or 18x of these?
 
tanstaafl said:
I don't think I agree with you. You're trying to reduce RDS to a single value when it varies with current and temperature. And in this case, those two ratings from the data sheets have different assumptions -- the IXYS part's max is given for 50A, the AOT for 20A. Looking at the plots of RDS at various currents and temperatures for the two parts makes me think the IXYS part is actually better, but it's hard to get too confident since I've not seen the exact same conditions given. Still, it sure looks to me that the IXYS part will have a lower RDS. But I'm just guessing here.

I don't see how the IXYS part could have lower RDS, unless you're getting confused with the normalized values given. Looking at how little RDS changes for the AOT290 between 0-30A, I think it's safe to assume it still doesn't change much up at 50A. I also find it kind of strange that only the maximum value of RDS is given.
 
Addy said:
tanstaafl said:
I don't think I agree with you. You're trying to reduce RDS to a single value when it varies with current and temperature. And in this case, those two ratings from the data sheets have different assumptions -- the IXYS part's max is given for 50A, the AOT for 20A. Looking at the plots of RDS at various currents and temperatures for the two parts makes me think the IXYS part is actually better, but it's hard to get too confident since I've not seen the exact same conditions given. Still, it sure looks to me that the IXYS part will have a lower RDS. But I'm just guessing here.

I don't see how the IXYS part could have lower RDS, unless you're getting confused with the normalized values given. Looking at how little RDS changes for the AOT290 between 0-30A, I think it's safe to assume it still doesn't change much up at 50A. I also find it kind of strange that only the maximum value of RDS is given.

The AOT290 has a lower Rds based on the spec sheet. I know what you mean...spec sheets are quite often incomplete. I wish they all listed min, max and average values and rarely is that the case.

I found the IXYS mosfet on Digikey, but you have to buy them in quantities of 50. They are $4.35 each. 50 is a bit more than what I need to do a swap out in a 12 or 18 fet controller, but oh well, if they work out like I expect, they will get used else where too. They are about a dollar more than the AOT290, but if they are a much better mosfet, that's not really a big deal. The proof will be putting them in a controller and seeing what happens. 50 of them will cost me about $218.
 
ElectricGod said:
The AOT290 has a lower Rds based on the spec sheet. I know what you mean...spec sheets are quite often incomplete. I wish they all listed min, max and average values and rarely is that the case.

I found the IXYS mosfet on Digikey, but you have to buy them in quantities of 50. They are $4.35 each. 50 is a bit more than what I need to do a swap out in a 12 or 18 fet controller, but oh well, if they work out like I expect, they will get used else where too. They are about a dollar more than the AOT290, but if they are a much better mosfet, that's not really a big deal. The proof will be putting them in a controller and seeing what happens. 50 of them will cost me about $218.

I still think that the IXYS mosfet is not an improvement, but I agree the best way to know is to test it out. Can't you get it at another distributor? It's a bit of a gamble having to buy 50 of them.

I was looking at mosfet alternatives before, but I haven't checked them in depth. Here are a couple other promising alternatives:
CSD19536KCS
IPP023N10N5
 
Alan B said:
For controllers that parallel the FETs it may be helpful to match the units that are physically in parallel. So you could take advantage of the extra quantity to do that.

I haven't done this matching myself, but it has been discussed in various threads and done in some of the controller upgrades.

I've done mosfet matching on my 12 fet PV controllers that I upgraded to AOT290's.

Supposedly IXYS mosfets are really good. I've had AOT290's that were really average and others that were very good. Quality seems to vary a fair bit. The last batch of them I got was not as good as the first batch I bought. That was a bit disappointing. If the IXYS mosfets are simply more consistent than the AOT290, then that's already an improvement. I've had quite a few AOT290's that had an Rds far higher than the specs. I'll try these mosfets and see how it goes.

Keep in mind that the AOT290 is a great mosfet, but for the whiney and picky folks out there that want perfection, get extras so you can sort through all of them to find the best of the batch. That's realistically true of any mosfet.
 
ElectricGod said:
Supposedly IXYS mosfets are really good. I've had AOT290's that were really average and others that were very good. Quality seems to vary a fair bit. The last batch of them I got was not as good as the first batch I bought. That was a bit disappointing. If the IXYS mosfets are simply more consistent than the AOT290, then that's already an improvement. I've had quite a few AOT290's that had an Rds far higher than the specs. I'll try these mosfets and see how it goes.
that might have been the reason why the 12F from my friend died with only 48V battery and a small DD hub. the quality or consistency seems to be a problem.
Keep in mind that the AOT290 is a great mosfet, but for the whiney and picky folks out there that want perfection, get extras so you can sort through all of them to find the best of the batch. That's realistically true of any mosfet.
i just thought about how many FET's you would need to buy than to find similar ones for a 12F for instance.
It is unacceptable.
 
I'm not convinced these parts are that inconsistent, unless they are counterfeit knockoffs which are common.

Many people do not use proper static control and damage these static sensitive parts while handling them. Static damaged parts may work for awhile and then fail later.

You only need 6 matched pairs for a 12 FET controller. The pairs can be different, just the two in parallel should be well matched. If they all come from the same batch they may be fairly well matched already.
 
Alan B said:
I'm not convinced these parts are that inconsistent, unless they are counterfeit knockoffs which are common.

Many people do not use proper static control and damage these static sensitive parts while handling them. Static damaged parts may work for awhile and then fail later.

You only need 6 matched pairs for a 12 FET controller. The pairs can be different, just the two in parallel should be well matched. If they all come from the same batch they may be fairly well matched already.

I made 2 orders from Verical/Arrow. I just went back and looked at my order history. I buy from Digi-key as well so I was confused until I looked at my past invoices just now. One order was in March 2017 (qty 30) and the other was in May 2017 (qty 40). Anyway, both companies are pretty legit...so I doubt they are fake mosfets.

I have had results while testing AOT290's that vary a lot. For example I have never had Vgs(th) be lower than 3.4 volts despite the minimum being 2.9 volts. 3.6 volts was common in my first batch, but I had several that were 4.1 or 4.2 volts too. Just about all of the second batch have a Vgs(th) of 4.1 to 4.3 volts which is the max or higher and none were at or near the typical rating.

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've never scrutinized a mosfet as much as I have the AOT290 so I can't speak from experience if this is typical or not. I'm just saying, I want to try something else...aka the IXFP180N10T2 and scrutinize it like I have the AOT290. It costs a dollar more per mosfet. It had better be at least a more consistently similar mosfet!
 
ElectricGod said:
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.

it's milli ohms, not micro ohms. How do you know they are well over 3.5 if you don't even know the correct notation?
 
If the parts are out of specified ranges, any major customer could return them to the manufacturer. This is very costly for the manufacturer and the customer so it only happens rarely. If it happens much then Digikey will stop carrying the part.

For this reason I suspect they are almost always within spec, and if not then they are counterfeit, or they are damaged, or the measurement is not done it the way the manufacturer does it. Measuring milli-ohms is tricky, for example and any connections carrying the main current cannot be part of the voltage measurement, a 4 wire technique is required. Instruments must be accurate in the manner used.

There are cases where the conterfeit parts do get into the standard distribution channels, but this should be rare. It happens from time to time. But not often.
 
Addy said:
ElectricGod said:
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.

it's milli ohms, not micro ohms. How do you know they are well over 3.5 if you don't even know the correct notation?

So that's the best you can do...make note of a minor mistake? Really? Why bother!
 
Alan B said:
If the parts are out of specified ranges, any major customer could return them to the manufacturer. This is very costly for the manufacturer and the customer so it only happens rarely. If it happens much then Digikey will stop carrying the part.

For this reason I suspect they are almost always within spec, and if not then they are counterfeit, or they are damaged, or the measurement is not done it the way the manufacturer does it. Measuring milli-ohms is tricky, for example and any connections carrying the main current cannot be part of the voltage measurement, a 4 wire technique is required. Instruments must be accurate in the manner used.

There are cases where the conterfeit parts do get into the standard distribution channels, but this should be rare. It happens from time to time. But not often.

My first step in testing is to pull out my component tester and run all of the mosfets through it. This finds all the mosfets that turn on at too high a voltage. It happens quite a lot. All of the second batch of AOT290's Vt is too high. Maybe I should send them back and get replacements. This particular mosfet is a random scavenged mosfet just to show results on the parts tester. I could post 20 pictures of my AOT290's as proof. I think most manufacturers don't test the mosfets they receive at all until they are in an actual controller and then it's more about just seeing the controller run. Little or no concern is given for picking out the best matching set.

Mosfet%20test.jpg
 
I don't think reputable component manufacturers ship parts that are out of specifications very often. They test parts before they ship them. If they don't test every single part, they test a representative sample that is designed to be adequate to insure the batches of parts are within spec. On more costly parts, every single part is tested in some way. Not every spec will be tested, but a test designed to weed out the bad parts, and they know how to detect bad parts efficiently. They have to.

If you are seeing parts out of spec then something fishy is going on. Counterfeit parts, ESD damage, different test conditions, something.

The typical values are not specs, but the min and max values generally are, the datasheet has to be looked at carefully to understand what's guaranteed vs not.
 
I would be suspect of the test conditions. Gate threshold voltage is not that important as long as they are matched fairly closely.
 
ElectricGod said:
Addy said:
ElectricGod said:
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.

it's milli ohms, not micro ohms. How do you know they are well over 3.5 if you don't even know the correct notation?

So that's the best you can do...make note of a minor mistake? Really? Why bother!

Because it reveals a lot. If you actually understood how to accurately measure resistances on the order of 3.5mOhm, you wouldn't mess the units up. No, an ebay component tester can't accurately measure this.
 
Addy said:
Because it reveals a lot. If you actually understood how to accurately measure resistances on the order of 3.5mOhm, you wouldn't mess the units up. No, an ebay component tester can't accurately measure this.

from my experience those measuring devices from ebay are not bad.
I own a resistance meter (4-wire method) and it measures very precise (like a calibrated DMM) ;)
 
Manufacturer direct should minimize the opportunity for counterfeit parts to get in. You'll have to ask them if they will guarantee the parts are from the same silicon wafer or the same run of wafers. I haven't heard that they will do that, but they could. The quantities involved might be large for direct purchasing.

Matching is another story. Same Batch is helpful, same wafer is much better, but how do you know what you have. Testing is the only way to know. The paralleled sets of FETs should be matched for gate turn-on voltage and for saturation resistance, and for the reverse diode characteristics if possible. The goal is to equalize current sharing and heat production among the two to four physically paralleled devices. So the whole 12 or 24 FETs don't need to be identical. The diodes are probably harder to match, and they don't have a temperature coefficient that helps equalize like the FETs do, so they may be more critical, depending on how the firmware runs the FET and whether the diodes are carrying much power.

Or just run them very conservatively and hope for the best, which is what most manufacturer's do since matching parts costs too much for a production line. Then we come along and hotrod the box and wonder why the FETs are failing.
 
As far as I could tell, the controller uses synchronous rectification, so the diodes won't matter much.
 
Addy said:
ElectricGod said:
Addy said:
ElectricGod said:
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.

it's milli ohms, not micro ohms. How do you know they are well over 3.5 if you don't even know the correct notation?

So that's the best you can do...make note of a minor mistake? Really? Why bother!

Because it reveals a lot. If you actually understood how to accurately measure resistances on the order of 3.5mOhm, you wouldn't mess the units up. No, an ebay component tester can't accurately measure this.

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.
 
madin88 said:
Addy said:
Because it reveals a lot. If you actually understood how to accurately measure resistances on the order of 3.5mOhm, you wouldn't mess the units up. No, an ebay component tester can't accurately measure this.

from my experience those measuring devices from ebay are not bad.
I own a resistance meter (4-wire method) and it measures very precise (like a calibrated DMM) ;)

This is the second one of these types of devices I've purchased. Both measure components despite being very different designs nearly identically. Despite the fact that I paid $25 for it, it has been a very capable tool that I've used and used and used. The only reason why I retired my original one is the ZIF connector finally gave out and buying another one was so cheap, there was no point in replacing it. I'd recommend these things to anyone. They aren't perfect and there's things I wish they did...like measure Rds, but they do make identifying parts very quick and easy.
 
fechter said:
As far as I could tell, the controller uses synchronous rectification, so the diodes won't matter much.

How do you know this? No back diodes is a good thing right? If the back diodes are not really used, why can't I use mosfets that don't have them?
 
What is the indicator of synchronous rectification on the PV controller?

True that controllers not using the diodes would not benefit from matching them.

I've been working on my own controller firmware, and I may shift between synchronous and async depending on the current level. It is tricky to avoid slip regen with full synchronous rectification, so at low currents it might be easier to use the body diodes.

However it is still true that they would not matter much in terms of diode matching in that case, as they would not be dissipating much heat due to the current being low. The FET gate voltage and fully on resistance would still be important of course.

The back diodes are inherent in these powerfets, they are not optional.
 
ElectricGod said:
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?
 
Status
Not open for further replies.
Back
Top