PowerVelocity controller review

Status
Not open for further replies.
The cheap 2 wire resistance meter arrived. As I expected, it measures resistance just fine in or out of a capacitor. It has 2 alligator clips which it is currently zeroed out for. It also has a plug in connector for small components. Since the connections in the connector are much shorter than the test clip leads are, going from one to the other requires re-zeroing the meter depending on which one you use. The shunt, depending on how I clamp onto it's legs reads much closer to the .005 ohms that it really is. Any losses obviously mess with the measured results.

Milli%20ohm%20meter%20-%20cap.jpg

Milli%20ohm%20meter%20-%20shunt.jpg
 
ElectricGod said:
The shunt, depending on how I clamp onto it's legs reads much closer to the .005 ohms that it really is. Any losses obviously mess with the measured results.

"where is the problem if it does not have 4-wire resistance measurement"
Well, now you know :lol:
 
madin88 said:
ElectricGod said:
The shunt, depending on how I clamp onto it's legs reads much closer to the .005 ohms that it really is. Any losses obviously mess with the measured results.

"where is the problem if it does not have 4-wire resistance measurement"
Well, now you know :lol:

Well..since I don't have a 4 wire low ohm meter, how would connecting 2 test leads to the same exact place on the shunt somehow produce different results? 2 test leads or one connected to the exact same place are electrically identical. I'll probably replace the alligator clips for better ones. It's just getting a good bite that makes the difference here and causes the losses.

Anyway, it DOES read resistance in or out of a capacitor and that was supposedly "a problem", but I was fairly certain it wouldn't be and it isn't.
 
Putting the excitation current and the reading currents through the same contact point causes the meter to be unable to differentiate the contact resistance from the shunt resistance, so it cannot accurately measure low values of target resistance. Four wire meters separate these two currents, and by using a very high impedance amplifier on the voltage readout they make the contact resistances disappear from the measurement.
 
Alan B said:
Putting the excitation current and the reading currents through the same contact point causes the meter to be unable to differentiate the contact resistance from the shunt resistance, so it cannot accurately measure low values of target resistance. Four wire meters separate these two currents, and by using a very high impedance amplifier on the voltage readout they make the contact resistances disappear from the measurement.

Contact resistance from 2 sets of alligator clips is going to be less than one set of clips. Same for good clips vs cheap ones. Two sets of clips are going to be attached to the part side by side. Assuming equally good connections, the resistance difference between the 2 side by side alligator clips is going to be negligibly different. If it was say 1 ohm different, then that's going top badly skew the measured result and invalidate any measurement. The chances of current separation between side by side alligator clips is also equally negligible. I would think that contact resistance can be zeroed out just by clipping the 2 leads together and zeroing the meter. Then, whatever you get when attached to a component should be all component resistance. 2 or 4 wires shouldn't make a difference since 2 wires electrically in parallel is going to be the same as a single wire.

Notice how I expose anything I discover. For example, depending on how the meter is connected to the shunt, I get different readings on the meter. It may be that better alligator clips significantly reduce this issue. I have some on order and they will be in my hands come January. I have no ego to protect and as I've said from the beginning...the results will be the results. I was also able to measure .005 ohms like my precision shunt is valued at. The picture when I took it shows .012 ohms, but just a second before, I was looking at .005 ohms after zeroing the meter.

Of course anything I post that I find will get used as "proof" that I'm doing something wrong whether it is or isn't wrong.

Anyone with a 4 wire meter...
Please do a test for me for comparison. Take a length of solid copper wire, a shunt, a 6-32 screw, leg to tab on a transistor...whatever you easily have on hand. Connect it to your meter. Don't cram the clips into the object under test, just clip them on. Measure the results. Now shift your leads around. Do you consistently get the same resistance? My suspicion is that you don't. My suspicion is that a 4 wire or 2 wire meter is going to suffer from contact resistance variations equally. And especially so at these low resistance levels.
 
ElectricGod said:
Anyone with a 4 wire meter...
Please do a test for me for comparison. Take a length of solid copper wire, a shunt, a 6-32 screw, leg to tab on a transistor...whatever you easily have on hand. Connect it to your meter. Don't cram the clips into the object under test, just clip them on. Measure the results. Now shift your leads around. Do you consistently get the same resistance? My suspicion is that you don't. My suspicion is that a 4 wire or 2 wire meter is going to suffer from contact resistance variations equally. And especially so at these low resistance levels.


I suggest you do some reading about 4-wire sensing: https://en.wikipedia.org/wiki/Four-terminal_sensing

There's no need to experiment or guess here, these are well understood and well documented concepts. With 2-wire sensing, contact resistance is in series with the measurement resistance so the reading you get is the sum of these. With 4-wire sensing, the voltage sense lines will still have contact resistance in series, but these lines are only measuring voltage, the current flowing is very small, so the contact resistance is insignificant.
 
ElectricGod said:
Notice how I expose anything I discover. For example, depending on how the meter is connected to the shunt, I get different readings on the meter. It may be that better alligator clips significantly reduce this issue. I have some on order and they will be in my hands come January. I have no ego to protect and as I've said from the beginning...the results will be the results. I was also able to measure .005 ohms like my precision shunt is valued at. The picture when I took it shows .012 ohms, but just a second before, I was looking at .005 ohms after zeroing the meter.

Of course anything I post that I find will get used as "proof" that I'm doing something wrong whether it is or isn't wrong.

Better alligator clips won't help much because the difference still would be way to high for comparing Rds on of mosfets.
I would keep this meter for measuring ESR of caps and other things, but would buy another one with 4-wires which can measure down to micro ohms like this one:
https://www.ebay.com/itm/NEW-ViCHY-...045780&hash=item588f6ae566:g:iBwAAOSwe-FU3YyC
Nobody wants to look for errors in what you are doing. You may see it like this, but my intention is to help you and to prevent of publishing measured values which are not correct. Learning never stops!
 
Thanks for doing the test for me guys...NOT! That's pretty much what I expected. No one wants to do anything at all except offer advice. LOL! Well thanks for that, but it's cheap and lazy on your parts!

Madin...thanks for the totally independent back-up on that 4 wire meter. I was already looking at it as a possible candidate...if I get a 4 wire meter. I'm not done with the 2 wire meter yet. It may not be perfect or exactly capable of what a 4 wire can do, but if it's good enough, then it's good enough. It is limited to .001 ohms and that may not be accurate enough. However, typical mosfets never have an Rds that low. I'll post my findings once I have the new alligator clips.

Madin...I found it for a bit cheaper...I think I'l go ahead and buy one. It's just money and since no one else will give me corroborating or differing results, then I can publish them myself! I expect someone will look at the meter and make comments about how it's a POS. LOL!

https://www.ebay.com/itm/VC480C-3-1-2-Digital-Milli-ohm-Multimeter-with-4-Wires-and-power-supply-Cord-US/232526662788?hash=item3623ab4c84:g:zMoAAOSw43hZ5cik
 
ElectricGod said:
Thanks for doing the test for me guys...NOT! That's pretty much what I expected. No one wants to do anything at all except offer advice. LOL! Well thanks for that, but it's cheap and lazy on your parts!

Why do a test if you're already certain what the result will be? You're free to spend your time learning things on your own that are already common knowledge, but it's not surprising that no one else feels the need to do these experiments.
 
Addy said:
ElectricGod said:
Thanks for doing the test for me guys...NOT! That's pretty much what I expected. No one wants to do anything at all except offer advice. LOL! Well thanks for that, but it's cheap and lazy on your parts!

Why do a test if you're already certain what the result will be? You're free to spend your time learning things on your own that are already common knowledge, but it's not surprising that no one else feels the need to do these experiments.

It took you longer to write your original reply than to do the actual test for me. Then you had to read my reply and write another reply. By then you could have tried this out on 10-20 different conductors and posted your actual results. LAME!!!
 
ElectricGod said:
It took you longer to write your original reply than to do the actual test for me. Then you had to read my reply and write another reply. By then you could have tried this out on 10-20 different conductors and posted your actual results. LAME!!!

I didn't do the test because I don't have a 4-wire meter and I don't plan to buy one right now because I don't think I need it.

On the same note, you could do some reading and gain some understanding of these concepts instead of yakking on about how you're the only one testing and everyone else must be lazy.
 
Addy said:
ElectricGod said:
It took you longer to write your original reply than to do the actual test for me. Then you had to read my reply and write another reply. By then you could have tried this out on 10-20 different conductors and posted your actual results. LAME!!!

I didn't do the test because I don't have a 4-wire meter and I don't plan to buy one right now because I don't think I need it.

On the same note, you could do some reading and gain some understanding of these concepts instead of yakking on about how you're the only one testing and everyone else must be lazy.

Well... 4 wire meter purchased for comparison purposes and a couple sets of decent grade gold plated kelvin clips purchased for the 2 wire meter. We will see what happens soon enough! If the 4 wire meter blows the socks off the 2 wire meter...fantastic...if it doesn't...fantastic. The results will be the results. That's infinitely more than anyone else can say.

Those IXYS mosfets are still sealed in their bag. I've upgraded my grounding (just in case) and added my ground strap. I am preparing for actual tests.

Addy...
I'm going to guess that you don't have any test equipment at all beyond maybe a DMM. That explains a lot! I can't imagine why you didn't just post that to begin with. It is perfectly OK to NOT have scopes and meters and function generators and what not. I've spent money on this stuff over the years and obviously am still willing to buy more of it. My best scope cost me several thousand dollars in the 90's. I'd love to replace it with something current, but it still works great so what's the point? You should have just said to begin with that you don't have test gear. Why bother hiding the fact that you can't test and that you are taking this all on faith? I own the test equipment so I CAN test. LOL! Seriously...that was totally weak. Now we all know why you don't ever test anything...you can't test...plain and simple. You have no choice but to take what you read on faith. And that's perfectly OK...just say so. Why was that so hard? There was quite literally nothing to hide except for the fact that you can't prove anything you say with actual testing. I on the other hand can test and can back up what I say with real live testing. Good god man! What a waste of time posturing and pretending when honesty was the best approach!

Not bragging, but I have 3 o-scopes, 2 function generators, 3 component testers, DMM, 2 wire low ohm meter and soon a 4 wire meter. I've built various test tools too. I have probably 15 soldering irons and a Hakko solder station. I also have various adjustable power supplies, bread boards and zillions of components. I've done electronic component level repair since the late 80's. You use this sort of equipment for that kind of work. I currently work as a software developer since 2010. Prior to that, 25 years in sysadmin and prior to that component level repair as a job for 4 or 5 years while in school and a while afterwards. I went to school for an EE in the 80's. I never worked a day in the field because I discovered I liked computers so much more and hated all the math an EE needs. Electronics have been a hobby of mine since I was 10 years old or so. I still have that yellow book written by Forrest Mims my mom bought me as a kid!
 
ElectricGod said:
Now we all know why you don't ever test anything...you can't test...plain and simple. You have no choice but to take what you read on faith. And that's perfectly OK...just say so. Why was that so hard? There was quite literally nothing to hide except for the fact that you can't prove anything you say with actual testing. I on the other hand can test and can back up what I say with real live testing. Good god man! What a waste of time posturing and pretending when honesty was the best approach!

Seems like you're more interested in conjecture than technical discussion.

ElectricGod said:
I never worked a day in the field because I discovered I liked computers so much more and hated all the math an EE needs.

Well, I can see why you don't bother learning the theory.

I'm not as old as you but I have worked in the field longer, apparently. I have two diplomas from studying EE, one in computer control and the other in industrial power. I have been working for the past 3 years since graduating. I design and test electronics for a company that manufactures well-logging equipment.

At home and at work I have scopes, DMMs (portable and bench models), function generators, bench power supplies and various other equipment. It's not that I lack equipment to perform tests, but I learn and understand concepts without needing to prove everything with a test.
 
Addy said:
ElectricGod said:
Now we all know why you don't ever test anything...you can't test...plain and simple. You have no choice but to take what you read on faith. And that's perfectly OK...just say so. Why was that so hard? There was quite literally nothing to hide except for the fact that you can't prove anything you say with actual testing. I on the other hand can test and can back up what I say with real live testing. Good god man! What a waste of time posturing and pretending when honesty was the best approach!

Seems like you're more interested in conjecture than technical discussion.

I'll comment on this one thing...
Yup...that's probably what I intended. Why did I bother?

OK great so you have test equipment and won't use it...like I said...LAME! And yeah you take things on faith! You said as much. Theory is great, facts however based on real tests on the other hand are real. I still stand by my test results on AOT290's and you are still trusting in your faith and not on actual facts from testing. LAME LAME LAME!!!
 
Addy said:
ElectricGod said:
Theory is great, facts however based on real tests on the other hand are real. I still stand by my test results on AOT290's

I'd rather trust theory from reputable sources, than the postings of some guy on the internet who doesn't understand the limitations of his testing.

Fortunately for me, I'm not worried if you trust my tests or not. I trust them. Beyond that, if I help others, great. If people get in the way of that, than shame on them.

Since I'm such a numb skull that doesn't know his butt crack from a fence post, how about you show me how to do it right? Set up all that fancy equipment you say you have and do the tests yourself. OH wait...that's right...that will never happen. It's soooo much better to disrespect someone who is actually doing something than to do the same work yourself. If you tested too...with all that engineering experience and know how and equipment, then you would have nothing to say except to let the chips fall as they may. Right or wrong, good or bad. Just like me! So go ahead and be critical and condescending. I really don't care. I'm actually doing something. I back my words with actions. What do you have...ummm give me a minute...oh that's right...nothing but theories and no actual proof.

Get off your lazy condescending butt and actually put your actions where your arrogance is! TEST!!!
 
ElectricGod said:
Fortunately for me, I'm not worried if you trust my tests or not. I trust them.

You are publishing your test results in an open forum, writing statements that AOT290 FET's have big abnormalities.
But you don't know how to measure them properly (at least at the beginning there was the thing with the to low gate voltage), and you do not understand BASIC things about measuring resistance in the milli ohm range.
Honestly, instead of discussing endlessly about same things, you should be thankful for the help you get here.

As mentioned i own a 4-wire meter and this week i received the AOT290 FET's for repairing my 12F PV controller.
I will measure the Rds on from all of them with 10V on the gate and post the results :wink:
 
Today i repaird the defect 12F AOT290 controller.

There was a short between battery minus and the green phase (thus the low side Fet's), but the QUESTION is why this happened with only 48V battery and a small 9C DD hub motor?

From what i found out the reason could have been a poor heat transfer, melted insulation washers and a short over the bolts and heatsink.
check out the pics:

There was an aluminum chip near the bolt hole which led eventual to a gap and poorer heat transfer. This was behind the FET which died (on the far right where the green phase wire is soldered).

9Yw0EC.jpg


some of the insulation washers i removed (most of them also from the FET's in this area).
As you can see they probabaly got to hot and melted which led to a short through the bolts.

Ksydlw.jpg


measuring Rds on with 10V on the Gate:

I purchased the AOT290 from digikey, and they all had about 2,56mOhm (+- a few percent which should be in tolerance).

first one:
CYMg07.jpg


second one:
50sqmp.jpg


the one i unsoldered from the controller had much higher resistance (but still within specs of <3,5mOhm). Note that it looks differnt:

BzL1yG.jpg


on the left the one from digikey:

Gh3cE5.jpg


The meter is really accurate. The shunt of that controller is declared to have 2.00mOhm and this meter measured exactly the same.

As summery:

Vadim, or the producer of powervelocitiy controllers should absolutely do better quality control.
Aluminum chips leading to a gap between the heatsinks, messy attachment of the kapton tape, insulation washers which melt and causing a short, and loose BT modules flying around inside....
MAN, maybe i had bad luck or just got "monday" models, but in order to avoid these messy-flaws it would really not need much :roll:

regarding the FET's:
It is a fact that there are different batches around from the AOT290.
It makes a huge difference if you have ones with 3,4mOhm or 2,56mOhm in your controller. Latter means 25% less resistive losses or about 15% higher current!
 
Madin88...

I have no issues with you. With Addy...that's a different story.

Thank you for actually testing and presenting your results. This is exactly what I have asked for and NOT gotten from anyone else...just excuses and criticism.

In your resistance test, what voltage are you feeding the gate? I would be curious to find out why the digikey mosfets are under Rds spec. I'm not complaining...less resistance is a good thing. And why the one from the controller was dead on typical spec. Variations do occur and if you had three mosfets and one was significantly different than the other 2, then it's also possible for that difference to swing the other way as well. Like I have said in the past.

Mosfet variations...
I have AOT290's from 2 purchases from digikey and they look different from each other and then I have some on a few BMS and in in these controllers and there's always the "T290" on them, but the fonts vary or other details. Sometime the package is a little different. Hopefully they are all legit components, but maybe not?

Quality control...
Vadym gets the controllers from China prebuilt. He then adds a programming/telemetry module, CA cables and a few other things depending on customer requests and tries them out briefly. He has specified to the manufacturer because of me to use AOT290's instead of IRF4110's like they originally had in them. I too had a PV controller...brand new that under no load, just spinning a motor popped 2 mosfets in the same phase. When I took the controller apart, I didn't find anything like you did. All I had was 2 bad mosfets. After I replaced them, the controller has worked fine ever since. I do have to agree, damaged insulators are going to create problems. I had a damaged kapton insulator in a Grintech controller. It ran for 20 minutes or so under no load and then killed 3 mosfets. I found 3 spots in the insulator where there was no insulation between the mosfets and the heat spreader. It happens, but is not the norm...same for bad or less than optimal mosfet runs.

kapton%20insulator_zpsbjl5k7mj.jpg
 
Madin88...

Once you are happy that the controller is running well make some changes to it.

1. These controllers can handle down to .001 ohms on the shunts. I don't know what value the shunts are from the factory, but I know that they limit the wattage of the controller to about 2500 watts. On any of my controllers, I use precision .005 ohm shunts and then put 4 of them in parallel. That keeps me a smidge above that .001 ohm lower limit so the MCU can still read current via voltage drop, but dramatically increases the current flow through the shunts. Without this one change I typically get 40-45 amps at my watt meter. Changing to 4 parallel shunts I see 60 amps plus. The AOT290 can handle the increased wattage and amps so I'm not worried about cooking them. I run my 12 fets at 4kw continuous and they "warm up" to ambient air temperature. I put 2 shunts in the through holes and 2 more on the underside of the board.

2. Upgrade the phase and battery wires. I use a 1/2" long section of solid core copper wire that is the largest that I can cram in the through holes. I think that's 12 awg. I crimp and solder that to 10 awg silicon wires and heat shrink all of it. The end of the copper wire gets soldered into the controller.

3. Add an LED each to P2 and P3. Now you have status LEDs for the controller.

4. Add an additional heat spreader inside the shell. This made a big difference in dealing with heat when running at 4kw continuous. I do this to my 12 and 18 fet controllers.

Heat%20spreader%202_zpsm8emddvg.jpg

Heat%20spreader%203_zpslvemvgwy.jpg


5. I reflow all the solder on any power traces. On my 18 fet controller, the solder laid down on the power traces was not done with sufficient heat and so it was clumpy and dull looking. I also look for places where the added copper reinforcing may not be continuous and patch that too. I found that on the same 18 fet controller. This is the main solder traces from the factory. I don't blame Vadym for this. None of this was his doing and frankly it would have been fine as is. THe lower left corner of the large outer trace had a gap between the 2 pieces of copper. I added another section of copper here to bridge it.

Cold%20solder%20work_zpsknshdpf0.jpg


6. I add a thin piece of fiberglass sheet between the bottom of the board and the shell. It's probably not needed, but there's no way I'm going to short against the inside of the shell with it in place.


The quality of these controllers has in fact improved. As I find things or make changes, quite often those things make it back to the factory and then after a while we see them in new controllers. The much improved soldering, power busses and shells are examples of that.

This is the board in the original 12 fet controller. I have 2 of these. It has an early protoype module in it in the blue heat shrink. I bet I built this one.

Bottom%20of%20board%2010%20awg%20wires_zpsicuvz3f2.jpg

top%20of%20the%20board%20-%20original%20wiring%202_zpsh5ukri9r.jpg


This is the new generation of these controllers. I currently have this in a 12 fet and 24 fet. The 18 fet version will be arriving soon. The power busses are moved and all copper now. That's the biggest change, but the boards are a lot cleaner and less crowded now too. It arrived to me like this. I was impressed! They changed solder formulas too. Notice how much shinier the solder is here compared to above. That usually means more silver in the solder.

New%20board%20bottom%20with%20better%20parts%20spacing.jpg

Copper%20buss%20closeup.jpg


The shells have gotten thicker too. This is the original 12 fet shell. ignore the added heat spreader.

Heat%20spreader%203_zpslvemvgwy.jpg


And the new shell. All the way around the walls are thicker. It's difficult to see the difference between these 2 pictures since the first one is much closer in than the second, but the walls are about 25% thicker now. So far this new 12 fet has never had an added heat spreader and runs 4kw just fine.

12%20fet%20new%20shell.jpg


The 24 fet controller implements similar things. Greatly cleaned up board layout and solid copper power rails. The shells on these things is impressive.

Board%20bottom.jpg

batt%20-%20power%20buss.jpg

Shell.jpg
 
I've been trying to find shells like that for my own controller. Couldn't seem to find anything that nice on ebay, banggood, digikey or mcmaster-carr. I have some leads to look through on aliexpress still, but yeah I would really like to know if these kind of cases can be sourced online.
 
ElectricGod said:
In your resistance test, what voltage are you feeding the gate? I would be curious to find out why the digikey mosfets are under Rds spec. I'm not complaining...less resistance is a good thing. And why the one from the controller was dead on typical spec. Variations do occur and if you had three mosfets and one was significantly different than the other 2, then it's also possible for that difference to swing the other way as well. Like I have said in the past.

I fed the gate with 10V
The one i unsoldered from the controller had no short. It was working properly (at least as far as i could tell with my measurment equipment).

Once you are happy that the controller is running well make some changes to it.

I guess this would only make sense regarding the first versions as far as i could see?
My 12F controllers have already the larger case and larger wires. But a smaller case with improved heat path would be nice but it's not an issue...

Vadim, PM sent.
 
Status
Not open for further replies.
Back
Top