Starting with the PCB or code doesn't matter worth a darn. The PCB can be designed with all the features in mind, if the code isn't be written to support all the hardware features..that's not at all a problem. When a team works on a project, the hardware guys don't wait around for the software to be finished and that method has worked well for ages.
Not simple BLDC controller It RUNS! :)
- Thread starter Arlo1
- Start date
miniebiker
100 mW
- Joined
- Aug 5, 2011
- Messages
- 35
bigmoose said:
Is that what goes on here ?
So many times I click on something and it has been removed.
I feel blessed when I click on a picture and it shows !
bigmoose
1 MW
Naw... the pictures were corrupted in either a server hard drive crash or a migration about 4 or 6 months ago...That killed about 1/3 of the older pix.
Then just last week, when the board changed servers again, with the new ownership, we got "hit" again. There seems to be a programming error in the phpBB software (what the forum runs) backup function. The board computer guru MRVass is busting his tail trying to recover from the big "tarball" that is the current backup. We are hoping he can sort the bits out...
I'm just being a crab-_ss ol' guy today. Kids are home from college, wanting me to fix things, taking "things" for granted, and the ol' guy is more than a bit tired, having driven 20 hours straight to make my brother in law's funeral yesterday... Steve was a great guy, I am going to miss him a bunch... he played a good hand of ol' time poker, straight and 3 card draw. I lost another good poker partner, and I don't have that many left...
Then just last week, when the board changed servers again, with the new ownership, we got "hit" again. There seems to be a programming error in the phpBB software (what the forum runs) backup function. The board computer guru MRVass is busting his tail trying to recover from the big "tarball" that is the current backup. We are hoping he can sort the bits out...
I'm just being a crab-_ss ol' guy today. Kids are home from college, wanting me to fix things, taking "things" for granted, and the ol' guy is more than a bit tired, having driven 20 hours straight to make my brother in law's funeral yesterday... Steve was a great guy, I am going to miss him a bunch... he played a good hand of ol' time poker, straight and 3 card draw. I lost another good poker partner, and I don't have that many left...
miniebiker
100 mW
- Joined
- Aug 5, 2011
- Messages
- 35
Ouch. Sorry to hear that. Hang in there.
Steve
Steve
bigmoose
1 MW
jdb identify the specific p channel you want to use, and all the peripheral parts to make it "outside world" safe. Identify the parts price per qty(100) from digikey (for example) for your proposal and the one on the schematic.
We are all open to see the results, both performance, cost and board real estate.
I'm going to work this the same way I have for 30 years... don't just comment, make your comment AND provide an alternative. Show that your alternative OUT PERFORMS the base line design, and it WILL BE the new baseline!
That is how I have run design reviews, and that is how we have evolved many designs...
We are all open to see the results, both performance, cost and board real estate.
I'm going to work this the same way I have for 30 years... don't just comment, make your comment AND provide an alternative. Show that your alternative OUT PERFORMS the base line design, and it WILL BE the new baseline!
That is how I have run design reviews, and that is how we have evolved many designs...
From a guy who saidLebowski said:
listen man this forum would be nowhere with out guys like bigmoose!!! We really need to watch all our tongues we keep track of ourselves. You wont find the usual pissing contest here that you would on a sport bike forum or one of the latest sports car forums.... I have to apologize for my first comment to you as well but I was offended with all the work Burtie did on his awesome and cheep board! You have some out of the box thinking and I love it, but you will find Dave to be one of the most educated/nicest people here so lets listen to him because it will save us all a life time of work to get where we need to go!!!Lebowski said:
bigmoose said:
Sorry, I wasn't clear. This is not something that I'm familiar with, and I was guessing at your rationale for preferring PNP over PMOS. I was not challenging your decision, I was trying to guess the answer to your open-ended question.
So, um, yeah. Was I warm or cold?
I love discussing design, and there seems to be less and less opportunities to do it...bigmoose said:1) I have a "standard" RS232/CAN/USB module that I interface so things are set up for what I like to do. By bringing the pins out the way I did, everyone can roll their own interface if they want/need it. You don't need the RS232 for typical "dumb" controller operation. You can add your opto's off board. Remember PCB real estate cost money! I don't like to put "everything" on the board. There are layered economic reasons for doing it this way in the commercial world.
Njay: Well, at least we shouldn't be calling it RS232, because that implies certain voltage levels and negative logic, and someone may end up connecting a real RS232 master on there which isn't healthy. When you do lots of projects with RS232 it makes sense to off-board this interface, but for a one-off as I suppose would be the target for this controller it is not the simplest solution - and I don't like "special cables", but anyways it's just a personal preference.
2) Note from the data sheet on the LM2931:in the TO-263 package, θJA is 73°C/W; If the TO-263 package is used, the thermal resistance can be reduced by increasing the P.C. board copper area thermally connected to the package: Using 0.5 square inches of copper area, θJA is 50°C/W; with 1 square inch of copper area, θJA is 37°C/W; and with 1.6 or more square inches of copper area, θJA is 32°C/
W.
I am sinking to 0.17 sq in of copper. Using your numbers, 0.219 W * 73 °C/W = 16 °C Not a problem! Besides, don't know if this matters, but I have used this regulator many times before...
3) Yes. Think it through to the next level, do the FMEA. This is the more reliable version with respect to shutdown, I believe.
I wasn't seeing it right, the culprit that I wasn't connecting to was "GND to inhibit". So yes, PNP is the choice.
4) Thanks! ... raaats,guess I left in some "special sauce" that's going to get me whacked by the "corporate" Chief Engineer!
Just make sure the PCB traces can handle the current until the user realizes the mistake
Keep 'em coming! Can't believe I used to hate design reviews 30 years ago! :wink:
Sorry about your loss. Get some rest.
bigmoose
1 MW
jdb, no problem with challenging me! I do it all the time, and my colleagues do it relentlessly. Early on, I thought I knew more than I know I know now... (don't know if that came out right.)
I have to tell a story, like all ol' salts do... Back in 1984 I was responsible to make the decision on whether to commit to a "mission" based on suspect spot welds in a LOx tank. The chief metallurgist was a good friend of mine and John and I poured over X-rays through the night, until 4 am or so I believe. The next day we had one of the most heated debates known in the review meeting. My boss thought we hated each other... but John and I had dinner together that night as we did many times until his death in 2004. I lost a good friend and the best metallurgist I have ever known. We just "sharpened" each other with our dialogue. I always called it the "search for truth..."
I have learned to listen. So never hesitate to speak... or to ask a question as you have. It is the responsibility of us ol' guys to pass on what we know to those younger, and willing to learn... always with the knowledge that we do still make a lot of mistakes!
I like FETs when they are big, for power; or when they are small in a protected environment, like in a DSP for low power dissipation. When you look at the little bit of gate charge needed to change state on a FET it troubles me in a safety critical application. Can we make it work? Sure. Let's source a P channel FET like a ZVP2110A -100V 230 mA, Gate capacitance of 100 pF and it has a Turn on delay time of 7nS and a rise time of 15 nS... can we say FAST! It's so fast, it's dangerous. We need to slow it down. We add a resistor from Source to Gate and then need to bypass that with a zener so we don't pull too far down. Add another resistor to our switch source from the gate. What about debouncing now with a 22 nS response time!
Now look at the 2n3906 -40V and 300mA 60 ish hfe 35nS delay time, 65 nS turn on time, 35 ns rise time = 135 nS response time. Still too fast for me, but better. I really wanted to add a little resistor and a cap to slow the base response down, but I didn't have the board space for it. That is one of the things I would test with the scope and the real world lay out. I might go back and try to squeeze in a filter to slow it down a bit on turn on now that you made me think about it. I really want to slow down the release more than the attack. The release is around 600 nS to let go and shut off. That is really the time I want to slow down to debounce the response. Should be down in the 10mS range for letgo.
Now for the economic analysis:
ZVP2110A $0.61 per 100
2 resistors 0.02 per 100
1N5243 0.17 per 100
PFet solution $0.80
2n3906 $0.07 per 100
2 resistors 0.02 per 100
PNP solution $0.09
Note that this is just my offering of a design. You are free to use parts of it to roll it the way you want it. You may want to put the MAX232 on your board... stuff like that.
So feel free to evolve it, CAD up your design and run with it. Hopefully this got things going... and folks thinking.
I have to tell a story, like all ol' salts do... Back in 1984 I was responsible to make the decision on whether to commit to a "mission" based on suspect spot welds in a LOx tank. The chief metallurgist was a good friend of mine and John and I poured over X-rays through the night, until 4 am or so I believe. The next day we had one of the most heated debates known in the review meeting. My boss thought we hated each other... but John and I had dinner together that night as we did many times until his death in 2004. I lost a good friend and the best metallurgist I have ever known. We just "sharpened" each other with our dialogue. I always called it the "search for truth..."
I have learned to listen. So never hesitate to speak... or to ask a question as you have. It is the responsibility of us ol' guys to pass on what we know to those younger, and willing to learn... always with the knowledge that we do still make a lot of mistakes!
I like FETs when they are big, for power; or when they are small in a protected environment, like in a DSP for low power dissipation. When you look at the little bit of gate charge needed to change state on a FET it troubles me in a safety critical application. Can we make it work? Sure. Let's source a P channel FET like a ZVP2110A -100V 230 mA, Gate capacitance of 100 pF and it has a Turn on delay time of 7nS and a rise time of 15 nS... can we say FAST! It's so fast, it's dangerous. We need to slow it down. We add a resistor from Source to Gate and then need to bypass that with a zener so we don't pull too far down. Add another resistor to our switch source from the gate. What about debouncing now with a 22 nS response time!
Now look at the 2n3906 -40V and 300mA 60 ish hfe 35nS delay time, 65 nS turn on time, 35 ns rise time = 135 nS response time. Still too fast for me, but better. I really wanted to add a little resistor and a cap to slow the base response down, but I didn't have the board space for it. That is one of the things I would test with the scope and the real world lay out. I might go back and try to squeeze in a filter to slow it down a bit on turn on now that you made me think about it. I really want to slow down the release more than the attack. The release is around 600 nS to let go and shut off. That is really the time I want to slow down to debounce the response. Should be down in the 10mS range for letgo.
Now for the economic analysis:
ZVP2110A $0.61 per 100
2 resistors 0.02 per 100
1N5243 0.17 per 100
PFet solution $0.80
2n3906 $0.07 per 100
2 resistors 0.02 per 100
PNP solution $0.09
Note that this is just my offering of a design. You are free to use parts of it to roll it the way you want it. You may want to put the MAX232 on your board... stuff like that.
So feel free to evolve it, CAD up your design and run with it. Hopefully this got things going... and folks thinking.
parabellum
1 MW
Every idea is a rewarded with initiative. Gust start!Lebowski said:
Edit: Just startSorry for your loss Dave. I have always tried to take it all in as much as I can. I hope Steve did to. This is one of my things when I go I will not be worried because I tried to get as much out of this life as I could! Life is not about how many breaths you take but how many things take your breath away!bigmoose said:
So on the power stage there is no reason we cant run more then three H bridges!!! What I mean is if I split colossus into two motors and run it as two separate motors with 9 teeth on one side as one motor and 9 on the other as the other motor this will give 2x the inductance with the same kv. So if there was 6 H bridges three for each section of the motor it would be like having 2 separate controllers running perfect timing. And because the resistance in each section of the motor is 2x and more importantly the inductance is 2x all the fets will have better control. I would rather do this then just add more fets in parallel!
Hey Bigmoose.bigmoose said:
This looks aweomse. I knoticed that you have the brake just run to the fault pin. Looks like it will not have any pins left for regen. What are the plans for AN0 and AN1? Im guessing phase current sensing?
Lebowski
10 MW
Arlo1 said:
I'm not sure. Splitting it in 2 like this will work, I have the same with my bike motor. However, if you want a certain Kv on one set of windings
it means you will need the same amount of turns (but on half the number of poles) as previously when it was not split in 2. Same amount of
turns means same inductance as before, I think your resistance will double though as you'll need to use thinner wire (half area ?) as original...
bigmoose
1 MW
Arlo1 said:
Arlo I am sort of "playing" with this design and "open source" sort of a test case of Luke's mentoring of me... therefore I am presenting in steps, without everything fully developed.
AN0 and AN1 likely have intentions for battery voltage monitoring and "some type" of current control if the user wants. The DtoA is there for a quick current comparator topology I have in mind also.
I accept this is a building block only in this design, and not the final product. The dsPIC30F3010 is a few pins short for a "full function" design with bells and whistles. Think of this as a raw, bare bones platform. Like Luke's "pinky" bike. This isn't the CrMo eMotorycycle killer.
So this is more like a bigMoose "doodle" drawing and not an ISO9000 certified D size drawing.
I wanted to get a "starter platform" (but a capable one) out there for folks to start fiddling with.Start reading up on the software, and start getting your thoughts together. If it's open source, someone else needs to lead the work in that area... We will evolve the board when it reaches limitations...
With respect to the regen, I am thinking along the lines of this not being everything from the initial release. It will be iterative. Better to get a low inductance, low resistance motor running, then work up to regen and all the rest. A step at a time. Sadly, I envision a long journey here...
Yes I planned to leave the bells and whistles off on the first try as well. Get the motor running with out wrecking stuff. I was thinking yesterday what this could use is a spin/learn mode where you spin the motor at a constant rpm in the direction you want it to spin with an out side source of something driving it. And the processor can learn the Hall to motor phase sequence! Instead of hooking all the wires up and trying to turn it then switching them till it works. I tried the second way two days ago and blew all the fets I just replaced in the controller I was testing with. I started my pic lessons and hope to spend a little time on them everyday till I have it understood enough to make it run.
Im not sure you understand what I meen.Lebowski said:
If you look at colossus it has 18 stator teeth there is 9 on top and 9 on the bottom the top and bottom are a copy of each other so when I would this motor I wound 8 turns a tooth and did one phase on one side of the motor then brought the wires out. The I did the same on the other side of the motor and connected them in parallel which gives the same 75kv as before and the same 8uH as before but now I can hook them up in series or parallel to help save controllers from hard starts. But what I just suggested is keep the top half run to one set of H bridges and the bottom have to another set of H bridges. The inductance will be 2x because when you hook 2 equal inductors up in parallel the inductance is 1/2! The resistance will do the same which is fine and somewhat good. But the inductance is what really matters!
Would using a full sine-wave controller reduce the high current spikes that are caused by these low inductance motors? I am thinking that the transient currents would be lower because the hard corners on the trapezoidal signal of the phrase current switching are smoothed.
And, once again, how hard could it be to copy out the infineon controller design and re-print the pcb with slightly larger spacing for the mosfets so you can use TO-247 package mosfets? I hope I'm not hopelessly behind in the controller development discussion...
And, once again, how hard could it be to copy out the infineon controller design and re-print the pcb with slightly larger spacing for the mosfets so you can use TO-247 package mosfets? I hope I'm not hopelessly behind in the controller development discussion...
J-aprilia-N
100 µW
why is so still on this topic
nieles
10 kW
well "stil" is the dutch word for quiet/silent
so i think he means why is is it so quiet on this topic
so i think he means why is is it so quiet on this topic
Oh sorry... Im just geting into the PICkit lessons. I am almost done the other e-bmx I built for my borther as well so once thats all done I will be back at this big time. I have to do a list of what I need to order to build up a couple bords in the next couple days as well and then I suspect I should be able to built this with in a couple weeks and start blowing shit up!