Oh no... 50ns... how ever will you cope...
*looks at his 1000ns rise times and cries*
Printable View
Oh no... 50ns... how ever will you cope...
*looks at his 1000ns rise times and cries*
Lol, the majority of shiity Hobbyking-breed ESCs are like 5us, so you're orders of magnitude better than the average!
Ellis, I have a secret for you. The semi indestructable Tz85's, are basic "el cheapo Hobbyking ESC's".
Logic boards are done and shipped on Wednesday so should be here sometime next week.
I have had very little time to work on this the last few weeks with 2 weddings! So should make some better progress this week.
My plan is to have it running a small motor by the end of the week and gradually increase the size until I get it to run the 150cc rotomax.
We have also decided to redesign Triforce slightly and will be using the 50cc rotomax for our drive. This means Triforce will be running 4 of these speed controllers when we apply for next series! So should be a good test for them if we get on :D
Turns out I had another error in my spreadsheet so in actual fact this controller should handle 450A 16s just fine which puts it peak output at over 30kW lol. Talk about overkill. :D
Should be good for my Neumotors then :D
It seems I want to try one.
Managed to get round to working on it some more tonight. I found out I don't have any wire thin enough to solder onto the tiny pads on the vesc so I just wrote some simple bldc commutation code for arduino to see if it worked.
One thing I figured out is that I need a bigger powersupply lol.
https://youtu.be/gvgNHs0L0WM
With a slow ramp up it does get reasonably fast.
https://youtu.be/HUAgkX2r-6g
I also settled on calling it the A450S
Logic boards arrived. 40 is quite a lot...
http://imgur.com/h7CusDG.jpg
I have now made up one of the logic boards and started testing it out. Should hopefully get it connected to the power board tonight!
As you might have guessed this is going to be a two part ESC, but with a difference. The logic board is actually going to be on a longish cable ~30cm so that you can keep all your logic boards, receiver, etc in the same place, reducing cable clutter. It also makes the power boards smaller and therefore it is easier to fit the controller into a bot.
They are connected by a microfit cable which I can supply in 150mm or 300mm.
They look like this: Attachment 7143As they are locking they should stay on through big hits just fine.
That sounds like a bad idea. Usually with high speed signals you want everything nice and close, not wrapped around all the noise sources you get in a robot.
It's not really high speed... I will be testing it out before to see how bad the noise is. If needed I can add a shield to the cable which would remove the problem.
Going to use some shielded cat6 cable instead since they connectors are smaller and include a nice shield for each twisted pair.
Managed to figure out a problem that was driving me crazy. As it was only outputting high side pwm. Until I found a little option that was in the GUI... fml.
Made a new version of bldc tool with support for my silly currents. (Old version only lets you set 200A max)
https://i.imgur.com/paX037G.png
With some luck I should be able to get a test board fully working by the end of the week!
I also decided to just put the logic board stuff onto the power board since it was proving too awkward connecting them. So will need to design a new board for all that. Which I was planning on doing anyway.
So it turned out that it was not that option after all, since after turning it on it still didn't work!
Turned out that I kinda forgot to connect the pins on my board... oops!
https://i.imgur.com/SppBlaX.jpg
Here is the bodge I had to do to fix it.
https://i.imgur.com/vk0F4kf.jpg
After fixing that mishap I plugged everything in and hozza it moves the motor! Although more jumps back and forward would be a better way to describe it. This is because I haven't yet connected the phase voltage sense wires and current sense, once I add that tomorrow it will hopefully have a better chance of completing a circle!
Have you looked at vesc tool rather than bldc tool. The new software runs much better for me. Www.vesc-project.com
Also you might want to look at testing off a battery so that regen doesn't blow up your psu.
Yeah I am just looking at it now, didn't actually notice it was finally out. Should hopefully improve performance on FOC with the new firmware. It also supports 3 current shunts which is a welcome addition.
I don't really want to test from a battery yet as any shorts will destroy the whole board. I just turn off regen in the settings.
Have a bunch of 70A+ power supplies for testing before I move on to using a battery.
Large bag of 8mm bullet connectors arrived today!
https://i.imgur.com/A1hKMKc.jpg
Spent some time putting proper wires to the board with some connectors to make my life a bit easier. Then put on the shunt resistors and the.. oh the current shunt amplifiers are the wrong size... fail. So I have ordered the correct ones, which don't get here till monday :( So in the meantime I am going to just look more closely at the pwm outputs and phase voltage sampling.
Over the weekend I will layout the hopefully final board for the A200S version (3 Fets per phase leg(18 total), 1 driver per 6 fets (3 total)) which is going to be the first board. I don't personally have a use case for the A450S yet and with a new series of RW coming up I need to get the lower current one working first so people can actually have something in hand.
https://i.imgur.com/KZCqzVC.jpg
Tonight I have been doing the layout for the new board for the A200S version. I have finished component placement and just moving on to routing which should be done by the end of the weekend.
The new version is a massive improvement on what I had before, with lots of improvements from my testing of the first set of boards.
https://i.imgur.com/B99MDu0.png
https://i.imgur.com/dLjnEwE.png
Is that one driver per 6 FETs as in 3 high, 3 low, or 6 high, 6 low? If it's the latter your switching times may be sad..
Yeah it saves a lot of board space too. With the 22r gate resistors it works out at 1.6A for a set of 3 so well within the 4amp peak on the drivers.
The VEX Victor BBs have 3 FETs running off a 1A driver and have had them running at 300A, so you should be grand. I see it's only single sided FETs; if you move the gate resistors to the same side as the FETs you'd be able to get heatsinking on both sides of the board...
There is going to be copper bars on that side anyway so its not a problem as the heatsink has a 2mm thick thermal pad anyway.
Decided to finish off the pcb layout instead, so here is the completed board! The exposed parts will have 3mm copper on them. The cutout is for capacitors. With the whole board sandwiched between a set of aluminium plates.
https://i.imgur.com/dyZLput.png
https://i.imgur.com/Rgp4O6W.png
New pcb is now with the fab after I got the schematics checked over by some other people. Thanks to them!
Ordered the parts that changed for the new version, mostly capacitors and more TVS diodes for ESD protection.
Also getting the case prototype made by protolabs so that should arrive soon.
Copper bus bars were cut at the same time as my order for Triforce and look pretty good.
Just waiting on stuff arriving to test it out!
All parts have now arrived. The boards were dispatched on Monday so should be here sometime next week.
I got the bottom half of a case done by protolabs and the copper bus bars were cut by kcut.
https://i.imgur.com/MUIHEN8.jpg
https://i.imgur.com/pQOoscn.jpg
Been working on this slowly over the last few months to sort out the problems. I have just sent off to get the next revision of the board manufactured.
https://i.imgur.com/owmLdQU.png
https://i.imgur.com/EncqR8c.png
That's a LOT of vias! What sort of issues have you run in to?
Yeah, I doubt the PCB manufacturer likes me...
When I originally designed it I was working from the VESC4 schematic, so the VESC6 software didn't really work properly. So this version follows the VESC6 schematic so that the software will work properly.
Also added lots of ESD protection and other protection devices, changed the board shape a bit and changed all the values for bootstrap capacitors and gate resistors.
I have a few weeks off around Christmas so I will be testing this board then.
New boards have arrived and been soldered up, testing to follow.
https://i.imgur.com/1X1NUZP.png
https://i.imgur.com/DGKNNGw.png
https://i.imgur.com/aU5UCH3.png
I spent a lot of time testing over the last few weeks. It is now possible to run a motor using FOC.
Sending a prototype unit to Rory Mangles to see what he thinks.
Currently I have tested:
100000+ ERPM - Seems fine, tested at 105000ERPM, will try testing with a higher Kv motor and see what the limit is.
100A continuous - Seems fine, motor got too hot and started smelling. Will continue with larger motor in tank of water.
Motor Tracking (Coasting motor) - Working up to 4000ERPM, working to fix at the moment.
BLDC - Not working, probably related to motor tracking issue.
FOC - Works fine, but detection does not work (Motor tracking). So you have to enter motor parameters manually from the detection.
Still to test:
PPM input
150A continuous
200A continuous (3min)
High Enertia load
CANBus
Hall Sensors
Encoder
Once I get motor tracking working I will be gearing up for a first batch.
I designed a new heatsink which can be waterjet, hopefully keeping the price down. The design allows multiple controllers to be stacked to allow them to share heatsinks. The plates are 10mm aluminium so you can just drill and tap holes wherever you like to mount them.
https://i.imgur.com/MEoSXYU.png
Here is a video showing it running. (Excuse my brain fart on the ERPM graph)
https://www.youtube.com/watch?v=VwPTPv4K_v0
The heatsinks also arrived and look great, thanks again KCut! So have now tested at 190A (Max my powersupply can do) so far it manages to run for 25s before hitting the temperature limit and throttling down to 135A continuous. Although this is clearly because the thermal transfer between the board and heat sink is not good enough. So I have some better thermal pads on the way to try out.
I also have some temperature sensors and hall sensors turning up soon to test out throttle on motor temperature and sensored mode.
Started producing the first batch of 10 boards. Don't have enough FETs to populate them all but it at least gives us spares.
https://i.imgur.com/nZ3hp5b.jpg
https://i.imgur.com/sfhPviD.jpg
https://www.youtube.com/watch?v=tvLkA1hXbYI
That looks a piece of expensive kit.
But I have use for it. Remind me, whats the max voltage of these monsters?
Pity, as we have a 18S 60Ah pack for the motorbike.
But for transients, isn't that the job of the capacitators and coils to smooth those out?
You might manage at 18s if you stick a load more bus caps on it, but it will be pretty marginal. If I remember right the simulation only showed 1-2V of ringing with 6 of those big caps. I can probably do a special for higher voltage with different FETs, drivers and shunt amplifiers if you really want it. Should be just a small layout change for the shunts. For robot combat 80V seemed a sensible max since robot wars rules have 75V maximum.
What current is your motorbike estimated to use?
Actually looking at the parts, I already was lined up to use 100V FETs, the gate drivers handle 120V boot, the linear regulators are 100v. So its just the current shunt amps that are limiting it (they say 90V maximum so they might work), which could probably be fixed by dividing the voltage inputs to them. I will add some resistors and solder jumpers to add that for higher voltage use.
We're to use a 18S 60Ah pack. This should be limited to 600A peak, but the motor we use shouldn't draw more than 200A in "insane" mode (a 250A fuse will be installed on the main powerline).
Normal streetuse would be very mild with 70A , as that is the legal limit on the machine we're building.
Need photos of the bike :D