Simple BLDC controller

[pelle242]

When you have the time, it would be great if you could post the final (MK3 ?) version of the PCB layouts!

A couple of questions:

That you boost the driver signals to 12v is that being prudent or is it from experience? Would 5v be OK most likely?

Can you spell out the pinout of the 10 pin connector between the boards?

Also where did you get the the shunt resistors?

Thanks!

 

[texaspyro]

That you boost the driver signals to 12v is that being prudent or is it from experience? Would 5v be OK most likely?

There are high power MOSFETs with logic level (5V) gate drive, but logic level fets are more expensive and more fragile than standard MOSFETs that use a 10V drive level. They also tend to have a higher gate capacitance and on-resistance which makes them less efficient.

 

[pelle242]

That you boost the driver signals to 12v is that being prudent or is it from experience? Would 5v be OK most likely?

There are high power MOSFETs with logic level (5V) gate drive, but logic level fets are more expensive and more fragile than standard MOSFETs that use a 10V drive level. They also tend to have a higher gate capacitance and on-resistance which makes them less efficient.

Yes but it is the signals to the drivers that are boosted. So its TTL->Discrete transistor drivers to 12v->FET driver chip->FET. At least that is how I understand the schematics...

 

[Jeremy Harris]

When you have the time, it would be great if you could post the final (MK3 ?) version of the PCB layouts!

A couple of questions:

That you boost the driver signals to 12v is that being prudent or is it from experience? Would 5v be OK most likely?

Can you spell out the pinout of the 10 pin connector between the boards?

Also where did you get the the shunt resistors?

Thanks!

The drive signals from the MC33033 are 12V, not TTL, so I'm just using them as they are. The extra three transistors are just inverting the high side drives to get them the right way around, because the MC33033 has them in the opposite sense to the low side drives, for some reason. The NCP5181 accepts 12V input signals OK, although it does have TTL-like switching thresholds.

The attached pics show the layout and the connection details for the two boards, viewed from the component side.

The shunts are these: http://uk.farnell.com/welwyn/oar1-r005fi/resistor-1-0r005/dp/1200357 I'd guess that something with a similar spec should be available from most places, like Mouser, Digikey etc.





Jeremy

 

[pelle242]

Thanks!

I have this plan (long term) to use your power board with something uC based so I am just trying to understand the circuit.
And I am in europe to so farnell is perfect.

 

[grindz145]

Very clean!

 

[liveforphysics]

This controller does every feature I've ever wanted in a controller. I've created an all inclusive list below:

1. Running the motor.



Are you happy enough with this version that you would like me to CNC up a batch of boards and send some your way my friend?

 

[Arlo1]

This controller does every feature I've ever wanted in a controller. I've created an all inclusive list below:

1. Running the motor.



Are you happy enough with this version that you would like me to CNC up a batch of boards and send some your way my friend?

I got one more thing for the list
Running the motor and
Stoping the motor lol.

Send me some or well give a couple bords to a guy who will send them to me like I meen someone who can go got the post office without getting arested
I would be very interested in a populated bord from someone as well for I am getting buisy very buisy!

 

[gwhy!]

Running the motor and
Stoping the motor lol.

I think the next controller chip in the series (MC33035) has a motor brake funnily enough.

 

[Kingfish]

Digikey lists the MC33035 as obsolete. The Braking pin does not equal Regen.

Dynamic motor braking allows an additional margin of safety to be designed into the final product. Braking is accomplished by placing the Brake Input (Pin 23) in a high state. This causes the top drive outputs to turn off and the bottom drives to turn on, shorting the motor−generated back EMF. The brake input has unconditional priority over all other inputs. The internal 40 kΩ pull−up resistor simplifies interfacing with the system safety−switch by insuring brake activation if opened or disconnected. The commutation logic truth table is shown in Figure 20. A four input NOR gate is used to monitor the brake input and the inputs to the three top drive output transistors. Its purpose is to disable braking until the top drive outputs attain a high state. This helps to prevent simultaneous conduction of the the top and bottom power switches. In half wave motor drive applications, the top drive outputs are not required and are normally left disconnected. Under these conditions braking will still be accomplished since the NOR gate senses the base voltage to the top drive output transistors.

~KF :|

 

[gwhy!]

Digikey lists the MC33035 as obsolete. The Braking pin does not equal Regen.

Dynamic motor braking allows an additional margin of safety to be designed into the final product. Braking is accomplished by placing the Brake Input (Pin 23) in a high state. This causes the top drive outputs to turn off and the bottom drives to turn on, shorting the motor−generated back EMF. The brake input has unconditional priority over all other inputs. The internal 40 kΩ pull−up resistor simplifies interfacing with the system safety−switch by insuring brake activation if opened or disconnected. The commutation logic truth table is shown in Figure 20. A four input NOR gate is used to monitor the brake input and the inputs to the three top drive output transistors. Its purpose is to disable braking until the top drive outputs attain a high state. This helps to prevent simultaneous conduction of the the top and bottom power switches. In half wave motor drive applications, the top drive outputs are not required and are normally left disconnected. Under these conditions braking will still be accomplished since the NOR gate senses the base voltage to the top drive output transistors.

~KF :|

I know it doesnt do regen which is a shame , I think the dil version has become obsolete but the smd is still obtainable, as far as I can tell.

 

[liveforphysics]

I bought a 20 pack of them on ebay a few months back when Jeremy first started playing with them.

Check around. It wasn't a rare chip by any means.

 

[pelle242]

Just trying to think forward a bit...
Wouldnt it be a good idea to connect Vbatt+ to the NC pin on the connector?
That way you could have a 12v step down regulator on the control board without a extra battery cable and Vbatt+ can be monitored by the uC for low voltage cut off...

 

[texaspyro]

Vbatt+ can be monitored by the uC for low voltage cut off...

What uC? This controller is a wonderfully simple brain-dead/fully lobotomized hardware based controller with no microcontroller on it.

It would be nice to have the option of an on-board regulator for making the +12V from the battery pack when used with smaller systems.

 

[pelle242]

Vbatt+ can be monitored by the uC for low voltage cut off...

What uC? This controller is a wonderfully simple brain-dead/fully lobotomized hardware based controller with no microcontroller on it.

It would be nice to have the option of an on-board regulator for making the +12V from the battery pack when used with smaller systems.

The uC that I will put om MY control board Im a software guy, its not wonderful if it cant be reprogramed, recompiled and reassembled into a coffemaker...
I dont know zap about analog power electronics but thanks to this wonderfull design, I get to play to

 

[Jeremy Harris]

Are you happy enough with this version that you would like me to CNC up a batch of boards and send some your way my friend?

Thanks for the offer, my only concern is that there may be a problem under high load operation that doesn't show when the controller is running under light loads. I'd like to try and get some high current tests done before declaring that the design is frozen, but there's nothing stopping people taking a punt and trying one as-is, with no guarantee from me that it will work as I expect. I've not seen anything that makes me think that it won't work just as well at max load as it does at low load, the FET signals all look excellent, better than those on a XieChang by a fair bit, so the risk of it not working OK at high power is probably pretty low.

Jeremy

 

[Jeremy Harris]

Just trying to think forward a bit...
Wouldnt it be a good idea to connect Vbatt+ to the NC pin on the connector?
That way you could have a 12v step down regulator on the control board without a extra battery cable and Vbatt+ can be monitored by the uC for low voltage cut off...

Yes, that would work OK. I only used ten pads because I happened to have some ten way right angle Molex headers............

I don't need a 12V supply on the controller for my application, as I already have a 20A 12V DC DC converter that powers all the other 12V stuff, so it made sense (for me, at least) to just use that to power the controller. That way the "ignition" switch (which powers up the DC DC converter and 12V rail) will also power up the controller and the kill switch on the handlebars will just cut the 12V feed.

Jeremy

 

[liveforphysics]

Are you happy enough with this version that you would like me to CNC up a batch of boards and send some your way my friend?

Thanks for the offer, my only concern is that there may be a problem under high load operation that doesn't show when the controller is running under light loads. I'd like to try and get some high current tests done before declaring that the design is frozen, but there's nothing stopping people taking a punt and trying one as-is, with no guarantee from me that it will work as I expect. I've not seen anything that makes me think that it won't work just as well at max load as it does at low load, the FET signals all look excellent, better than those on a XieChang by a fair bit, so the risk of it not working OK at high power is probably pretty low.

Jeremy

I will see if I can create decent code to get them cooking, and if they suck, no biggy, I'm only out a little time and materials.

 

[Jeremy Harris]

This controller is a wonderfully simple brain-dead/fully lobotomized hardware based controller with no microcontroller on it.

Gotta love that description! Maybe I ought to rename this thread to the "Brain dead BLDC controller"...........

One really big virtue of a dumb hardware controller like this is the extremely fast response to an over-current problem. One of the things (apart from the simplicity) of this chip that made me choose it was the cycle-by-cycle current limit capability; it can cut back the PWM pulse width on the fly if the current increases beyond the set limit. This means it will potentially current limit within about 1uS or so, probably fast enough to protect the FETs from pretty much anything, up to and including a dead short on the phase wires.

A very fast uC can do this, but it needs a heck of a lot of speed, plus some slick code, to make it work OK. Ricky's pretty sophisticated design can do it OK, but it's not something that a run of the mill 20MHz 8 bit uC is going to handle well, unless something like the crude XieChang interrupt driven external limiter system is used.

Jeremy

 

[oldpiper]

Maybe I ought to rename this thread to the "Brain dead BLDC controller"...........

the BDBLDCC ?

That's a mouthful.

Cameron

 

[Jeremy Harris]

Maybe I ought to rename this thread to the "Brain dead BLDC controller"...........

the BDBLDCC ?

That's a mouthful.

Cameron

Maybe we'll just call it "lobo" for short ................

Jeremy

 

[liveforphysics]

lobo is perfect.

 

[Thud]

I am still trying to keep from gettting too excited untill we get some more testing, but I am failing horribly.
A LOBO to power my HOBO! perfect!
Jeremy,
If you have the artwork on AutoCad, I would be happy to make the Scaling & pdf conversions so any one wanting to attemp the Laser-jet/etching proccess can down load it & start printing....I admit my reasons are selfish...I will also post a drilling code for cnc boring of the board & even the engraving code if anyone else wants that also...anything I can do to make it as accessable as possible, count me "on board". (pun intended :lol: )

LFP shares my requirments for a motor controller...I have the controller chips in bound..& am starting to add the rest of the required parts to my shopping lists.

Also at the risk of sounding redundant, Thanks again for this highly informative thread. Your contributions here have immeasurable value.

 

[liveforphysics]

For folks needing the chips:

http://cgi.ebay.com/5pcs-DIP-IC-MC33035P-MC33035-/220662376719?pt=LH_DefaultDomain_0&hash=item336080a90f#ht_265wt_833

5pcs for $12.99 shipped.

 

[liveforphysics]

I would be happy to make the Scaling & pdf conversions so any one wanting to attemp the Laser-jet/etching proccess can down load it & start printing....I admit my reasons are selfish...I will also post a drilling code for cnc boring of the board & even the engraving code if anyone else wants that also...anything I can do to make it as accessable as possible, count me "on board". (pun intended :lol: )

Thud- If you make the g-code for my retarded ass, I've got a big stack of blank PCBs. Just laying down double-side tape, my machine's table could likely cut about 20-30 PCB's in a batch. I would be more than happy to cut/drill as many boards as the forum members need, and simply ship a big stack of them to somebody who wants to handle distribution/shipping to other folks, as I personally am not the right guy for doing things that involve responsibility and/or shipping.