Questions about populating a homemade driver board

Hello,

I’m trying to replicate a certain motor driver that someone on Make Magazine made:

If you follow that link you can click on ‘7. Drill holes in the PCB’, there’s 3 pictures that you can look at…or, I posted them below.

If it’s relevant, a very brief background summary:
Ultimately, I plan on doing the same thing that he is - control 2 independent wheelchair motors with (2) 12V wheelchair batteries (in series), so at 24V…via Arduino of course. :wink:

Anyway, I’m pretty much stuck at the 2nd photo - the transistors. I don’t know which transistors go where and how they’re oriented. I think all the heatsinks for his transistors make it hard to tell.

His BoM called for the following transistors:
(12) N-Channel MOSFET 52A 60V - Digikey Part#FQP50N06L
(12) P-Channel MOSFET 47A 60V - Digikey Part#FQP47P06
(4) N-Channel MOSFET 60V 200mA - Digikey Part#2N7000FS

Photos from the Make Magazine article/tutorial:

Photo #1:

Photo #2:

Photo #3:

My photos from my board that I’m making:

Photo #1:

Photo #2:

Photo #3:

Photo #4:

Photo #5:

His schematic:


Note - since this is showing up so small, I also ‘attached’ it to this post.

I’ll break down my big/vague/general question into 5 individual, focused questions about populating this board:

Question #1 (reference ‘my’ photo #2) - Do I have the transistors in the correct orientation? (Note - I came up with the orientation that I have it in just by referencing his photos…I did not come up with what I did from the schematic)
In ‘my’ photo #1, I broke down the board into 8 rows to make things a little easier to reference. I just put 1 transistor in each row to show which way I think how the transistors in that row should be oriented. (Don’t worry, I didn’t solder those yet)

Question #2 (reference ‘my’ photo #2) - Of the big transistors, which ones go where? I’m trying to figure out the schematic, but just can’t decipher/translate it into hooking them all up onto the board. There’s (12) N-Channel (52A 60V) and (12) P-Channel (47A 60V), but have no idea where they all go. Do they alternate by row? Do they go in ‘quadrants’ (1 kind in Rows 1 and 2, the other kind in Rows 3 and 4, 1st kind in 5 and 6, etc.)?

Question #3 (reference ‘my’ photo #3) - Do I have these little transistors oriented correctly? Again, this is how I have them from referencing his photos. It’s kind of hard to tell for sure if I have all 4 correct.
(the blue circles will be part of my next question)

Question #4 (reference ‘my’ photo #3) - I think I drilled 4 total extra holes on the board (where the small transistors go). So not only am I not sure about the orientation (see previous question), but I’m not sure about the placement. If you look at ‘my’ photo #2, you can see that there are 4 holes there for the 4 different little transistors instead of 3. Sooo, in photo #3 I circled in blue the un-used holes that I have next to each little transistor. Basically I used the 3 holes that are closest to the ‘inside’ of the board for each small transistor. But, I don’t know if it’s the correct 3 holes for each of the 4 small transistor locations.

Question #5 (reference ‘my’ photo #5) - If you look at his photos #1 and #2, it looks like he’s got some (2 wires in each location) bare-metal ‘jumper’ wires near each small transistor. If you look at ‘my’ photo #5 I circled in blue the places where I believe they go. Am I correct about this? Is that just bare wire acting as jumpers?

I would be leaving a comment on his tutorial for all these questions, but he hasn’t responded in a long time now, and everywhere I look for him it looks like he’s MIA. So I’m turning to the Arduino community. I’m actually going to be controlling this motor driver board with an Arduino ultimately anyway. :wink:

I'm no expert but the P channel MOSFETs on that schematic appear to be drawn upside down, their body diodes would be a short circuit.

Hello,

Thank you for the reply outsider.

Anyone else want to chime in?

The schematic does indeed show the p-fets upside down.

I have a hard time making sense of the pictures but should be pretty straight forward. The source pins for all N-channel mosfets should all be attached to ground.

The source pins for all the p-channel fets should all be attached to Positive.

If you just try and build from pictures it will frustrate you more than reading and understanding the schematic. Just beep out the paths with a multimeter from power or ground to each mosfet and make sure they are each going the right way.

That circuit can't work at 24V, its a 12V only circuit. What's more it has shoot-through so it'll be
generating big current spikes direct from the battery every time it switches. This may be enough to
pop your MOSFETs (especially with PWM) unless the power supply is current limited.

Also the PSB traces probably aren't really up to handling the current level - standard PCBs have
copper only 0.035mm thick. You should be considering bolstering the current handling by soldering
thick tinned copper wire along the high current traces so they don't act as fuses!

And you'll need a suitable fuse of course, if using lead-acid or lithium batteries. Such power sources
can generate enough current if shorted to incinerate your wiring and damage themselves.

For 24V volt you should be thinking "this is the 21st century, one uses MOSFET driver chips to
drive MOSFETs", and add the datasheets of chips like the HIP4081A to your reading list.

MarkT:
What's more it has shoot-through so it'll be generating big current spikes direct from the battery every time it switches.

Could you please explain. I don't see shoot-through if only one input is PWM-ed and the other input stays LOW.

I agree with using a mosfet driver.
Could also be unhealthy for an Arduino pin to PWM ~5nF gate capacitance through a ~15ohm current limiting resistor.
Leo..

This circuit does not shoot through, The low gate is tied to the high gate on the other side.

It might actually work at 24v as well. The N-channel FETs seem like they will turn on from 5 volts (barely) and the p-channel FET actually has a gate tolerance of 25v. Its too close for comfort though and 12v is more realistic.

alka:
This circuit does not shoot through, The low gate is tied to the high gate on the other side.

Yes it does. Bare MOSFETS turn on faster than they turn off. The high and low will be on at the same time, during the moment of switching. Thousands of times per second. That's why MOSFET driver chips have diodes and other clever circuitry to manage this for you.

Both of the PCB designs shown in the photos look difficult. The MOSFETs aren't lined up so they can be bolted to a common heatsink. Assembling lots of tiny heatsinks and making sure they don't touch each other or anything else is going to be a hassle. All the weight of those heatsinks falls on the component leads. They will eventually snap off the MOSFETS if there's any vibration. When I'm laying out a PCB like that, the first consideration is getting all of the big power transistors into a line so they will bolt up to one heatsink. (With electrical isolation washers.) The flow of electricity is less important than the flow of heat.

Or I would use a specialist motor driver like the VNH2SP30. That can run some seriously big motors with no heatsink.

MorganS:
Yes it does. Bare MOSFETS turn on faster than they turn off. The high and low will be on at the same time, during the moment of switching.

No shoot through if you only switch/PWM the PWM_1-A input, and allow some dead-time before you break/reverse with the PWM_1B input.
One leg of the H-bridge is just a P-channel fet, the motor, and the N-channel fet in series.
Not important how fast/slow/different they switch.
Leo..

MorganS:
Yes it does. Bare MOSFETS turn on faster than they turn off. The high and low will be on at the same time, during the moment of switching. Thousands of times per second. That’s why MOSFET driver chips have diodes and other clever circuitry to manage this for you.

This type of circuit will only cause shoot if you change direction without a dead time as wawa said. In normal operation the upper fets on one side are kept fully off while the bottom fets on that side are switched at the same time as the upper fets on the opposite side of the bridge.

No shoot through is even possible as long as one input is held low.

alka:
This circuit does not shoot through, The low gate is tied to the high gate on the other side.

Ah, yes you’re right - but there is no protection against shoot through either…

It might actually work at 24v as well. The N-channel FETs seem like they will turn on from 5 volts (barely) and the p-channel FET actually has a gate tolerance of 25v. Its too close for comfort though and 12v is more realistic.

The p-channel gates would need zeners to make that work, and you’d have to be careful about heat generation in them and the dropper-resistors. But no-one does things like that today, standard MOSFET driver chips
are how its done, the right tool for the job.

Wow, thank you everyone for the replies. Unfortunately, I'm even more confused than before though. It sounds like some are saying this is a bad circuit design, others are saying it's going to work?

I'm not sure what I'm going to do at this point...

Motor control is almost a solved problem. There's lots of chips available out there which will do what your PCB does, with more additional features than you can think of. More safety, more feedback, more reliability...

Start by looking at what's available in Arduino shields and other modules that are easy to wire up. If you don't find anything you like, then look at what chip the next-best one uses and use that same chip in your own design.

Consider something like this Pololu driver: 18V (recommended, 24V possible) at 25 amperes continuous, without a heat sink or fan.

@MorganS - I don't suppose you have a couple part/model#s for examples of some chips you're referring to?

@jremington - I like that option, but it says use of 24V is not recommended (I'm going to be using 2 12V wheelchair batteries in series, so I guess I can't really use that).

But I also found a couple boards from Sabretooth:

Sabertooth 2X12 regenerative dual motor driver ($80, drives 2 motors, up to 12A each)

and

Sabertooth 2X25 V2 regenerative dual motor driver ($125, drives 2 motors, up to 25A each)

I'm still trying to figure out which one I need. I'd rather go with the $80 one since it's cheaper, but not sure if my motors need more than 12A.

It sounds like it would be a lot easier if I just use one of those boards, but they're so expensive. That makes me more curious about the chips that @MorganS was referring to.

My favourite is the VNH2SP30. You can get two of them on an Arduino shield from Sparkfun. It's called the Monster Moto Shield, or something like that.

Thanks for the link MorganS. It looks great, but just one problem - it's maximum voltage is 16V. I do appreciate the info though. :slight_smile: