Enraged wrote:what about running multiple small standard size, off the shelf heatsinks rather than one large custom on?
Actually i'm only considering standard sized heatsinks, custom ones are only cost effective if you need thousands of them - not there yet i think
The 0.9"x 0,9" is the smallest / cheapest usable heatsink i've found by far, so i just ordered a bunch of them. they're less than 1/10 the cost of the smaller "pin-fin" ones.
The latest layout was more of a design test / reality check-
i don't think I'll try to make it that way because there's too much risk of ground noise and other assembly issues as a result of the crowding.
and I'm not planning on running any batches of other 4-layer boards soon where i could just throw them into the same prototype run.
and i really rather get rid of that pot altogether.
It was useful to help figure out what the optimum size and arrangement might be.
In the end if i did such a module i would really need to add a few more parts on the board for robustness and i'd space things enough to allow a 2-layer only layout.
Size will be about 1" x 1.5" to 2" (if including screw terminals) x 1/2" thick (with heatsink)
Here's what i'm thinking - please comment on this, if enough people like it or show interest will get made, if there's no feedback i'll just drop it for now
1- "Pololu" compatibility - too restrictive and inconvenient for this one, so i will likely drop that in favor of a more optimum arrangement - especially considering the currents- (see below)
2- Orientation:
I'm thinking of going with a connector on one edge of the board with pins sticking out in line with the board surface, so the module installs vertically (perpendicular to the motherboard)
This has several benefits:
- much less board area used on the main board, which can be more compact;
- much better natural air convection through the heatsink and around parts on the opposite side;
- Mounting the screw terminal blocks (pluggable) directly on the module:
this i especially like as it makes the whole control system much more compact especially for 4 or more outputs.
In an enclosed box with a fan the box can be 3cm high and the space around the terminal blocks acts as exit for warm air (heatsink rotated 90 degrees)
3- Interface:
I really dislike having configuration dip switches, shorting blocks, jumpers, and especially potentiometers that need to be manually adjusted by the user.
its just too error prone and finicky especially for anyone not "expert" in electronics. One slip and you short pins with your screwdriver and smoke your board.
I would add a small micro or circuit with either SPI or I2C interface to configure it including a DAC for the reference voltage (max current level)
so it could be automatically set up correctly by the host processor.
I like the idea of making it "software compatible" with the first 3-5A 30V driver for serial configuration, if going with a SPI interface.
I2C is simpler electrically though just 2 wires.
One of the shortcomings of this stepper controller IC (still the best available) is the peak current reference has a relatively small range, 0.8 to 2V,
so if 2V = 8 Amps then 0.8V = 3.2A. If you want to be able to select a a lower peak current then you have to change the or mess with the current sense resistor value.
That's why on my test design i had put 4 sense resistors in parallel. So i have 3 choices on how to handle that:
1- follow KISS principle, live with 3.2A to 8A range. If you need less than 3A, you should get the other 100mA to 5A 30V module instead.
2- add shunts / jumpers to select a combination of sense resistor values to expand that range down;
3- add a programmable gain current sense amplifier for the sense resistor.
this has added advantage that lower value resistor can be used thus reducing power dissipation.
the downside is added complexity, a little extra cost (not much) and mostly the risk of greater susceptibility to noise and transients.
Lastly, the other thing I'm tempted to try is adding extra high side high-voltage mosfet drivers to allow much higher motor voltage - up to 180V.
Since these stepper IC's run at relatively low frequencies with very long "dead times" this is unlikely to create problems with shoot through or current sense blanking.
Does that significantly increase the appeal for such a product?
What is your opinion on all this?
Thanks for your feedback!