[lihaijiang]

Looking at the picture, Toshiba are really at a miss to provide an unambiguous determination of pin #1.

Generrally,the layout of pins are defined from left to right or by counter-clockwise direction.

[James Newton]

James,

Please see the picture,which is the chip dimension. It shows the space between the pin is 2.54mm,not 2mm. You can find it from datasheet page41.

Not the pin spacing... that is right... the spacing between the pin /rows/. They are further apart on the real chips than shown on the datasheet. Not a huge deal, but for the layout, it's good to have that extra room.

[H500]

It's quite normal to manufacture through hole parts with the pins further apart than specified. During assembly, the pick up device can squeeze them to the exact size. If they targeted the exact size, tolerance would cause some to be bigger and some to be smaller.

[James Newton]

It's quite normal to manufacture through hole parts with the pins further apart than specified. During assembly, the pick up device can squeeze them to the exact size. If they targeted the exact size, tolerance would cause some to be bigger and some to be smaller.

Ah, that makes sense. Still, it's nice to have that extra space when the board is designed for hand assembly. Perhaps the datasheet should specify a range?

[H500]

I usually work with 0603 surface mount parts. All through hole parts look gigantic to me.

[James Newton]

I usually work with 0603 surface mount parts. All through hole parts look gigantic to me.

Nice problem to have! I wish more people felt that way, as I love SMT components as well, but the vast majority of people are scared to death of doing SMT work. If you want to get something into most peoples hands, and have them try to actually build it, it has to be through hole.

I've tried a few times, and I was trying with this one, to combine the SMT and through hole pads so that it can be built either way. Hopefully that will allow a partiar pick n place run if there is enough interest, or for people who want to try a little SMT soldering to give it a shot if they like.

But then I saw this the other day: PCB Layout: Why Vias Under Pads are a Bad Idea « adafruit industries blog which makes me think twice about trying to combine them... except where the SMT pads would fit well inside the thru hole pads.

I've DONE SMT on thru hole pads before, and I didn't have any of the problems he points out, but then maybe I'm just special? ,o)

[lucas]

Hello all,

a few weeks back I decided to make my own design (have some experience with my own designs of the TB6560, the CPLD drive and others recuperated from printers).
So I ordered some chips and started the design, all through hole, very basic using parts I had laying around. Nice wasn't a priority, just to evaluate the chip.
But I did want a "full" driver with opto's included, on board voltage regulation and no compromises on quality.
Here's the very primary result, ready for testing.

One major problem arrised during assembly: Which heatsink to use and how to fix it, this will need some serious investigation.

Enough rambling, some pics:

[boldford]

Hello all,

a few weeks back I decided to make my own design (have some experience with my own designs of the TB6560, the CPLD drive and others recuperated from printers).
So I ordered some chips and started the design, all through hole, very basic using parts I had laying around. Nice wasn't a priority, just to evaluate the chip.
But I did want a "full" driver with opto's included, on board voltage regulation and no compromises on quality.
Here's the very primary result, ready for testing.

One major problem arrised during assembly: Which heatsink to use and how to fix it, this will need some serious investigation.

Enough rambling, some pics:

Do you intend releasing the PCB layout to the great unwashed on this forum?

[lucas]

I don't know yet if I will release it, normally I do but first I want to be sure it's ok, then we'll see.

It's doable for a good DIY in electronics but not easy, homemade PCB without plated holes: some of the pins are difficult to reach for soldering on the components side. I might DIY the 2nd version again for testing and verifying component placement but when it's finalised I will have them made by a PCB house, it's just so much easier, better and good looking.

There's still quite a list with questionmarks wich need an answer:
1: Power rating for the current sense resistors?
2: Wich heatsink? This is pretty important, a big one needs fixation to the PCB.
3: Freewheel diodes?? There aren't any now but they will reduce the chips power dissipation due to switching losses and thus a smaller heatsink but higher component cost.
4: Probably (certainly) some more issues to come during testing.

I have tested the driver with 24V supply and a 2.5 Amp stepper, works fine, nothing is very hot, still touchable.
But at 4 Amp (still 24VDC) the chip and CS resistors are hot, probably still within their spec's but too hot for me, heat reduces components lifetime. Active cooling could help but I don't like it.

Will do some more testing with higher supply tomorrow and see what that gives.

A T-style heatsink would be ideal, it would allow components behind the chip also and not be like the tower I have now.

Suggestons, remarks?

Regards,
Luc.

[SteveWill]

Power rating for the current sense resistors?

I would think that .2 or .25 Ohm resistor with 4.5 amp would need to be 4 to 5 watt and with a 50% duty cycle 2 to 2.5 watt. I am just guessing but would expect the 2 to 2.5 watt and yes a lot of heat. I would lift the resistor off the board about 1/4 inch or more and if it does not turn brown after an hour run time, would not worry about it failing anytime soon. or another way is to put a little water on your finger and touch the resistor, it should not sizzle/boil (212* temp).

Just thinking out loud.

Good looking board

Steve

[boldford]

. . . .put a little water on your finger and touch the resistor, it should not sizzle/boil (212* temp). . . . .

Also known as the "spit test". If you spit at it and it spits back it's too hot.

[lucas]

I'm using SMD sense resistors because I couldn't find through hole, low inductance ones. Regular through holes are usually wirewound and not suited for this application. They do exist in low inductance but are expensive.
SMD's are easy to find and cheap but can't be cooled by free air.

The drive is running on 40V and 4Amps now, no smoke yet, just too hot for me. I looked into freewheeling diodes but these are huge, adding 8 of these will increase the PCB size seriously. I have some laying around and will check the difference in temperature.

I'm working on a new layout with much bigger sense resistors, this should solve their heat issue.

Regards,
Luc.

[lucas]

Quick update:

The new layout is done, prototype build and tested, the remainig issues I had are solved.

The bigger current sense resistors are on order. For this proto I used 5 x 1Ohm 1206 in parrallel.
I moved the SMD CS resistors to the top and wanted as much copper as possible for cooling but by using thermal relief on the pads (my mistake) this was not achieved, hence the big solder blobs around them to bridge those reliefs.

I like the result: a compact drive with everything on board: power supply regulation, optocoupled galvanic isolation, only 2 connectors.....

Thus this is my final version, a few components need to be shifted and some other cosmetic changes.

Will order real PCB's next week.

If there's interest in this design : let me know and I will open a new thread to not clutter this thread any further with my rambling.

Luc.

[boldford]

If there's interest in this design : let me know and I will open a new thread to not clutter this thread any further with my rambling.

Luc.

To me, it looks like you've produced a nice compact design that includes pretty much all the features the vast majority of folk would want.

I'm sure others would join me in asking you to keep posting your progress reports.

[lihaijiang]

Quick update:
The bigger current sense resistors are on order. For this proto I used 5 x 1Ohm 1206 in parrallel.

Dear Luc,

Sorry for the delay reply,duo to the business trip.

I think you'd better to use 2X0.47ohm 2512 in parallel instead of 1206.

If necessary, you should use schottky diodes or Fast recovery diodes(below 100ns)to reduce the chips power dissipation .


If you want to reduce the temperature, you can add the fast decay proportion or add the decay time .That means you can increase the resistor value which is connected with PIN23.

[lucas]

To me, it looks like you've produced a nice compact design that includes pretty much all the features the vast majority of folk would want.

Thanks, here we go:

http://www.cnczone.com/forums/open_s...cb_design.html

[lihaijiang]

Hey James,

How about your board? Is it on working?

[James Newton]

Hey James,

How about your board? Is it on working?

LOL... I never wanted to make a board, but was just doing it because no one else was. As soon as others started working on it, I dropped back.

My goal is to provide a nice PCB (and maybe a kit) for hobbist to solder up, and that is very close to happening.

Anyone interested should know that the PCB's will be available Real Soon Now.

[James Newton]

The new kit is /just/ about ready to ship. Preliminary documentation is at:
THB6064 MassMind Stepper Motor Driver Kit

Let me know if you see anything wrong with that page?

[H500]

I intend this as constructive criticism..... If I did a web search and found the page, I would have no motivation to read the wall of text just to find out what the site was about. I cannot quickly tell if you were selling something or just presenting information.