Just IN- Should I build these drives? 120V DC, PWM current limit, 20 Amps

[leop]

I need some expert advice on whether I should build and sell these drives.

I recently built my own stepper drives for my Bridgeport mill upgrade because I wanted to keep the original stepper motors (NEMA 42, 8.2A, 56V) and couldn’t find a Gecko or Larkin drive that could handle the current and voltage requirements for them.

My driver design runs 120V DC (no need for a transformer on the supply) and up to 20 amps. They will do ½ step, ¼ step and full step. They will cut current to a user selectable value after 10 ms of hold. Current control is via PWM control.

If you make limit switch connections to the controller it will allow you to jog off of the limit switch but sound an alarm if you try to jog past them.

The drivers are:
1) Built on a double sided board
2) Include manual fuse protection
3) Allow you to change the output transistors if you blow them

I will include a power supply that can drive up to four of these drivers.

I HAVE NOT ADRESSED mid-band resonance damping in my design, but so far the prototypes have not had an issue with mid-ban resonance.

MY Question to you experts is:

1) Should I build more of these and sell theme here on the CNC forum?
2) 2) If you think I should build them what should I charge?
3) Is it better to offer them as a kit or as a finished board?
4) Do I need address mid-ban resonance?

I am not able to find work so building and selling these drives would help. Thank you for your advice.

[battwell]

show them off im sure there is a market!

[GMitchell]

Hi, I like the idea of having a set of easily replaceable drive transistors!

I guess that there will be a market, although from my understanding (which could be wrong) there are a lot of hobby CNC people on this site (including myself) and i assume that hobby CNC machines will mostly (but not all) be using much smaller and less powerful steppers / servos and drives. There probably is a "bridgeport replacement" market, but I don't know how you will go, put up lots of pics and see what happens. Most hobbyists have low budgets, but you do get what you pay for. Most people freak out when they have to part with much over the magic figure of $99.95 $100 sounds like lots more, But I know that your drive transistors may cost $10 or $15 each? depends on switch speed and capacitance etc etc... I guess $200 is not bad, I would work on $10 per watt of motor power output?

Please make sure that the drives are opto Isolated, scary things happened to my PC once (my ex PC)... I don't want to talk about it.

Maybe can you incorporate a small breakout board (with a couple of transparent latches in sockets so that they can be replaced too) into your design, and all I/O on board, it just makes it easier for the non electrical guru to plug in a parallel cable, and then have it all work right away with e-stops, spindle contactor and coolant relay etc rather that figure out sourcing and sinking, termination, pull up resistors, diodes across relay coils... ,make it simple for the hobbyist to "plug it in and go"

I don't know? on the other hand, there are also lots of electronics guru's who would love to have a crack at building their own kit form drive, and want to do the whole project starting with a blank PCB, and a pile of components?

Maybe serious metalies have big bridgeports, and would love a big drive pre assembled, and by contrast electronics geeks (like me) who don't need such big drives (because I have a much smaller machine for routing PCB's myself) would be happy to solder their own kits, but may need smaller drives for routers, and not bridgeports?

WHO KNOWS, I DON'T.

It is risky though, big voltages and large filter caps can be disasterous for hobbyists to make themselves, especially with high currents etc on the board. Maybe it is safer to sell a finished ready to bolt on product?

Does someone who needs 120V, 20 Amps (a MASSIVE 2400 Watts peak!) use a servo instead of steppers for that much power? If I needed 2400 watts per axis, and I was building a new machine from scratch, I personally would go for a Servo, maybe even three phase servo because of constant torque etc. I don't know.

I wish you well, I know I have toyed with the idea of making my own rival for the Gecko servo drives, (a bit like the UHU servo drive out of germany, the guy has sold thousands of his programmed chips!) but I have been unable to find the time to perfect the PID software.

I will watch with interest.

Good luck!

[keebler303]

I think you will find a limited market for what you have come up with.

A gecko is only slightly underpowered for the bridgeport motors. Many have used them on their bridgeports and they work fine from what I have heard of them. That means that best case, you have to be cheaper than a Gecko to sell them. Then you have to have a reputation. People buy Gecko because of their reputation, without that, its tough to make it selling anything right now.

Rectified mains power will give you around 170V DC, not 120. Surely you know this though. Mains powered drives also need to be adequately guarded, isolated, etc. to make them safe to use, I don't know if you have done this or not. I think using a 170V bus will cause excessive heating for most hobby type stepper motors, how many commercial stepper drives use a bus voltage that high?

As stated before, the specs on your drive are overkill for most hobby applications and not advanced enough for commercial applications. Where is the market?

To answer your questions:

1) I wouldn't.
2) If so, has to be less (significantly) than a Gecko or it won't sell.
3) Whose fault is it if the assembled kit doesn't work? It gets expensive paying for others mistakes in assembly. If you don't support them, then you get a bad rap even if you did nothing wrong.
4) Bridgeport didn't, you haven't had to, depends on the application I guess.

My $0.02
Matt

[leop]

All very good input and much to think about.

No desire to compete with Gecko for sure!

Please keep the input coming if you have more.

Leo

[Larken]

There is a market for a big drive for Bridgeport retrofits.

The drivers are:
1) Built on a double sided board
Thats all you need if your ground path is properly layed out. 4 layer is overkill unless your processor is running over 100mhz.
2) Include manual fuse protection
One for each bridge makes testing a lot easier
3) Allow you to change the output transistors if you blow them
Also replaceable driver chip because it always blows with the mosfets,

MY Question to you experts is:
1) Should I build more of these and sell theme here on the CNC forum?
yes, but not too cheap. You have bills to pay !
2) 2) If you think I should build them what should I charge?
A decent price for all your work. You arn't going to sell thousands, so don't base your profits on quantity. You are going to have to give up sleep for a few years as well
3) Is it better to offer them as a kit or as a finished board?
A kit is a lot of tech support.
4) Do I need address mid-ban resonance?
Don't need it, a lot of the schemes out there don't do much anyway, some even cause loss of steps if the incoming pulse stream isn't perfect.

Is this going to be a chopper or PWM ?
Chopper isn't quite as smooth, but the motor runs a lot cooler, is more reliable and you can run at a much higher voltage. There are ways to make a chopper smooth though.

[leop]

Hi Larry,

You know I would not be building these boards if you would start making more Cobra drives! Yes I know the servo drives have you very busy!!

1) I have been doing two layer boards for a long time. People do go overkill with the multi layers all the time now. It's because they let the auto-router take over!

2) Ok, I'll add a fuse per winding.

3) Ok I'll put the drivers in a socket too

I was thinking that $299 would be fair but is that too much?

I have done both chopper and PWM. Chopper was easy, but now that the PWM code is done it doesn't matter. I still don't understand why chopper runs cooler. Average current is average current, but you can't argue with the temp probe I guess!

Do you think the Bridegport drives can take 170V DC over the long term with good power control? I hate transformers!!!!

By the way I will not build and sell these things if you intend to bring back the Cobra drives! This is not a full time gig for me! I am just a design engineer with some friends in the PCB business and i like building stuff and taking tech support calls. In fact if you do bring back the Cobra drives I'll order three right now for my other mill LOL!

Would you be willing to sell these drives on your site as well?

Leo

[keebler303]

Originally Posted by leop I have done both chopper and PWM. Chopper was easy, but now that the PWM code is done it doesn't matter. I still don't understand why chopper runs cooler. Average current is average current, but you can't argue with the temp probe I guess! Leo

I think it is because a chopper turns off the power as soon as the current limit is reached, with PWM, you have to wait until the next cycle to make any adjustment, so more power is delivered to the motor because it has to wait until the next period versus acting instantly.

Matt

[leop]

That makes sense, but the Micro is running at 16MHz and the PWM cycle is on a high priority interrupt. Those motors have so much inductance that the ramp time is super slow. You would think that a few uS on a slow ramp wouldn't put that much more power into the motor! Perhaps I should try cutting the power early in the PWM circuit to make up for the delay.

[keebler303]

At 20KHz you have worst case 50uS when the current limit is exceeded. This is if the current surpasses the limit 1 instruction after the start of the PWM cycle and isn't remedied until the start of the next cycle. The time to service the interrupt might also make it miss being updated on time. Say your ISR takes 10uS to complete, that makes your worst case 60uS before the current is reduced. If this starts happening every few PWM cycles, you can see how it could make a substantial difference in motor heating.

Matt

[amplexus]

I think if you cannot compete with gecko on a feature and price basis there is little point, Even if you can offer more features better specs, lower price etc it will still be hard. Also note many new businesses fail du to lack of capital. It is different if you have a steady job to fall back on. Read Mariss' comments on business startups. Board costs, parts in quantity, liability ins. marketing costs, it all adds up. We succeed in the rare plant business because we offer species and clones that no one else in the world offers, we don't try to be the cheapest , just the best. It helps to find a niche. Websites alone are not enough, you need to develop industry contacts to move quantities of product.
Amplexus Ender

[leop]

Originally Posted by amplexus I think if you cannot compete with gecko on a feature and price basis there is little point, Even if you can offer more features better specs, lower price etc it will still be hard. Also note many new businesses fail du to lack of capital. It is different if you have a steady job to fall back on. Read Mariss' comments on business startups. Board costs, parts in quantity, liability ins. marketing costs, it all adds up. We succeed in the rare plant business because we offer species and clones that no one else in the world offers, we don't try to be the cheapest , just the best. It helps to find a niche. Websites alone are not enough, you need to develop industry contacts to move quantities of product. Amplexus Ender

I totally agree with you and have no desire to compete with Gecko. However Gecko drives don't have the current capacity to handle the Bridgeport motors. If you want to use the original motors Gecko is not an option. This is what i would be building for, not to compete with Gecko.