[Fish4Fun]

Jeffs555,

You are quite correct, I am talking about clamping the 'reverse pulse'. On page 5 of this thread, kreutz stated:

A simple power dump circuit is described in the attached application note. As I told you I use an halogen lamp that is highly an overkill (150W 24 Volt), you won't feel it getting warm. If the mosfet driving it gets shorted You will notice it immediately. I also use a similar circuit to the application note's, driving one interrupt pin on the at90s2313 and controlling the lamp's current by PWM. (I use that at90s2313 already as a safety monitor)

When using integrated bridge drivers it is more uncertain what the safety margin must be, excessive de-rating will create an under-performing driver. A common solution is adding Snubbers, that will be effective if you knew beforehand what motor the user is going to drive, otherwise it is a compromise solution (will work within a range of motor inductances, but can be ineffective in some cases).

As I stated, I am certain there is a reason for considering a snubber, or a regenerative circuit combined with a power dumper that makes those choices better than using a simple TVS. The focus of this thread appeared to be primarily about tweaking a design for robustness and affordability, and so I thought I would ask why a simply TVS clamp was not an appropriate choice for protecting the bridge. My assumption is that it is not an appropriate choice, I am just curious why.

Fish

[jeffs555]

Sorry, I hadn't read this whole thread. What he is using the power dump circuit for appears to be the reverse emf generated by the motor as it continues to rotate due to inertia during deceleration. This is different from the reverse pulse from the collapsing field of the coil. Without a power dump circuit, they will both be dumped back into the power supply capacitors through the body diodes in the mosfets. The difference is that the pulse from the collapsing field is relatively short so while the peak power may be high, the average power is relatively low. With a high inertia high speed motor, the average power generated during deceleration could be significant. The problem I see with using a TVS for this is that most TVS devices are designed to handle short pulses with high peak power, but their continuous power ratings are rather low.

[kreutz]

Hello guys,

I am back!, It will take a little time to read all the new postings and answer them. I also need to dust and organize my working environment, it is a mess.

Regards,

Kreutz.

[kreutz]

As I stated, I am certain there is a reason for considering a snubber, or a regenerative circuit combined with a power dumper that makes those choices better than using a simple TVS.

Using a TVS is a simple and effective way to protect against very short time, low repetition rate, high energy surges. Each power mosfet internal diode has a clamping voltage and some capacity to withstand medium energy (in the hundreds of mJ) short time surges too, but aren't as fast, the Snubbers have double function of limiting dV/dt, and limiting the peak surge voltage during the reverse recovery time interval after turning off the mosfet with an inductive load, helping to control the voltage spike before clamping occurs, in the worst case scenario, while dissipating part of that energy.

The problem with a TVS only protection, in our case, is that the power IC published absolute maximum voltage is too near the working voltage (52 volts and 45-48 volts respectively), TVS are not designed to withstand long time over-voltages (and clamping voltages have a tolerance range), then there is the dv/dt issue that is also easily solved with the snubber.

For discrete H bridge output, I would opt for the TVS and mosfet's gate resistance controlled dv/dt.

The energy fed back into the power supply when quickly decelerating a motor (capacitor voltage increases due to this additional energy), is best controlled by de-rating the working voltage and/or using a power dump load if it becomes an issue. For low inertia loads I use a 300 ohm 15 Watt resistive load continually attached across the power supply capacitor's terminals.

[mugabe]

The energy fed back into the power supply when quickly decelerating a motor (capacitor voltage increases due to this additional energy), is best controlled by de-rating the working voltage and/or using a power dump load if it becomes an issue. For low inertia loads I use a 300 ohm 15 Watt resistive load continually attached across the power supply capacitor's terminals.

Just this one wont be enough ?

[kreutz]

Just this one wont be enough ?

Mugabe;

Interesting concept. I will test it as soon as I have some time left...

Thanks,

Kreutz.

[Jojje Bergan]

http://www.hv4all.com/cnc/RETURNED%2...Y%2520DUMP.pdf

[mugabe]

http://www.hv4all.com/cnc/RETURNED%2...Y%2520DUMP.pdf

To be honest , i did not know this link. The only concern here is time the bjt will stay in active region - that will define power dissipated on the transistor.

[Jojje Bergan]

And http://www.hv4all.com/cnc/Power%2520...%2520Bleed.pdf

They are coming from Geekodrive from the beginning i think.

[mugabe]

And http://www.hv4all.com/cnc/Power%2520...%2520Bleed.pdf

They are coming from Geekodrive from the beginning i think.

imho you should be careful with this scheme due to high resistance on opamp inputs, as those could be influenced by rectifier diodes reverse leakage currents which widely varies with temperature.

[Cristal]

Where is the schematic?

[kreutz]

Manually populating the 202 components on the 2 axes prototype board is a slow task, I haven't started yet because I should start and finish within few hours in order to do the reflow the same day. I am waiting for a weekend break to start.

The job is going to be easier due to a gift from Pminmo, I will post a photo of my SMD dispenser gadget based on his gift.

Most of the components are 0603 size.

[David_Web]

Any news?

[kreutz]

Any news?

I have been overloaded in my job, no weekends and working everyday till late.
I will continue working on that board as soon as my workload comes back to normal.

Regards,

Kreutz

[jwentzco]

Any news?

Guys and girls,

This is my first post on this forum and, chances are, I'll be banned for it.

I've been following this thread for at list 2 months and, like everybody else, very much interested on it.

Unfortunately, I think commercial interests got on the way and we'll not see the $140 end product.

Kreutz stated in Aug 2006, "The whole idea came out while trying to help a fellow engineer, who was laid off last week (and needs the income). In order to make sure this is going to help her, I need to know if it has some demand in our community."

Well, it's hard to believe that populating and testing such a nice design (with the help of a unemployed engineer) takes such a long time. As an engineer myself, Id be very tickled and pressed to do it myself ASAP, that's the nature of engineering - does not matter how many jobs you have - it's pleasure!

So, here I am, ready to be booted from this forum, but feeling firm (and hopefully wrong) that this is another good intention gone wrong. We'll not see this board near the stated starting price at all.

John

[atlab]

Guys and girls,

This is my first post on this forum and, chances are, I'll be banned for it.

I've been following this thread for at list 2 months and, like everybody else, very much interested on it.

Unfortunately, I think commercial interests got on the way and we'll not see the $140 end product.

Kreutz stated in Aug 2006, "The whole idea came out while trying to help a fellow engineer, who was laid off last week (and needs the income). In order to make sure this is going to help her, I need to know if it has some demand in our community."

Well, it's hard to believe that populating and testing such a nice design (with the help of a unemployed engineer) takes such a long time. As an engineer myself, Id be very tickled and pressed to do it myself ASAP, that's the nature of engineering - does not matter how many jobs you have - it's pleasure!

So, here I am, ready to be booted from this forum, but feeling firm (and hopefully wrong) that this is another good intention gone wrong. We'll not see this board near the stated starting price at all.

John

Hi, John.
If you are engineer, you can take L6203, project "Mardus-Kreutz Unipolar Microstepper, a simpler alternative." and DIY.
Or you can wait Kreutz, or buy Geckos, etc. Isn't it?

P.S. To put all eggs in one basket isn't good idea...

[kreutz]

Guys and girls,

This is my first post on this forum and, chances are, I'll be banned for it.

I've been following this thread for at list 2 months and, like everybody else, very much interested on it.

Unfortunately, I think commercial interests got on the way and we'll not see the $140 end product.

Kreutz stated in Aug 2006, "The whole idea came out while trying to help a fellow engineer, who was laid off last week (and needs the income). In order to make sure this is going to help her, I need to know if it has some demand in our community."

Well, it's hard to believe that populating and testing such a nice design (with the help of a unemployed engineer) takes such a long time. As an engineer myself, Id be very tickled and pressed to do it myself ASAP, that's the nature of engineering - does not matter how many jobs you have - it's pleasure!

So, here I am, ready to be booted from this forum, but feeling firm (and hopefully wrong) that this is another good intention gone wrong. We'll not see this board near the stated starting price at all.

John

John;

The problem is that I am doing it all by myself. My fellow engineer has two small children (on vacation) and, since she is still unemployed, she is not sending them to a daycare facility. Do you know what two small, hyperactive, noisy children, can do to your concentration? I don't want them around while I am working...

The other issue is that, since I began to work again (from my vacation), I haven't stopped working at my job, even during the weekends.

Regarding the commercial interests, it is true that the original target price might not be reachable, but I haven't abandoned it yet. What is certain is that it is not a DIY project anymore, and what is going to determine the final price is the cost of professional populate and reflow solder the boards in "small quantities" until an accurate production quantity per order could be determined.

You are not going to be banned for stating your honest opinion. I really appreciate people who sincerely say what they think.

Thanks,

Kreutz.

[jwentzco]

Kreutz,

Please do not take my posting as any kind of pressure to finish the project you started so altruistically. It was just an observation, be it right or wrong.

I really admire your humanism and engineering skills and wish you the best of luck on this and all of your endeavors.

And, although I already got my drives, I'm still anxiously waiting for the release of yours (Give the children passion-fruit juice, it will calm then down...).

Thanks, John

[kreutz]

The water is taking its normal level at work, so I will begin to enjoy my weekends and weekdays nights from now on. Retaking the project from where it was left in July, here is the 2 axes prototype's composite layout:


BTW: The latest revision of the PCB is Rev 3.0j, some minor changes on a connector and a couple of vias.

[kreutz]

Update.

I ordered a stencil from pololu.com a couple of days ago, it is on a 0.4 mylar sheet and a quick test fit on top of the PCB came out perfect (it arrived today). I will let you know how it went after the reflow soldering on my modified toaster oven. Now that applying soldering paste is going to take less time, the test process can start.