PDA

View Full Version : PWM lead screw drive info



Evan
10-20-2005, 05:57 PM
See here:

http://vts.bc.ca/workshop/pwm/pwm.htm

topct
10-20-2005, 06:25 PM
Well done. http://bbs.homeshopmachinist.net//smile.gif

------------------
Gene

Rustybolt
10-20-2005, 06:29 PM
Thank you very much Evan.

jkoper
10-20-2005, 07:09 PM
Great job Evan! That is just the circuit I need to upgrade the drive on my homemade sinker edm. Thanks for posting it!

Jim

Evan
10-20-2005, 07:34 PM
Thanks guys. I left out something important in the construction notes which I have now added:

IMPORTANT NOTE: The FET mounting tabs are at drain potential. Either the devices must be mounted with insulating spacers and washers or the entire heat sink must be insulated from the chassis.

topct
10-20-2005, 07:59 PM
Noted.

------------------
Gene

Alan in Oz
10-20-2005, 10:06 PM
Evan,
Thank you for taking the time put together the info, much appreciated. All the best.

uute
10-20-2005, 10:14 PM
I shouldn't have been suprised! But I was.

Evan, your publishing is as nice and high quality as your machine work!!

Thank you.
uute

darryl
10-21-2005, 01:28 AM
Thanks for putting that together for us, Evan. I wish I still had the urge to do projects like that. Maybe my enthusiasm for electronic projects will return- this is the exact project I need for my tadpole, a control circuit for the motor. I will have to go multiple fets for sure, as you mentioned. Probably six 50 amp devices. I'll have to pay attention to the combined gate capacitance as far as switching times and the available output current from the timer chip.

I got a chuckle out of seeing those snubber parts- I don't think I've seen an old style carbon resistor for years now. Rescued from a certain death in the 'assorted parts' drawer- http://bbs.homeshopmachinist.net//smile.gif

Mcgyver
10-21-2005, 07:25 AM
thanks very much Evan - nice job.

I've a long way to go on the learning curve and wonder if you could answer a couple of questions?

if you moved to an irf740 (400V 10 amps) wouldn't that work for the treadmill motor?, they are also cheap and common. you mention a dozen ganged up but do you think one would one suffice?

I understand that the first FET, when on, allows current to flow through the motor to common. but what does the second do? it seems connected to itself with an rc in the middle

Is there a way to get the shematic in a higher resoltion?

thanks again

Your Old Dog
10-21-2005, 08:54 AM
Nice job Evan!

Darryl, I have a Motor Guide Pulse Width 50 footlbs thrust trolling motor that "may" work in your application. They are a bit pricey but it already comes with the foot control for gas pedal!!

This motor will produce enough torque to damn near put you in the drink if it's in reverse and you're leaning way over the bow when somebody punches it http://bbs.homeshopmachinist.net//biggrin.gif http://bbs.homeshopmachinist.net//biggrin.gif

J Tiers
10-21-2005, 10:39 AM
<font face="Verdana, Arial" size="2">JT,

I cobbled up the circuit on a breadboard first and swept the frequency to see how the motor behaved. It really lost torque at higher frequencies. I have no specs on it so I went with what worked best. I also am working with one hand tied behind my back as my good triggered sweep scope is down right now (until I repair it). All I have for a working oscilliscope is an old POS RCA educational kit I picked up at a garage sale. It's good for at most low sonic frequencies. I guess I should fix the good one.
</font>

The 555 is somewhat limited in current.... At higher frequencies the 0.1 uf (C8) will load the output enough that it may not fully turn-on the fets. (The fets have a low gate charge requirement, so that is likely not the problem.)

The R-C time constant of C8 and R11 is good to around 700 Hz, so much higher than that and you will no longer be discharging C8 before the next pulse. Maybe lower than that depending on duty cycle.

If you get rid of C8, or drastically lower it, you should be able to work to higher frequencies. It shouldn't be required.

That circuit depends on R11 do discharge both the gates and C8 before the next pulse. A more robust driver would be able to pull down the gate fast and allow higher frequency operation.

Even a number of paralleled sections of CMOS inverter would help. I have seen a number of drives use that, although a "real" gate driver has more drive. You don't need a lot with 20 nC of gate charge, the inverters would work, especially at a few kHz.

You could still isolate the PWM portion from teh inverter using a diode and resistor as now, and also a diode for its power supply. That would give you the drive and the isolation, both.

A higher frequency could reduce vibration a lot, and also move the noise out of the audible range.

The motor shouldn't be a limit. Remember, VFDs do use higher frequencies, up to and above 20 kHz, and they have relatively high inductance motors to run.

Evan
10-21-2005, 11:09 AM
Mcgyver,

The FETs are in parallel if you study the circuit. FETs have a positive temperature coefficient so that as they warm the on resistance goes up. As long as they are on the same heat sink this means they share load automatically.

Right clik the link below the schematic image and save it to your computer for the high res schematic.

JT,

The 555 can source and sink about 200 ma at 12 volts. Of course in this application it is only sourcing but it should be enough to very firmly put the FETs into full conduction as the 720 exhibits full turn on at about 8 volts on the gate at full current. For higher frequencies the value of R8 can be reduced considerably and C11 reduced as well. I like to keep a bit of time constant as the 555 has a turn on/off time on the output of only 100ns. Slowing that down a bit will reduce the chance of ringing in the gate circuit as well as soften the motor pulses and noise.

If more drive is needed then the 555s can be operated at 15 volts.

[This message has been edited by Evan (edited 10-21-2005).]

sch
10-21-2005, 11:15 AM
[QUOTE]Originally posted by Mcgyver:
[B]thanks very much Evan - nice job.
if you moved to an irf740 (400V 10 amps) wouldn't that work for the treadmill motor?, they are also cheap and common. you mention a dozen ganged up but do you think one would one suffice?
I understand that the first FET, when on, allows current to flow through the motor to common. but what does the second do? it seems connected to itself with an rc in the middle

Is there a way to get the shematic in a higher resoltion? QUOTE

There is a download link for the larger resolution schematic just below the schematic itself. The result prints nicely on an 8x10 sheet. The FETS are paralleled for increased current capacity. Nice thing about power fets is that they can be paralleled without any concern over one device hogging all the current. They play together nicely and linearly increase current capacity with number though there are some minor limits.
I have driven an "18amp" treadmill motor from 110vac with two paralleled IRF740 FETs with no temperature rise in the FETS, until the uncooled motor got too hot to touch. Speed control was smooth over the range of the DC motor controller I used (a www.mpja.com (http://www.mpja.com)
6067-kt dc motor controller with a similar ckt to Evan's sold originally with a FET now with a power darlington. Just sub the FETs of choice.)
Steve

J Tiers
10-21-2005, 11:34 AM
Careful with that spec.....


The National spec shows that for pull-UP, the 200 mA corresponds to typical 12.5V DROP on the output. In other words with a 12V supply, it can provide that 200 mA only into a short. 12-8=4, and at 4 volts drop, it can only supply a lower current, which isn't specified.

I have reason to think that may be a misprint in this copy of the catalog, but there is some drop, and the pull-up is not as strong. And, teh current data sheet at National's website also shows the same data. And their graph vs tabular data are inconsistent.

However, it isn't unusual for pull-up to be weaker......

So you probably won't get as strong a pull-up as you might expect.

The multi-inverter drive is cheap and very effetive.

Evan
10-21-2005, 11:50 AM
"The National spec shows that for pull-UP, the 200 mA corresponds to typical 12.5V DROP on the output. In other words with a 12V supply, it can provide that 200 mA only into a short."

My understanding (which agrees with the graph) is that is what the output drops to, not how much it drops. That also agrees with my experience.

Mcgyver
10-21-2005, 11:52 AM
was that hi-res link there before or am I really that blind http://bbs.homeshopmachinist.net//biggrin.gif thanks guys for the help. i'll continue the study until the ball drops on the fets in parallel

I actually surprised myself by designing one that was very similair to this and almost worked, but with no transformer and a bigger cap. used a dual 555 to drive a 740. worked pefectly on a light bulb load but would blow up with the motor. only recently discover i wasn't using fast a enough diode - i'd used a 4001 iirc instead of the Schottky

anyway, thx, the design as encouraged my to try again

Evan
10-21-2005, 04:19 PM
I just added a bit more info to the web page and am posting it here as well.

----------------------------------
Scrounging Parts

Dead computer power supplies are a good source of parts for a project such as this. In particular, heat sinks and diodes as well as mounting hardware are easily obtained. Just go to your nearest computer repair shop and ask for a few dead power supplies, I am sure they will be happy to give you some.

Here is an example of what is often available in a dead supply. These parts will not usually be damaged. Especially useful is the dual schottky diode in this picture. It is rated at 2x15 amps at 100 volts DC.

http://vts.bc.ca/workshop/pwm/parts1.jpg

topct
10-21-2005, 10:03 PM
Those old power supplies for parts? Just make sure you can and know how to discharge the filter caps. Man that hurts.

------------------
Gene

Ryobiguy
10-21-2005, 10:28 PM
Nice looking project! I noticed the mid-project workbench behind the lathe has been blurred out. http://bbs.homeshopmachinist.net//smile.gif

I think I noticed a typo on the output parts list:
R3, R4 100 ohm gate resistor for FET. See note 1.
R3 1k, 1/2 watt

Is the second R3 supposed to be R5?

Evan
10-22-2005, 02:18 AM
Yep, typo. Thanks, I have fixed it.

J Tiers
10-22-2005, 02:27 AM
<font face="Verdana, Arial" size="2">Originally posted by Evan:
"The National spec shows that for pull-UP, the 200 mA corresponds to typical 12.5V DROP on the output. In other words with a 12V supply, it can provide that 200 mA only into a short."

My understanding (which agrees with the graph) is that is what the output drops to, not how much it drops. That also agrees with my experience.</font>


That's why I mentioned that I have reason to doubt the actual words used in the data sheet description ....

In any case, a driver with bidirectional drive would reduce mosfet dissipation and allow you to reach higher frequencies easily without reducing the resistor to impractical values.

And you could still maintain your diode isolation at a point prior to the driver.

For higher voltage drives as has been asked about, an optocoupled gate driver such as mentioned in teh original thread, will allow control via a control circuit which is at low voltage and NOT directly connected to the line voltage (for 90V DC motors, etc). Optos with UL /VDE isolation and the requisite speed are available.

A transformer also will do it, but has a serious problem with wide range PWM unless a lot of trouble is taken. The opto has no such problems.

Evan
10-22-2005, 02:53 AM
Yes, if I were going to drive at a higher voltage I would use an opto-iso. They are cheap by the bagfull and fast. As far a xformers for this sort of application, that is beyond my level of expertise to design.

hoffman
10-22-2005, 03:53 PM
Thanks for taking the time to put that together Evan. I have a few projects in mind that will require controlling motor speed and I may try my hand at some electronics work.

John Stevenson
10-22-2005, 08:16 PM
Just picked up MEW 110 today from the newsagents.
In there is an article By Peter Rawlinson who goes on the describe a power feed for the tailstock as hes'got problems with his hands,

He goes about this slightly different to Evan in that he uses a stepper motor and driver to achieve the motion.
Using this method he gets forward and reverse, variable speed and jog.

By playing with the step multiplier on the stepper driver and the pot you can get a very wide range of speeds.
Going to the extreme Peter found that he could get 1 motor rev in two minutes, hardly viable but it shows what is achievable.

Sir John .

sch
10-23-2005, 02:43 AM
And I just got MEW 109 last week. Must come by container via Suez..
Steve

Evan
10-23-2005, 02:47 AM
John,

I am trying to imagine using a power feed on the tailstock. You do mean the quill feed, right? I can see it for some drilling jobs but not others. I change tooling so often in the tailstock it seems that it would be in the way a lot.

John Stevenson
10-23-2005, 06:27 AM
Evan,
This was done on a Myford that had the capstan handle conversion fitted.
These have a thru hole so you can drift tooling out without having to wind back to auto knock out.
The guy was saying that with his hand problems he couldn't press on anything with his palms hence the conversion.

I have often thought that it might be nice to have power feed on a tailstock on the larger lathes as punching a 3" drill thru takes a bit of grunt but I don't know without doing it how useful it would be as a fixed attachment.

Some lathes have a limk that can fit between the carriage and tailstock to allow the carriage to tow the tailstock but we are getting off the subject.

The purpose of mentioning the article was to bring attention to the use of stepper motors to the same application.
This is only then one step away from CNC threading with a single slot opto switch on the spindle.

John S.