Variable speed press drill

[Susquatch]

The only real downside to these simple ones is that it’s a single pickup and less accurate at lower RPMs.

This is a problem I have noted too. What do you think the right number is?

While not nearly as easy as buying a tach off the shelf, the Arduino stuff can clobber this problem by balancing the update time with an averaging display.

But I'm with you. It's hard to beat something with multiple sensors. My DRO has rpm and can handle multiple sensors. I just need to take my mill apart to install them....... Gunna be a while before that happens.

 

[jcdammeyer]

This is a problem I have noted too. What do you think the right number is?

While not nearly as easy as buying a tach off the shelf, the Arduino stuff can clobber this problem by balancing the update time with an averaging display.

But I'm with you. It's hard to beat something with multiple sensors. My DRO has rpm and can handle multiple sensors. I just need to take my mill apart to install them....... Gunna be a while before that happens.

I've not really seen issues with lower RPM unless it's a DC controlled motor. The AC motors used in most drill presses are sync'd to the power line and that doesn't really vary that much.

Now VFD driven motors might be another issue but any decent RPM measuring device will average anyway. In fact think about it. If you have 60 teeth and it's turning 120 RPM that's 2 RPS and that's an edge every 8mS. Slowest update rate for Cine film is 24 FPS (~42mS) in order to not have flicker so we'd see 5 teeth and report a value. But if the RPM is changing within one revolution then the digits on the display will turn into a blur.

In fact for instruments like meters etc. the update is normally no more than about 5 Hz to avoid the flicker effect. And then you have to ask why it's important to see the bottom digit flicker based on uneven RPM within one revolution.

So one PPR is actually more than adequate for reporting RPM if it's averaged over several revolutions. Higher resolution won't improve that.

 

[Susquatch]

I've not really seen issues with lower RPM unless it's a DC controlled motor. The AC motors used in most drill presses are sync'd to the power line and that doesn't really vary that much.

Now VFD driven motors might be another issue but any decent RPM measuring device will average anyway. In fact think about it. If you have 60 teeth and it's turning 120 RPM that's 2 RPS and that's an edge every 8mS. Slowest update rate for Cine film is 24 FPS (~42mS) in order to not have flicker so we'd see 5 teeth and report a value. But if the RPM is changing within one revolution then the digits on the display will turn into a blur.

In fact for instruments like meters etc. the update is normally no more than about 5 Hz to avoid the flicker effect. And then you have to ask why it's important to see the bottom digit flicker based on uneven RPM within one revolution.

So one PPR is actually more than adequate for reporting RPM if it's averaged over several revolutions. Higher resolution won't improve that.

Might be just an issue associated with the cheapo tachs. Your logic is sound. I'm not sure what is up. I bought 2 on Amazon that were junk at low rpms jumping all over the place but ok at higher speeds. At first I thought it was a sensor issue. But nothing I did fixed it. No problems I could see with a scope on the sensor signal. So I added more sensors and bang - it worked - rock solid - albeit reading 4x as high as it should with no way to correct. I returned them as they were useless to me.

Edit - tried them on various motor types. Made no difference. They just don't like low rpms.

FWIW, I have the same problem with a cheap optical digital laser tach.

 

[whydontu]

I have the 120VAC tach on my lathe, and an Arduino version on my mill. The lathe tach is fit for use. It is reasonably stable, maybe one in twenty readings bounces. It's accurate with a single sensor. For my thinking, for machine spindle speeds, we're not concerned about super-accuracy or resolution, as long as it's close that's all we need. Having 5% error is irrelevant (and both my tachs are well within a 5% error).

My lathe sits at a fixed speed when in use. Having a tach is just window dressing. The mill is a bit different, it is variable speed, but even then I ran it for decades without a tach and just relied on the RPM values on the control dial. The mill tach is useful now because the Arduino code calculates SFM vs. cutter size in real time, and with this information I find I'm running cutters at higher RPM than I used to.

 

[Marc Moreau]

@Susquatch - my mill is an older KBC machine so there is lots of room to play with. I'll start a thread when I do the install.

View attachment 29565


I called TECO about this and they said the DC output is only good for 20mv.

@Marc Moreau - if you are meaning the tach install here are a couple of pics. It's running off a 220v wall wart.

View attachment 29566View attachment 29567
The knob (pot) controls the speed from the VFD the little switch between the knob and the tach is on/off power to the tach.
The picture below shows the sensor and magnet position. I milled a flat spot in the collar and JB welded the magnet in place, the collar is held in place by three set screws.

View attachment 29570

Ho yessss I like this Thank you I will problably do this on my lathe.

 

[Marc Moreau]

@Susquatch - my mill is an older KBC machine so there is lots of room to play with. I'll start a thread when I do the install.

View attachment 29565


I called TECO about this and they said the DC output is only good for 20mv.

@Marc Moreau - if you are meaning the tach install here are a couple of pics. It's running off a 220v wall wart.

View attachment 29566View attachment 29567
The knob (pot) controls the speed from the VFD the little switch between the knob and the tach is on/off power to the tach.
The picture below shows the sensor and magnet position. I milled a flat spot in the collar and JB welded the magnet in place, the collar is held in place by three set screws.

View attachment 29570

This milling look my Craftex

 

[jcdammeyer]

Might be just an issue associated with the cheapo tachs. Your logic is sound. I'm not sure what is up. I bought 2 on Amazon that were junk at low rpms jumping all over the place but ok at higher speeds. At first I thought it was a sensor issue. But nothing I did fixed it. No problems I could see with a scope on the sensor signal. So I added more sensors and bang - it worked - rock solid - albeit reading 4x as high as it should with no way to correct. I returned them as they were useless to me.

Edit - tried them on various motor types. Made no difference. They just don't like low rpms.

FWIW, I have the same problem with a cheap optical digital laser tach.

That's an issue with cheapo tachs. My ELS threads anything from really coarse (more like spirals) threads to extremely fine threads with 1 PPR. RPM is averaged though; over a few seconds. Multiple pulses per rev will usually also average.

In fact you can usually set up a scope to not erase on each trigger. If you set it up to see the trigger edge and then the same edge after one revolution you'd be amazed at how much the second pulse varies from the first even with an AC driven belt drive system. I suspect the bumps on the belt and pulleys contribute to that.

Then go to a lathe like my South bend with a V belt from motor to half shaft and then the flat belt to the spindle and it's amazing that I can cut threads at all. But over time the overall average is stable.

I also added a feature to my ELS to investigate this sort of thing. There's now a 't' command that returns the last 16 values while the spindle was turning.

For the Gingery Lathe which has COTS zinc multi step pulleys. The values are the number of 50uS ticks between spindle index pulses. Yet the 282 RPM display is stable.
SpinRate 4252,
RPMAverage Table:,
4245, 4266, 4235, 4278, 4236, 4265, 4247, 4247, 4263, 4238, 4268, 4238, 4260, 4251, 4251, 4243, 282, 75,

Here's one from the South Bend lathe after it stopped.
SpinRate 0,
RPMAverage Table:,
1198, 1194, 1199, 1194, 1195, 1197, 1194, 1197, 1199, 1192, 1198, 1197, 1192, 1198, 1198, 1194, 1003, 10000,

It's better but there's still a 92..99 variance or 350uS. So 1199 is 1000.834 RPM while 1192 is 1006.71 RPM.
The way the hardware works there is no way there could be 350uS latency where the software somehow disables interrupts for that long and is late detecting the edge. So what we're really seeing is a general drift based on pulley diameters and belt shape that appears to be cyclic.

 

[Marc Moreau]

@Susquatch - my mill is an older KBC machine so there is lots of room to play with. I'll start a thread when I do the install.

View attachment 29565


I called TECO about this and they said the DC output is only good for 20mv.

@Marc Moreau - if you are meaning the tach install here are a couple of pics. It's running off a 220v wall wart.

View attachment 29566View attachment 29567
The knob (pot) controls the speed from the VFD the little switch between the knob and the tach is on/off power to the tach.
The picture below shows the sensor and magnet position. I milled a flat spot in the collar and JB welded the magnet in place, the collar is held in place by three set screws.

View attachment 29570

Ho I just saw how you did it I like this way very nice job.

 

[Susquatch]

That's an issue with cheapo tachs. My ELS threads anything from really coarse (more like spirals) threads to extremely fine threads with 1 PPR. RPM is averaged though; over a few seconds. Multiple pulses per rev will usually also average.

In fact you can usually set up a scope to not erase on each trigger. If you set it up to see the trigger edge and then the same edge after one revolution you'd be amazed at how much the second pulse varies from the first even with an AC driven belt drive system. I suspect the bumps on the belt and pulleys contribute to that.

Then go to a lathe like my South bend with a V belt from motor to half shaft and then the flat belt to the spindle and it's amazing that I can cut threads at all. But over time the overall average is stable.

I also added a feature to my ELS to investigate this sort of thing. There's now a 't' command that returns the last 16 values while the spindle was turning.

For the Gingery Lathe which has COTS zinc multi step pulleys. The values are the number of 50uS ticks between spindle index pulses. Yet the 282 RPM display is stable.
SpinRate 4252,
RPMAverage Table:,
4245, 4266, 4235, 4278, 4236, 4265, 4247, 4247, 4263, 4238, 4268, 4238, 4260, 4251, 4251, 4243, 282, 75,

Here's one from the South Bend lathe after it stopped.
SpinRate 0,
RPMAverage Table:,
1198, 1194, 1199, 1194, 1195, 1197, 1194, 1197, 1199, 1192, 1198, 1197, 1192, 1198, 1198, 1194, 1003, 10000,

It's better but there's still a 92..99 variance or 350uS. So 1199 is 1000.834 RPM while 1192 is 1006.71 RPM.
The way the hardware works there is no way there could be 350uS latency where the software somehow disables interrupts for that long and is late detecting the edge. So what we're really seeing is a general drift based on pulley diameters and belt shape that appears to be cyclic.

On the cheapos, readings jump around up to 25% or so at low RPM. Not at all the same as you describe. Its not reasonably possible to figure out what's going on in the software. My guess is incompetent programming or design. Who knows. I'd love to test a 110v unit. They probably required more talent to complete so might be more reliable on the sensor interface side too.

I bought my last cheapo.

 

[jcdammeyer]

On the cheapos, readings jump around up to 25% or so at low RPM. Not at all the same as you describe. Its not reasonably possible to figure out what's going on in the software. My guess is incompetent programming or design. Who knows. I'd love to test a 110v unit. They probably required more talent to complete so might be more reliable on the sensor interface side too.

I bought my last cheapo.

It's also possible that the type of sensor they are using is sensitive to speed. So if the speed is too slow it doesn't cause it to trip. Kind of like the magneto that doesn't fire the spark plug if you rotate the crankshaft slower than a certain speed.

On my ELS it's done this way which is why a very narrow sensor won't work. This approach is also very resilient to electrical noise spikes.
// We time in Pulse Clocks how long it takes for one revolution. // The time is from rising edge to rising edge for a one slot per rev counter. // New Levels need to be active for INDEX_DEBOUNCE pulse clocks before the state change // is considered valid. At 50uS per PULSE_CLOCK the default is 4 so noise glitches // less than 200uS are ignored.

 

[Susquatch]

It's also possible that the type of sensor they are using is sensitive to speed. So if the speed is too slow it doesn't cause it to trip. Kind of like the magneto that doesn't fire the spark plug if you rotate the crankshaft slower than a certain speed.

On my ELS it's done this way which is why a very narrow sensor won't work. This approach is also very resilient to electrical noise spikes.
// We time in Pulse Clocks how long it takes for one revolution. // The time is from rising edge to rising edge for a one slot per rev counter. // New Levels need to be active for INDEX_DEBOUNCE pulse clocks before the state change // is considered valid. At 50uS per PULSE_CLOCK the default is 4 so noise glitches // less than 200uS are ignored.

Entirely possible. It's just a magnet. The read head was sending out a very reliable looking signal even practically stopped. The signal integrity goes bad inside the counter itself somehow. I still think it's crappy design or programming - most likely the latter.

Anyway, I can't recommend these units. I'd prolly try the 110V version. It would be a nice add something cheap and simple like this to a mill or variable speed lathe with no dro. Not everyone wants to do any programming but might be willing to solder a few wires together.

 

[jcdammeyer]

Entirely possible. It's just a magnet. The read head was sending out a very reliable looking signal even practically stopped. The signal integrity goes bad inside the counter itself somehow. I still think it's crappy design or programming - most likely the latter.

Anyway, I can't recommend these units. I'd prolly try the 110V version. It would be a nice add something cheap and simple like this to a mill or variable speed lathe with no dro. Not everyone wants to do any programming but might be willing to solder a few wires together.

I've been mulling over this for a while here. It's possible the software will think the unit is stopped if the magnet is in front of the sensor for too long or long relative to no magnet. So turning slowly keeps fooling the system. Maybe it counts up to 255 and if it rolls over to 0 then it assumes stopped and just starts counting again until it sees the magnet gone. Then it resets the RPM count to 0 and counts until the magnet is seen again. That would give wild values at low speeds.

With multiple tabs per rev the ratio between magnet there and not is smaller so it may not see the spindle as stopped.

I know it doesn't matter but it's a curiosity.

 

[Mcgyver]

A good option for fraction HP machines that need variable speed, is the little Consew sewing machine motor.

https://www.amazon.ca/Genuine-CSM1000-Machine-SM550-1-PLKHG484UY5151/dp/B016EJ1WB6/ref=sr_1_9?crid=2HIER1OFRFGO7&keywords=consew&qid=1674043626&sprefix=consew%2Caps%2C210&sr=8-9

They're a servo motor meaning they have feedback and really delivery a lot of low end torque. They are tiny and I question whether its really 550 watts, but when dailed down they perform much than another setup I have with a 1.5hp on a VFD, to the extent that at the lowest speeds, one will do work and the other you can grab with your hand.

Seems a couple of years ago they were $110, but still, not a bad package at 230. You do have to do a bit of a hack to convert to a potentiometer control (its supplied with a lever intended for foot control, change is via positioning a hall sensor relative to a magnet ) and be careful, absurdly the control voltage is from the DC bus and is 160V....past that, they are solid little performers

 

[Susquatch]

I've been mulling over this for a while here. It's possible the software will think the unit is stopped if the magnet is in front of the sensor for too long or long relative to no magnet. So turning slowly keeps fooling the system. Maybe it counts up to 255 and if it rolls over to 0 then it assumes stopped and just starts counting again until it sees the magnet gone. Then it resets the RPM count to 0 and counts until the magnet is seen again. That would give wild values at low speeds.

With multiple tabs per rev the ratio between magnet there and not is smaller so it may not see the spindle as stopped.

I know it doesn't matter but it's a curiosity.

I'm glad you are curious. I do many many things out of curiosity alone. Even to the point of spending money I'll never see again.

Your explanation makes GREAT sense in all but one way. The readouts on the two units did not read zero. They typically read higher than actual.

There are two ways that rpm can be derived. One is to count pulses for a given amount of time and the other is to count time between pulses. It is as though they didn't stop counting the time between pulses.

In this latter scenario, your guess makes even more sense than in the other.

As you cause me to think about it more. This business of error detection creeps in too. As I recall, the chinglish in the description strictly talked about engine rpm. Nobody using them in an engine would ever see rpm below 150 or so. Huge engines do go much slower, but those characters would never use a $20 tach.

So ya, perhaps the signal error detection kicks in and resets or something and allows a higher count.

I'm gunna see if I can find out what I bought and get another one to take apart and play with. I think I'll get a 110V one too. It's way cheaper than flying south to go golfing.....

Edit - HEY. I just thought of something else. They were 12V units (which implies automotive). Clearly 110V units are not automotive!

 

[Rauce]

There are a wide variety of industrial counters/tachometers out there. They aren’t nearly as cheap ($200 is sort of the low end) but they will work with different types of sensors and you can calibrate them for any number of pulses per rotation up to a maximum frequency.

For example if I go this route for my lathe project I could have a proximity sensor read off of a gear on the spindle in the headstock.

A good option for fraction HP machines that need variable speed, is the little Consew sewing machine motor.

https://www.amazon.ca/Genuine-CSM1000-Machine-SM550-1-PLKHG484UY5151/dp/B016EJ1WB6/ref=sr_1_9?crid=2HIER1OFRFGO7&keywords=consew&qid=1674043626&sprefix=consew%2Caps%2C210&sr=8-9

They're a servo motor meaning they have feedback and really delivery a lot of low end torque. They are tiny and I question whether its really 550 watts, but when dailed down they perform much than another setup I have with a 1.5hp on a VFD, to the extent that at the lowest speeds, one will do work and the other you can grab with your hand.

Seems a couple of years ago they were $110, but still, not a bad package at 230. You do have to do a bit of a hack to convert to a potentiometer control (its supplied with a lever intended for foot control, change is via positioning a hall sensor relative to a magnet ) and be careful, absurdly the control voltage is from the DC bus and is 160V....past that, they are solid little performers

I’ve been considering one of these for a toolpost grinder build. The other option I’ve considered is a DC motor and power supply. Variable speed isn’t critical but a small/light motor with at least 1/4hp is what I’m after, and the faster the better.

 

[Mcgyver]

I’ve been considering one of these for a toolpost grinder build. The other option I’ve considered is a DC motor and power supply. Variable speed isn’t critical but a small/light motor with at least 1/4hp is what I’m after, and the faster the better.

Coincidentally, that is also a current endeavour. I've got a really nice Cincinnati internal grinding spindle I've made a mount for on the lathe. The consew style had quit low vibration but by playing around with some plasticine stuck on the pulley was able to get it silky smooth. My first attempt was with timing pullies as that was what the grinder had on it. That has proven a pain, that one that was on there was proprietary, so I went with XL and changed both. Wasn't really happy, got some vibration. Most was probably from my cast small pulley, but timing belts can create some vibration. I think i'll try again with shop made V pulleys and an O ring.

 

[PeterT]

@Mcgyver so you have a Consew motor? Does it look like this? If so, can you show your mounting steup? I was considering it for another future application. Not insurmountable but its single casting lug looks a bit different that typical frames, obviously intended for that specific application. Is the swivel for belt tensioning?

 

[Rauce]

Coincidentally, that is also a current endeavour. I've got a really nice Cincinnati internal grinding spindle I've made a mount for on the lathe. The consew style had quit low vibration but by playing around with some plasticine stuck on the pulley was able to get it silky smooth. My first attempt was with timing pullies as that was what the grinder had on it. That has proven a pain, that one that was on there was proprietary, so I went with XL and changed both. Wasn't really happy, got some vibration. Most was probably from my cast small pulley, but timing belts can create some vibration. I think i'll try again with shop made V pulleys and an O ring.

I was planning on O ring as well. I’m making the spindle using a straight shank 20mm ER arbor and a pair of ABEC7 angular contact bearings preloaded with Belleville washers.

I’m just acquiring parts at moment. I plan to line bore the spindle housing in place on my lathe and the lathe is still in pieces.

Vevor has servo sewing machine motors as well, and the price is a bit better. There’s also a 750w version.

VEVOR Servo Motor CS1000 Sewing Machine Servo Motor 110V 0.75HP 550W Servo Drive Motor Max Speed 4500RPM for Industrial Sewing Machine | VEVOR CA

Shop VEVOR Servo Motor CS1000 Sewing Machine Servo Motor 110V 0.75HP 550W Servo Drive Motor Max Speed 4500RPM for Industrial Sewing Machine at lowest price, 2-day delivery, 30-day returns. Shop now at VEVOR.

m.vevor.ca

 

[Mcgyver]

Peter, this a hefty pivot between the base and motor. I forget the word for it, but there are radial serrations on pivot so when tighten its locked and doesn't slip. I've used them a few times now, some shots betlw that better show the motor

 

[Mcgyver]

Vevor has servo sewing machine motors as well, and the price is a bit better. There’s also a 750w version.

I used them three or four times and at least one wasn't a Consew. It did seem a bit less refined, and the Consew is bit down and dirty to begin with, at least electronics wise.