Voron TAP comparisons

[ChicagoKeri]

Hi everyone!

Let's start a thread for comparisons of various Voron TAP solutions.

Here's my gaggle of TAPS....  a Mellow CNC TAP, a Chaoticlab CNC TAP and a Printed TAP, left to right.

 

I built my 2.4 with the Printed TAP, which has been working alright.  After a recent Plastic Blob of Death (PBOD) event, which severely distorted the plastic TAP front, I decided to try a metal TAP, as I was never happy with the magnetic holding action of Printed TAP and also nearly all of the bolts had loosened.  

I liked the Mellow CNC TAP, but the estimated arrival was over a month away.   So, I put one on order and ordered a Chaoticlab CNC TAP locally.

Well, surprise, a few days later they both arrived on the same day!  So a comparison is in order.

Printed TAP.  This is my baseline.  Printed TAP R6 has been working well, even as most of the bolts had loosened over time. First layers are better than my PRUSA Mk3s+ had been until I sold it. Printing a new Front after the PBOD and tightening all the bolts surprisingly made little difference.  

I am not fond of the magnetic holding action of R6 TAP, nor am I crazy about the use of relatively soft ABS in the critical Extruder-to-X Carriage mount, but it has been reliable with Quad Gantry Leveling usually completing within 2-3 passes and with good first layers.  By the way, Printed TAP is about 18mm thick in the Y axis, shifting it towards the front of the printer about 3mm, which can cause the Stealthburner to hit the doors.  My R6 Printed TAP weighs in at about 85 grams as pictured, including sensor.

Mellow CNC TAP.  This is a beautifully made and packaged CNC machined aluminum version.  It is dimensionally very similar to Printed TAP, uses the same Linear Rail, a very similar magnet arrangement and the same optical sensor.  It comes with the sensor and necessary wiring and connectors, which is compatible with 5v and 24v.  The magnetic action is very much stronger than my R6 Printed TAP. Mellow CNC TAP is also 18mm deep in the Y axis, similarly shifting the Y axis forward.  Mellow TAP weighs in at about 75 grams, as pictured.

Chaoticlab CNC TAP.  This is a very nicely made CNC machined aluminum version of TAP.  Chaoticlab has changed the design considerably, using a smaller Linear Rail and a non-standard microswitch style sensor that is 5v ONLY.  It also uses a single magnet working against the Linear Rail instead of the Printed TAP dual magnet arrangement. The magnetic action is much less than that of the Mellow but considerably more than my R6 Printed TAP.    

The radically different design is much thinner in the Y axis, at only 13.6mm, meaning that the Stealthburner will no longer strike the doors and also reduces the moment arm of forces experienced during rapid movements. If one intends to use the original X endstop switch, a small adaptor must be printed out and attached, as seen in my picture. Chaoticlab TAP weighs in at only 65 grams, including the X endstop adaptor, as pictured.  

Instructions for Chaoticlab CNC TAP are not in the box but are easy to find on Github.  Critically, one should follow the instructions on lubricating the linear rail before installation.

A side view of Mellow, Chaoticlab and Printed TAPs left to right:

 

So, in summary:

Printed TAP works and is the standard arrangement. However, it is less rigid and heaviest at 85g and can be molten during a PBOD event. It also shifts the Y axis forward a few millimeters due to its 18mm thickness. On the other hand, replacement parts or updates can be easily printed out and implemented. 

Mellow CNC TAP is an all metal adaptation that remains very close to the Printed TAP arrangement. It is more rigid and lighter than Printed TAP at 75g,  uses the same Linear Rail and Optical Sensor and has a much stronger magnetic holding action. It shifts the Y axis forward the same amount as Printed TAP.

Chaoticlab CNC TAP is a radical departure.  Like the Mellow TAP, it is all metal but takes advantage of this material to radically change the linear rail, sensor and magnet design and reduce weight still further to only 65g, including the optional X endstop adaptor. In addition, it is about 4mm thinner in Y-axis, negating the TAP shift forward as well as reducing the moment arm of forces acting upon it.

 

Right now, I have chosen the Chaoticlab CNC TAP to replace the venerable Printed TAP R6 on my machine.  I will post my impressions after some time getting to know it.

However, any one of these choices are good. They each have their own strengths and weaknesses.  

 

Happy Printing!

 

 

 

Edited October 1, 2023 by ChicagoKeri
mis-spelled "Chaoticlab"

[mvdveer]

Great write up and comparison.  Will certainly help community members in making an informed decision on which TAP to use.

[fedgez]

The y shift is why I have not gotten tap (among other reasons).  Nice to know there is a cnc part that avoids that issue.

[Sly]

There is an improved version 2 of Chaoticlab CNC Tab available. At the moment I'm using version 1. Working very well - no issues so far. Only during installation - the screws were very hard to fix. They fall out/down very easy. This is fixed with version 2. There the screws are fixed with tiny rubber rings. I bought V2 also and it arrived today. I'll install it in the future. They changed the voltage of the sensor from only 5v to 5v - 24v. Also, the behavior regarding resonance should be better. They switched over to MGNH7. More info can be found on their webpage. 

[7milesup]

Well @ChicagoKeri, because of your review I ordered a Chaotic Labs V2 Tap from West 3D.  Looking forward to installing it once it arrives.

[ChicagoKeri]

On 10/26/2023 at 10:13 PM, 7milesup said:

Well @ChicagoKeri, because of your review I ordered a Chaotic Labs V2 Tap from West 3D.  Looking forward to installing it once it arrives.

Cool!  Please let us know how it works... I feel that the V1 could use a more stable and stronger magnet arrangement.

[ChicagoKeri]

Well, my verdict is in on the ChaoticLab V1 CNC TAP.

Unfortunately, is is not a good verdict.

After only four weeks of use, with some high-ish temperature 5-8 hour prints, the ChaoticLab CNC TAP has been getting worse at Quad Gantry Leveling. Also, print quality has worsened, especially top layer on certain prints.

OK, so maybe some bolts have loosened, as they did on the Printed TAP.  So, apart it comes and all of the bolts are still tight.

However, the single central magnet has lost virtually all of its magnetism!  It will no longer even hold the microswitch sensor closed against its internal spring, with the TAP lying flat on its face.  Four weeks ago, I had noted that the single central magnet seemed stronger than my Printed TAPs two magnets. Apparently, the magnet has gone weak due to heat.

The environment to which it was exposed:

A closed Voron 2.4 300mm with a Nevermore circulating fan / filter.

Exhaust fan set at 0.2 speed, then disabled for the past week.

Hotend fan set to 0.8 speed

Printing ABS, with nozzle at 245-250C and bed at 90-100C

 

The Mellow Fly SB2040 CAN Toolhead Board, located near the TAP, reports 55-65C chamber temperatures during a long print, but it also reads 30-35C with the printer cold and the bed and hotend agreeing on 22-25C .  So, the Chamber temperature is probably a lot lower than the indicated 55-60C.

 

So, today in went the Mellow CNC TAP with no installation issues whatever.  Soon, I will be able to give some impressions of it as well.

 

 

By the way, there are "Community Mods" available on ChaoticLab's GitHub  that include an add-on dual magnet arrangement for the V1 CNC TAP.

https://github.com/Chaoticlab/CNC-Tap-for-Voron/tree/master/Community Mods

As this is a manufactured product, it should't depend upon community fixes to work well, but at least there are some, and at least they put them on the github page...

 

[ChicagoKeri]

Some Resonances Data from all three of my TAPS:

All Resonance tests were performed shortly after installing the TAP, so bolts are all tight and everything is not yet broken in...

 

Printed TAP.

 

 

 

ChaoticLab CNC TAP V1

 

 

Mellow CNC TAP

 

 

It seems to me that the largest difference is that the Mellow CNC TAP has a much smaller reading for Power Spectral Density in the Y Axis,  notably 1.4 compared to 5.8 and 6.6.  The X Axis readings are much closer.  

 

 

Edited October 28, 2023 by ChicagoKeri
image resize

[eka157]

I have a v1 chaotic labs too, recently it started struggling to QGL and I narrowed my issue down to too much resistance in the linear rail. I took the rail apart and fully greased the balls and it's been much smoother and QGL works perfectly again. If you attempt this, be warned that it was pretty difficult to get the carriage back on without the balls escaping but it is possible. 

Edited October 29, 2023 by eka157

[7milesup]

On 10/29/2023 at 8:16 AM, eka157 said:

I have a v1 chaotic labs too, recently it started struggling to QGL and I narrowed my issue down to too much resistance in the linear rail. I took the rail apart and fully greased the balls and it's been much smoother and QGL works perfectly again. If you attempt this, be warned that it was pretty difficult to get the carriage back on without the balls escaping but it is possible. 

When I purchased the initial TAP setup, I believe it came with a plastic linear rail section that you bumped up against the metal linear rail and then slid the carriage off of the metal one onto the plastic one to keep the balls contained. 

[7milesup]

Ok, one issue with the new Chaotic Labs TAP.  The X-axis end stop switch is no longer activated with the Chaotic Lab.  My current setup has the X and Y axis end stops on one circuit board that is contained in the right corner carriage.  The Chaotic Labs one wants you to mount the X-axis end stop switch on the TAP whereas it will bump up against the right carriage.  I will have to come up with plan B for activating the X end stop.

[ChicagoKeri]

2 hours ago, 7milesup said:

Ok, one issue with the new Chaotic Labs TAP.  The X-axis end stop switch is no longer activated with the Chaotic Lab.  My current setup has the X and Y axis end stops on one circuit board that is contained in the right corner carriage.  The Chaotic Labs one wants you to mount the X-axis end stop switch on the TAP whereas it will bump up against the right carriage.  I will have to come up with plan B for activating the X end stop.

Would this XY switch adaptor help?

https://github.com/Chaoticlab/CNC-Tap-for-Voron/tree/master/STL/XY_Switch_Adapter

[mvdveer]

21 hours ago, ChicagoKeri said:

Would this XY switch adaptor help?

@7milesup - It certainly would - have used that on two of the conversions before converting to canbus. Works great

[7milesup]

5 hours ago, mvdveer said:

@7milesup - It certainly would - have used that on two of the conversions before converting to canbus. Works great

 

On 11/3/2023 at 6:35 PM, ChicagoKeri said:

Would this XY switch adaptor help?

https://github.com/Chaoticlab/CNC-Tap-for-Voron/tree/master/STL/XY_Switch_Adapter

Awesome.  Thank you for the link and the info.  Now, I have to "fake" the x-axis endstop since I already took the original off and installed the Chaotic Labs one.  Also, the connector for the 2-piece Hartk PCB is different than the Chaotic Labs connector, so I will need to change that out.  Hopefully, I will get to this tomorrow.

[7milesup]

I was able to fabricobble a Y-axis endstop extension, so that now works, BUT...

I cannot get it to do a z-axis limit probe.  The problem is I get the error "Endstop z still triggered after retract" message.  I am wondering if the magnets are not allowing the TAP to return to its resting state.  I am getting a blue light most of the time (open state). 

My setup is a Hartk 2-piece toolhead and all of that is wired stock, i.e: No Canbus.  I am utilizing PG15 on the Octopus board but I did have to change it to !PG15 for it to move in the correct direction, which I found odd that it would be different than the original TAP sensor.  I have messed with the "Sample Retract Distance" by changing it anywhere from 2.0 to 15.0 but it does not seem to make a difference. 

 

 

[mvdveer]

1 hour ago, 7milesup said:

"Endstop z still triggered after retract" message.

A few of things come to mind:

1. Wiring - a bad crimp or reversed wiring - e.g. pin swapped for either gnd or power, or power reversed.

3. Look at the printer.cfg [probe] section and see if there is a ^ before the pin. If so, change this, if not add one.

3.  A problem on the board connection - is it wired to the HartK tool board or directly to the octopus board?,Try a different pin is the voltage set to that of the TAP probe 5V vs 24V? If it wired to the tool head board(hartk), try running the wires directly to the Octopus

 

The LED should alter before blue and red. 

[7milesup]

I think I figured out why it is behaving the way it is.  It is a mechanical issue.  Now, to figure out how to correct it.

The cause for my issue is that either the magnets are not allowing it to settle back to its normal state, or the sensor is sitting just slightly low in its mount, thus preventing the sensor from going into an untriggered state.  Unfortunately, I will have to disassemble the dang thing to determine the root cause. 

@mvdveer I am curious about your suggestion in line 2 (Ok, it actually says line 3, but I know what you meant).  What does that caret do in the probe line?  So you are suggesting  "pin: ^!PG15" instead of "pin: !PG15?" 

[mvdveer]

1 hour ago, 7milesup said:

What does that caret do in the probe line?

It enables or disables a hardware pull up resistor on the probe.

 

[mvdveer]

1 hour ago, 7milesup said:

"pin: ^!PG15" instead of "pin: !PG15?" 

It is worth a try yes. But if you think it is a hardware problem where the probe is not triggered, this won’t make a difference. If you think the probe does not engage on the sensor, it can be tested to see if you physically move the probe whether it changes the colour of the LED. Or take the PCB with the switch off and manually trigger this to see if it functions as expected. Chaoticlab uses a micro switch rather than an optical sensor, as you know.

[claudermilk]

Isn't the pull up resistor on the Octopus?

I feel your pain on the mechanical issues. I struggled with that when installing Tap myself; I had the earlier beta parts printed with the slightly too short trigger tab. Gluing a piece of filament on top and only using one mount screw on the wired sensor I was using at the time finally resolved it. Now that I have the latest printed parts and the current 5-24V PCB sensor it's all working as it's supposed to. I also have the hartk tool head boards and Tap is wired into that.

[mvdveer]

8 hours ago, claudermilk said:

I had the earlier beta parts printed with the slightly too short trigger tab.

Had the same issue with the printed parts thought the Chaoticlab TAP, uses a micro switch, not an optical one - so this should not be an issue here.

@7milesup, does the LED colour change when you manually depress the micro switch when it is wired and off the machine?

[7milesup]

@mvdveer

I do get a positive indication of the switch being open and closed with some help.  As previously stated, I need to provide a slight amount of downforce when it probes the z-axis.

That being said, I have the V2 version of the Chaotic Tap, which does not have a mechanical microswitch but rather an optical microswitch. 

I will be able to get this solved, but my time has been very limited the last few days.  It stinks that I have to take the print head apart again, mostly because I had already trimmed the belts thus making the belt attachment to the Chaotic Labs Tap rather time-consuming and fiddley. 

[7milesup]

Well, I effed up my new Chaotic Labs Tap by modifying it.  LOL.  Oh well, I will probably have to order a new one, but it is definitely a mechanical issue and not software, which I regard as a plus.

[zogzogzog]

I ran a V1 CNC Tap and wasn't happy with it.  I needed to shim the switch down about .007" to get reliable actuation at all (I was getting only like 3 out of 5 asctivations with a manual full travel move of the tool head and that isn't going to work) but the activation (when it was working) wasn't really consistent enough for this application.  I also felt a little more magnetic tension would be a plus but I can't say that it made a difference overall.

I went with a V2 and I'm super happy.  The V2 uses an optical switch like the Voron design and allegedly, badnoob got involved in giving them some feedback (which if they listened, could be a good thing IMO).

The V2 has been super solid for me.  I've got first layers that make you cry for joy - every single time - without fail.  It seems to be plenty consistent; I'm getting probe accuracy test runs with SAMPLES=100 with std dev in the .0004 range.  I like the two color LED indication where blue is not activated and red is activated - a beef with the V1 was no indication of power so when I was debugging the fact that it didn't seem like it was triggering (because it wasn't) - it just stayed dark.

I ended up making non-trivial changes to my slicer start script (I use S3D mostly) and ended up splitting PRINT_START into two sections due to the nozzle temperature handling for probing.  I also do a major retract and nozzle wipe as part of PRINT_END before I turn off the heaters to keep the nozzle clean.  I worry about oozing impacting my Z zeroing ad this has worked really well for me.

In my experience, the general Voron Tap design together with the CNC Tap V2 from Chaotic has made my V2 much better.  I'm super happy with how things are running now - better than ever.  About 1,500m of printing joy so far with this upgrade and it's been great.

 

V2.2874

[zogzogzog]

Actually to be more specific, my V1 CNC tap would not settle fully and hence would not "untrigger".  Same issue as @7milesup.

It's been while since I messed with it and got my polarity reversed on that.  That's why I needed to shim the switch *down* to trigger it reliably in the lowest position so that part of my memory is correct.  The issue I had is not reliably UNITRIGGERING.  It thought the Z endstop was triggered even when the Tap carriage was in the lowest position about 20+% of the time.

Sorry for any confusion.

 

Edited November 25, 2023 by zogzogzog