Looking to Sub a $225 encoder w/ another...

[countryguy]

Morning all. Nice to be back home! Back to the YCM-30 and conversion from 80's CNC to a newer setup.

Step 1) I need to find .375 hollow shaft quad/diff Line driver (+5v) style. Not much luck in a cost effective line unless you know of something?

Step 2) the automation direct is hollow shaft or solid. I could do .375 solid and then just use a female to female .375 to join them. Also finding they are all 6mm to 8mm will need to join thse to my shaft.... Anyway--- Easy stuff there....
LINKS:

A) http://ecatalog.dynapar.com/downloads/F18_DS_702368_2_B.pdf
http://www.ajaxcnc.com/wp-content/uploads/2011/07/comp_encoder_large.jpg

B) http://www.automationdirect.com/static/specs/encoderld.pdf

Step 3) electrically, Anyone see any major diff w/ the automationdirect Item -vs- the dynapar F18. The electrical output of the F18 is twisted pair on a 16pin sip... -vs- 9pin on the ad unit. I'll need to check what connects up to the ajaxbox but I think it's only 1 line of the twisted pair ( they use the F18 and it's cost is $$2x more than the ad unit.) trying to move the costs downwards if possible. It's $375 in savings if I can sub the encoder to under $100.

 

[JimDawson]

Step 3) electrically, Anyone see any major diff w/ the automationdirect Item -vs- the dynapar F18. The electrical output of the F18 is twisted pair on a 16pin sip... -vs- 9pin on the ad unit. I'll need to check what connects up to the ajaxbox but I think it's only 1 line of the twisted pair ( they use the F18 and it's cost is $$2x more than the ad unit.) trying to move the costs downwards if possible. It's $375 in savings if I can sub the encoder to under $100.

It looks like the F18 unit is way overkill, I don't even know what all the extra wires are for. The inputs to the ajax box are A, /A, B, /B, Z, /Z. The AD encoder has these outputs, and meets the other specs.

 

[countryguy]

Thanks Jim! :thumbzup3: I thought so too. twisted pair on the output is all I could find on the doc. so I suppose noise immunity ? The savings are indeed nice. I'll get into the print today and get this show on the road w/ an order. (crazy me.... More stufff yet again! LOL> but it's going to be a grrrreat Mill when we're done. Jeff and Mitch.

It looks like the F18 unit is way overkill, I don't even know what all the extra wires are for. The inputs to the ajax box are A, /A, B, /B, Z, /Z. The AD encoder has these outputs, and meets the other specs.

 

[JimDawson]

By it's nature, the line driver output is pretty much noise immune. I have had problems with encoders that only output A, B. But have never had a problem with a line driver output. Also note that a 1000 PPR quad/diff encoder is read by the controller as 4000 PPR because of the 4 channels.

 

[countryguy]

Kewl bean's (as someone once said???) I zoomed way in on the PDF of the F18 unit. There are Inverse outputs (not) as a line above the other signals. The objective is to get to 8000ppr via quad/diff on a 2000ppr encoder. The DSP unit will create the 8K as you note on the the channels note. I'm going ahead w/ the AD units.

I do have one question- What if you saw the 3500ppr units on sale. Would those be of any advantage? Is the 4 channel multiplier start to become of no use once you meet/exceed the limits of the screw turning anyway? Just curious. A lot of theory which does fascinate me... just so little time w/ the other full time job. :-0 Need to remedy that somehow? about 15years or so.

By it's nature, the line driver output is pretty much noise immune. I have had problems with encoders that only output A, B. But have never had a problem with a line driver output. Also note that a 1000 PPR quad/diff encoder is read by the controller as 4000 PPR because of the 4 channels.

 

[JimDawson]

With a 5 pitch lead screw, a 1000 line encoder will give you 0.00005 theoretical accuracy, if you have a 4 pitch lead screw, you would still have 0.0000625 theoretical accuracy. Either case would far exceed the accuracy of the lead screw, or your mill in general.

I normally use the lowest resolution encoder that will provide the accuracy needed. To get better accuracy, you need to have the position reader on the load, not on the driver, in other words, a linear encoder on the table. That could be any type of linear encoder, magnetic, glass scales, or one of several other systems.

Take a look at this thread: http://www.hobby-machinist.com/showthread.php/30957-linear-encoder?p=264769&viewfull=1#post264769

 

[countryguy]

2500ppr quad/diff's Was: ooking to Sub a $225 encoder w/ another...

Jim, I know that thread well! Tx for all those GREAT Pics too! I have several LM10 styles on the hook for us as soon as we get this beast all together. The Ajax has a extra input set for these scales I'm told. Not sure if there are 2 or 3, but it's on the list!

So- The automationdirect side went down for maint. last night.... and BOOM All the 2k ppr quad/diff encoders in my shaft solution were replaced w/ 2500ppr. So that's what I have now..... Either pay $$ or go w/ the 1/4" LineDriver (5V) in the 2500ppr output range. Nice price too! $97 -vs- $225 for Dynapar's What would you say to the 2500's over the 2Ks for a bit more overkill on the screw to encoder resolution.... :noidea: Waiting to order pending your review Sir. Funny to as I had my CC in hand last night when this all went down.

With a 5 pitch lead screw, a 1000 line encoder will give you 0.00005 theoretical accuracy, if you have a 4 pitch lead screw, you would still have 0.0000625 theoretical accuracy. Either case would far exceed the accuracy of the lead screw, or your mill in general.

I normally use the lowest resolution encoder that will provide the accuracy needed. To get better accuracy, you need to have the position reader on the load, not on the driver, in other words, a linear encoder on the table. That could be any type of linear encoder, magnetic, glass scales, or one of several other systems.

Take a look at this thread: http://www.hobby-machinist.com/showthread.php/30957-linear-encoder?p=264769&viewfull=1#post264769

 

[countryguy]

Re: 2500ppr quad/diff's Was: ooking to Sub a $225 encoder w/ another...

Just got a note back from Ajax on this Q:
2000 line (8000 ppr) is the min recommendation. I just installed 10,000 line quad/diff encoders on some sem motors that were on a bport ez trak. Works great. I



[

QUOTE=JimDawson;266226]With a 5 pitch lead screw, a 1000 line encoder will give you 0.00005 theoretical accuracy, if you have a 4 pitch lead screw, you would still have 0.0000625 theoretical accuracy. Either case would far exceed the accuracy of the lead screw, or your mill in general.

I normally use the lowest resolution encoder that will provide the accuracy needed. To get better accuracy, you need to have the position reader on the load, not on the driver, in other words, a linear encoder on the table. That could be any type of linear encoder, magnetic, glass scales, or one of several other systems.

Take a look at this thread: http://www.hobby-machinist.com/showthread.php/30957-linear-encoder?p=264769&viewfull=1#post264769[/QUOTE]

Jim, I know that thread well! Tx for all those GREAT Pics too! I have several LM10 styles on the hook for us as soon as we get this beast all together. The Ajax has a extra input set for these scales I'm told. Not sure if there are 2 or 3, but it's on the list!

So- The automationdirect side went down for maint. last night.... and BOOM All the 2k ppr quad/diff encoders in my shaft solution were replaced w/ 2500ppr. So that's what I have now..... Either pay $$ or go w/ the 1/4" LineDriver (5V) in the 2500ppr output range. Nice price too! $97 -vs- $225 for Dynapar's What would you say to the 2500's over the 2Ks for a bit more overkill on the screw to encoder resolution.... :noidea: Waiting to order pending your review Sir. Funny to as I had my CC in hand last night when this all went down.

 

[JimDawson]

The 2500 line should do it. I don't see a problem, based on what Ajax says.