So I began to build my Lynxy. (Final will be Hex Extreme 3 with robot arm). First step was the gripper. All went fine, but I think the securing of the gripper axes in step 14 to 17 does not work. The acorn nuts will not secure the axes, either they are too tight or to loose. IMHO two interlocking standard nuts (counternut) would much better serve this purpose. I will replace all the axes with 3mm metal screws and the appropriate nut/countenut combos. ( 
âŠwhen will you folks go metric finally⊠yes, Iâm from Europe ). I prefer metal to nylon anyway, nylon has a very nice black look but the threads are often of minor quality. Of course, weight might also be a consideration. Another mod I made was replacing the HS-85BB servo with a HS-85MG.
to be continuedâŠ
Sounds like fun! 
Yes, the decision to use nylon hardware on the gripper was mainly a weight issue. The arangement should work well, but Iâm sure there are tolerance issues that effect this. Personally, I have used a tiny drop of glue to the nuts when I find a finicky connection. I will consider changing it to metal hardware. We are using some metric hardware, although a complete conversion will be difficult. I appreciate your feedback.
Sorry, untat, but we Americans are too pigheaded to change to metric.
Why measure with a simple, streamlined system of decimals, when we can blunder around with complicated fractions?
I personally prefer to see things like 15/64".
I mean, everyone can convert 15/64 into decimal form in their head, canât they?
The gripper is finished!
Finally I decided to use M3 nylon screws, because I had no aluminium screws available and steel or brass seemed to heavy.
But I put 2 metal washers on the top and the bottom and secured all axes with nut/counternut combos (also nylon) and (very few) loctite between them. This and the broader head of the M3 screws resulted in much more stability of the gripper. The arm is also finished, the yellow side fits not very well to the servo hinges, it is very tight so this part is under pressure and bends at both ends. I fitted HS-5945 digital servos for higher torque, they are about 17g heavier (56 vs. 39g) but torque is 128 cNm vs. 54 cNm, a big difference. I will use the same servos for the double shoulder joint at the base, there I think the better parallelism due to the digital design will be beneficial.
The arm parts are designed to fit the HS-475 servo specifically. The fit is overly tight because you are using different servos. I could make the parts fit the slightly larger servos at the expense of making the assembly overly loose for the HS-475 servo. Because 99.9% of the people building this arm will not swap the servos out for larger ones I will not make any changes.
From the sound of it, you should be building with the Servo Erector Set. It is designed specifically for the âpower builderâ with a no compromise approach for sturdy precision assemblies.
The arm is finished completely now. It is very nice and working well
The bending of the Lexan(BTW is this the same a Makrolon?) parts could be very easily remedied with some careful file strokes, no problem with that. BTW, the HS 5945 Servos are exactly the same size than the HS475.
The tape connecting the hinges with the servos is a bit thick, I substituted a thinner one which fits perfect now.One really has to be very careful with those servos, they are very fast at full speed and when they reach their end positions could break very easily because of the hard break. I would suggest putting in a command in the firmware of the SSC-32 to slow down all those movements. This would be very useful for testing and adjusting the mechanics. The elbow servo is strained a lot, it makes very loud buzzy sound, I have to look to relieve that a little, with some springs or counterweight I suppose. Another important issue is the gripper. The servo gear of the HS81 is very weak and I have already broken one. If you need to pick up several objects of different sizes, there has to be a way of finding out when the grip is closed and cannot go any further. This can only be done with a sensor for the force of the grip closing or a switch of some sort. Iâll try to find a good solution for that.
Meanwhile, building the hexapodâŠ
Lexxan is a polycarbonate.
I donât know about the plastic that you mentioned, though.
There is another way â perhaps a better way, for your needs â to detect the position to stop your handâs grip pressure.
Check out Mike Thompsonâs Ultraservo kits.
Or, you could take a look at andylippettâs website.
Both of them use the servosâ internal pots to measure the strain that the servo is encountering, as well as to detect the actual position of the servo.
I believe this would be more helpful than a pressure sensor (which you can find at phidgetsusa.com).
Oooh kay, but not. The base of the 475 is 1.52" x .77". The 5945 is 1.58" x .79". We make the parts to fit snugly⊠60 thousanths is enough to make it a squeeze.
Again⊠The height of the 5945 is 45 thousanths taller than the 475âŠ
Arghhhh! A command to slow the servo speed down. Hmmm⊠You think? Well you could probably use the âSâ (speed) command, or maybe the âTâ (timed) command. Sorry if Iâm missing something here.
This is documented in the assembly informationâŠ
The gears can be replaced with metal gears if required. The grippers are fine as long as you know the size of the object you need to manipulate. We encourage modification to incorporate feedback into the gripper. Let me know if you are sucessful with that.
Thx Jim for your comments,
I can agree with all of them. On Hitecs German page all dimensions of their servos are given only within 1mm accuracy, so youâre right on this point.
What I meant with the slow down command, was a global parameter, so you donât have to put the S and T commands during testing, but thatâs only comfort.
I have ordered a 5995 Hitec robot servo for the elbow, that should be strong enough if I got my maths right
The Lynxy is nearly finished now, except for the clean wiring of all those servos. I had to replace all the digital servos in the robot arm with the standard HS475BB, the reason was that the digitals (3* 5954 shoulder and wrist, 5955 elbow) where much to noisy. It was impossible to get rid of this whining sound when the servos came under load, even only a little.
SO those digital servos seem not to be the right choice for a robot, even for their much higher torque. Electronics will be the SSC-32 or the ServoPod or a combination of the two, I have yet to decide. Perhaps also one of the two and a totally different Atmel based board for the brain.
Also planned is an easy radio transceiver to the PC.
Iâm still looking for a good strategy for the startup sequence. All servos are power down, Lynxy lies on his body with all legs up. Servo comes to live, go to position ??? and then Lynxy should rise slowly. Have you done such a startup already?
Right Nick! Instead we have both American AND Metric! My toolbox has both American and Metric tools! Arrrrg!
I see it this way:
Metric for small sizes/measurements
American for large sizes/measurements
Digital servos whine when under load. This is normal and thereâs not much that can be done about it.
Read this tutorial to help you decide what to use. Although it uses a Madcatz wireless game controller as the remote control device, it should be easy to edit the program for serial control of the robot from a wireless link to a PC.
lynxmotion.com/images/html/build097.htm