Forum Replies Created

Viewing 3 posts - 1 through 3 (of 3 total)
  • Author
    Posts
  • in reply to: Scaling Issue #1224

    Ashok Mathur
    Participant

    Hello All,

    I visited Fab-lab Kothrud today and looked at the problem of scaling Hackberry hand. The issue is, we have the CAD model with certain(79) palm width and now we want two additional proportionately scaled hands with palm different widths (eg 85 and 90). We can think of following approaches:

    1. CAD scaling: Scale the entire assembly. This will scale the holes, springs, shafts, screws as well, which we do not want. To bring them back to the original dimensions, idea was to suppress these features before the scaling and un-supress them after them scaling. But after inspecting the part feature trees, I found that these features are not separate ones, but part of extrude sketches. So they cannot be suppressed independently. Editing dimensions of just these sketch holes, would be very tedious and highly error prone as part may not update well. Also, the original modeling looks non-conducive for such changes. Thus, this option looks cumbersome, labor intensive, time-consuming and error prone.

    2. CAD scaling + hardware adjustments: Scale the entire assembly. Let the holes, springs, etc scale. But then, by drilling or by putting padding, bring these holes etc back to the required dimensions. Need to evaluate this option thoroughly, because there are certain dimensions like Servo motor screw alignments, which can not be corrected by any post processing. We will need to edit relevant parts to bring back the original dimensions. Thus, this option is also cumbersome and error prone.

    So, unless we have readily scaled assembly files available from the original creators, I am not seeing any quick and full-proof solution as of now. Can some one ask at the forum, if these creators have readily scaled CAD models?

    Thanks
    Ashok
    (For Yogesh)

    in reply to: Scaling Issue #1222

    Ashok Mathur
    Participant

    Dear All

    Yogesh and I are wanting to scale the Hackberry Hand so that we can best meet the needs of the Indian population.

    We have no experience in such an endeavor and therefore welcome any help/guidance in any manner that any one can give.

    From the reading of the literature, we find that palm width and finger lengths are closely correlated and using of palm width as the basis of scaling of the hand may be appropriate. If that is not the case, please let us know what is an appropriate measure for scaling.
    We have measured the palm width of the existing Hackberry hand as per Attachment 1.
    This measure came to 79 mm.
    From the survey of literature on anthropological data for the Indian population, we found the following table (Attachment 2) to be relevant.

    The above table shows that the present Hackberry hand has the same palm width that 50 percentile of Indian Women, but is very much smaller than 50 percentile of man’s hand whose palm width is 90 mm.

    In the long run, we will need to scale the Hackberry hand to the following sizes to meet the needs of the Indian population.

    Size (in mms) 5 percentile 50 percentile 95 percentile
    Men 85 90 98
    Women 71 79 85

    Let me explain as to why we are not considering continuous scaling fitting each patient with the hand that fits his size. Most of the patients come from outside the city where fitting takes place.
    In a system where the hand is manufactured to fit that person’s exact size, there is a process of about 7 days between measuring his hand size and 3d printing, assembling the hand and then testing it before fitting it. During that time the patient has to stay in the city at his own expense.

    But if we can categorize his hand size in anyone of the five sizes given above, then a pre-printed and pre-assembled hand can be fitted the same day saving him considerable costs with similar performance of the hand.

    At the moment we are trying to scale the Hackberry hand to 85 mm and 90 mm palm width and do the other two palm widths at a later date.

    I am aware that CAD software from Adobe family store individual elements in a database that is directly addressable by any language of the Visual Studio family. If any one can guide us to Open Source software where we could import the Inventor file and have continuous scaling, that would be great.

    For sacking we are planning to be the following process; If you can suggest ways to improve it, you are welcome.
    We study all the punctures on the surface of a part and will decide if we should erase a particular puncture as we do not want to scale it ( for instance a hole for a 2mm rod).
    A careful list will be kept of the parts that were erased as they have to be restored at appropriate location later on.
    The entire drawing will be scaled.
    For most parts there is no point in increasing the thickness of various parts( difference between the outer and inner skins). So the thickness will have to be restored to the one before.
    Erased punctures will be have to be restored with old dimensions but new locations.
    The middle finger and the last two fingers are joined by a bar (linking mechanism). Its new length will have to be measured from new drawings.

    If you can assist us in any way, we will be grateful.

    If you will examined the scaled drawings and point out any improvements/flaws that we have made, we will share the files with you.

    Later on the scaled images will be shared with everyone.

    Regards

    Ashok

    in reply to: TPR-105 IR Sensor in the US #787

    Ashok Mathur
    Participant

    1) I understand that initially hackberry started with two sensors i.e. GL55 photo resistor and infrared sensor pj835.
    Then it has switched to a single sensor TPR 105 or its equivalent.
    Can you explain in few words why this change was made?

    2) When the one second calibration is performed, what values are stored in the memory?

    3) Would adding a pre-amplifer to the output of TPR 105 before Arduino processes it, help?
    We are building a hackberry hand.
    Thanks
    Ashok

Viewing 3 posts - 1 through 3 (of 3 total)