CMC/Harken Clutch and CMC MPD are very different devices. MPD weighs and costs about 40% more and clearly has different markets and usage. But there are similarities, as well. And I’m able to study both.
My method of measuring pulley efficiency has no problems with PCD devices like these, since it does not require the pulley to work in both directions.
The sheaves
The sheave dimensions of these devices are very similar, but Clutch has notches in the groove walls. The groove also seems a bit deeper and steeper. I guess these sheaves are designed a bit differently from typical pulleys in that the rope is not intended to rest against the groove base, but the contact is similar to a v-belt drive. Anyway, these sheaves need to have some friction against the rope in order to activate the braking mechanism. The effective sheave diameter is about 55mm in both cases.
The braking mechanisms
This is an interesting difference between the devices. Both are based on some kind of ratchet mechanism that blocks the sheave in other direction, so that the pulley becomes a bollard or descender bar.
In MPD, the braking mechanism is a cam that rotates along with the sheave, if the sheave is forced to rotate in the capturing direction. Thus, the brake cam tilts and squeezes the rope against a counterpart.
Of course this alone does not apply a sufficient braking force, but it merely acts as an initial “holding tension” which is magnified by the capstan drum that the now blocked sheave constitutes. This is described by the capstan equation.
Clutch works in a similar manner when it comes to the capstan effect, but the initial clamping is caused by a different kind of mechanism.
There’s no cam, but the entire sheave tilts towards a brake pad, eventually squeezing the rope. This is again due to moment caused by the rope trying to rotate the sheave in the blocking direction. The deviation at the tip of the device adds leverage and probably magnifies the effect. Notice the green arrow’s direction compared to vertical.
The deviation also increases the angle swept by the rope around the capstan. This is a similar solution to Wild Country Revo, in which the rope sits in an omega like shape. This is most likely to add the capstan angle and to make sure the rope will bite into the groove as fast as possible.
Test results and discussion
I measured both devices three times at 1kN output force. The rope used was once again Petzl Grillon rope. The results:
To be honest, I was not surprised by the difference. I’m quite sure the deviation has a lot to do with it. But there’s another candidate for a partial explanation that cannot be simply dismissed: when used as a pulley, the resultant of the two force vectors still tilts the sheave against the brake block:
That is, the rope drags against the brake surface all the time, if there’s any tension on the load side.
Conclusions
Any of this should not be taken as there’s something wrong with Clutch. As said, it’s one third cheaper and lighter and otherwise more like a descender. It’s not “only” 75% efficient; as a pulley it’s in completely different category than any of the conventional descenders.