The Graham Escapement

... for almost 300 years it has been the horologists favourite toy. The reason being is, that even though it has got a simple construction it is so outrageously good. And for this outrageousness to really show up it needs to be understood and used properly.

And now?
Let us pretend that we have got an escapement in front of us that we do not trust an inch. Which means, that we are suspicious and prepared for every dirty trick. Now it should be our aim to adjust the escapement thoroughly; adjust it in such a way that it runs well again. Outrageously well.
Checking the pitch and truth in the round of the escape wheel:
(this fault cannot be simulated above)
Maybe this is a bit of an exotic fault, but when it occurs one is left looking a bit stupid. To discover the fault we act as follows: tooth by tooth we move the escape wheel through the anchor and by doing so we keep an eye on the drop of the entry pallet (respectfully exit pallet). During one circulation of the escape wheel it has to be absolutely level - thus, each tooth. If it is not, than there is a fault in the pitch which leaves us with either changing or re-working the escape wheel or we simply have to live with the fault.
Let us move the teeth through again and observe, how far the tooth of the escape wheel falls on the rest at one side (either entry or exit). Then we can check the truth in the round. If this fault shows up it is not as tragically as with the fault in the pitch. It can actually be worked over again.
Checking the thickness of the pallets:
(this fault cannot be simulated above either)
This fault also belongs to the nasty type. Therefore, it occurs rather seldom and if it does show up it is the evidence of ending common decency of horology. Here we measure the width of the entry and exit pallet only to discover with delight that it measures the same. If the delight fails to materialise we just have to make sure that the pallets measure the same width. Or we simply fight it out with our conscience and just leave it as it is.
Checking the angle of lift:
The angle of lift at entry and exit pallet should be equal. To catch on to this fault we simply have to lead a few teeth of the escape wheel through the anchor and observe if both of them fall on rest (this works rather well in the model shown above, if the rest is set to almost zero). Now even the slightest tolerance is easily to be noticed. However, in case of the rest not being equal on both sides we have to work over the angle of lift of one pallet so it measures the same as the other one.
And to anticipate all prophets that might want to say: "Yes but, these faults could also occur due to tilted pallets or the wrong distances of the pallets from the pivot to the anchor..." I do have to agree to it, and ask for a mild verdict.
Adjusting the distance between anchor and escape wheel:
Now we will lead a few teeth of the escapement through the anchor again and observe if the drop of the entry and exit pallet is equal. If it isn't we will change the distance between anchor and escape wheel. If after adjusting the drop on both sides is too little the pallets must be too wide, if it is too big then they are too small. But this should not occur very often.
Adjusting the rest:
Due to moving the pallets we can now adjust the depth lower for the tooth of the escape wheel falls on rest far enough. Not too much and not too little. Here, among other things, it depends on the quality of the escape wheel and the conditions of putting up the clock. (Vibrations and stuff like that require more rest from the escapement for the clock to run smoothly.)
Adjusting to beat:
In the model shown above the anchor can be turned nicely and therefore the beat can be adjusted nicely, too. On both, the entry and exit side the beat should be equally far away from the position of rest of the pendulum.
If we act in the above mentioned order, than we should have pursued the way that enabled us to discover all mistakes and to eliminate one after the other without having to touch the thing twice. Of course it is a different story that the entire escapement has got to be in top mechanical order. (condition of the impulse plane and locking plane and so on). But every theory is grey. Therefore, do not show any false inhibition, but press the random button and play until all of it works almost by itself.