Hi @rac,
I use the EAF on a C8 exclusively for astrophotography and find that if I maintain the same back focus to the camera I don't need to move the focus very much at all. Even when I add a focal reducer, it's pretty much all in the same focuser range.
Are you a visual astronomer? Do you need to move the focus over a larger range than 64k steps?
I'm not against the idea, just interested in why.
Steve
I do astrophotography too, but the different optical configurations that I use require the mirror to be moved through the range of 15 to 38 turns. In fact a 3:1 ratio, whilst not covering the full 45 turns, would meet my requirements.
I use HyperStar (f/2), 0.5x reducer (f/5). 0.63x reducer (f/6.3) and 2.5x Powermate (f/25). Most of these are in the in the 15 to 25 turns range and I could just about get by with the EAF covering that with one operating range. The 38 turns comes in if I want to use 0.63 reducer with a cooled camera because that requires the introduction of a 2" diagonal to avoid the risk of backend collisions.
At the moment I have the Celestron Focuser which also has the 65,000 limit but it has 1,000 counts per turn, potentially 60 turns, so it handles the full 45 turns easily. The problem that I have is that it tends jam, spoiling what is otherwise a pretty satisfactory solution. This I believe, is caused by the fact that it bolts rigidly to to the rear cell of the OTA. The EAF has a flexible coupling and of course a belt drive would be similarly flexible.
The higher top speed (be it 3x or 4 x) would also be appreciated when moving between say 15 and 38 turns.
EDIT: A 3:1 with the upper limit at 45 turns would provide 12 to 45 turns, comfortably covering all the focus positions that I have ever needed.
OK, that makes sense.
I was thinking visual observer because I thought you must be swapping a diagonal in and out. 
@tech@zwo#47881
To summarise my earlier posts. When using a focus motor to move the primary mirror of an SCT the requirements are different than those of a Crayford type focuser, longer travel distances and faster rotational speeds are appropriate.
@tech@zwo#47881 Ok, so let me ask what is ZWO’s solution to the fact that the entire travel range of the EAF only covers about 15-20% of the range of a typical SCT. Is going into the EAF settings to reset the step count 5-6 times during a refocusing session because of a configuration change ZWO’s permanent long term solution to this? If so there should be some kind of note added to the sales page so that SCT owners are aware of this and can make an informed purchase.
I think the long term solution is for ZWO to figure out a way to give the EAF a 300k-500k step range, or just have a “turn until I tell you to stop” manual option , or offer some kind of (affordable) gear reduction add-on. My two cents.
The solution suggested but WaterT is one that I have considered too. It represents a reasonably simple software solution. If we were to call each passage through 65,536 counts a "Zone", it would be helpful if there was an additional Zone parameter defaulting to 0. When the count reached 65,535 fir the first time, the next count would be 100,000 instead of 0 and so on. This would avoid having mentally keep track of the number of resets.
The only problem remaining would be the end to end travel time. I have been told that the EAF completes a revolution (turn) in 12 seconds at high speed, 5 RPM. For 45 turns on my SCT that would be 9 minutes. That's longer than I would like, but could live with it if the "Zones" were counted. By comparison the Celestron Focus motor takes 2 minutes and a slower focus motor that I once used took about an hour.
I agree with WalterT's comments but should point out that it requires speed multiplication not speed reduction. I agree with his point that this issue should be made clear in the product specifications so that SCT owners could make an informed decision. I only discovered when I downloaded (a read) the Manual, after placing an order.
I forgot to mention the simplest software solution, change the position counter from 16 bits to 32 bits.
Another thought, the existing, direct coupled SCT bracket, could remain on offer, with a note regarding potential speed and back-end clearance issues, and a new timing belt system could be offered as an alternative. I suspect that it would be more expensive. I think that a software fix is required regardless.
rac Rac, I agree with all your points. ZWO needs to come up with a whole SCT focusing solution. One thing about the EAF right now is that it’s slow and it doesn’t have enough step range or travel range. And as slow as it is now it produces terrible mirror shift when Autofocusing, causing the stars to appear way larger than they are for a second or two, so addressing your issues (which I hope they do) will still leave the EAF unable for auto focusing, which is the only reason I, and I suspect a lot of other people bought it for. Based on your OP I don’t know if you are using it with an ASIAIR or not, if you are you will discover that the Autofocus doesn’t work because of the mirror shift introduced.
I think ZWO should invest on a used SCT telescope and conduct some real world tests on the EAF, OAG, and ASIAIR and other ZWO products and so that they can see how they actually perform. Increasing the steps, making it faster, or increasing the turn ratio would be an empty victory if the thing can’t actually be used to auto focus with the AAP if it is not also updated some how.
I did suggest that they implement a “wait” period before taking each AF measurement in the AF routine, like NINA and other programs do, to allow for the mirror shift to settle down before the star measurement is taken (which could be an option so that refractor users don’t need to enable it) but I don’t think that will ever happen.
Not trying to hijack your thread, just wanted to add that they need a whole SCT / RCT focusing solution which entails not only the focuser range and speed but also addressing the mirror shift that IT creates.
rac I have been told that the EAF completes a revolution (turn) in 12 seconds at high speed, 5 RPM. For 45 turns on my SCT that would be 9 minutes.
The usual limiting factor of stepper motors is losing torque at high speeds.
They could always limit the "Zone" traversal to when the OTA is set horizontally (least torque needed to move the focuser), or even ask you to remove the camera, OAG (even worse, an ONAG), dispersion corrector and filter wheel. I.e., don't let you move by zones to actually focus, but just to pre-focus.
The 65536 number is pretty artificial (and limited to it because they are too cheap to use more than 16 bit arithmetic) because 5760 does not evenly divide 65536.
Chen
WalterT I use the Focus Assistant in Sharpcap but not auto-focus. I suspect that mirror shift could always be a problem for active autofocus, not to mention excessive backlash.
I would like to add an unrelated comment regarding belt tensioning. Rather than two sliding plates, it might be better to have a single adapter plate and slotted holes for the EAF motor.
Further to my post #13, it has occurred to me that the 16 bit position register might well reside in non volatile memory in the EAF. Changing it from 16 bit to 32 bit would require a firmware upgrade for the EAF. If 32 bit integers aren't available, the firmware would probably have make use of an additional 16 bit register and emulate a 32 bit number.
Another option could be leave the 16 bit register in tye EAF untouched and implement the higher 16 bits on the PC side, maybe in the ASCOM driver, to report positions beyond 65,535.
Maybe the firmware could accept an additional "steps per count" setting, defaults to 1, but to get range over 40+ turns you could set it to say 4? You'd be able to customize the range to the focus mechanism you're using.
There are many possibilities, but only the ZWO team know what is do-able. There is probably a restriction that we know nothing about that make 64k steps the only option. Maybe it's baked into ASCOM or INDI. What is the type that they pass?
Steve.
I am quite sure that something can be done, once the need for it has been recognised. It really just needs one decimal digit (10x) or 3 bits (8x) in front of the current 16 bits.
Technically the purest solution is to either use a 32 bit integer put a 16 bit integer in front of the current 16 bit integer and process the combination as a 32 bit integer. The required would be an integer can go over 4 billion, a little more than we need. As I said, thismay have to be done in the NEAF firmware, which would mean a firmware update.
Another option could be to prefix the decimal digit (0 .. 9) or the three bits (000 .. 111) in the ASCOM DRIVER on detecting a step change, from 65,535 to 0 or 0 to 65,535, depending on the direct of travel.
StevenEvan I only just noticed this post. This sounds very plausible to me. At 5 steps per count, it would be 5760/5 = 1,152 counts per turn (revolution) about the same as the Celestron focus motor at 1,000:counts per turn. It would achieve much the same result as a timing belt, without any mechanical changes. I if it could be done at driver level so much the better.
The timing belt solution also greatly improves back end clearance which is very tight on my Evolution 8 but just manageable.
+1 for this request, i just bought 2 EAFs for my SCTs and found after installing them found this issue and got here to this forum. I agree the best solution is the x steps per count, in fact some others focusers use that. In my C11 i had to use 2 steps per count to cover the whole range.
Just an example, this other focuser has also a limit of 65k values, however it implements the configurable steps per tick value, and also in my case has to be 2 to cover a range of 114800 steps (59000-1600)*2 to cover the required turns. It originally defaulted to 1 but if you needed to expand the range you an apply an arbitrary integer multipler at the expense of decrease of precision) I think this kind of solution just requires really small development and can easilly be implemented in EAF Firmware, as other focusers do.
Hi all.
My biggest problem is the placement. It would be nice to have it more to the side.
Tried moving the bracket to the side but it hits the body of the scope, maybe a small spacer could fix that problem, you would also have to shorten the long slotted bracket but thats an easy fix. Maybe some slotted holes to fix the ring to the scope so you have more adjustability than the 3 holes it has now. Also a shorter coupling would be nice.
