|
Post by ianholder on Jan 14, 2022 18:28:20 GMT
I'm sorry this is going back three months, but I've been catching up on some of the threads on here. I have in the past built two locos with crank axles, a GWR 1400 ( not Dart ) and a Great Eastern T26. I did not like the method of cross drilling for pins fixing the webs to the rest of the assembly and instead drilled axially to fit pins as sometimes used to stop wheels moving on axles. You need to plan the assembly in order to do this, centre, inside webs and crankpins as the first stage then the two outside webs and outer sections of the axle. I used two pins on each seat at 180 degrees. It is then possible to assemble the complete axle and drill for the last set of pins in the outer webs to crankpins. To keep everything in line I had drilled through each axle section 3/16" and the final assembly was done on a length of silver steel. The outer sections were left a few thou over size for final machining between centres after assembly. The assembly was carried out using Loctite 638 High Strength Retaining Compound, all parts being machined to a push/slide fit, no "cotton reeling". This is just not necessary, you do not need "room" for the Loctite, it is designed to fill the microscopic voids to lock together the components that are being assembled. If you look at the data sheets they quote a maximum clearance that the various grades will fill and I think this is where the myth of needing room for the Loctite comes from. They also state that you can use the product with a press fit, but this seems a bit belt and braces to me. I have used this method on all axle and wheel assemblies as well as crankaxles over the years without any failures.
|
|
|
Post by andyhigham on Jan 14, 2022 21:19:48 GMT
In his books on building the Shay, Kozo Hiraoka describes how to turn a 3 cylinder crankshaft with eccentrics for Stephenson valve gear from one solid piece of bar. That is 10 axes in total! He breaks it down into a number of simple operations
|
|
|
Post by palmsticks on Jan 16, 2022 1:47:07 GMT
Dave Applogies for the tardy reply, we have been away and just got back home. Now you mention a spreadsheet I did recal a spreadsheet from Don. However, searching my files he didn't send me one with his Dart design. Searching deeper I find I do have a spreadsheet from him, it appears to have data for Hall, Grange and Abergaveny. I have no idea where I got it from but it wasn't part of the email exchange from him. I may have downloaded it from his website. Is that any use to you or, knowing of the existence of a DA spreadsheet, do you already have it? Also I am afraid I am unable to help with your detailed questions. When I was talking to Don and asked him about differences between his design dimensions and Martin Evans originals his reply was along the lines "You want it to work don't you" I just used what he gave me and put his book back on the shelf. regards Pete Hi Pete, Thank you for the reply. Apologies for the late reply to your reply - real life got in the way! The spreadsheet you mention sounds like it could be the standard download from here: This features Purley Grange (launch links), Purley Grange (launch links) @180deg, ABERGAVENNY (with loco links), [no name] (with loco links)@180deg.
For completeness, please would you mind saving the excel file you have to your google files so I can just double check?
I realised I must have got something wrong and after I have read through everything again and while I still have some questions, I think I know a bit more as to why Don did what he did. Don managed to keep the weighshaft in the same position as the Martin Evans design yet still get pretty good equiv cut offs which is not necessarily what one would do if one had complete freedom of all dimensions - clever stuff - I wish I knew how he did it. I'll come back to you with specific requests on dims of the setup of the eccentrics since the simulation and drawings do not agree and indeed Don's drawing has a (small) conflict depending if you marked out to angle of advance or "x" and "m". Will be much clearer to understand with a diagram which I shall mark up together with my adaptation of Don's spreadsheet.
Thankyou once again.
Dave
|
|
|
Post by palmsticks on Jan 16, 2022 1:54:23 GMT
Dave, Having corresponded with Don about my own (A1) locomotive, I found that he just used the spreadsheet as a starting point. After that he tweaked values based upon his vast experience. Therefore, in my case at least, the final values did not match the original spreadsheet. Reg Hi Reg,
Thank you for the reply. I think you are correct. He has done something clever to get everything to work with the weighshaft in an unaltered position. i wonder if he had a technequie to hold these dimesnsions and work backwards to alter the lead, expansion link dimensions and eccentric rod length ecc. to do this? Such a clever man. I wish I had had the knowledge to ask him what I knew I didn't know when he was still with us.
Anyway, shall report my findings here as it will aid me in my design and maybe others in the future also seraching for answers using this wonderful forum tool.
Best regards,
Dave
|
|
|
Post by doubletop on Jan 16, 2022 20:22:28 GMT
Dave Applogies for the tardy reply, we have been away and just got back home. Now you mention a spreadsheet I did recal a spreadsheet from Don. However, searching my files he didn't send me one with his Dart design. Searching deeper I find I do have a spreadsheet from him, it appears to have data for Hall, Grange and Abergaveny. I have no idea where I got it from but it wasn't part of the email exchange from him. I may have downloaded it from his website. Is that any use to you or, knowing of the existence of a DA spreadsheet, do you already have it? Also I am afraid I am unable to help with your detailed questions. When I was talking to Don and asked him about differences between his design dimensions and Martin Evans originals his reply was along the lines "You want it to work don't you" I just used what he gave me and put his book back on the shelf. regards Pete Hi Pete, Thank you for the reply. Apologies for the late reply to your reply - real life got in the way! The spreadsheet you mention sounds like it could be the standard download from here: This features Purley Grange (launch links), Purley Grange (launch links) @180deg, ABERGAVENNY (with loco links), [no name] (with loco links)@180deg.
For completeness, please would you mind saving the excel file you have to your google files so I can just double check? I realised I must have got something wrong and after I have read through everything again and while I still have some questions, I think I know a bit more as to why Don did what he did. Don managed to keep the weighshaft in the same position as the Martin Evans design yet still get pretty good equiv cut offs which is not necessarily what one would do if one had complete freedom of all dimensions - clever stuff - I wish I knew how he did it. I'll come back to you with specific requests on dims of the setup of the eccentrics since the simulation and drawings do not agree and indeed Don's drawing has a (small) conflict depending if you marked out to angle of advance or "x" and "m". Will be much clearer to understand with a diagram which I shall mark up together with my adaptation of Don's spreadsheet. Thankyou once again. Dave
Dave
It will be the same spreadsheet, Don didn't send it to me and there is nowhere else I could have got it from. Are you sure the weighshaft is in the same place as Martin Evans located it? Here are the questions I sent to him at the time that elicited the reply "You want it to work don't you?" > Further questions. > > Your eccentric rods are 5.57" and Martins 5.5625. (5,9/16") The set > back of the pivot pins is 7/16" being the same as the arc in the link > (6"). The set back dimension was omitted from your drawing. So if I > used my existing links would I use your dimension for the eccentric > rods or Martins? I would suspect Martins which no doubt affects your > design as well? > > I also now notice that the way shaft is moved forward 0.185" but your > lifting arm is 0.443" shorter than Martins (2.307" vs 2.75"). which > does seem large variation compared to the other dimension changes. > The combined distance from the axle centreline is 3.395"+2.307"+0.264" > = 5.966" and not 6" of the expansion link. Is the 0.034" difference > material? > > Both of these points combined would put the expansion link on the end > of the eccentric rods 0.075" further forward and the lifting arm pivot > point 0.034" further back using the 6" arc as the datum. > > If the requirement is to keep the lifting arm parallel to the gear > centre line at the mid point. I could just drop the pivot point by > 0.044" making it 1.664" below the axle centre line? I've got to > re-make the brackets anyway to push the bearing forward by 0.185" > > The eccentric throw on Martins design is 0.315" you have 0.334" I've > included the sheet so you have all the relevant parts. > > I miss read 6deg it is 5degAnd Don's replys >Hi Pete,
>Don't fall into that trap! Never mix two designs. >I know all about flailing arms in the simulator - it is easily done and sometimes a headache to discover why!
>Don.
>Hi Pete - oooooh, this could have been dangerous but fortunately all is well. >Expansion link centres of 1.375 is fine.
>Lifting links of 2.0 is ok with a caveat. My drawing shows the lifting arm parallel to the gear centreline, and the gear trunnion on that line, so the lifting link's length comes out in the wash at 1.956. If >you use 2.0 the trunnion is not then central. Who cares? Well, moving to full forward gear, by pulling the expansion link down through 17 degrees of arm travel, leaves us a bit short of good starting cut off >and you really need another 2 degrees. For full back gear you will have 2 degrees too much. Of course, all this positioning has to be produced by the reverser at the cab end. The 2 307 must not be altered. >Exact full gear is not a requirement as long as the starting cut off is around 75% but 70% is a bit too 'iffy' for comfort.
>'Additional 0.19 offset of eccentrics?' Where is this? I don't have original drawings. Anyway, you say no problem.
>6 degree inclination - well mine is 5 degrees but taken as the horizontal. Does not concern the gear veracity. The weighshaft is shown dimensioned square to the frame as usually found on drawings (blow the >photo up for figures). Too small to have significance anyway - just ensure that the reach rod does the right things from the cab.
>Ted Gowan was ex-Canadian National Railroad engineer, very old fashioned but fastidious with principles and pencil and graph paper! I learned a good deal from him in the 1960s. His style is a bit unforgiving >at times. He laid down correct laws without explanation as you say, and just expected conformity. It was extremely rare for FS engineers to use the wrong link type, whereas the model engineer, not >understanding the principles, often used launch links in the wrong gear and vice versa. Dart is an example and it can play havoc with the valve events unnecessarily. Dart could have been immaculate. As it >is, the suspension offset at 0.264 is doing its best under the circumstances but introduces its own faults. >The real answer to your 'compromise on my design' is that mine is already a compromised forced by using the wrong link type for the driveline.
>Cheers, >Don. Pete
|
|
|
Post by palmsticks on Jan 17, 2022 0:03:45 GMT
> > I also now notice that the way shaft is moved forward 0.185" but your > lifting arm is 0.443" shorter than Martins (2.307" vs 2.75")
Hi Pete,
Ha! Now you come to mention it, you did say you had to introduce an offset into the weighshaft brackets within this thread, and chop a brake bar to miss - so clearly he did move it. However, my understanding of the position being unchanged is probably down to a difference between the Martin Evans drawings for the locomotive valvegear (which I don't have) and the dimensions that are in the simulator file for DART EVANS.L2? When I compare the two simulator files for DART EVANS.L2 [original] vs. DART.L2 [DA design], the dimensions for the weighshaft are as follows:
Dimension | DART EVANS.L2
| DART.L2 | ARx Axle to Reverse Shaft
| -3.395" | -3.395" | ARy Axle to Reverse Shaft | -1.943" | -1.943" | LR Lifting Arm
| 2.307"
| 2.750"
|
Don did say that he didn't have original drawings, so perhaps he simulated both "original" and his modified design in the same position - or I have corrupted the files mucking about? I'm tempted to go back to the DART EVANS.L2 simulation and update that one with the weighshaft dims you started with to see how that changes the original valve events although I'm not sure what I will learn other than how original Dart compares to your Dart. Must focus on the task in hand which is to understand the techniques and what influences design choices so I can apply it to my GER Claud.
More mysteries as to how he went about getting to what is clearly a design that works well with all the caveats that he stated about incorrect driveline.
Thank you so much for your continued help and providing new nuggets of information.
Hope all is well in NZ.
Best regards,
Dave
|
|
|
Post by doubletop on Jan 18, 2022 8:45:47 GMT
Dave
This may help a snip from the original drawing in ME You will see that the weighshaft centre is 3.375" and 1.6875" from the driver CL All is well in my part of NZ thanks, I'm on the south of the North Island well away from the sad situation in Tonga Pete
|
|
|
Post by palmsticks on Jan 18, 2022 12:14:40 GMT
Hi Pete, Thanks for the drawing that is most helpful. Glad all is well. Couldn't remember how far away NZ was from Tonga - far. I don't know if you've come across Scott Manley on youtube? He makes some interesting videos on space and his latest on the Tonga eruption is worth a watch compiling the satellite imagery and weather station pressure measurements. Absolutely terrible for the people who live near. www.youtube.com/watch?v=zoMRwyNhqJ4Regards, Dave
|
|
jma1009
Elder Statesman
Posts: 5,901
|
Post by jma1009 on Jan 18, 2022 23:37:26 GMT
I think we are getting a bit off topic, but so far as my own contact with Don Ashton, I don't ever recall him referring to his website spreadsheet.
I can tell you what my experiences of dealing with Don were, and it was nice to see Pete quoting his emails with Don. I am pretty sure that Don discussed with me Pete's situation in our regular weekly phone calls.
The essential problem with Dart and Claud is that for outside admission slide valves and direct drive with Stephensons valve gear, neither in full size as per the prototype (I can really only comment on this as to Dart), you wouldn't have launch type links. Martin Evans designed a valve gear for Dart that was a total mess and rubbish. Whenever Don Ashton came across these errors by Martin Evans, he would do his best to correct matters. There were limitations imposed. I have no doubt that we discussed the superiority of using loco links as per prototype.
|
|
|
Post by palmsticks on Jan 19, 2022 0:32:22 GMT
I think we are getting a bit off topic, but so far as my own contact with Don Ashton, I don't ever recall him referring to his website spreadsheet. I can tell you what my experiences of dealing with Don were, and it was nice to see Pete quoting his emails with Don. I am pretty sure that Don discussed with me Pete's situation in our regular weekly phone calls. The essential problem with Dart and Claud is that for outside admission slide valves and direct drive with Stephensons valve gear, neither in full size as per the prototype (I can really only comment on this as to Dart), you wouldn't have launch type links. Martin Evans designed a valve gear for Dart that was a total mess and rubbish. Whenever Don Ashton came across these errors by Martin Evans, he would do his best to correct matters. There were limitations imposed. I have no doubt that we discussed the superiority of using loco links as per prototype. Hi Julian, thankyou for your reply. I shall come back to you on your points soon. best regards, Dave
|
|
|
Post by palmsticks on Jan 19, 2022 23:35:32 GMT
Dave
This may help a snip from the original drawing in ME You will see that the weighshaft centre is 3.375" and 1.6875" from the driver CL All is well in my part of NZ thanks, I'm on the south of the North Island well away from the sad situation in Tonga Pete
Hi Pete, comparing the M.E. drawing to the simulator dimensions reveals that the Weighshaft is definitely wrong in the DART EVANS.L2 simulator file. See drawing below. This probably has no significance to your engine other than when the comparison simulation is re-run of the Evans design vs. the Ashton design, the Ashton I bet is going to look like an even better improvement then Don originally gave it credit for. Haven't done this yet as I have been mostly concentrating on checking dimensions and re running Don's spreadsheet which is now broadly agreeing with the dimensions he suggested to start with(!) - but at least I understand why now!
|
|
|
Post by doubletop on Jan 19, 2022 23:40:45 GMT
I think we are getting a bit off topic, but so far as my own contact with Don Ashton, I don't ever recall him referring to his website spreadsheet. I can tell you what my experiences of dealing with Don were, and it was nice to see Pete quoting his emails with Don. I am pretty sure that Don discussed with me Pete's situation in our regular weekly phone calls. The essential problem with Dart and Claud is that for outside admission slide valves and direct drive with Stephensons valve gear, neither in full size as per the prototype (I can really only comment on this as to Dart), you wouldn't have launch type links. Martin Evans designed a valve gear for Dart that was a total mess and rubbish. Whenever Don Ashton came across these errors by Martin Evans, he would do his best to correct matters. There were limitations imposed. I have no doubt that we discussed the superiority of using loco links as per prototype. Julian Probably "on topic" and more from Don Hi Pete,
Must check that we are both using drawing 961 sheets 5 and 6 of 15. The 14xx had a poor end-suspended gear. My simulation is for a launch link. I have no cylinder drawing but believe my dimensions are ok. Just keep an eye cast.
Don.Which says to me he wasn't happy with 14xx loco linked design and opted for launch link, like Martin. Which raises the question of John Smith's 14xx as he has followed the prototype with loco link. Later Don provided the explanitory document and drawings for which I supplied in the link here Don Ashton Dart (I am convinced I had posted this link in the past but can't find it in this thread) In the document Don said Basically Martin Evans discarded the awful 19th century prototypical gear and used a launch link without understanding the consequences. Stephenson’s gear with a launch link occasions three errors: 1) That of the main crank’s making in the conversion of linear to rotary motion, 2) the similar but smaller error from the eccentric, 3) and the rather virulent ‘knuckle-joint’ of the pins behind the link slot.No doubt opening up more of a debate.
Pete
|
|
|
Post by palmsticks on Jan 20, 2022 0:16:55 GMT
Pete, Here is Don's Stephenson's valvegear excel spreadsheet filled in by me so I could better understand the dimensions chosen in the simulator. This is an FYI rather than anything for you to do! The following link is to my Onedrive.
The dimensions are pretty close now. Will illustrate with a drawing soon and results of comparison of simulation but near bed time now zzz.
One last thing, I know dimensionally this is only small and is probably not of consequence but it tripped me up (probably because i'm not as used to working in Imperial). Could you take a look at the eccentric settings drawing below - which one did you use? Comparing Don's documents to the simulator reveals a small rounding error depending if one is following Angle of Advance or "x" and "m" dimensions. Basically, the simulator requires an angle and the drawing and setup in Don's book refers to using distances rather than accurately making to a angle, however the angles between the three sources (word file, pdf drawings, simulator) do not agree.
Best regards,
Dave
|
|
|
Post by palmsticks on Jan 20, 2022 0:51:28 GMT
I think we are getting a bit off topic, but so far as my own contact with Don Ashton, I don't ever recall him referring to his website spreadsheet. I can tell you what my experiences of dealing with Don were, and it was nice to see Pete quoting his emails with Don. I am pretty sure that Don discussed with me Pete's situation in our regular weekly phone calls. The essential problem with Dart and Claud is that for outside admission slide valves and direct drive with Stephensons valve gear, neither in full size as per the prototype (I can really only comment on this as to Dart), you wouldn't have launch type links. Martin Evans designed a valve gear for Dart that was a total mess and rubbish. Whenever Don Ashton came across these errors by Martin Evans, he would do his best to correct matters. There were limitations imposed. I have no doubt that we discussed the superiority of using loco links as per prototype. Hi Julian, May I firstly say thank you for your inputs elsewhere on this forum, the contributions on here from all the forum's members are incredibly useful and interesting. The excel spreadsheet in question is the tool Don used to replicate the graphical method of equalising leads on the eccentric rods, finding offset of the expansion link trunion and best place the suspension to be for equalising events at 50% cut off using Solver to iteratively find suitable variables. I wanted to find this spreadsheet with an example that had drawings and simulator files to understand the origin of some of the dimensions in the simulation as Don built a handy tab for gathering the relevant data. I had (have) gaps in my knowledge and had made mistakes in previous attempts to make it all work. Regards correct driveline: launch links may have been chosen in Dart instead of the preferred loco links with direct drive to a slide valve because dimensionally it may not be possible to have fit the loco links without making a recess in the boiler for full reverse and digging up all the sleeper and ballast when in full forward! As an exercise I was going to put a proposed design together with the spreadsheet and simulator, but it might not fit in the loco! Regards the Claud, that is a subject for a different time and different thread. My version is not Martin Evans and uses the correct driveline but is different to the prototype. Regarding Don more generally, do you have many designs that he worked on with you and do you know if there is a repository of his work? Does anybody have his files from his computer? I do regret never getting in touch with him directly. He seemed such a wonderful and talented man. Best regards, Dave
|
|
jma1009
Elder Statesman
Posts: 5,901
|
Post by jma1009 on Jan 20, 2022 22:57:45 GMT
I will attempt to reply.
Don was always dismissive of the Armstrong/Dean top suspended loco links, but I always reminded him that 'City of Truro' went 100mph with this arrangement and with slide valves and no superheat. Whether you believe 'City of Truro' went 98mph or 99mph or 100mph or 101 or 102mph, down Wellington Bank, the undeniable fact is that 'City of Truro' went incredibly fast, as so did 'Duke of Connaught' afterwards with pretty much the same valve gear.
Don's first book dismissed top suspended loco links. His final book that he very kindly sent me and signed, showed a section on top suspended loco links and how they can be made to work. My view would be that if you were building 'Dart' from scratch you would be better to follow the full size arrangement with a few tweaks, of top suspended links, and loco links. If that requires a shortening of the firebox, to accommodate the larger loco link eccentrics and sheaves, so be it.
There is a considerable repository of Don's legacy on this forum. Many of us had a considerable email correspondence with Don plus regular phone calls, plus letters via the postal service.
What you ought not to assume is that Don's solution to Pete's problem was the optimum solution, that would have required drastic remaking of parts. Actually, this was what Pete did. And perhaps Don's 'short fix' wasn't ultimately the best solution, but it met a 'short fix' to Martin Evans' rubbish and mess he made of the 'Dart' valve gear.
|
|
|
Post by doubletop on Jan 21, 2022 4:18:16 GMT
I will attempt to reply. Don was always dismissive of the Armstrong/Dean top suspended loco links, but I always reminded him that 'City of Truro' went 100mph with this arrangement and with slide valves and no superheat. Whether you believe 'City of Truro' went 98mph or 99mph or 100mph or 101 or 102mph, down Wellington Bank, the undeniable fact is that 'City of Truro' went incredibly fast, as so did 'Duke of Connaught' afterwards with pretty much the same valve gear. Don's first book dismissed top suspended loco links. His final book that he very kindly sent me and signed, showed a section on top suspended loco links and how they can be made to work. My view would be that if you were building 'Dart' from scratch you would be better to follow the full size arrangement with a few tweaks, of top suspended links, and loco links. If that requires a shortening of the firebox, to accommodate the larger loco link eccentrics and sheaves, so be it. There is a considerable repository of Don's legacy on this forum. Many of us had a considerable email correspondence with Don plus regular phone calls, plus letters via the postal service. What you ought not to assume is that Don's solution to Pete's problem was the optimum solution, that would have required drastic remaking of parts. Actually, this was what Pete did. And perhaps Don's 'short fix' wasn't ultimately the best solution, but it met a 'short fix' to Martin Evans' rubbish and mess he made of the 'Dart' valve gear. I think Don would agree that his design was a compromise, within the constraints of the Dart design, but it works well and sounds great in the flesh. The videos don't do it justice. What we need now is somebody as competent as Don to take up the baton to keep moving valve gear design forward. They would need to be able explain the changes and prove they were an improvement on the previous versions, as Don was capable of doing. I think that he would be very pleased to know that his legacy was being maintained. Pete
|
|
jma1009
Elder Statesman
Posts: 5,901
|
Post by jma1009 on Jan 21, 2022 23:58:01 GMT
I am not sure that the 'baton' needs to be passed on.
Don Ashton did so very much to help everyone. He corrected all the major errors in miniature and a few in full size. A compendium of his work ought really to be compiled and published over and above his own published books. This could be something readily accessible and of interest.
We are coming up to the second anniversary of Don's death. Others are looking after his website currently. There are lots of people interested in Don's work. And perpetuating his legacy.
|
|
|
Post by doubletop on Jan 22, 2022 20:46:08 GMT
I am not sure that the 'baton' needs to be passed on. Don Ashton did so very much to help everyone. He corrected all the major errors in miniature and a few in full size. A compendium of his work ought really to be compiled and published over and above his own published books. This could be something readily accessible and of interest. We are coming up to the second anniversary of Don's death. Others are looking after his website currently. There are lots of people interested in Don's work. And perpetuating his legacy. Julian I agree, but surely a line can't be drawn under Don's work? I was suggesting that if then it would require somebody with skills similar to Don to attempt an optimum solution. Progressive re-engeering if you like otherwise we'd still be stuck in the 1800's or even 1980's when it comes to Martins work. Pete
|
|
jma1009
Elder Statesman
Posts: 5,901
|
Post by jma1009 on Jan 22, 2022 22:47:54 GMT
Hi Pete,
I partially agree with you, especially with some designs such as Dart and Pansy. All top suspended loco links in full size.
Don did a workable solution to Dart, and when Don was ill we have a thread on here when he did the same with assistance from myself and Jim Scott with Dave's Princess of Wales.
I did a re-design of the Pansy valve gear over 20 years ago. Retaining the top suspended loco links. If scaled up for Dart it would achieve something prototypical to Dart in full size. Actually, it was inappropriate, for Pansy, as the rocker arm reversed the motion so it was an ideal candidate for launch type links for the 'LBSC' design!
|
|