Deleted
Deleted Member
Posts: 0
|
Post by Deleted on May 4, 2018 2:26:18 GMT -5
Before going on with my layout, I wanted to know if I could achieve "continuous" acceleration/deceleration, to see where I have to put the insulators (distance from the stations ? ), in conjunction with my detection sections ... So I used (on a Veroboard) : - A voltage source, in my case an analog multiplexer with several trimmers, controlled via a rotary encoder. - The output is buffered with 1/2 L272M (connected as a unity gain follower), see further. - Then a RC integrating cell ; several adjustable resistors will be selected in the final version via a second analog multiplexer, one for each loco, and a "minimum" one to achieve precise stopping. - And finally the second halfth of the L272M, connected like the first one, as a power driver. Yes, I know, 700mA capacity for trains needing less than 100mA  ... But it was the only power Op-Amp that I found for a reasonable price. Two 1N4148 are used as freewheel diodes (between the output, and the power lines), and I put 2 1N4007 in series with the tracks (they will be used by the inverting relay and the detection circuits). Now, when I switch from 0 to maximum speed, the engine DOES NOT start like a rocket, as I read in a former topic ... It performs a wonderful smooth start ! I think this is a problem related with break-in. So I think that testing is better than believing.  Only one locomotive (an AZL GP7) starts less smoothly (my AZL GP9, with the same chassis, starts slowly from stop ! ) |
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on May 7, 2018 3:02:25 GMT -5
Unfortunately, if the trains run over the turnouts at a too slow speed, I noticed that they can stall very easily ; so there is a minimum speed for reliable operation. And when they stop abruptly (as I wrote, in a "precise" position), this cancels the positive effect of progressive deceleration, because of a residual speed which is too high. So my system could apply to a "hand controlled" railroad (I think the most of them are controlled in such a way), but with difficulty to an "automated" one ...
|
|
|
|
Post by dazed on May 7, 2018 10:02:41 GMT -5
It seems to me that you are describing what is effectively a fixed-value version of Momentum w/Braking.
I think you would just need to shorten the length of the deceleration "process" or curve. (If you can't do that, then perhaps your isolation section is too short.) It would take some practice but you should be able to get it to stop fairly precisely, assuming you are using the same or similar locomotives every time. (even the real engineers don't hit the same exact stopping point every time.) This is one of the advantages of DCC....being able to setup all the locomotives to perform similarly in these types of situations.
I can't recall what types of turnouts you are using, but obviously make sure you have sufficient feeders all around.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on May 7, 2018 16:03:40 GMT -5
On my former layout (DB, N scale), all trains that would stop or run over deviant turnouts were switched (without momentum) to "slow" at the entrance signal ; I made some tests with deceleration ... over 35 years ago (PWM with microprocessors), and noticed that the behaviour of the locos varied dramatically when they got "warmed up", and so I gave it up. All the way I will switch the train "off" without momentum at the "stop" position on my future layout, and if the residual speed is too high, as I wrote, the effect of progressiveness is destroyed. I don't wanna need to correct parameters all the time due to warming up. It was a test, it could be interesting for people building their own (hand) controller, they don't need PWM to drive their trains, my selector could be replaced with a simple potentiometer, but momentum does not fit my specifications, I am (and was) a geek and want automation, but such one that I can build, maintain and repair myself. The insulator will be about 220 to 300 mm before the first turnout, so I will first switch to half speed for about 1.5 s or 2 s (I calculated that the train will run less than the distance between the insulator and the turnout) and then to slow (switching) speed .... but fast enough so the train runs over the turnouts without issues ... and then switch power off when it reaches its stop position. In the past there were only relay switches (On/Off) or contacts inside of the signals (Märklin, etc) for automation, delays in stations were controlled via a "thermo-switch" and people were satisfied with them. When we children saw such an "automated" (HO) layout around 1960 in a store window at Christmas, with trains stopping and starting as if by magic, we almost thought it came directly from Santa's workshop.
|
|
|
|
Post by BAZman on May 7, 2018 22:39:51 GMT -5
Rather than a linear power amp, I would use variable period (PWM) so that no matter what % duty cycle, you are 'coercing' that loco to a dead stop, especially where a uP can program a curve that lets it slow, then crawl, then stop.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on May 8, 2018 0:46:56 GMT -5
PWM is not good for motors (MTL writes explicitely "no pulse" )... and the harmonics (sent by the whole layout acting as an antenna) could disturb the radio receivers in the neighbourhood, as I read.
I had a long "discussion" with Greg about this subject in the past.
The problem is the curve, it should be individually programmed for each loco ... and as I wrote, vary when loco is warming up ! The only reliable solution in my eyes is a position detection which stops the locos, with no care of the residual speed. So I am sure that my Budds will stop with all their doors in front of the platform.
|
|
|
|
Post by BAZman on May 8, 2018 8:20:15 GMT -5
PWM is no good on MTL  ? We run 10000's of locos on DCC which is . . . .wait for it . . . PWM driven. The pulse power MTL related to was the 1970's 120 Hz sinusoidal pulse in DC power. |
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on May 9, 2018 2:19:36 GMT -5
I can't believe that MTL didn't modify ( = edit .... with a simple DTP program ! ) their user guide since the 70's ! Mine was included into a GP35 and at the foot of the page I found "Rev. 8/05" ...  |
|
... But it was the only power Op-Amp that I found for a reasonable price. Two 1N4148 are used as freewheel diodes (between the output, and the power lines), and I put 2 1N4007 in series with the tracks (they will be used by the inverting relay and the detection circuits).
Only one locomotive (an AZL GP7) starts less smoothly (my AZL GP9, with the same chassis, starts slowly from stop ! )
? We run 10000's of locos on DCC which is . . . .wait for it . . . PWM driven.
