|
Post by Paul Harman on Dec 19, 2013 16:44:06 GMT
Looking at the PCB the two output solenoid has some tracking errors around the diodes - the + and - is shorted by a jumper and there are some inverse parallel diodes.
Think about using a 16-pin outline for the opto isolators on the 8-output board - you never know there might be a PS2501-4 available like the TLP621-4 that can be a possibility then.
As for the power track for the CDU supply if you are considering using 0.2" pitch plug in terminal strips there will be enough PCB beyond the terminal pads to run a power track along the edge for the negative supply from the CDU. The positive CDU supply is only low current for the spike protection diodes and you can run a track for this under the opto isolators to reduce the number of links required. The alternative is to mount the diodes all vertically and run a wire along the top
|
|
|
Post by phillby on Dec 20, 2013 5:47:06 GMT
Hi Paul, I think I corrected the error a mystery track must have got in there. The single unit was copied and modified to suit the four unit. As i said Mystery! I checked my supplier and there is a PS2502-4 but I think it will cause hassels with the layout needing more jumpers. The bottom one is laid out like a 16 pin device already the two "2's" are cheaper than a "4" by the princley sum of about 5p. I am looking longingly at the space left on a square board and with 1 jumper I could put a CDU that I have already drawn an built. It will suit me nicely as for complicated reasons. I will only have four Solenoid sets of points. The rest will be DCC concepts Points motors or similar. Its a pitty that the Accessory Decoder cannot pe triggered with a pushbutton and display on leds to mimic points positions at the track and on a display. I guess the leds could be done with Macros but not sure about the push button. Long way to go and so little time. The units have been rearranged on the page and should be 1:1. I will still be doing more work when we have what we want. Question is it Ok to rave on like I tend to do if its close to Topic or should I stay closer to Subject? Cheers Brian the Ozzie in 38 degree C heat. Attachments:
|
|
|
Post by Paul Harman on Dec 20, 2013 9:51:45 GMT
Brian
Since you have sorted the operation problem it will be nice to start a new thread so that your decoders will have more visibility.
I prefer to layout the PCB for the TLP621-4 because that way there will be more flexibility for people to use the smaller parts as an alternative if required. The links can be eliminated by repositioning the diodes.
The DIY Decoder CDU only uses five components and will fit easily in the available space so it is well worth laying out for - you don't have to fit it.
With a solenoid decoder it is possible to fire the solenoids with local buttons directly on the coils as with traditional analogue if required. As to lamps to show the point position, either use another decoder in the panel set to the same addresses - or use LEDs connected to the frog of the point to show the actual point position.
|
|
|
Post by phillby on Dec 20, 2013 22:19:31 GMT
Hi Paul, First another question sometimes the controller reads the CV No and most times it does not. Could some thing not be quite right, remember at rhis stage I am using a BC547. I have done another revision using 16 pin optos removing all the links and adding the CDU to the PCB. It's Big about 150x100. It will however do my job and shoud be useful to others. I have not started another thread yet prefering to get close to completion so there will be less confusion. I woud caution that users read the CDU page taking note of the AC requirements and noteing that the board layout is inverted to the veroboard version. Cheers Brian the Ozzie a little cooler today 22 degrees at 9 AM Attachments:
|
|
|
Post by Paul Harman on Dec 21, 2013 9:46:27 GMT
Brian
Check that you are getting at least 13.2V across R16 when the BC547 is turned on. If the gain is a bit low you might have to reduce the value of R17 to compensate.
Back to your question about local control buttons and indicator lights, Suzie's board has a solution to it, but the reason it has been left off of the base design is because it would need a bigger processor (PIC16F131) which would need some quite big changes to the firmware (which is a bit of a squeeze anyway) to be compatible and would probably be too big to easily fit on stripboard using the current layout. Local control buttons for the most part are probably more suited to more compact designs using double sided PCBs and surface mount components.
If you want to have a go porting to the PIC16F131 it will not be too hard - the structure is the same, just a few of the registers have moved around. Other 20-pin alternatives become possible too when you get to the bigger processor with a bit more memory in them which I guess might help.
|
|
|
Post by phillby on Dec 23, 2013 12:05:17 GMT
Paul, Thanks for all your help and suggestions. I was wondering how to turn the transistor on. Should I put 5 volts to pins 4 & 5 to ensure it is turned on to check the voltage or will that harm the PIC. I will be going to the Big Smoke ( Sydney) fort the next few days, (it's Christmas you see) to visit the Daughter Son-in Law and the grand kids. The five year old has an HO Layout (his father drives Surburban Trains). He visited recently and was able to master the NCE Power Cab with 3 Locos in three days. Good going if I may say so when he has not been recognising number fo too long.
Merry Christmas and Happy New Year to all. Brian the Ozzie in Track Pants and Jumper it's 12 degrees at 11pm
|
|
|
Post by Paul Harman on Dec 23, 2013 12:49:04 GMT
Happy Christmas Brian
Take out the PIC and then link pin 1 to pin 5 to switch the transistor on. Don't take too long to take the measurement or the resistor will overheat.
looks like you will be able to spend some time getting the grand children fully ready for DCC then. Excellent.
|
|
|
Post by phillby on Dec 27, 2013 6:26:44 GMT
Hi Paul, I'm back I had a great Christmas I hope yours was the same.
I have had a chance to change R17 to 3K 9 and it didn't do much It is now 9.37V DC. The total volts from the top of R16 to ground (the zero rail is 12V when Q9 is off. Reducing it 820R gives 9.25 V across R17 and 9.32 to Gnd, 0.7V across Q9 I don't think I should go further. The Power Cab produces 14.14 Volts AC at the Input to the decoder. I hope this makes some sense
Thanks, Brian the Ozzie back in the Train Cave (my workshop is in a kind of basement.
|
|
|
Post by Paul Harman on Dec 28, 2013 10:22:16 GMT
Brian
It looks like your BC547 is not working correctly. Do you have the correct pinout for it? Base pin is in the middle.
Not sure I understand the voltages, 9..37V + 0.7V = 12V, the sums are not adding up. The voltage across R17 should never be more than 4.4V or you are likely to damage the PIC.
Voltage at the top of R16 to 0V should be pretty much the same as across R16 when it is turned on
|
|
|
Post by phillby on Dec 28, 2013 20:15:22 GMT
Paul, I have confused you. The rail volts drops to 9.32 Volts when base resistor is 820R so voltage 9.32 = 9.25+0.7. Don't know why the 3v aprox drop when the transistor is conducting. even with the 4k7 it was 9.37 across R16. I haven't measured asross R17 but was suppling 5V to the pin 5 end without the pic. Cheers Brian The Ozzie
|
|
|
Post by Paul Harman on Dec 29, 2013 16:48:18 GMT
Brian
It looks like you might have to decrease the value of R16 a bit. I would try 100R to start with and see what you get. R=V/I where V is the voltage you see across the bridge rectifier and I is 0.06A
Reducing R16 might need a corresponding proportional reduction of R17 to compensate.
|
|
|
Post by phillby on Dec 31, 2013 6:30:41 GMT
Hi, I reduced R16 to 100R and R17 to 820R and the voltage across R16 was 7.2V. Only guessing but it looks like something is current limiting. The only things in circut are the small signal diodes 1N914's.
I put the resistor back to R17=4k7 and R16=220 and the transistor was changed to a BC182B. The voltage is back to 9.35 across R16. With this configuration I have noticed that if I do a read of a CV and get a mesage that it can't be read and then read it again it's contents are then read. I think also reading CV 1 2 and 3 because they are close and were read one after another they gave up their values. It's like the double hit provides sufficient energy to allow the read.
Happy New Year.
Cheers Brian the Ozzie
|
|
|
Post by jjbrisvegas on Aug 8, 2014 10:48:13 GMT
Hi all, If I read the above correctly, by changing CV3-6 to lets say 1, and CV33 to 15, you would expect the decoder to send a 10ms pulse? With tis config on my Lenz the outputs both stay on. Changing CVs3-6 back to 0 give the expected toggling. Am I missing something ?
|
|
|
Post by bigsteve on May 17, 2015 21:17:00 GMT
Hi, I am new to the site and have been trying to catch up with the conversation, but you have left me hanging. How did you fix the 60ma acknowledgement pulse? It sounds like you are dropping four volts across the low power 1N914 switching diodes. They are designed to operate at 10 or 20ma and may be doing some strange things at sixty. Is this thread still open?
|
|
|
Post by bigsteve on May 18, 2015 3:06:37 GMT
Hi again, I just had a thought. Since a 60ma pulse appears to be the design goal for the acknowledge signal, why not make Q9 a current generator by adding an emitter resistor? My proposal is to remove or reduce R17, reduce R16 to 30 ohms, and select a value for the emitter resistor that provides a 2.0-volt pulse across R16 (probably, about 68 ohms).
I still think you have isolated the cause of the voltage drop to the diodes or the programmer. Did you measure the 14 volts from the Power Cab while you were drawing current?
I apologize for sticking my nose into your conversation. Sometimes I can't help myself.
Steve
|
|
suzie
Keen DIYer
Posts: 30
|
Post by suzie on May 18, 2015 14:51:14 GMT
I just had a look at the spec for the Vishay 1N914, and it is capable of handling a lot of current, 300mA on DC, presumably double this on DCC (50% mean duty cycle), and up to 1A for a pulse of up to 1 second long, so no problem handling the >60mA ACK pulse. www.vishay.com/docs/85622/1n914.pdfI have always used 1N4148 but cannot tell what the difference is between 1N4148 and 1N914.
|
|
|
Post by Paul Harman on May 22, 2015 10:01:58 GMT
Yes, 1N914 should be OK.
|
|