Another thing....If you see random oddities, it's possible the CPU is doing housekeeping tasks. I dont know about Arduino, but definitely on a Raspberry Pi the Linux OS will periodically do mandatory OS functions and it will definitely affect timing loops. When I was analyzing performance on a parallel I/O port I implemented in software, I found random cases where the OS would take a few milliseconds away from my tight-timing loop.
Although I use a RasPi in all of my projects since 2017, when I need timing-critical work done, I offload it to an FPGA and use software to interact with the FPGA in a non-timing-critical manner. The other advantage of the FPGA is the logic is programmable, so it's easy to make changes, fix bugs, etc. On Thursday, January 7, 2021 at 11:42:56 PM UTC-8 Mateusz Dziuba wrote: > @Bill van Dijk It's 1x6 multiplexing actually. Posted schematic is not > mine, I used it just as reference. It seems someone mixed tube connections > as well. I tried adding resistors before, but without success. > > @gregebert I am not using PWM yet, but I am planning to do so, to make > automatic dimming of nixies. I am not sure how it will interact with > multiplexing. > > Taking into account all your ideas I worked on timing again. First I added > 700ms delay inside interrupt before switching next anode on. I know it is > huge no-no, so I rewrited all multiplexing to the main loop and add some > delays - experimentally found I need 1ms after switching off all anodes and > 3ms before switching them off. > > Thank You for all your time and help. I may return to this post if > implementing pwm will challenged me. > > czwartek, 7 stycznia 2021 o 21:09:11 UTC+1 gregebert napisał(a): > >> It looks like you are using PWM dimming. Can you disable it for now, and >> then focus on optimizing dead-time between cathodes and the anode timing. >> >> I'm fairly certain you want to turn OFF the anode before changing the >> cathode signals, allowing some dead-time, then turning on the next anode. I >> cant tell from the waveforms above if there is any dead-time or >> overlap/underlap. >> >> On Thursday, January 7, 2021 at 11:05:27 AM UTC-8 Bill van Dijk wrote: >> >>> This is a 1x4 multiplex, each tube is lit 25% of the time if no “dead >>> space” is built in the software. I have not checked the data sheets, but it >>> looks strange that the 0 output from the 74141goes to the 1 of the tube, >>> and so on. Anyway, with your problem, program a space between switching >>> digits, length to be experimentally done since different tubes switch >>> faster than others. The other thing to try is a resistor of a few hundred K >>> between the tube anode and ground to dissipate the space charge in the tube >>> quicker. (not sure if that works; I have not tried that myself). Generally >>> speaking it is not recommended to multiplex Nixies, it surely dims them. >>> >>> >>> >>> *From:* [email protected] [mailto:[email protected]] *On >>> Behalf Of *Mateusz Dziuba >>> *Sent:* Thursday, January 07, 2021 1:27 PM >>> *To:* neonixie-l <[email protected]> >>> *Subject:* Re: [neonixie-l] IN-4 ghosting, once again. >>> >>> >>> >>> I did not develop specific circuit diagram for my clock, but it is build >>> around this project: (please don't mind wrong polarity of opto. Someone >>> draw it wrong.) >>> [image: e44b2492e18895184e259a89f71556fc.jpg] >>> >>> czwartek, 7 stycznia 2021 o 18:32:35 UTC+1 Bill van Dijk napisał(a): >>> >>> Can you post the schematic? First I notice that a bright 1 has a ghost >>> 2, and it seems every number has a ghosted lower number. Could be timing >>> (discussed many times), or perhaps the KD1551s are too slow depending on >>> your multiplex speed. What is you multiplex sequence, i.e. 2x3, 3x2, 1x6 ? >>> >>> >>> >>> Bill >>> >>> >>> >>> *From:* [email protected] [mailto:[email protected]] *On >>> Behalf Of *Mateusz Dziuba >>> *Sent:* Thursday, January 07, 2021 12:21 PM >>> *To:* neonixie-l <[email protected]> >>> *Subject:* [neonixie-l] IN-4 ghosting, once again. >>> >>> >>> >>> Hi all, >>> >>> >>> >>> I am working on nixie clock and experiencing ghosting between >>> neighboring tubes. Please see attached photo. >>> >>> >>> >>> [image: 1.jpg] >>> >>> >>> >>> I am using 6xIN-4 nixies, Arduino with Atmega328, 6xTLP627 transoptor >>> for driving anodes and K155ID1 chip for driving cathodes. I am multiplexing >>> nixies in interrupt. Code attached. >>> >>> >>> >>> What i tried so far: >>> >>> 1. I connected 10x 1N4148 fast switching diodes from cathodes bonding >>> points to mid voltage of 90V. It didn't help a bit and increased power >>> consumption only. >>> >>> 2. I removed the diodes and tried "center-pull" all anodes and cathodes >>> bonding points through 500k ohm resistors. Didn't help, although reduced >>> decay of anodes after switching off. >>> >>> >>> >>> Some photos below: >>> >>> [image: 4.jpg] >>> >>> [image: 5.jpg][image: 2.jpg][image: 3.jpg] >>> >>> >>> >>> What comes to my mind, I might try to increase/rewrite blanking interval >>> but i am not sure how to do it inside interrupt. Maybe I should rewrite it >>> to the main loop. >>> >>> >>> >>> All help, comments, suggestions are welcome. >>> >>> -- >>> You received this message because you are subscribed to the Google >>> Groups "neonixie-l" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to [email protected]. >>> To view this discussion on the web, visit >>> https://groups.google.com/d/msgid/neonixie-l/ed921226-078c-4a66-bae2-bf79fad2b9b8n%40googlegroups.com >>> >>> <https://groups.google.com/d/msgid/neonixie-l/ed921226-078c-4a66-bae2-bf79fad2b9b8n%40googlegroups.com?utm_medium=email&utm_source=footer> >>> . >>> >>> -- >>> You received this message because you are subscribed to the Google >>> Groups "neonixie-l" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to [email protected]. >>> >>> To view this discussion on the web, visit >>> https://groups.google.com/d/msgid/neonixie-l/e3a07d84-34e0-4f4a-9ab1-65ff8e128d2dn%40googlegroups.com >>> >>> <https://groups.google.com/d/msgid/neonixie-l/e3a07d84-34e0-4f4a-9ab1-65ff8e128d2dn%40googlegroups.com?utm_medium=email&utm_source=footer> >>> . >>> >> -- You received this message because you are subscribed to the Google Groups "neonixie-l" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/a9de3c84-6a4d-4ae1-a77b-f89197a43173n%40googlegroups.com.
