IN-8-2 Modular Rev 3 build - feedback and questions

More
1 year 3 months ago #9749 by Ian
A frequency of the crystal oscillator will indeed cause the MOSFET to get all hot and bothered. Part of the calibration is to set the pulse on time so that we just about don't go into saturation. We do this by choosing the shortest possible on time that doesn't cause an increase in output voltage. (Are there too many negations in that sentence)?

Here's a thought. If you try a factory reset and tell it to go into the calibration routine, what is the number it counts up to? Usually it should get to around 170-180. Anything outside of that range is anomolous.

Please Log in or Create an account to join the conversation.

More
1 year 3 months ago #9774 by Phil_Mod3
Ian: I couldn't get the HV calibration to run when I had the error, since the power-supply switched off after only a few seconds (presumably due to over current, which would be 3A).

Anyways, I have it running again. I started replacing parts in the following sequence and tested after each step:
crystal -> caps driving it -> HV cap -> HV inductor -> clock functions properly. So I suspect that the inductor has a short (don't know how to test it, since i don't know what the DCR of the original part was).

I replaced the inductor with a part, which has a lower DCR (based on how it looks, image old on the left, new on the right), resp. higher maximum current (new part 0.19 Ohms, 900 mA). However, the HV generation is now much more noisy than it used to be. Not only produces noise, but even leads to varying brightness of the tubes. A reset doesn't help, HV cal get's me the same operating point as before ( OCR1A = 183 ).
What I currently do: I manually set the PWM ON time to 120. Since this is too low to reach my target HV (175V), I assume that the µC adjusts PWM TOP to the lowest allowed value ( ICR1 = 300 ) and by this is able to reach ~172V. So effectively, I prevent that the algorithm can follow some fluctuations. By this, the clock is quiet and the brightness does not flicker. I tried other PWM ON times, but at 130, the clock is already able to reach target HV and becomes noisy again.

I'd like to look into this, so a few questions:
-Is there an easy way to get the HVS readings + PWM TOP times while to clock runs?
-Has anybody experience what happens if PWM TOP is not adjusted during runtime? (given that an initial calibration has been carried out)
-Could the automatic adaption of PWM TOP cause unintended oscillation/noise in the HV generation?
-Is it correct that the boost converter runs in discontinuous condcuction mode? (the duty cycle that is needed to achieve the HV points in this direction)

Thanks for reading and looking forward to the discussion.


Side note: I lifted one pad of the input cap, it seems that it is still connected but no idea how this will turn out on the long term, any advice/ideas? Better to replace the board?


Attachments:

Please Log in or Create an account to join the conversation.

More
1 year 3 months ago #9780 by Ian
Hmm I've never seen one of my inductors "without its clothes on", it's not that pretty is it?

I picked these way back when, and have stuck with them ever since, because they get the job done, and don't cause trouble, except, apparently, when they do.... :-/

I don't think the noise so coming from any type of oscillation of the PWMTop point. This is smoothed and moves quite slowly. It's really only there to handle long term changes.

This is why when the digit is off, the HV generator is off. Perhaps this is what you mention as "discontinuous". The idea behind this is that turning the digits on and off changes the load very rapidly, especfially when there are changes in the dimming, or when we do slots mode. The generator code just can't keep up with those changes. It is probably this fairly high frequency modulation that you can hear.

If we only turn on the HVG when there is load to it, it effectively provides a stable load to the HVG.

I have not tried running the HVG with the PWMTop held fixed. It should not be a problem. It is the PWM ON that is the tricky one...

Just a thought: Did you try holding the inductor still? Often the noise is purely mechanical, and nailing the component down well can deal with it,
The following user(s) said Thank You: Phil_Mod3

Please Log in or Create an account to join the conversation.

More
1 year 3 months ago #9786 by Ty_Eeberfest
Okay, I'm gonna suggest something quite radical... be gentle with your replies guys! ;)

I have played around quite a bit with the "ongoing" (not initial calibration) HV regulation code. I was unhappy with the little bit of flickering / pulsating that my perfectionist eye kept noticing, so after finding the "ideal" resistors for my neon separators and still being able to see flicker I dug into the regulation code. I am not totally unfamiliar with closed loop regulation schemes, so I was not hacking totally blindly. I tried many things that affected sensitivity, gain, deadband, etc. - basically "how much, how fast, under what conditions" it would vary HV parameters. I didn't keep records of what all I did, but I did a lot of experimenting.

My best results that I'm running to this day? I commented out Line 899, disabling ongoing regulation completely! Basically I let it initially calibrate the HV as it sees fit, leave all settings alone, and don't let it regulate while running. And it works great!

Admittedly I am one user testing one board, so this may not be representative research, but it works for me!

Also, it runs quiet this way though I never had any really significant amount of inductor noise anyway.

Ian: regarding discontinuous mode, it refers to how the conduction of the inductor is managed by the controller:

During continuous conduction mode, the inductor current in the energy transfer never reaches zero value. In the case of the discontinuous conduction mode, the inductor current falls to zero level which is very common in DC-to-DC converters.

Taken from: www.allaboutcircuits.com/technical-artic...f-simple-converters/

Look into it later when the dust is clearing off the crater.
The following user(s) said Thank You: Phil_Mod3

Please Log in or Create an account to join the conversation.

More
1 year 3 months ago #9790 by Phil_Mod3
Thank you for the input. It is important for me that nobody gets offended by my questions/posts, since this is definitely not my motivation.

Does anbody has some ideas on these?
  • Is there an easy way to get the HVS readings + PWM TOP times while to clock runs?
  • I lifted one pad of the input cap, it seems that it is still connected but no idea how this will turn out on the long term, any advice/ideas? Better to replace the board?

I'll borrow an oscilloscope hopefully by end of this week and I'll try to find out more about the source of noise and flickering brightness of my clock.
I'll most likely test Ty's solution and comment out the checkHVVoltage().

Also I found this interesting and will try to find out if it is relevant: "The ICR1 Register is not double buffered. This means that if ICR1 is changed to a low value when the counter is running with none or a low prescaler value, there is a risk that the new ICR1 value written is lower than the current value of TCNT1. The result will then be that the counter will miss the compare match at the TOP value. The counter will then have to count to the MAX value (0xFFFF) and wrap around starting at 0x0000 before the compare match can occur." (from the 328P data sheet)

Please Log in or Create an account to join the conversation.

More
1 year 3 months ago #9795 by Ty_Eeberfest
Regarding displaying HV and PWM Top: how much programming are you up for? I'd look at the "temp modes" code in processCurrentMode(). General area: Line 1420. You could pick a couple low numbered modes (like maybe the ones that display the IP address) that you don't care too much about. Use loadNumberArrayConfInt to feed the value you want to see to the tubes instead of the IP address or whatever.

Something like: loadNumberArrayConfInt(pwmTop);

I'm assuming you don't need these values displayed forever so just comment out the original lines of code and put it back when you're done analyzing.

Regarding the lifted pad: somebody will disagree with me on this, but my solution to lifted pads is to install a "bodge wire". That is, a thin (28ga ?) wire with one end soldered to the pin that has the lifted pad and the other end soldered to where the trace goes. I'm not sure what the "input capacitor" is so I can't be more specific on where to put the bodge wire.

I don't think the ICR1 update issue will have any meaningful effect on HV performance but hopefully Ian will discuss this with us...

Look into it later when the dust is clearing off the crater.
The following user(s) said Thank You: Ian

Please Log in or Create an account to join the conversation.

Moderators: AccutronTy_EeberfestIan
Time to create page: 0.211 seconds

Search

Tube Suppliers

Go to top
JSN Boot template designed by JoomlaShine.com