Printable View
HI everyone,
This is something I've been talking about for a while, and even wrote an article on it, and the article has some photos, but video conveys a whole lot more, and a whole lot better than words. See the brief video I made here which shows what can happen to projected images if the connections are not made in the most optimal way:
https://www.youtube.com/watch?v=Vm0GISQi9fE
Bill
Excellent video! Thanks for posting that.
+1!Quote:
Originally Posted by JohnYayas
...I think I need to take another look at the wiring in my projectors...:o
Thank-you Bill!
Me too! +15 charQuote:
Originally Posted by Stuka
David
Useful info ,will need to find a grid pattern to check
Excellent demonstration, very informative, thank you! When I build my projector I will definitely keep this in mind. Two questions, because I can't help being curious:
1, Do this effect and solution manifest regardless whether the amps are running differential inputs or single-ended? Or, is there different guidance for differential vs. single-ended amp installations?
2, Is this another one of those situations that capacitors won't fix? Clearly this is a complex system, so I'm trying to understand the circumstances of the amps... In the video, I hear that under acceleration amps "draw a lot of current," but also in the past I have heard "instantaneous currents are not that high." So, please forgive me for splitting hairs, but I am just attempting to gain a better understanding of the nuances. Obviously the best solution is to let that connection float, not to tie it to ground and then add capacitors to the system (nor would I suggest doing so!) but these are mysterious things until they are understood, so the instinct is to keep asking why things work the way they do. :)
Thanks!
Mike
Insidious instantaneous currents can reach 6 to 9 amps or more on a modern scanner. It is more worrisome as scanner speed increases.
This is the back EMF and forward current surges getting into the control signal ground plane on the circuit board. Good board layout practices mean much more expensive boards, so when cost cutting during design occurs, the split ground plane needed to mitigate this disappeared from clone amps.
Say you have 0.1 ohm of resistance in your amp's ground plane and ground wire to the PSUs. At 6 amps of surge, you just lifted the board ground six tenths of a volt off ground. That is bad.
Single ended control does not work with Bill's solution, it needs a ground reference for the signal, making it susceptible to hum. With single ended the answer is to use analog power supplies with huge surge capability and "Stiff" grounds, as well as reverse biased diodes on the supply rails to allow surges to return to the rail. Care is then taken to keep the distances from the signal source to the amplifier very short. As well as use shielded cables or cables with driven shields. "Home Run" or "Star" grounding is used with multiple ground wires to a central point. Control portions of the amp board have their own power supplies and own ground plane
While I urge the use of external stiffening capacitors on switching power supplies, caps don't fix all of the ground return issues. Neither do reverse diodes dump all the EMF. Differntial signaling gets closer to zero error then anything short of multiple ground planes on the amp board.
Eye Magic provides for the ground jumper on their older amps, but that does not cure all. I have Cambridge Amps for very large galvos with split ground planes, the noise content is noticeably smaller when scanning.
Incidentally Bill's technique is somewhat different from DJs "lifting" the AC safety ground on their equipment, so please do not call differential signaling "ground lifting". Your projector should still have the safety ground on the AC line cord connected to case. "lifting" does little to solve the problem in this case, and is hazardous.
Follow Pangolin's PDF when wiring projectors. It works.
Steve
No, you definitely need differential inputs, because that's the thing that actually solves the problem. During a period of high-current flowing through the X amplifier, *probably* the Y amplifier's "ground" will become "lifted". If you have differential inputs, that "lift" will be experienced by both the + and - input, and the common-mode-rejection capability of this input will thus, ignore this "lifting" action and not interfere with the actual images. This is the real "secret".Quote:
Originally Posted by tribble
HOWEVER, one thing I didn't mention and that you can verify is -- that your amp actually has good "common-mode-rejection" in the first place. When I first started working on these amps, they didn't! The test is -- feed the same signal (let's say Y+) into both the + and - inputs. You should see substantially NOTHING -- no scanning. If you see scanning, then the common-mode-rejection is crap, and you'll have to solve that problem first.
In the case of the amps I was working on, they had 100K resistor in three (out of the four) locations of a difference amp, and 200K effective resistance in the fourth place. So I had to solve that problem. Once I got that solved, the images looked really good. Amps that started out really concerning me in terms of quality graphics, wound up delivering some of the best images I had ever seen (after all the modifications).
It could be that capacitors placed in some certain location would really help. But those capacitors would do largely the same thing that connecting the two amp grounds together with a huge piece of wire does. But doing it "my way" also has other benefits, and it's free. So, the way I think about it -- let's see -- huge expensive capacitors placed in some strategic location, or free. Your choice? ;)Quote:
Originally Posted by tribble
Bill
Ah yes, output or supply currents in the input path... Current errors.
Designing a projector wiring topology is no different to audio amplifiers really.
Steve, Bill, thanks for the detailed responses. I'm just going for, you know, more Capt. Kirk. vs. Lt. Savvik when it comes to knowing "why star (projectors) do what they do." :D
Well, I stated my (obvious) choice in my original message, but one still likes to know the rationale behind the right decision so one can make the right decision in similar, but different, circumstances. :DQuote:
Originally Posted by Pangolin
Just making sure you don't have a voyage on the Kobayashi Maru Mr. Tribble...
Besides a scanner amp maker might read this and get rid of the ^&%$# flaw!
Steve
great, will recheck my wirings
Resurrecting an older thread here. It's good to keep newer people on the forum informed.
My question though, is what if you have a dual ch amp that ties the ground together via board traces? Curious if it would be better to break the trace on the board and keep the ground isolated to just the power ground.
I never saw this and had thought my tuning sucked and I could never solve it. Bang. What a difference.
ok this is extreme but what if you used separate power supplies or used a supply twice as large as need current wise. Yea I hear you why free is free. Just learning.
Yeah, same here. I could never get rid of the "wobbly" lines on the Laser Media pattern. I always thought it was just because I wasn't using CT scanners or equiv.
The PSU I am using is plenty big enough. It's specifically designed for A/B audio amps up to 400W, so 4A constant on each rail and easily 9A spikes shouldn't be a problem. Part of the potential problem I see is the ribbon cable that delivers power to the amp. I believe the ribbon cable is composed of 28awg wire.. Granted, it's using two wires each for -/+/G but I don't think that's enough for large jumps and quick current spikes. I suppose I could modify the amp use larger power wires but I am still wondering about the grounds that are linked. I might try breaking the trace on the board that connects the two signal grounds. If it has an adverse effect, I'll just jump them back together.
Two separate power supplies can be used, as long as they meet at a common star point. I am sure this is shown in my diagram.
AHA dual channel amp -- no need to worry about modifying the amp itself. The singular ground of the dual channel amp should -- itself -- not be a problem.
The main point is -- just don't connect ANYTHING to the "signal ground" (third pin on an input connector). Use only the + and - inputs to the driver (which should be differential receivers).
Also make sure there are no "phantom" grounds, which might happen if -- for example -- the X-Y scanner block is bolted directly to the projector chassis which is also grounded. It's best to use an insulator (for example "sil pad") under the X-Y mount and nylon screws holding it to the projector housing.
And lastly -- I'm not sure if I ever did this in my demo video or not. But you can see if there is "cross-talk" between axes by projecting content only on the Y axis and making sure you have exclusively a vertical line, and then do the same with the X axis and see if you have exclusively a horizontal line. HOWEVER -- don't be fooled because, if you've got crappy scanners or out-of-balance mirrors, this too will cause deflection -- particularly while exercising the Y axis (small horizontal "perturbations"...)
Thanks for the response, Bill. After hooking up and testing, there was no need to modify the amp. The grid patten and LM Test patterns were sharp and not wavy. The one thing I did learn is; in every single one of my setups, ILDA pin 25 *MUST* be hooked to "power" ground at the PSU output, which your diagram shows. I neglected to do this at first and my scanner output was wavy in a way I couldn't really describe. Once I hooked pin 25 to the PSU ground, all was good!Quote:
Originally Posted by Pangolin