I don’t have a close up picture. However these caps are very simple to fit. They either bolt in or have a solder flange. The bolt-in caps are easiest as you can use a diecast box from someone like Jaycar.
I mentioned a few posts back the number of caps required is 14. This allows for the USB connection to be filtered. Wiring is simple. All common leads are joined together either side of the cap. All +5v are joined together, same with +12v and +3.3 volt leads. This means one feed through for each voltage.
I finally received the FM-DAB radio module a few days ago and just finished writing the DAB-FM control software for my CAR –PC. I fitted the radio module up front in the car as I had USB, FM aerial and audio leads running back to the PC in the boot from there.
There is a slight audio noise that can be heard when the radio volume is low or the radio muted.
This noise is not from the long audio lines. I was surprised that I could turn the amp volume up and touch the active audio inputs to any ground or chassis point in the front of the car and have absolutely no PSU or PC noise through the speaker. – No earth loops of any kind.
Disconnecting the serial made no difference. The noise is simply some residual PSU noise on the USB +5v line that powers the DAB-FM radio module. When I get a chance I will add a filter to the USB supply line and that should remove this slight noise.
The FM aerial is in the Rear glass and even with the boot open I get all FM stations with no noticeable interference from the PSU or PC. The PC is only a few feet away and all the cables from the PC to the front of the car run within a foot of this aerial. The Feed through caps and shielded enclose really make a difference here.
I will be mounting a digital aerial up front with auto-switching it between FM and DAB aerials once I find a suitable dig aerial.
thank you Mickz,i ll give it a shot if i cant find another solution.
Originally Posted by Mickz
Quick Update – The residual noise getting into the audio lead when the USB lead is plugged in to the DAB-FM module was NOT caused by the M4 – So the mods I carried on the M4 for noise really work well.
The noise is from the MB and varies with screen changes and general slight changes in CPU load.
The final proof was to remove the DAB- module and take it to the PC with a short 2 foot USB lead. I disconnect all cables from the PC except for a short audio leads and the USB leads to the DAB module. Noise was still there. Varying the supply to the M4 makes no difference when the SMPS changes frequency and operation point. It’s not the M4, video leads or LCD screen.
It is not coming down the +5v USB supply line, it’s on the USB supply return lead.
Filtering the return lead made some difference as did grounding the DAB- module with a heavy ground strap to the system ground point in the centre console.
The real fix was to cut the USB return lead and combined with the ground strap the system is silent.
There is still a +5 return via the audio lead to the PC and via the aerial. However it was only the USB return lead that injected the residual digital noise.
(One system crash half way through writing my last post....lets try again)
I took the advice of Mickz, and wanted to show my version of his M4 power filtering using pass through caps. It's not quite finished, but isn't far off, just need another couple of evenings to finish soldering up the parts, and mount the M4 into the chassis.
I decided to make up some bus bars to group the voltage leads into single outputs to each side of the pass through caps, and also to allow for strain relief, so it doesnt pull the leads off the caps.
I just used some circuit board sheeting, the stuff with the fiberglass panel with a layer of copper sheet on one side. All I did was mark out and drill the holes for the wires, and then use a grinding bit in a dremel to grind the copper away between each sections of the parts.
This will all mount inside a chassis picked up from Jaycar, as this way it keeps the pass through caps safe from being knocked and damaged.
Things to note:
1) My soldering sucks big time....please ignore the horrible mess!
2) This takes a long time, be prepared for it to take a few weekends lol
3) This is expensive. The passthrough caps cost me $150AUD, so over twice the price of the M4 itself.
4) The copper sheeting ends up pretty messy and dirty, if I did it again I would pick up some cleaning solution and then do a final coat with something to protect it...maybe...or maybe not :love:
As I said, I haven't quite finished yet, and I havent included pass through caps for the USB lead, as currently I am not going to use it, but I have space to add it in later on if needed (I think!).
So, the pics!
One of the bus bars, there is one each for the input and output side of the pass through capacitors:
The mounting for the pass through caps:
The passthrough caps installed:
The input busbar all soldered up except for the motherboard header lead and the amp turn on lead:
The output side, all soldered up (2 pics):
And my workstation currently....a bit of a mess!
Just to understand this correctly:
1) holes are drilled thru the copper sheets and pass through capacitors insatalled?
2) each wire gets one pass through capacitors installed to it?
3) I still don't understand why the m4 is encased in copper/alum? Help?
Do you have a part # for these pass through capacitors? I have seached high and low and I can't find it? Or maybe my searching skills suck b/c I don't exactly what they are? Do you plan to write a how-to on this (like a dumbed down version?) b/c I still have a hard time understanding the concept of how this works? Thanks
Originally Posted by itwanabe
FTC = Feed through capacitor with a value around 1000pf (.001). Four of the caps should have a current rating of at least 10A (+5, +3.3, +12 out and +12v in) the rest can be around 1A.
A: The case can be aluminium, made with copper clad PCB or an aluminium diecast box.
B: If you use bolt in FTC’s then get a diecast box to fit.
C: If you use solder in FTC’s you need to make the case with copper (to solder the caps to) OR use a diecast box with a piece of double sided copper printed circuit board bolted to one side wall with the caps soldered to that.
Every wire entering the box needs to do so via a FTC to filter the noise.
That is NOT the same as saying every wire on the PSU needs a FTC.
All + 5v connections can go to and from ONE FTC (10A or more).
All +12v as above.
All + 3.3v as above.
All grounds wires bolt to the inside of the case via a common earth bolt that passes through the case, NO FTC needed.
The extra effort put into the construction in the pictures above was to relieve the strain of multiple heavy current wires being soldered to both sides of a small FTC.
The reason for the FTCs’ and the case is explained earlier on in this thread.
You only need to do this if you have a PSU that is “definitely proven” to be causing interference to your systems such as: GPS is desensitised; FM and or DAB radio is desensitised or has interference. Noise is getting into the audio system.
FTC’s used to be cheap, now they are expensive and harder to find. There are links earlier in this thread to a few sources for FTC’s.
See pages 17 of some FTC links, read pages 14 onwards.
Apologise for the short reply, kind of busy at the moment and for the next week or two.
Should they also work? They're very small and there's no ampere statement. What happens when I use more than one FTC parallel for one voltage?
They are a small feedthrough used for low voltage RF bypassing. I have fed around an ampere through these in the past – don’t know about long term – they are easily broken and need to be soldered in.
You would not join the supply rail wires together with these. Instead one FTC on each lead of the 5v, 3.3v and 12v lines would be required. I can’t say how they would hold up with excessive current.
IMHO they “may” be ok. I can see no reason why paralleling them would be a problem in this situation.
I don't know much about electronics but I did some research and I was thinking to myself why not just filter the source? @ 250watts rating, I figured 100,000uF capacitor btwn the battery and the M4-ATX would do the job? Again, this is just a thought and I dont even know if it would even work... Any thoughts?
1. Because the source of interference is not the 12v line from the ATX.
Originally Posted by itwanabe
2. And simply stated, at higher RF frequencies that capacitor looks like an open circuit.
It’s a good question though.
BTW the cap will not work connected like that, unless it a dual cap with a hidden chassis ground.