if there are two routes for a signal circuit then that extra route creates a loop.
Interference on that loop sets up an interference current.
That current times the resistance of parts of the loop creates an interference voltage.
if any interference voltage gets added to a signal then you cannot remove that later.
You must ensure there is ONLY one route for the Flow and Return current to pass from Source and back to Source. That route MUST be low LOOP AREA.
That is why I keep repeating
Interference on that loop sets up an interference current.
That current times the resistance of parts of the loop creates an interference voltage.
if any interference voltage gets added to a signal then you cannot remove that later.
You must ensure there is ONLY one route for the Flow and Return current to pass from Source and back to Source. That route MUST be low LOOP AREA.
That is why I keep repeating
Yeah there will be only one route for the signals (both speaker and input) -- the one that's one the traces on the amp pcb's. As I said I'll wire these two CHG together onto one screw on the chassis. The rca's and speaker terminals won't be touching the metal case (using plastic washers).
So I have been working on mounting the DC supply board to the case. Rather than drill four new holes, I decided to make use of an existing four holes that were already in this reclaimed chassis.
So I made an adapter plate for the dc supply board, with integrated beefy plastic standoffs. I designed it quickly in OpenSCAD and then printed it out on my 3D printer in PLA plastic (ignore the rough wavy lines, as my printer just started doing this and I need to perform some simple maintenance to correct this -- however it doesn't affect the functionality of this adapter plate.) Here is the photo of it. Notice the existing 4 holes on the case and how the adapter attaches to it:
I took calipers to the board when designing this in OpenSCAD. I made the standoffs 13.5mm high which is exactly the height of the AC input terminals -- that way the terminals are supported by the case as well:
Here it is installed. The plate screwed onto the case with hex nuts on the bottom. The dc supply board is screwed into the plastic standoffs with self tapping screws.
Here is an overall photo of what I have done so far:
The heatsink will be here on Sunday. I got this one:
It is six inches long and about 1 inch thick including fins. About 2 1/2 inches tall if I recall. I'll be installing it parallel to the dc supply board with the fins facing the dc supply board. Front of the heatsink will be flush with the front of the vent in the baseplate. Intake air will come up the center of the heatsink. And will have more escape paths above. The input signals will be low in the case so the heat sink can help shield the signals from the transformer and dc supply board. Twisted pairs will run from the dc supply board , probably over the heatsink, or maybe I'll route a path for them through the bottom of the heatsink.
So I made an adapter plate for the dc supply board, with integrated beefy plastic standoffs. I designed it quickly in OpenSCAD and then printed it out on my 3D printer in PLA plastic (ignore the rough wavy lines, as my printer just started doing this and I need to perform some simple maintenance to correct this -- however it doesn't affect the functionality of this adapter plate.) Here is the photo of it. Notice the existing 4 holes on the case and how the adapter attaches to it:
An externally hosted image should be here but it was not working when we last tested it.
I took calipers to the board when designing this in OpenSCAD. I made the standoffs 13.5mm high which is exactly the height of the AC input terminals -- that way the terminals are supported by the case as well:
An externally hosted image should be here but it was not working when we last tested it.
Here it is installed. The plate screwed onto the case with hex nuts on the bottom. The dc supply board is screwed into the plastic standoffs with self tapping screws.
An externally hosted image should be here but it was not working when we last tested it.
Here is an overall photo of what I have done so far:
An externally hosted image should be here but it was not working when we last tested it.
The heatsink will be here on Sunday. I got this one:
An externally hosted image should be here but it was not working when we last tested it.
It is six inches long and about 1 inch thick including fins. About 2 1/2 inches tall if I recall. I'll be installing it parallel to the dc supply board with the fins facing the dc supply board. Front of the heatsink will be flush with the front of the vent in the baseplate. Intake air will come up the center of the heatsink. And will have more escape paths above. The input signals will be low in the case so the heat sink can help shield the signals from the transformer and dc supply board. Twisted pairs will run from the dc supply board , probably over the heatsink, or maybe I'll route a path for them through the bottom of the heatsink.
Last edited:
DC supply is hooked up now and in place.
The screw on terminals for the ac inlet of dc supply really suck. They don't hold the wires in.. they fall out. Maybe I did something wrong but I gave up. I took them off the PCB and then soldered the secondaries directly to the board with no problems. Much better connection anyways.
The screw on terminals for the ac inlet of dc supply really suck. They don't hold the wires in.. they fall out. Maybe I did something wrong but I gave up. I took them off the PCB and then soldered the secondaries directly to the board with no problems. Much better connection anyways.
An externally hosted image should be here but it was not working when we last tested it.
Cool to see there are like-mindedI find the cases incredibly pricey considering the chimpamp.com lm3886 kit is relative cheap along with the transformer. It's like I'd spend more on a case than all of what's inside! Crazy business!
Anyways I have an idea for you for a front plate. I might do this. Take a piece of 1/4" thick MDF and cut it appropriately. Then spray paint flat black several coats. Then use a satin or glossy sealer over that and give it a nice shine
Or you could use a nice looking piece of wood instead of the MDF. HERE
I replaced the aluminum end on THIS $15 (shipped) Chinese case with a piece of wood I got at a dollar store; it's now a $16 case.
Thanks I ended up getting a machine piece of aluminum bar for it just the right dimensions for $15 plus shipping on ebay
Actually 2 pieces of the same -- shipping was $10. so $25 total for 2 pieces, or $12.50 a piece.. I'll have a spare for another project.AndrewT,
I did as you said and connected the purple screen wire directly to chassis with as short as possible distance. Thanks for the tip:
I did as you said and connected the purple screen wire directly to chassis with as short as possible distance. Thanks for the tip:
An externally hosted image should be here but it was not working when we last tested it.
That's the conventional way to do it.
It will take pretty high frequency interference into the enclosure.
I asked some while ago if turning the toroid upside down and using an even shorter screen lead would perform even better with our RF riddled homes where >2GHz interference is very local to our equipment.
Only one came back and used some supporting inductance calculation to suggest that upside down variation would offer no performance improvement.
Crucially he did not suggest a long screen lead out would be good enough.
It will take pretty high frequency interference into the enclosure.
I asked some while ago if turning the toroid upside down and using an even shorter screen lead would perform even better with our RF riddled homes where >2GHz interference is very local to our equipment.
Only one came back and used some supporting inductance calculation to suggest that upside down variation would offer no performance improvement.
Crucially he did not suggest a long screen lead out would be good enough.
So this is how the amp boards are going to be oriented and about in that exact location. The heatsink they will be attached to is 6 inches long and about 60mm tall. (1 inch thick).
Since I've read to keep RCA signal input as far away as possible from AC, would it make sense to locate these all the way to the right and low? Reason I say LOW is because there is a 60mm tall heatsink behind the amp boards shielding the signal from the transformer (front left side of case) and dc supply board (front right side of case).
And put the speaker terminals above those?
Here is what I am suggesting. Green are RCA inputs and Blue are speaker outputs. Photo below. If you don't mind, I'd love some feedback about this.
Since I've read to keep RCA signal input as far away as possible from AC, would it make sense to locate these all the way to the right and low? Reason I say LOW is because there is a 60mm tall heatsink behind the amp boards shielding the signal from the transformer (front left side of case) and dc supply board (front right side of case).
And put the speaker terminals above those?
Here is what I am suggesting. Green are RCA inputs and Blue are speaker outputs. Photo below. If you don't mind, I'd love some feedback about this.
Would the heatsink running front to back fit above that group of slots?
Then put the chipamps at the far right and the heatsink fins pointing towards the mains cabling on the extreme left.
You could stack the chipamp boards one above the other. The power and grounds could be 1" to 1¼" long heavy straight copper wires/rods linking the boards together. There are no adjustments on the PCB that need attention after assembly. Have you already had these working and proved to be non faulty?
Then put the chipamps at the far right and the heatsink fins pointing towards the mains cabling on the extreme left.
You could stack the chipamp boards one above the other. The power and grounds could be 1" to 1¼" long heavy straight copper wires/rods linking the boards together. There are no adjustments on the PCB that need attention after assembly. Have you already had these working and proved to be non faulty?
Yes.
Yes. I was thinking about doing it this way with a 3 inch by 3 inch heat sink and stacking them one above the other. But thought it'd be better if the heatsink was between both the transformer and dc supply .. i.e. the dc supply in the orientation you suggest would be within line of site of the amp boards -- but I guess this doesn't matter? Shielding from AC mains is most important?
Perhaps I need to order a 3 x 3 heatsink instead? and do this?
(In the orientation I suggested with the 6 inch long heat sink, I also had in mind that I put might a block of aluminum or something between the mains and amp boards. Kind of creating a boxed in area for the amp boards.)
Yep. Been using them for months now.
---
If I do as you suggest where should I locate the rca inputs vs speaker outs on the back?
AndrewT,
I am really liking your suggestion and I think I might be able to use my existing 150mm long heat sink. It's 60mm tall and each board needs 27.1mm (or let's say 28mm clearance). So there is an extra 4 mm of clearance.
I really like that the heat sink would be right over the vent for most of the length on the bottom vents, and virtually all of the length of the top vent. (Turns out the tape service door to vent conversion I did is going to be useful afterall.)
So I took a photo (with the cover of the case behind the unit for reference to illustrate airflow). Doing what you suggest would this be about right? Photo below:
Yellow: heat sink
Red: extra shield added (just a rectangular piece of some metal).
Green: signal input
Blue: speakers/output
Amp boards stacked one on top of other.
The heat sink would be butted up against the backplate. Therefore the rca signal input would be well shielded from the mains.
The 150mm x 60mm x 25mm heatsink I'll have tomorrow:
EDIT: I could put a thermal pad (I have some extra) between the heatsink and the backplate it'll be butted up against and screw it into the backplate with a couple screws from the backside. That way the case could help dissipate heat as well.
I am really liking your suggestion and I think I might be able to use my existing 150mm long heat sink. It's 60mm tall and each board needs 27.1mm (or let's say 28mm clearance). So there is an extra 4 mm of clearance.
I really like that the heat sink would be right over the vent for most of the length on the bottom vents, and virtually all of the length of the top vent. (Turns out the tape service door to vent conversion I did is going to be useful afterall.)
So I took a photo (with the cover of the case behind the unit for reference to illustrate airflow). Doing what you suggest would this be about right? Photo below:
Yellow: heat sink
Red: extra shield added (just a rectangular piece of some metal).
Green: signal input
Blue: speakers/output
Amp boards stacked one on top of other.
The heat sink would be butted up against the backplate. Therefore the rca signal input would be well shielded from the mains.
An externally hosted image should be here but it was not working when we last tested it.
The 150mm x 60mm x 25mm heatsink I'll have tomorrow:
An externally hosted image should be here but it was not working when we last tested it.
EDIT: I could put a thermal pad (I have some extra) between the heatsink and the backplate it'll be butted up against and screw it into the backplate with a couple screws from the backside. That way the case could help dissipate heat as well.
Last edited:
Any small gap/slot lets RF through.
Imagine hiding with your AM radio behind a big metal corrugated iron barn. The AM radio still picks up the transmissions from the other side of the barrier.
It is the same for your amplifier.
An effective barrier has to enclose the receiver. That what your metal case does. It attenuates RF from outside the box.
But internal interference passes around obstacles.
It's the rejection/cancellation of twisted pairs that reduces emi emission and reduces emi reception.
Make the aerials smaller !
Imagine hiding with your AM radio behind a big metal corrugated iron barn. The AM radio still picks up the transmissions from the other side of the barrier.
It is the same for your amplifier.
An effective barrier has to enclose the receiver. That what your metal case does. It attenuates RF from outside the box.
But internal interference passes around obstacles.
It's the rejection/cancellation of twisted pairs that reduces emi emission and reduces emi reception.
Make the aerials smaller !
Well I figure aluminum is highly heat conductive (not as good as copper) but still good. So that heat will dissipate laterally by conduction before dissipating via convective air flow. And convection will dissipate heat upwards along the whole length of the heatsink; but like you say more in the middle.
If I do them laterally is it okay that one pair of signal wires will be longer than the other -- same with the speaker wires?
EDIT: Oh I can't do them laterally, not enough room.. ONe of the amp boards would be right up against the backplate in the space needed by rca signal input.
EDIT #2: well maybe enough room but the rca would be like millimeters from the board's cap.
EDIT #3: I could put an aluminum plate spacer between the backplate and the heatsink to bring out the heatsink more to make room if doing laterally.
Last edited:
That heat sink I got stinks.. wish it was twice as beefy. I should probably buy another. It only cost me $9.
My brother is a machinist and probably has a scrap piece of aluminum laying around that he can cut into 150x60mm .. say 3/8" thick. Then that could be bolted to this other cheaper heat sink with conductive paste in the middle, to make a beefier heatsink. Would this work well?
My brother is a machinist and probably has a scrap piece of aluminum laying around that he can cut into 150x60mm .. say 3/8" thick. Then that could be bolted to this other cheaper heat sink with conductive paste in the middle, to make a beefier heatsink. Would this work well?
Last edited:
A spreader plate does indeed spread the heat.
But it also oncreases the Thermal Resistance due to the thickness of the added metal.
There is a better way but it adds considerably to the build.
Attach the two chipmaps directly to the spreader plate using a good thermal compound. Even better if you clamp the chipamp with a 2bolt and bar fixing.
Then electrically isolate the spreader from the heatsink and it's bolts, but using a large surface area film.
You gain a lot with the direct connection chipamp to spreader and then can afford to lose a little due to the spreader thickness resistance.
It's the electrical isolation between heatsink and bolts to the spreader plate that is the difficult bit, but worth the thought exercise before discarding it.
But it also oncreases the Thermal Resistance due to the thickness of the added metal.
There is a better way but it adds considerably to the build.
Attach the two chipmaps directly to the spreader plate using a good thermal compound. Even better if you clamp the chipamp with a 2bolt and bar fixing.
Then electrically isolate the spreader from the heatsink and it's bolts, but using a large surface area film.
You gain a lot with the direct connection chipamp to spreader and then can afford to lose a little due to the spreader thickness resistance.
It's the electrical isolation between heatsink and bolts to the spreader plate that is the difficult bit, but worth the thought exercise before discarding it.
AndrewT,
Is this what you were thinking?
What do you all think ? Ignore the soldering, I need to redo some and clean up the rosin.
CHG are connected as well and directly to the chassis via the fork terminal you see in the photo. What do you think of that? It's basically the star.. voltage grounds, speaker return & signal return are all connected to it on the PCB's then directly below to chassis.
What do you think about the proposed inputs (green) and outputs (blue) ? :
Made the heatsink screws extra long incase I want to replace those washers with a 1/4 to 3/8 inch thick aluminum bar the width and height of the heatsink.. It'd pick off some of the heat from the fin tips at least I'd think.
Is this what you were thinking?
What do you all think ? Ignore the soldering, I need to redo some and clean up the rosin.
An externally hosted image should be here but it was not working when we last tested it.
CHG are connected as well and directly to the chassis via the fork terminal you see in the photo. What do you think of that? It's basically the star.. voltage grounds, speaker return & signal return are all connected to it on the PCB's then directly below to chassis.
What do you think about the proposed inputs (green) and outputs (blue) ? :
An externally hosted image should be here but it was not working when we last tested it.
Made the heatsink screws extra long incase I want to replace those washers with a 1/4 to 3/8 inch thick aluminum bar the width and height of the heatsink.. It'd pick off some of the heat from the fin tips at least I'd think.
Last edited:
From diyaudio are some reasonably priced options. Free shipping. You can't beat $5, except unless you found it for free in the trash
Galaxy 3U w/Steel Covers – diyAudio Store (BETA)
Galaxy 3U w/Steel Covers – diyAudio Store (BETA)
Received the Aluminum bar yesterday, which happened to be just the right size. Got two of them, identical.. so I have a spare for some future project. $15 plus $10 shipping total for both.
My brother machined four holes and tapped them for 1/4-20 screws. Brought me the plate, 4 allen 1/4-20 screws and a drill bit to drill out my case the right size. Attached it today.
I really love the looks of it. One would never know it was a reclaimed commercial tape player enclosure lol. Here it is. Still need to do some sanding/polishing of the face plate -- just did it roughly before I handed it off to my brother to be machined :
Just needs my logo on the front now.
Before:
Anyways, I can't finish the amp until I get feedback from you guys about the locations of the rca inputs and speaker outputs in the previous post I just made. Also need to know that the way the amp boards are stacked and that the heat sink is sufficient.. and any other tips.
My brother machined four holes and tapped them for 1/4-20 screws. Brought me the plate, 4 allen 1/4-20 screws and a drill bit to drill out my case the right size. Attached it today.
I really love the looks of it. One would never know it was a reclaimed commercial tape player enclosure lol. Here it is. Still need to do some sanding/polishing of the face plate -- just did it roughly before I handed it off to my brother to be machined :
An externally hosted image should be here but it was not working when we last tested it.
Just needs my logo on the front now.
Before:
An externally hosted image should be here but it was not working when we last tested it.
Anyways, I can't finish the amp until I get feedback from you guys about the locations of the rca inputs and speaker outputs in the previous post I just made. Also need to know that the way the amp boards are stacked and that the heat sink is sufficient.. and any other tips.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.